The Enterprise of Science in Islam

1nr Lx+rrrrisr or Scirxcr ix Isi:x
Dibner Institute Studies in the History ot Science and 1echnology

ed Buchvald, general editor, Lvelyn Simha, governor
ed Z. Buchvald and AndrevWarvick, editors, Histories of the Lie.trou.lhe Lirth of
Mi.rophysi.s
ed Z. Buchvald and I. Bernard Cohen, editors, Isaa. êutou`s âturai Ihiiosophy
Anthony Gratton and Nancy Siraisi, editors, âturai Iarti.uiars. âture auo the Dis.ipiiues
iu Peuaissau.e Lurope
lrederic I. Holmes and 1revor H. Ievere, editors, Iustrumeuts auo Lxperimeutatiou iu the
History of Chemistry
Agatha C. Hughes and 1homas P. Hughes, editors, ystems, Lxperts, auo Computers.lhe
ystems pproa.h iu Mauaemeut auo Luiueeriu,1orio1ar II auo fter
an P. Hogendijk and Abdelhamid I. Sabra, editors, lhe Luterprise of .ieu.e iu Isiam.
êu Ierspe.ti:es
N. I. Sverdlov, editor, u.ieut strouomy auo Ceiestiai Di:iuatiou
1he MI1 Press
Cambridge, Massachusetts
Iondon, Lngland
1nr Lx+rrrrisr or Scirxcr ix Isi:x
Nev Perspectives
edited by
an P. Hogendijk and Abdelhamid I. Sabra
2003 Massachusetts Institute ot 1echnology
All rights reserved. No part ot this book may be reproduced in any torm by any elec-
tronic or mechanical means (including photocopying, recording, or intormation storage
and retrieval, vithout permission in vriting trom the publisher.
1his book vas set in Bembo and 1imes Nev Poman by 1he MI1 Press.
Printed and bound in the Lnited States ot America.
Iibrary ot Congress Cataloging-in-Publication Data
1he enterprise ot science in Islam : nev perspectives edited by an P. Hogendijk and
Abdelhamid I. Sabra.
p. cm. (Dibner Institute studies in the history ot science and technology,
Includes bibliographical reterences and index.
ISBN 0-262-19482-1 (hc. : alk. paper,
1. Science, Medieval. 2. ScienceIslamic countriesHistory. I. Hogendijk, . P. II.
Sabra, A. I. III. Series.
Q124.97 .L37 2003
309.17671dc21
2002029322
Cox+rx+s
Ix+roiuc+iox vii
Cox+riiu+ors xx
Cross-Cultural 1ransmission 1
1nr 1r:xsxissiox or Hixiu-Ar:iic Nuxrr:is
Prcoxsiirrri 3
Paul Kunitzsch
1nr 1r:xsxissiox or Ar:iic As+roxox. vi: Ax+iocn
:xi Pis: ix +nr Srcoxi Qu:r+rr or +nr 1vrir+n
Crx+ur. 23
Charles Burnett
1ranstormations ot Greek Optics 33
1nr M:x. Asrrc+s or Arrr:r:xcrs: Ar:iic Or+ics
+o ,o :i 33
Llaheh Kheirandish
Iix :i-H:.+n:x`s Prvoiu+iox:r. ProJrc+ ix Or+ics:
1nr Acnirvrxrx+ :xi +nr Ois+:cir 83
A. I. Sabra
Mathematics. Philosophy and Practice 119
M:+nrx:+ics :xi Pniiosorn. ix Mriirv:i Isi:x 121
Gerhard Lndress
1rx+n-Crx+ur. M:+nrx:+ics +nroun +nr L.rs or
Aiu S:ni :i-Kuni 177
. Iennart Berggren
I
1
2
II
3
4
III
3
o
Numbers, Geometry, and Architecture 197
Qu:ir:+us Mir:iiiis 199
acques Sesiano
C:icui:+ix Surr:cr Arr:s :xi Voiuxrs ix Isi:xic
Arcni+rc+urr 233
Yvonne Dold-Samplonius
Seventeenth-Century 1ransmission ot Astronomy 267
1nr P1IDDH^lP] or Ni+.:x:xi: 269
David Pingree
Ox +nr Iux:r 1:iirs ix S:xJ:q D:r`s 2I] LHPII 283
ulio Sams
Science and Medicine in the Maghrib and al-Andalus 307
A P:xor:x: or Prsr:rcn ox +nr His+or. or
M:+nrx:+ics ix :i-Axi:ius :xi +nr M:nrii
ir+vrrx +nr Nix+n :xi Six+rrx+n Crx+urirs 309
Ahmed Djebbar
Axo+nrr Axi:iusi:x rrvoi+ Iix Pusni`s Cri+iqur or
:i-Kixii`s IHPMCCLCICL CCMIl 331
Y. 1zvi Iangermann
Ixirx 373
Cox+rx+s vi
IV
7
8
V
9
10
VI
11
12
Between AD 800 and 1450, the most important centers Ior the study oI what
we now call 'the exact sciences were located in the vast multinational Islamic
world. The sciences denoted by this name included the mathematical sci-
ences oI arithmetic, geometry and trigonometry, and their applications in var-
ious felds such as astronomy, astrology, geography, cartography, and optics,
to mention only some oI the more prominent examples. During the eighth
and ninth centuries, the bulk oI Greek science, medicine and philosophy, and
much oI Indian and pre-Islamic Persian science, were appropriated by Islamic
civilization through a complex process oI translations Irom Pahlavi, Sanskrit,
Greek, and Syriac, in the course oI which Arabic became the language oI a
rich and active scientifc and philosophical tradition Ior many centuries. In the
eleventh and twelIth centuries, many Arabic scientifc works and Arabic ver-
sions oI Greek scientifc and philosophical texts were translated into Iatin, and
in turn were appropriated into the Iatin medieval culture. These translations
were crucial Ior the rise oI the 'renaissance oI the twelIth century in Europe
and they later played an important part in the development oI the exact sci-
ences during the Renaissance oI the sixteenth century. However, only a small
part oI the total Islamic accomplishments in science was transmitted to medi-
eval Europe. The scientifc endeavors in the Islamic world oI course remain as
an important subject to be investigated in its own right as a distinctive aspect
oI Islamic culture.
The Islamic tradition in the exact sciences continued well into the nine-
teenth century, and abundant source material is available in the Iorm oI unpub-
lished manuscripts in Arabic, Persian, and other languages in libraries all over
the world. In the last decades, many researchers have worked on the Islamic
scientifc tradition, and our views oI this tradition are rapidly changing as a
result oI recent discoveries. This process will, hopeIully, continue, because
important sources have not yet been identifed and studied. Hence the time is
not yet ripe Ior a reliable survey oI the entire feld.
The twelve chapters in this book discuss some oI the new perspectives
on the Islamic scientifc tradition emerging Irom recent historical research. The
Ix+roiuc+iox
Ix+roiuc+iox viii Ix+roiuc+iox ix
chapters are revised and updated versions oI contributions by experts in the his-
tory oI Islamic science at a conIerence on New Perspectives in Islamic Science
held at the Dibner Institute in November 1998. The emphasis oI the conIerence
was on the mathematical sciences, and the chapters in this volume represent a
cross-section oI the feld. Not all important topics could be included, and the
reader will notice the absence oI contributions on, Ior example, Islamic astro-
nomical instruments. Nevertheless, we believe that the present volume will
transmit to the reader a view oI an exciting and rapidly developing feld oI his-
torical research.
The editors have divided the twelve chapters into six groups oI two, under
the headings: Cross-cultural Transmission; TransIormations oI Greek Optics;
Mathematics: Philosophy and Practice; Numbers, Geometry and Architecture;
Seventeenth-century Transmission oI Astronomy; Science and Medicine in the
Maghrib and al-Andalus. This subdivision does not do Iull justice to the multi-
ple ways in which the twelve chapters are connected. Most oI the chapters are,
in one way or another, related to transmission oI scientifc knowledge, either
Irom one culture to another (Kunitzsch, Burnett, Kheirandish, Sabra, Endress,
Berggren, Pingree), or within the medieval Islamic world itselI (Kunitzsch,
Samso, Djebbar). Three chapters discuss mainly astronomy (Burnett, Pingree,
Samso), fve chapters are entirely or mainly on mathematics (Kunitzsch, Berg-
gren, Sesiano, Dold, Djebbar), two chapters are on optics (Sabra, Kheirandish),
while the chapter by Endress concerns the philosophy oI the mathematical sci-
ences. Several chapters are concerned with the relations between the exact sci-
ences and other felds, such as natural philosophy (Kheirandish, Endress),
architecture (Dold), medicine (Iangermann). Some chapters concern outsider`s
views on mathematics and its use (Endress, Iangermann), and medieval debates
on scientifc methodology (Berggren, Sabra). Two chapters discuss the individ-
ual mathematicians al-Kh (Berggren) and al-Ksh (Dold), whose styles and
attitudes turn out to be very diIIerent. The chapters by Djebbar, Iangermann,
and Samso concern geographical areas, which will be indicated by their medi-
eval Islamic names: al-Andalus is Islamic Spain and Portugal, that part oI the
Iberian peninsula that belonged to the medieval Islamic world, and the Maghrib
is Northwestern AIrica, that is, modern Morocco, Algeria, and Tunisia.
The chapters in this volume are mainly based on the analysis oI (oIten
unpublished) sources in Arabic and other languages. These documents are
usually scientifc, but may also include literary sources and biographies
(Kunitzsch, Djebbar). In some oI the chapters, the authors use special methods
to argue their new insights, ranging Irom the philological analysis oI technical
terms in diIIerent languages (Burnett, Kheirandish), via the analysis oI diIIer-
ent systems oI numerals (Kunitzsch, Burnett), to computer analysis oI numeri-
cal tables (Samso).
Ix+roiuc+iox viii Ix+roiuc+iox ix
Several chapters summarize many years oI research work by their authors,
as the Iollowing three examples show. Berggren`s characterization oI al-Kh as
a mathematician is the outcome oI his project to publish a complete edition oI
al-Kh`s works; Sabra`s analysis oI the argument oI Books IVII oI Ibn al-
Haytham`s Optics is based on his Arabic edition and English translation oI this
large and Iundamental work; Djebbar`s survey oI mathematics and astronomy
in the Maghrib and Andalusia includes his own research and that oI his pupils.
Several chapters deal with striking examples oI originality oI the Islamic tradi-
tion (Sabra, Dold, Sesiano).
Not surprisingly, the chapters in this volume support the view that the
Islamic scientifc tradition was even richer, more proIound, and with more
complex relations to other cultures than had been thought hitherto. As more
sources become gradually available, our picture also becomes more detailed.
Thus, Berggren is now able to characterize Ab Sahl al-Kh as a conserva-
tive, 'old-guard mathematician among his contemporaries, Djebbar is able to
list diIIerences between the algebraic traditions in al-Andalus and the Maghrib,
and so on. Most oI the chapters also conIront the reader with unsolved histor-
ical problems. Thus, Kunitzsch reminds us that it is still unknown when and
where the Hindu-Arabic number symbols evolved Irom the Eastern Islamic
Iorms (now used in the middle East) to the Western Islamic Iorms, which are
virtually identical to the modern Iorms 1,2,3,4,5,6,7,8,9,0. In spite oI all the
progress, much about science in the Islamic tradition is still unknown.
Since some oI the authors have diIIerent preIerences Ior systems oI reIer-
encing and on the precise ways to transcribe or write Arabic names, the editors
have decided not to interIere with these preIerences. Thus the same name may
appear in diIIerent chapters in slightly diIIerent Iorms, Ior example, Abu`l-
Rayhn, Ab al-Rayhn, and Ab ar-Rayhn; al-Kh, al-Qh.
The 'new perspectives in each chapter will now be briefy outlined in
the rest oI this introduction, under the six headings mentioned above.
Cross-cui+ur:i 1r:xsxissiox
In his chapter, 'The Transmission oI Hindu-Arabic Numerals Reconsidered,
Paul Kunitzsch strives to distinguish between, on the one hand, what can be
accepted as Iairly established Iact in regard to this widely discussed subject,
and, on the other, what still remains uncertain and in need oI Iurther study.
Kunitzsch accepts the usually cited evidence presented by the Arabic and Syr-
iac sources in support oI the thesis that the nine numerals plus a symbol Ior an
empty place initially came to the Arabs Irom India in the eighth century. He
notes that variable Iorms oI the nine numerals continued to be used well into
the eleventh century, but rejects the hypothesis, once proposed by S. Gandz,
Ix+roiuc+iox x Ix+roiuc+iox xi
that the dust board utilized by the Arabic mathematicians Ior perIorming their
calculations later became the model Ior the Iatin abacus.
Kunitzsch argues in the second part oI his chapter that the term dust
numerals (Arabic: hurf al-ghubr), rather reIerred to the written fgures, as
opposed to the non-written numerals signifed in processes known as men-
tal reckoning or fnger reckoning. The majority oI the numerals current in
the Arabic/Islamic West (the Maghrib and al-Andalus) could be obtained, so
Kunitzsch remarks, Irom their Eastern counterparts. Thus, while the general
line oI borrowing oI the numerals appears quite clear, Kunitzsch emphasizes
the need Ior Iurther research to establish the existence oI the Hindu-Arabic
numerals in the Western-Islamic sources prior to AD 1300. This should help
to clear up the process oI early Iatin borrowing by way oI the twelIth-century
Arabic-Iatin translations.
In his chapter, 'The Transmission oI Arabic Astronomy via Antioch and
Pisa in the Second Quarter oI the TwelIth Century, Charles Burnett studies a
hitherto neglected channel oI transmission oI the exact sciences Irom the East-
ern Islamic world to Christian Europe beIore Toledo emerged as a center oI the
translation industry, and the frst translation oI Ptolemy`s Almagest was pro-
duced in Sicily Irom Greek. Burnett`s frst source is a medieval Iatin manu-
script oI Book IIV oI the Almagest oI Ptolemy (ca. 150 AD), now in Dresden.
By a meticulous analysis oI the technical terminology and the numerals that
were used, Burnett shows that this 'Dresden Almagest was translated Irom an
Arabic version around 1120 in Antiochia (now in Turkey), and that the transla-
tor wanted to give the impression that he had translated the work Irom Greek.
Burnett then introduces a second source, a cosmological work in Iatin enti-
tled Liber Mamonis, which is written in the same technical terminology as
the Dresden Almagest. Burnett shows that this work was inspired by Arabic
sources, and he suggests that the title is derived Irom one such source, the
Jerihed Astronomical 1ables (Al-Zj al-Mumtahan) by the astronomers around
Caliph al-Mamn in the early ninth century. Burnett identifes the author oI
the Liber Mamonis as Stephanus the philosopher, who originated Irom Pisa
but who worked in Antiochia around 1120. Burnett argues that the 'Dresden
Almagest was written in the same milieu. In the last part oI his chapter, he
studies a third work, the 1ables of Pisa, which is based on the lost astronomi-
cal tables oI the Iranian astronomer al-SI (d. 986), and which has oIten been
related to the Jewish astronomer Abraham ibn Ezra (d. ca. 1160). By an anal-
ysis oI the systems oI numeration used in the 1ables of Pisa, Burnett estab-
lishes a clear connection with the Liber Mamonis, and hence with Antiochia.
Thus, Antiochia is shown to be an important center Ior the early transmission
oI astronomy IromArabic into Iatin.
Ix+roiuc+iox x Ix+roiuc+iox xi
1r:xsrorx:+ioxs or Grrrk Or+ics
The science oI optics (Arabic: ilm al-manzir, Greek: h optik techn) is per-
haps unique as a Greek mathematical discipline that received in the Islamic
middle ages a radical transIormation which ultimately succeeded in launching
it on an entirely new course. As a mode oI inquiry consisting oI a combined
mathematical and experimental approach to visual perception, optics was rep-
resented in Greek antiquity by works oI Euclid (ca. 300 BC), Ptolemy (second
century AD), and Theon oI Alexandria (Iourth century AD), all oI which carried
the title Optika, and at least two oI which, namely those oI Euclid and Ptolemy,
came to be known in Arabic translations, some made as early as the ninth cen-
tury, and given the title al-Manzir. The radical transIormation took place in
the frst halI oI the eleventh century, in the large seven-part work oI al-Hasan
ibn al-Haytham, Kitb al-Manzir. This work had the good Iortune oI fnding
its way to Muslim Spain where a Iatin translation was made oI it probably in
the late twelIth century, thereby securing an entry into the main stream oI Euro-
pean scientifc, mathematical, and philosophical thought where it is known to
have exerted considerable infuence that lasted all the way up to the seven-
teenth century.
In her chapter, 'The Many Aspects oI Appearances, Elaheh Kheirandish
Iocuses on transIormations occurring already through the ninth-century Ara-
bic assimilation oI central Greek optical terms, beginning with the Greek and
Arabic names Ior the discipline itselI. She takes as her starting point a chap-
ter in al-Erb`s Catalogue of the Sciences (frst halI oI the tenth century), a
work later also widely known in Europe in a medieval Iatin translation. Al-
Erb was a philosopher, and not himselI one oI those who contributed to
the mathematical science oI vision. But his chapter oIIers a valuable picture
oI the discipline and a set oI operative terms in it as understood by an intelli-
gent witness Irom the century immediately preceding that oI Ibn al-Haytham.
Kheirandish Iurther elaborates her analysis by critical comparisons with terms
and phrases Iound in the writings oI the ninth-century Arabic writers and their
Greek sources.
Eollowing the order oI exposition in al-Erb`s chapter, Kheirandish
organizes her discussion in sections under fve headings: I. Veracity and Accu-
racy oI Vision, in which she notes a particular emphasis on two aspects oI
appearances, namely veracity and accuracy, and a greater emphasis in al-Erb
on the latter than on the Iormer; II. Justifability and Variety oI Demonstra-
tion, noting the concern oI optics with explanation understood as a demonstra-
tive procedure, and the connected quantifcation oI visual clarity, by which she
does not oI course mean to imply any kind oI measurement; III. Versatility and
Ix+roiuc+iox xii Ix+roiuc+iox xiii
Eallibility oI Applications, where she regrets the Iact that al-Erb`s recogni-
tion oI various applications oI optics does not lead him to a discussion oI mir-
rors and the principle oI refection, a neglect which in addition to the omission
oI surveying she fnds surprising, and one oI the problematic cases oI trans-
mission; and, fnally, IV. Elements and Mechanisms oI Vision, and V. Modes
and Mediums oI Operation, in which Kheirandish goes into the widespread
and, sometimes, substantive conIusion between optical refection and reIrac-
tion oI visual rays. This conIusion strangely persisted in the works oI Nasr al-
Dn al-Ts and Qutb al-Dn al-Shrz in the thirteenth century, and until Ibn
al-Haytham`s work became generally known to Arabic readers, thanks to the
comprehensive Commentary made by al-Shrz`s giIted student, Abu `l-Hasan
Kaml al-Dn al-Ers more than two hundred years aIter Ibn al-Haytham died
(ca. 1041), and probably more than a hundred years aIter the latter`s Kitb al-
Manzir began to be utilized by readers oI Iatin.
In the Iollowing chapter by A. I. Sabra, 'Ibn al-Haytham`s Revolution-
ary Project in Optics: The Achievement and the Obstacle, the author claims
to use 'revolution in the strict sense oI a conscious and radical transIorma-
tion oI a widely practiced and accepted approach to a whole scientifc disci-
pline, a transIormation that goes to the heart oI the basic assumptions oI the
traditional system. Sabra does not believe that Ibn al-Haytham`s departure
Irom tradition was a creatio ex nihilo, and accordingly he stresses elements
borrowed Irom earlier methods and doctrines, especially those derived Irom
Ptolemy`s Optics and Aristotelian physics; but Sabra tries to show that these
elements now serve diIIerent Iunctions in a new system. The chapter is actually
an outline oI the single, continuous argument which, according to Sabra, runs
through all the seven books that make up the Optics oI Ibn al-Haytham: Hav-
ing totally rejected, on the basis oI empirical evidence, the visual-ray hypothe-
sis as the Ioundation oI previous mathematical theories oI vision, and aligning
himselI (again on the basis oI experience) with the Peripatetic view oI vision
as the reception oI Iorms oI light and color, Ibn al-Haytham was led to accord
psychology a new, inevitable and Iundamental role never realized earlier in
the works oI the Greek mathematicians and their Arabic successors. The new
approach Iurther leads Ibn al-Haytham to reIormulate the existing explanations
oI rectilinear transmission, refection and reIraction oI light in ways that may
conIorm with the older geometry, but that also generate new problems posed
by his own treatment oI specular images, as Sabra notes.
The obstacle reIerred to in the title oI Sabra`s chapter has to do with the
doctrine, inherited Irom Galen, locating the sensitivity oI the eye in the crystal-
line humor. This was a IateIul error which Iorced Ibn al-Haytham to Iall back
on psychology in a last attempt (in Book VII) to explain crucial experiments
Ix+roiuc+iox xii Ix+roiuc+iox xiii
that he was the frst to describe in the history oI optics. The erroneous doctrine
was itselI again adopted as an empirically established Iact oI ocular physiol-
ogy, but this time Ibn al-Haytham put his trust in what was widely reported
and accepted by the medical tradition as empirical prooI. It was Irom here, as
Sabra remarks but does not discuss, that Kepler started his own researches that
led to his well-known breakthrough in the Paralipomena ad Jitellionem, pub-
lished in 1604.
M:+nrx:+ics: Pniiosorn. :xi Pr:c+icr
Gerhard Endress begins his survey chapter, 'Mathematics and Philosophy
in Medieval Islam, with a summary oI the views oI the two most important
ancient Greek philosophers on mathematics and astronomy. Plato believed that
number and mathematics were related to an eternal world oI ideas, which is
the essence and source oI the changing world in which we live. Aristotle, on
the other hand, believed that mathematics was abstracted Irom reality and not
directly related to the essence oI the real world. However, he believed that the
properties oI the real world could be deduced Irom a Iew basic principles oI
natural philosophy, in a way somewhat similar to the deductive reasoning in
geometry. UnIortunately, Aristotle drew up his cosmological principle oI uni-
Iorm rotation around the center oI the earth at a time when astronomy was
still in a qualitative state. When Ptolemy worked out his planetary theory in
the second century AD, he could only make his models correspond to reality
by introducing epicycles and equant points. The resulting contradiction with
the Aristotelian principle oI uniIorm motion around the center oI the earth was
never resolved in antiquity.
Endress then turns to the attitude oI the most important Islamic philos-
ophers with respect to the mathematical sciences. The early philosopher al-
Kind (ca. 830) and his Iollowers considered mathematics, in the Platonic vein,
as an intermediary between philosophy and the science oI nature. Al-Erb
(d. 950) accepted the Aristotelian view oI mathematics and natural science,
and the same is basically true Ior Ibn Sn (Avicenna, d. 1037). These philoso-
phers were not astronomers, but Ibn al-Haytham (d. ca. 1040) was a competent
astronomer who tried to bridge the gap between Aristotelian physics and Ptol-
emaic astronomy. He fnally came to the conclusion that astronomical models
had to be deduced by demonstrative reasoning (in the way oI Aristotle) Irom
basic principles diIIerent Irom those enunciated by Aristotle, and that the ad-
hoc devices used by Ptolemy were unsatisIactory. However, Ibn al-Haytham
did not work out a new astronomical system oI his own.
Endress then turns to the Andalusian astronomers and philosophers,
who distinguished between the natural philosophy oI Aristotle, dealing with
Ix+roiuc+iox xiv Ix+roiuc+iox xv
essences and causes, and the geometrical models oI the astronomers. These
Andalusian scholars believed that these mathematical models dealt only with
accidental properties, abstracted Irom reality, and thus one could not use them
to penetrate into reality in its own right. They believed that Aristotle had
reached human perIection, and they made a Iew hopeless attempts to pro-
pose Aristotelian alternatives to Ptolemy`s models in order to (re?)construct
a true Aristotelian cosmos. Endress presents an interesting quotation Irom the
Iamous Ibn Rushd/Averroes (d. 1198), expressing Irustration and unhappiness
with the results.
In the fnal section, Endress discusses the Jewish philosopher Maimonides
(d. 1204) and late Islamic thinkers, who decided that 'natural philosophy is
domain oI God, and that Aristotelian natural philosophy ought to be discarded
as contrary to religion. They argued that the astronomers were Iree to imagine
models that correspond to observation and make possible predictions, an atti-
tude that may strike us as quite modern.
Whereas Endress`s chapter is devoted to the views oI relative outsiders
(astronomers and philosophers) on the nature and use oI mathematics, J. I.
Berggren studies an insider`s view in his chapter 'Tenth-Century Mathematics
through the Eyes oI Ab Sahl al-Kh.As many other research mathematicians
in his time, al-Kh was Iascinated by geometry in the style oI the Hellenistic
geometers, and some oI his tenth-century contemporaries regarded him as the
'master oI his age in the art oI geometry. Berggren suggests, plausibly, that al-
Kh was Iascinated by the certainty oI geometrical knowledge. He preIerred
to work on problems discussed in or inspired by the ancient Greek geometry oI
Archimedes and Apollonius, such as geometrical constructions oI the regular
heptagon and other fgures by means oI conic sections, and the determination
oI the volumes and centers oI gravity oI solids including the paraboloid. Berg-
gren states that among tenth-century geometers, al-Kh was unique in fnding
and (generally) solving geometrical problems oI some real depth.
Al-Kh believed in the progress oI science and he says that 'the sci-
ence oI geometry will endure and will continue to grow, in contrast to man`s
liIe span, which comes to an end. He contributed to methodological debates
on analysis and synthesis, the mathematical concept oI a 'known ratio in con-
nection with the quadrature oI the circle, and the question to what extent it was
legitimate to use motion in geometry. Unlike most oI his colleague-mathema-
ticians, al-Kh was uninterested in arithmetic, algebra, and trigonometry. He
even regarded mathematics based on numerical approximations as 'bad math-
ematics because it is not based on exact and demonstrative methods. Berggren
concludes that al-Kh was rather conservative in his outlook, and that he may
have been the last mathematician to look on mathematics with the eyes oI the
great Hellenistic geometers.
Ix+roiuc+iox xiv Ix+roiuc+iox xv
Nuxirrs, Groxr+r., :xi Arcni+rc+urr
The Islamic mathematicians contributed not only to geometry but also to the
theory oI numbers, and they studied a combination oI the two in what they
called 'harmonious dispositions oI numbers (Arabic: add al-wafq), nowa-
days called magic squares. A magic square is a square array oI integer num-
bers, with the property that the sums oI all elements in each column, row, or
diagonal are equal. In modern terms, the number oI elements in each row, col-
umn, or diagonal is called the order oI the magic square, and the sum oI the
elements in each row, column, or diagonal is called the magical constant.
1
Although magic squares are probably oI pre-Islamic (Persian?) origin, there is
no evidence that they were seriously studied beIore the Islamic tradition. The
title oI Jacques Sesiano`s chapter, 'Quadratus Mirabilis, is the name he has
given to the most complex type oI magic squares that has hitherto been Iound
in the entire medieval Islamic tradition. The construction oI this type oI magic
square is explained in a text by the late tenth-century mathematician al-Antk,
but Sesiano points out that it may have been discovered earlier. Eor each integer
n, al-Antk constructs a magic square oI order n, composed oI the numbers 1,
2, . . . n
2
, with the Iollowing two properties:
1. The odd numbers 1, 3, . . . are placed in a rhombus in the middle oI the square
and the even numbers 2, 4 . . . in Iour triangular corners oI the square;
2. The magic square is 'bordered, that is to say that iI all 4n 4 numbers on the
outside are taken away, the remaining square is also a new magic square oI order
n 2, and iI all 4n 12 numbers on the outside oI this new square are taken away,
the remaining square is also a magic square, and so on, until one is leIt with a magic
square oI order 4 or 3.
Sesiano gives a transcription oI the construction oI this 'Quadratus
Mirabilis in modern notation as well as a literal translation oI the Arabic text
oI al-Antk. His chapter shows that the theory oI magic squares reached a
much higher level in the tenth century than had been thought beIore. The term
'magic square is European, and as Sesiano points out, magic squares were
mostly investigated in the Islamic tradition because oI their mathematical as
well as recreational interest. Most magical 'applications oI these squares are
relatively late and only utilize the simplest Iorms.
The relation between geometry, numerical computation and architecture
is explored in the chapter byYvonne Dold-Samplonius, which is entitled 'Cal-
culating SurIace Areas and Volumes in Islamic Architecture. She begins with
a concise introduction to Islamic architecture. Not much is known about the
relations between architecture and mathematics in the early Islamic tradition,
but the available evidence suggests that sophisticated mathematics entered
Ix+roiuc+iox xvi Ix+roiuc+iox xvii
Islamic architecture relatively late in the tradition. (OI course, Iuture research
may show this impression to be incorrect.) Dold lists a number oI rough calcu-
lations oI volumes and surIaces oI domes Irom various practical mathematical
works until the thirteenth century, and she compares these calculations with the
sophisticated approach in the Key to Arithmetic oI al-Ksh, who died around
1429 in Samarkand. Al-Ksh computed various coeIfcients which enabled
craItsmen to easily fnd surIaces and volumes oI various types oI domes Iound
in Central Asia. His computations are unrelated to earlier Islamic studies on the
volume oI parabolic domes (paraboloids) oI a more theoretical nature, which
Dold characterizes as 'highschool mathematics. One oI the authors oI these
more theoretical studies was al-Kh, the subject oI Berggren`s chapter in this
volume. Dold fnishes her chapter with a brieI look at al-Ksh`s discussion
oI the stalactite vaults called 'muqarnas, which are characteristic oI Islamic
architecture.
Al-Ksh had a talent Ior fnding user-Iriendly and highly accurate approx-
imative solutions oI diIfcult practical problems which cannot be solved exactly,
and this is why Dold-Samplonius calls him 'the frst modern mathematician.
Her view is supported by Iurther mathematical analysis oI al-Ksh`s calcu-
lations. Thus, al-Ksh and al-Kh represent two radically diIIerent types oI
Islamic mathematicians.
Srvrx+rrx+n-crx+ur. 1r:xsxissiox or As+roxox.
The chapters in this section by David Pingree and Julio Samso discuss con-
crete examples oI transmission oI Islamic science aIter the medieval period.
Both chapters are related to the tradition oI the Iamous astronomical hand-
book (Zf) oI Ulgh Beg (13931449), the ruler oI Samarkand who Iounded
an astronomical observatory and appointed al-Ksh as its director. As Pin-
gree explains in his chapter, 'The Sarvasiddhntarja oI Nitynanda, the Zf
oI Ulgh Beg was revised in the early seventeenth century at the court oI the
Moghul emperor Shah Jahn at Delhi, and the reworking was then translated
into Sanskrit by Nitynanda. Because the translation Iailed to fnd Iavor with
the traditional Hindu astronomers at Delhi, Nitynanda decided to write a San-
skrit apology Ior Islamic astronomy, entitled Sarvasiddhntarja. Pingree pres-
ents a detailed analysis oI chapters 2 and 3 oI this work, on the computation
oI the mean and true longitudes oI the planets. Nitynanda rephrased Islamic
astronomy in a language which Hindu astronomers could understand, and thus
he expressed the planetary mean motions in integer revolutions in a Kalpa oI
4,320,000,000 years. He also tried to show that the diIIerences between tradi-
tional Hindu astronomy and what he called the 'Roman Zj, that is, the new
Islamic system, were only small. Nitynanda gave only numerical algorithms
Ix+roiuc+iox xvi Ix+roiuc+iox xvii
and he did not discuss the geometrical background oI the Islamic (essentially
Ptolemaic) models oI planetary motion; since he also Iailed to defne many oI
his new technical terms, his colleagues must have Iound it hard to understand
his work. Pingree points out that some oI the new vocabulary in the Sarvasid-
dhntarfa was nevertheless used in other translations made at Delhi a Iew
years later, Ior example in the Sanskrit prose version oI the 1abulae Astronom-
icae oI Philippe de la Hire (16401718). Thus, Pingree`s chapter sheds light
on one oI the most neglected areas in the history oI the transmission oI Islamic
astronomy, namely its development and infuence in South Asia.
Julio Samso`s chapter, 'On the Iunar Tables in Sanjaq Dr`s Zf al-
Sharf, is related to the complex transmission oI astronomical knowledge
Irom the Eastern to the Western Islamic world. Erom the ninth through the
eleventh centuries, there existed a special type oI Islamic astronomy in al-
Andalus. One oI the characteristics oI this Andalusian astronomy is the exis-
tence oI special theories to explain the supposed phenomenon oI 'trepidation,
that is the oscillation in the ecliptic longitudes oI all fxed stars with respect
to the vernal point. These theories were also studied in the Maghrib. East-
ern Islamic astronomers, such as Muhy al-Dn al-Maghrib and Ibn al-Shtir,
rejected trepidation, and Irom the late 14th century on, their theories were also
transmitted to the Maghrib. Erom that time on, the astrologers in the Maghrib
continued to use the older Andalusian astronomy, while the more sophisticated
Eastern Islamic astronomy was used by the astronomers and muwaqqits (oIf-
cial time-keepers in mosques). The Zf oI Ulgh Beg was apparently transmit-
ted to the Maghrib in the 17th century, and Julio Samso studies the infuence oI
that work on the 17th-century Zf al-Sharf ('Noble Astronomical handbook)
by Sanjaq Dr oI Tunis. Samso briefy summarizes the complex (essentially
Ptolemaic) astronomical contents oI this work, and using computer programs
developed by Benno van Dalen, he analyzes the lunar tables, to establish the
dependency oI Zf al-Sharf on the Zf oI Ulgh Beg in detail. The technical
terminology and the concrete examples oI numerical tables in Samso`s chapter
may give the reader some idea oI the knowledge and skills which a medieval
Islamic astronomer needed in order to compile an astronomical handbook (Zf)
Ior his own city and time.
Scirxcr :xi Mriicixr ix +nr M:nrii :xi :i-Axi:ius
Ahmed Djebbar`s contribution to this volume, entitled 'A Panorama oI Re-
search on the History oI Mathematics in al-Andalus and the Maghrib between
the Ninth and Sixteenth Centuries, is a unique survey oI the research done by
modern historians between 1834 and 1980 on the history oI medieval math-
ematics and astronomy in those geographical areas. Djebbar treats not only
Ix+roiuc+iox xviii Ix+roiuc+iox xix
Western literature but also the research by Arabic scholars in the nineteenth
and twentieth centuries, and he comments on the historical and political moti-
vations oI these Arabic researchers. Eor the period aIter 1980, Djebbar sum-
marizes much oI the work oI his own research group at the Ecole Normale
Superieure in Algiers, which is now the most important center Ior the study oI
medieval Maghribi mathematics and astronomy. The survey shows the prog-
ress that has been made in the last decades in our understanding oI medieval
Maghribi science, as well as a number oI its more puzzling aspects, which will
have to be clarifed by Iuture research. Examples are the discontinuities in the
transmission Irom the Eastern Islamic to the Western Islamic world, the Iact
that mathematics in the Maghrib was limited to arithmetic Irom the late 13th
century onwards, the transmission oI Euclid`s Elements and other geometrical
works, and the silence oI the bio-bibliographical works on the algebraic tra-
dition in al-Andalus and the Maghrib. Iibraries in the Maghrib contain mas-
sive amounts oI unpublished Arabic manuscripts and it is expected that Iurther
study oI these sources will not only shed light on these open problems, but also
reveal many more details about the mathematical and astronomical traditions
in specifc periods and specifc areas oI the Maghrib.
In his chapter 'Another Andalusian Revolt? Ibn Rushd`s Critique oI al-
Kind`s Pharmacological Computus, Tzvi Iangermann touches upon the ques-
tion which some have raised with regard to intellectual developments in North
AIrica and Muslim Spain under the rule oI the ideologically driven Almohads/
al-Muwahhidn (524667/11301269), namely whether certain trends in sci-
ence, medicine and philosophy, as well as in Islamic law and Arabic grammar,
could be interpreted as maniIestations oI a general revisionist attitude toward
the established authorities in those felds in the Eastern Islamic world. Evidence
Ior the revisionist thesis in a scientifc feld has previously been expounded
with reIerence to the conscious and reasoned rejection oI the Eastern decision
in Iavor oI Ptolemaic, as opposed to Aristotelian astronomy, by Andalusian
scholars, including especially the Iamous Ibn Rushd/Averroes, who reIerred
to the Ptolemaic system oI eccentrics and epicycles as 'the astronomy oI our
time. Iangermann is concerned with another episode: Ibn Rushd`s vehement
attack, in his medical textbook, Kitb al-Kulliyyt, on the work oI the East-
ern ninth century polymath, al-Kind, in which the latter proposed a new, non-
Galenic computus Ior calculating the right quantities oI simple drugs, in order
to produce the desired degree oI their compounded elemental qualities: heat,
cold, dry, and moist. Averroes objected, among other things, that al-Kind was
causing conIusion in a medical subject by straying too Iar beyond the boundar-
ies and the rules or laws/qawnn proper to a natural inquiry, such as the art oI
medicine, into considerations oI numbers and music. It is known that Averroes,
Ix+roiuc+iox xviii Ix+roiuc+iox xix
in a commentary on Aristotle`s Meteorology, raised an exactly parallel objec-
tion against the role oI mathematics in the 'physical work oI another East-
erner, Ibn al-Haytham.
Iangermann here reIrains Irom directly answering the general, cultural
question: hence the question mark in his title. And, Iollowing a good maxim,
truth is in the details, he oIIers in the frst part oI his chapter a lucid and enlight-
ening analysis oI many oI the details involved in Averroes`s arguments, with
due emphasis on their immediate medical context in the writings oI authorities
Irom Galen to medical authors in the Arabic tradition, including members oI
the distinguished Andalusian Ibn Zuhr Iamily. In the second part oI his chapter,
devoted to the question oI Contexts (thus in the plural), Iangermann observes
the lack oI general interest in Kind`s computus, but reIers to a lost treatise by
the infuential Abu`l-Al ibn Zuhr (d. 1130), directly addressed to al-Kind`s
book. Why |then| was there so little interest in al-Kind`s book? Iangermann
asks, and delivers the most plausible answer: while pharmacologists (in the
East and the West) were spurred by practical applications, Ibn Rushd, as the
confrmed Aristotelian philosopher, was (doggedly?) concerned Ior what he
took to be Aristotelian methodology. Iangermann concludes with the advice
that Iull attention should be given to the persistent interest in pharmacology
in its own right. His fnal paragraph reIers to the great Ab Marwn ibn Zuhr
(Avenzoar, d. 1161), son oI Abu`l-Al and close Iriend to Ibn Rushd, who had
served as wa:r and physician to the Iounder oI the Almohad dynasty, Abd al-
Mumin (11301163), and who had written his medical work al-1aysr on the
advice oI Ibn Rushd. This reIerence calls up again the ghost oI an underlying
Andalusian ideological tendency whose many ramifcations Ior the intellec-
tual history oI Muslim Spain and the Maghrib still need to be defned and ade-
quately explored.
1he Editors
No+r
1. II the order oI the magic square is n and the consecutive numbers are 1. . . n
2
,
the magical constant is
1
n
(1
2
2
2
. . . n
2
)
1
2
n(n
2
1).
1. Lennart Berggren is proIessor oI mathematics at Simon Eraser University.
His research interests center on the history oI the mathematical sciences in
ancient Greece and medieval Islam, in particular geometry and geometrical
methods in geography and astronomy. He is translating and annotating the
extant works oI Ab Sahl al-Kh, and has authored or co-authored a number
oI articles and books, including Episodes in the Mathematics of Medieval
Islam (1986), Euclids Phaenomena (with R. Thomas, 1996), and Ptolemys
Geography. An Annotated 1ranslation of the 1heoretical Chapters (with A.
Jones, 2000).
Charles Burnett is proIessor oI the history oI Islamic infuences in Europe at
the Warburg Institute, University oI Iondon. He has written extensively on the
translation oI scientifc and philosophical texts Irom Arabic into Iatin in the
Middle Ages. His recent works include 1he Introduction of Arabic Learning
into England (1997), Scientihc Weather Forecasting in the Middle Ages. 1he
Writings of al-Kind (with Gerrit Bos, 2000), and Ab Mashar on Historical
Astrology. 1he Book of Religions and Dynasties (On the Great Confunctions),
2 volumes (with K. Yamamoto, 2000).
Ahmed Djebbar is ProIessor oI History oI Mathematics at Iille University.
His main research interests are in the history oI Arabic mathematics in North
AIrica and Muslim Spain. He has been secretary to the International Commis-
sion Ior the History oI Mathematics in AIrica since 1986. Recent publications
include Une histoire de la science arabe (2001) and La vie et luvre dIbn
al-Bann. un essai biobibliographique (2001).
Yvonne Dold-Samplonius is an associate member oI the Institute Ior Sci-
entifc Computing oI the University oI Heidelberg, where she directed the
video 'Qubba Ior al-Kash, showing al-Kash`s geometrical constructions Ior
determining the volumes oI domes and arches. With her team she is now work-
ing on a three-year Muqarnas (stalactite vaults) project. With international
cooperation they want to develop a computer method to create virtual muqar-
nas based on old plans, such as the Topkapi Scroll in Istanbul.
Cox+riiu+ors
Gerhard Endress is proIessor oI Arabic and Islamic Studies at the University
oI Bochum. His research Iocuses on the Arabic translation and transmission
oI the Greek sciences, and the history oI philosophical thought in Islam. He
has published textual editions and studies oI the Arabic translations oI Aris-
totle and late Hellenistic Neoplatonism, and oI the Arabic philosophers. His
publications include An Introduction to Islam (English ed. 1988, rev. 2001); A
Greek and Arabic Lexicon. Materials for a Dictionary of the Mediaeval 1rans-
lations from Greek into Arabic (with D. Gutas, since 1992); and Averroes and
the Aristotelian 1radition (with J. A. Aertsen, 1999).
1an P. Hogendijk is a member oI the Department oI Mathematics oI the
University oI Utrecht. His research interest is the history oI the mathematical
sciences in Greek antiquity and medieval Islamic civilization. He has recently
published an edition oI the mathematical works oI al-Jurjn (10th c.) and an
analysis oI the works oI Ab Nasr ibn Irq (ca. 1000) on sundial theory. He is
now working on an edition and English translation oI a ninth-century manual oI
geometrical problems by Nuaim ibn Ms.
Elaheh Kheirandish is a senior resident Iellow at The Dibner Institute Ior
the History oI Science and Technology, Massachusetts Institute oI Technol-
ogy. Her publications include 1he Arabic Jersion of Euclids Optics. Kitb
Uqldis f Ikhtilf al-manzir (1999), and articles on the history oI mathemati-
cal sciences in the Islamic world. Her current work includes two projects in
the history oI optics and mechanics involving the application oI the electronic
medium (the IOTA project at the Dibner Institute, and the Archimedes Project
at Harvard`s Classics Department).
Paul Kunitzsch was proIessor oI Arabic Studies at the University oI Munich
Irom 1977 to 1995. His main felds oI research are the transmission oI the sci-
ences, especially astronomy and astrology, Irom Antiquity to the Arabs and
Irom the Arabs to medieval Europe; Arabic knowledge oI the stars and con-
stellations; and Oriental infuences in medieval European literature. His recent
publications include 1he Melon-Shaped Astrolabe in Arabic Astronomy (1999)
and Claudius Ptolemaus, Der Sternkatalog des Almagest (19861991).
Y. Tzvi Langermann is associate proIessor oI Arabic at Bar Ilan University.
His research interests cover a broad range oI topics in medieval science and
philosophy, with special emphasis on the thought oI Moses Maimonides
and the intellectual history oI Yemenite Jewry. Recent publications include
1he Jews and the Sciences in the Middle Ages (1999); 'Studies in Medieval
Hebrew Pythagoreanism; and 'Criticism oI Authority in Moses Maimonides
and Eakhr al-Dn al-Raz.
Cox+riiu+ors xxi
David Pingree is a university proIessor at Brown University and chair oI the
Department oI the History oI Mathematics. His interests include the history oI
astronomy, astrology, and astral magic in ancient Mesopotamia, Greece, the
Roman Empire, India, the Islamic world, Byzantium, and medieval Europe, as
well as the transmission oI these sciences Irom one culture to another. Recent
publications include Arabic Astrology in Sanskrit (2001) and Astral Sciences
in Mesopotamia (1999).
Abdelhamid I. Sabra is emeritus proIessor oI the history oI Arabic science
in Harvard`s Department oI the History oI Science. His work in the history oI
science frst Iocused on seventeenth-century Europe, with a book on 1heories
of Light from Descartes to Newton (1967, 1981). He later (aIter 1955) turned
his attention to Arabic/Islamic science, with studies, editions oI primary texts,
and translations. His edition oI Ibn al-Haythams Optics, Books IJJ (comple-
menting his edition and translation oI Books IIII, 1983, 1989) is in press.
1ulio Sams is proIessor oI Arabic and Islamic studies at the University oI
Barcelona. He is a member oI the International Academy oI History oI Science
and Iormer president oI the International Commission on Science in Islamic
Civilization. His main research interests are in the history oI astronomy in
Muslim Spain and in medieval North AIrica. His publications include Las
ciencias de los antiguos en al-Andalus (1992) and Islamic Astronomy and
Medieval Spain (1994).
1acques Sesiano is a lecturer in the history oI mathematics at the Ecole Poly-
technique Eederale de Iausanne. His publications include Un traite medieval
sur les carres magiques (1996) and Une introduction a lhistoire de lalgebre
(1999).
Cox+riiu+ors xxii
I
Cross-Cultural 1ransmission
Eor the last two hundred years the history oI the so-called Hindu-Arabic nu-
merals has been the object oI endless discussions and theories, Irom Michel
Chasles and Alexander von Humboldt to Richard Iemay in our times. But I
shall not here review and discuss all those theories. Moreover I shall discuss
several items connected with the problem and present documentary evidence
that sheds lightor raises more questionson the matter.
At the outset I conIess that I believe the general tradition, which has it
that the nine numerals used in decimal position and using zero Ior an empty
position were received by the Arabs Irom India. All the oriental testimonies
speak in Iavor oI this line oI transmission, beginning Irom Severus Sbkht in
662
1
through the Arabic-Islamic arithmeticians themselves and to Muslim his-
torians and other writers. I do not touch here the problem whether the Indian
system itselI was infuenced, or instigated, by earlier Greek material; at least,
this seems improbable in view oI what we know about Greek number notation.
The time oI the frst Arabic contact with the Hindu numerical system
cannot saIely be fxed. Eor Sbkht (who is known to have translated portions
oI Aristotle`s Organon Irom Persian) Euat Sezgin
2
assumes possible Persian
mediation. The same may hold Ior the Arabs, in the eighth century. Another
possibility is the Indian embassy to the caliph`s court in the early 770s, which
supposedly brought along an Indian astronomical work, which was soon trans-
lated into Arabic. Such Indian astronomical handbooks usually contain chap-
ters on calculation
3
(Ior the practical use oI the parameters contained in the
accompanying astronomical tables), which may have conveyed to the Arabs
the Indian system. In the Iollowing there developed a genre oI Arabic writings
on Hindu reckoning (f l-hisb al-hind, in Iatin de numero Indorum), which
propagated the new system and the operations to be made with it. The oldest
known text oI this kind is the book oI al-Khwrizm (about 820, i.e., around
fIty years or more aIter the frst contact), whose Arabic text seems to be lost,
but which can very well be reconstructed Irom the surviving Iatin adapta-
tions oI a Iatin translation made in Spain in the twelIth century. Similar texts
by al-Uqldis (written in 952/3), Kshyr ibn Iabbn (2nd halI oI the 10th
1
1nr 1r:xsxissiox or Hixiu-Ar:iic
Nuxrr:is Prcoxsiirrri
Paul Kunitzsch
P:ui Kuxi+zscn 4 1nr 1r:xsxissiox or Hixiu-Ar:iic Nuxrr:is Prcoxsiirrri 3
century) and Abd al-Qhir al-Baghdd (died 1037) have survived and have
been edited.
4
All these writings Iollow the same pattern: they start with a des-
cription oI the nine Hindu numerals (called ahruf, plural oI harf; Iatin lit-
terae), oI their Iorms (oI which it is oIten said that some oI them may be written
diIIerently), and oI zero. Then Iollow the chapters on the various operations.
Beside these many more writings oI the same kind were produced,
5
and in later
centuries this tradition was amply continued, both in the Arabic East and West.
All these writings trace the system back to the Indians.
The knowledge oI the new system oI notation and calculation spread
beyond the circles oI the proIessional mathematicians. The historian al-Yaqb
describes it in his 1rkh (written 889)he also mentions zero, sifr, as a small
circle (dira saghra).
6
This was repeated, in short Iorm, by al-Masd in his
Murf.
7
In the Iollowing century the encyclopaedist Muhammad ibnAhmad al-
Khwrizm gave a description oI it in his Mafth al-ulm (around 980); also
he knows the signs Ior zero (asfr, plural) in the Iorm oI small circles (dawir
sighr).
8
That the meaning oI sifr is really 'empty, void has been nicely
proved by August Eischer,
9
who presents a number oI verses Irom old Arabic
poetry, where the word occurs in this sense. It may thus be regarded as beyond
doubt that sifr, in arithmetic, indeed renders the Indian snya, indicating a
decimal place void oI any oI the nine numerals. Exceptional is the case oI the
Fihrist oI Ibn al-Nadm (around 987, that is, contemporaneous with the ency-
clopaedist al-Khwrizm). This otherwise well-inIormed author apparently did
not recognize the true character oI the nine signs as numerals; he treats them
as iI they were letters oI the Indian alphabet.
10
He juxtaposes the nine signs to
the nine frst letters oI the Arabic abfad series and says that, iI one dot is placed
under each oI the nine signs, this corresponds to the Iollowing (abfad) letters
y to sd, and with two dots underneath to the remaining (abfad) letters qf
to z (with some deIect in the manuscript transmission). This sounds as iI he
understood the nine signs and their amplifcation with the dots as letters oI the
Indian alphabet. Even a Koranic scholar, Ab Amr Uthmn al-Dn (in Mus-
lim Spain, died 1053), knows the zero, sifr, and compares it to the common
Arabic orthographic element sukn.
11
(Eor all these authors it must be kept in
mind that the manuscripts in which we have received their texts date Irom more
recent times and thereIore may not reproduce the Iorms oI the fgures in the
original shape once known and written down by the authors.)
OI some interest in this connection are, Iurther, two quotations recorded
by Charles Pellat: the polymath al-Jhiz (died 868) in his Kitb al-muallimn
advised schoolmasters to teach fnger reckoning (hisb al-aqd) instead oI
hisb al-hind, a method needing 'neither spoken word nor writing; and the
historian and literate Muhammad ibn Yahy al-Sl (died 946) wrote in his
Adab al-kuttb: 'The scribes in the administration reIrain, however, Irom using
these |Indian| numerals because they require the use oI materials |writing-
tablets or paper?| and they think that a system which calls Ior no materials and
which a man can use without any instrument apart Irom one oI his limbs is
more appropriate in ensuring secrecy and more in keeping with their dignity;
this system is computation with the joints (aqd or uqad) and tips oI the fngers
(bann), to which they restrict themselves.
12
The oldest specimens oI written numerals in the Arabic East known to
me are the year number 260 Hijra (873/4) in an Egyptian papyrus and the
numerals in MS Paris, BNE ar. 2457, written by the mathematician and astro-
nomer al-Sijz in Shrz between 969 and 972. The number in the papyrus
(fgure 1.1)
13
may indicate the year, but this is not absolutely certain.
14
Eor an
example oI the numerals in the Sijz manuscript, see fgure 1.2. It is to be noted
that here '2 appears in three diIIerent Iorms, one Iorm as common and used
in the Arabic East until today, another Iorm resembling the '2 in some Iatin
manuscripts oI the 12th century, and a Iorm apparently simplifed Irom the lat-
ter; also '3 appears in two diIIerent Iorms, one Iorm as common in the East
and used in that shape until today, and another Iorm again resembling the '3
in some Iatin manuscripts oI the 12th century.
Eigure 1.1
Papyrus PERE 789.
Reproduced Irom Grohmann, Pl. IXV, 12
Eigure 1.2
MS Paris, B. N. ar 2457, Iol 85v. Copied by al-Sijz, Shrz, 969972
This leads to the question oI the shape oI the nine numerals. Still aIter
the year 1000 al-Brn reports that the numerals used in India had a variety
oI shapes and that the Arabs chose among them what appeared to them most
useIul.
15
And al-Nasaw (early eleventh century) in his al-Muqni f l-hisb al-
hind writes at the beginning, when describing the Iorms oI the nine signs, 'Ies
personnes qui se sont occupees de la science du calcul nònt pas ete dàccord
sur une partie des Iormes de ces neuI signes; mais la plupart dèntre elles sont
convenues de les Iormer comme il suit
16
(then Iollow the common Eastern
Arabic Iorms oI the numerals).
Among the early arithmetical writings that are edited al-Baghdd men-
tions that Ior 2, 3, and 8 the Iraqis would use diIIerent Iorms.
17
This seems to be
corroborated by the situation in the Sijz manuscript. Eurther, the Iatin adapta-
tion oI al-Khwrizm`s book says that 5, 6, 7, and 8 may be written diIIerently.
II this sentence belongs to al-Khwrizm`s original text, that would be astonish-
ing. Rather one would be inclined to assume that this is a later addition made
either by Spanish-Muslim redactors oI the Arabic text or by the Iatin translator
or one oI the adapters oI the Iatin translation, because it is in these Iour signs
(or rather, in three oI them) that the Western Arabic numerals diIIer Irom the
Eastern Arabic ones.
18
Another point oI interest connected with Hindu reckoning and the use oI
the nine symbols is: how these were used and in what Iorm the operations were
made. Here the problem oI the calculation board is addressed. It was especially
Solomon Gandz who studied this problem in great detail and who arrived at
the result that the Arabs knew the abacus and that the term ghubr commonly
used in Western Arabic writings on arithmetic renders the Iatin abacus.
19
As
evidence Ior his theory he also cites Irom Ibn al-Nadm`s Fihrist several East-
ern Arabic book titles such as Kitb al-hisb al-hind bi-l-takht (to which is
sometimes added wa-bi-l-ml ), 'Book on Hindu Reckoning with the Board
(and the Stylus). I cannot Iollow Gandz in his argumentation. It is clear, on
the one side, that all the aIorementioned eastern texts on arithmetic, Irom al-
Khwrizm through al-Baghdd, mention the takht (in Iatin: tabula) and that
on it numbers were written andin the course oI the operationswere erased
(mahw, Iatin: delere). It seems that this board was covered with dust (ghubr,
turb) and that marks were made on it with a stylus (ml). But can this sort oI
board, the takht (later also lawh, Iatin tabula), be compared with the abacus
known and used in Christian Spain in the late tenth to the twelIth centuries?
In my opinion, defnitely not. The abacus was a board on which a system oI
vertical lines defned the decimal places and on which calculations were made
by placing counters in the columns required, counters that were inscribed with
caracteres, that is, the nine numerals (in the Western Arabic style) indicating
the number value. The action oI mahw, delere, erasing, cannot be connected
with the technique oI handling the counters. On the other side, the use oI the
takht is unequivocally connected with writing down (and in case oI need, eras-
ing) the numerals; the takht had no decimal divisions like the abacus, it was
a board (covered with fne dust) on which numbers could be Ireely put down
(Ibn al-Ysamn speaks oI naqasha) and eventually erased (mahw, delere).
Thus it appears that the Arabic takht and the operations on it are quite diIIer-
ent Irom the Iatin abacus. Apart Irom the theoretical descriptions in the arith-
metical texts we have an example where an astronomer describes the use oI the
takht in practice: al-Sijz mentions, in his treatise F kayfyat sanat fam al-
asturlbt, how values are to be collected Irom a table and to be added, or sub-
tracted, on the takht.
20
Eurthermore it is worth mentioning that al-Uqldis adds
to his arithmetical work a Book IV on calculating bi-ghayr takht wa-l mahw
bal bi-dawt wa-qirts, 'without board and erasing, but with ink and paper, a
technique, he adds, that nobody else in Baghdad in his time was versant with.
All this shows that the takht, the dust board oI the Arabs, was really used in
practicethough Ior myselI I have some diIfculty to imagine what it looked
likeand that it was basically diIIerent Irom the Iatin abacus.
Iet me add here that the Eastern Arabic Iorms oI the numerals also pene-
trated the European East, in Byzantium. Woepcke has printed Iacsimiles oI the
Arabic numerals appearing in Iour manuscripts oI Maximus Planudes` treatise
on Hindu reckoning, Psephophoria kat Indous.
21
So Iar, at least Ior the Arabic East, matters appear to be reasonably clear.
But now we have to turn to the Arabic West, that is, North AIrica and Muslim
Spain. Here we are conIronted with two major questions, Ior only one oI which
I think an answer is possible, whereas the second cannot saIely be answered Ior
lack oI documentary evidence.
Question number one concerns the notion oI ghubr. This term, meaning
'dust (in reminiscence oI the dust board), is understood by most oI the mod-
ern authorities as the current designation Ior the Western Arabic Iorms oI the
numerals; they usually call them 'ghubr numerals.
It is indeed true that the termghubras Iar as I can seedoes not appear
in book titles on Hindu reckoning or applied to the Hindu-Arabic numerals in
the arithmetical texts oI the early period in the Arabic East. On the contrary, in
the Arabic West we fnd book titles like hisb al-ghubr (on Hindu reckoning)
and terms like hurf al-ghubr or qalam al-ghubr Ior the numerals used in the
Hindu reckoning system. The oldest occurrence so Iar noticed oI the term is in
a commentary on the Sefer Yesira by the Jewish scholar Ab Sahl Dunas ibn
Tamm. He was active in Kairouan and wrote his works in Arabic. This com-
mentary was written in 955/6. In it Dunas says the Iollowing: 'Ies Indiens ont
imagine neuI signes pour marquer les unites. Jài parle suIfsamment de cela
dans un livre que jài compose sur le calcul indien connu sous le nom de hisb
al-ghubr, cèst-a dire calcul du gobar ou calcul de poussiere.
22
The next work to be cited in this connection is the 1alqh al-afkr f
amal rasm al-ghubr by the North AIrican mathematician Ibn al-Ysamn
(died about 1204). Two pages Irom this text were published in Iacsimile
in 1973;
23
on page 8 oI the manuscript ( page 232 in the publication) the
author presents the nine signs (ashkl) oI the numerals which are called ash-
kl al-ghubr, 'dust fgures; at frst they are written in their Western Ara-
bic Iorm, then the author goes on: wa-qad taknu aydan hkadh |here
Iollow the Eastern Arabic Iorms| wa-lkinna l-ns indan al l-wad al-
awwal, 'they may also look like this . . . , but people in our |area| Iollow the
frst type. (It should be noted that the manuscript here reproducedRabat
K 222is in Eastern naskh and oI a later date.) Another testimony is Iound
in Sid al-Andalus`s Tabaqt al-umam (written about 1068 in Spain). In
praising Indian achievements in the sciences this author writes: wa-mimm
wasala ilayn min ulmihim f l-adad hisb al-ghubr alladh bassatahu
Ab Jafar Muhammad ibn Ms al-Khwri:m etc.,
24
'And among what has
come down to us oI their sciences oI numbers is the hisb al-ghubr |dust
reckoning| which . . . al-Khwrizm has described at length. It is the shortest
|Iorm oI| calculation . . . , etc. This paragraph was later reproduced by Ibn
al-QiIt in his 1rkh al-hukam (probably written in the 1230s), but here the
most interesting words oI Sid`s text were shortened; in Ibn al-QiIt it merely
reads: wa-mimm wasala ilayn min ulmihim hisb al-adad alladh . . . ,
'And among what has come down to us oI their sciences is the hisb al-adad
|calculation oI numbers| which al-Khwrizm . . . etc.
25
Erom these testimonies it is clear that in the Arabic West since the mid-
dle oI the tenth century the system oI Hindu reckoning as such was called 'dust
reckoning, hisb al-ghubrcertainly in reminiscence oI what the eastern
arithmetical texts mentioned about the use oI the takht, the dust board. It will
then Iurther be clear that the terms hurf al-ghubr or qalam al-ghubr (dust
letters or symbols) Ior the nine signs oI the numerals used in this system oI cal-
culation basically described the written numerals as such, without specifcation
oI their Eastern or Western Arabic Iorms. This is corroborated by some known
texts that put the hurf al-ghubr, written numerals, in opposition to the num-
bers used in other reckoning systems that had no written symbols, such as fnger
reckoning and mental reckoning. In Iavor oI this interpretation may be quoted
some oI the texts frst produced by Woepcke. One supporting element here is
what Woepcke derives Irom the Kashf al-asrr |or: al-astr| an ilm |or: wad|
al-ghubr oI al-Qalasd (in Muslim Spain, died 1486).
26
Eurther, in Woepcke`s
translation oI a treatise by Muhammad Sibt al-Mridn (muwaqqit in Cairo,
died 1527), where the author cites words Irom the Kashf al-haqiq f hisb
al-daraf wa-l-daqiq oI the Cairene astronomer Shihb al-Dn Ibn al-Majd
(died 1447), we read (oI Ibn al-Majd), 'Cependant (Chehab Eddn), . . . , sèst
etendu dans lòuvrage cite sur lèxposition de la methode des (mathematici-
ens des temps) anterieurs, en Iait du maftouh et du gobar.
27
Here the two sys-
tems, hisb mafth (mental reckoning) and hisb al-ghubr (Hindu reckoning,
with written numerals), are clearly set apart. In another paper Woepcke gave the
translation oI a treatise Introduction au calcul gobari et hawa (without men-
tioning an author or the shelI-mark oI the manuscript) where, again, the 'cal-
cul gobari (Hindu reckoning, with written numerals) is opposed to the 'calcul
hawa (mental reckoning, i.e., without the use oI written symbols).
28
Erom these testimonies it can be derived that the written numerals in the
Hindu reckoning system were called al-hurf al-tisa (the nine letters, or lit-
terae) or, in Mafth al-ulm, al-suwar al-tis (the nine fgures) or ashkl al-
ghubr (dust fgures, in Ibn al-Ysamn) and hurf or qalam al-ghubr (dust
letters) by other Western Arabic authors. The designation thus reIers to the
written numerals as such, as opposed to numbers in other reckoning systems
that did not use written symbols. I should think that, thereIore, it is no longer
justifed Ior us to call the Western Arabic Iorms oI the Hindu-Arabic numer-
als 'ghubr numerals. Rather we should speak oI the Eastern and the Western
Arabic Iorms oI the nine numerals.
The second, most diIfcult, question in connection with the Arabic
West concerns the Iorms oI the written numerals in that area, their origin and
their relationship with the 'Arabic numerals that came to be used in Iatin
Europe.
Here one might ask why the Arabic West developed Iorms oI the num-
erals diIIerent Irom those in the East. It is hard to imagine a reason Ior this
development, especially when we assumein conIormity with our under-
standing oI the birth and growth oI the sciences in the Maghrib and al-Andalus
in generalthat the Hindu reckoning system came to the West like so many
texts and so much knowledge Irom the Arabic East. About the mathematician
and astronomer Maslamain Spain, died 1007/1008Ior example we learn
Irom Sid al-Andalus
29
that he studied the Almagest, that he wrote an abbre-
viation oI al-Battn`s Zf and that he revised al-Khwrizm`s Zf (this work has
survived in a Iatin translation by Adelard oI Bath and has been edited); he also
knew the Arabic version oI Ptolemy`s Planisphaerium and wrote notes and
additions to it that survive in Arabic and in several Iatin translations.
30
Thus
he, or his disciples, will certainly also have known al-Khwrizm`s Arithmetic
and, together with it, the EasternArabic Iorms oI the numerals. Not quite a cen-
tury later Sid al-Andalus knows oI al-Khwrizm`s Arithmetic under the title
hisb al-ghubr, as we have just heard.
Certainly, in this connection one has to consider that also some more
elements oI basic Arabic erudition took a development in the West diIIerent
Irom that in the Arabic East: frst, the script as suchwe think oI the so-called
Maghrebi ductus in which, beyond the general diIIerence in style, the letters
f and qf have their points added diIIerently; second, the sequence oI the let-
ters in the ordinary alphabet; and, third, the sequence oI the letters in the abfad
series where the West deviates Irom the old Semitic sequence that was retained
in the East and assigns to several letters diIIerent number values.
31
As Iar as I
can see, linguists have also not brought Iorward plausible arguments Ior these
diIIerences.
That the Eastern Arabic numerals were also known in al-Andalus is dem-
onstrated by several Iatin manuscripts that clearly show the Eastern Iorms, Ior
example, MSS Dresden C 80 (2nd halI 15th century), Iols. 156v157r; Berlin,
Iol. 307 (end oI 12th century), Iols. 6, 9, 10, and 28; OxIord, Bodleian Iibrary,
Selden sup. 26; Vatican, Palat. lat. 1393; and Munich, Clm 18927, Iol. 1r,
where the Eastern fgures are called indice hgure, whereas the Western Iorms
are labeled toletane hgure;
32
the zeros are here called cifre.
However that may be, the evidence Ior the Western Arabic numerals in
Iatin sources begins in 976; a manuscriptthe 'Codex Vigilanuswritten in
that year and containing Isidor`s Etymologiae has an inserted addition on the
genius oI the Indians and their nine numerals, which are also written down in
the Arabic way, that is, proceeding Irom right to leIt, in Western Arabic Iorms.
33
The same was repeated in another Isidor manuscript, the 'Codex Emilianus,
written in 992.
34
HereaIter Iollow, in Iatin, the 'apices, the numeral nota-
tions on abacus counters, which render similar Iorms oI the numerals.
35
Here,
the Western Arabic Iorms are still drawn in a very rough and clumsy man-
ner. A third impulse came in the twelIth century with the translation oI al-
Khwrizm`s Arithmetic; Irom now on the Iorms oI the numerals become
smoother and more elegant.
36
UnIortunately, the documentary evidence on the side oI Western Arab-
ic numerals is extremely poor. So Iar, the oldest specimen oI Western Arabic
numerals that became known to me occurs in an anonymous treatise on auto-
matic water-wheels and similar devices in MS Elorence, Or. 152, Iols. 82r and
86r (the latter number also appears on Iol. 81v). Two other texts in this section
oI the manuscript are dated to 1265 and 1266, respectively (fgures 1.3ab).
37
Here we have the symbols Ior 1, 2, 3, 4, 5, 8, and 9. The fgures Ior 2 and 3
look like the corresponding Eastern Arabic Iorms and are not turned by 90 as
in other, more recent, Maghrebi documents. The meaning oI these numerals
in the present context remains unexplained to me. The numerals in two other
Maghrebi manuscripts that Iell into my hands (fgures 1.41.5)
38
resemble the
Iorms Iound in the specimens reproduced in Iacsimile by IabartaBarcelo
Irom Arabic documents in Aragon and Valencia Irom the 15th and 16th cen-
turies.
39
While specimens oI Western Arabic numerals Irom the early period
the tenth to thirteenth centuriesare still not available, we know at least that
Hindu reckoning (called hisb al-ghubr) was known in the West Irom the
tenth century onward: Dunas ibn Tamm, 955/6; al-Dn, beIore 1053; Sid
al-Andalus, 1068; Ibn al-Ysamn, 2nd halI oI the 12th century. It must be
regarded as natural that, together with the reckoning system, also the nine
numerals became known in the Arabic West. It thereIore seems out oI place
to adopt other theories Ior the origin oI the Western Arabic numerals. Erom
among the various deviant theories I here mention only two. One theory, also
repeated by Woepcke,
40
maintains that the Arabs in the West received their
numerals Irom the Europeans in Spain, who in turn had received them Irom
Alexandria through the 'Neopythagoreans and Boethius; to Alexandria they
had come Irom India. Since Eolkerts`s edition oI and research on the Pseudo-
Boethius
41
we now know that the texts running under his name and carrying
Arabic numerals date Irom the eleventh century. Thus the assumed way oI
transmission IromAlexandria to Spain is impossible and this theory can no lon-
ger be taken as serious. Recently, Richard Iemay had brought Iorward another
theory.
42
He proposes that, in the series oI the Western Arabic numerals, the 5,
Eigure 1.3a
MS Elorence, Or. 152, Iol. 82r
(dated 12651266)
Eigure 1.3b
MS Elorence, Or. 152, Iol. 86r
(dated 12651266)
6, and 8 are derived Irom Iatin models, 5 as rendering the Visigothic Iorm oI
the Roman v, 6 as a ligature oI vi in the same style, and 8 as the o oI octo with
the fnal o placed above. This might appear acceptable Ior the Arabic numer-
als used in Iatin texts. But since the Western Arabic numerals are oI the same
shape, that would mean that the Western Arabs broke up their series oI nine
numerals and replaced their 5, 6, and 8 by Iorms taken Irom European sources.
This seems highly improbable. The Western Arabs received their numerals
Irom the East as a closed, complete, system oI nine signs, and it would only
appear natural that they continued to use it in this complete Iorm, not breaking
the series up and replacing single elements by Ioreign letters.
When one compares the Eastern and the Western Arabic Iorms oI the
numerals, one fnds that they are not completely diIIerent. The Western Iorms
oI 1, 2, 3, 4, 5, and 9 can be recognized as being related to, or derived Irom,
the corresponding Eastern Iorms. Major diIfculty arises with 6, 7, and 8. It
may not be accidental that the oldest existing Iatin re-working made Irom
the translation oI al-Khwrizm`s Arithmetic mentions just these three fgures
(plus 5) as being diIIerently written.
43
As I have already said earlier, this notice
can hardly stem Irom al-Khwrizm himselI; rather it may have been added by
a Spanish-Arabic redactor oI al-Khwrizm`s text. He would have been best
equipped to recognize this diIIerence. The Iatin translator, or Iatin adapters,
would less probably have been able to notice the diIIerence between the East-
ern and Western Arabic Iorms oI these Iour numerals. We cannot explain why,
and how, the three Western fgures were Iormed, especially since we have no
Eigure 1.4
Rabat, al-Khizna al-mma, MS 321, p. 45 (aIter 1284)
Eigure 1.5
MS Ait Ayache, p. 192 (aIter 1344)
written specimens oI Western Arabic numerals beIore the thirteenth century.
Eor Iurther research into the matter, thereIore, the discovery oI older, or old,
documents remains a most urgent desideratum.
Iastly, I want to mention a curious piece oI evidence. Somebody in the
Arabic West once Iound out that the Western Arabic Iorms oI the nine num-
erals resemble certain letters in the Maghrebi script and he organized their
description in a poem oI three memorial verses (in the metre kmil). The poem
is reported by the Spanish-Arabic mathematician al-Qalasd (died 1486) in
a commentary on the 1alkhs f amal al-hisb oI Ibn al-Bann (died 1321 or
1324) and, aIterwards, by Husayn ibn Muhammad al-Mahall al-Shf (died
1756, an Eastern Arabic author) in a commentary on an arithmetical work oI
al-Sakhw (died aIter 1592, also an Eastern author). The two loci are cited by
Woepcke.
44
The text oI the poem is as Iollows:
alifun wa-yun thumma hiffun |wa-| badahu
(
awwun wa-bada l-awwi aynun tursamu
hun wa-bada l-hi shaklun zhirun
(
yabd ka-l-khutth idh huwa yursamu
sifrni thminuh wa-alifun baynah|um|
(
wa-l-wwu tsiuh bi-dhlika yukhtamu
That is, 1 is compared to an alif, 2 to a fnal y (but to h in al-Mahall;
both comparisons are possible), 3 to the combination h-fm, 4 to the combina-
tion ayn-ww, 5 to ayn, 6 to (an isolated) h, 7 to a khuttf (i.e., an iron hook),
8 to two zeros above each other and linked by a stroke, and 9 to a ww. These
memorial verses may be much older than al-Qalasd`s time. They seem to have
become a topic since they are cited even by an Eastern Arabic author. Perhaps
one can conclude Irom this standardized description that the written Iorms oI
the Western Arabic numerals were less variable than the Eastern ones.
To sum up, we can register that the history oI the transmission oI the
Hindu numerals and Hindu reckoning to the Arabs in the East appears to be
clear. Eor the Arabic West it is known that all the cultural and scientifc achieve-
ments oI the East were transIerred there. In the stream oI this cultural move-
ment the knowledge oI Hindu reckoning and the nine numerals must also have
passed there. The oldest known testimony Ior the acquaintance with the Hindu
system is documented Ior 955/6 in Kairouan. So Iar no written evidence oI
Western Arabic numerals Ior the tenth to the thirteenth centuries have been
Iound; documents are only known Irom the thirteenth century on. But these
numerals must have existed earlier since the frst evidence in Iatin sources
which took up these numerals Irom the Arabs in Spaindates Irom 976. The
most important task Ior Iurther research would thereIore be to fnd older West-
ern Arabic material Ior the knowledge and use oI the Hindu numerals in that
region.
Arrrxiix
An inspection oI microflms oI the manuscripts oI Ieonardo oI Pisa`s Liber abaci
(AD 1202) shows that a group oI older manuscripts has numerals similar in shape to
those in the NewYork MS oI al-Khwrizm`s Arithmetic as visible in the Iacsimiles
oI its recent edition (Eolkerts 1997): MSS Elorence, BN, Conv. Sopp. C.1.2616
(beg. 14c.? Here the series oI the nine symbols, at the beginning oI the text, looks
diIIerent, more 'modern; but in the text itselI and in the diagrams and tables etc.,
they are oI the Khwrizm-MS N-type. This manuscript was used by Boncompagni
Ior his edition, 18571862); Siena, Bibl. Publ. Comm., I.IV.20 (2nd halI 13c.);
Elorence, Magliabecchi XI, 21.
On the other hand, more recent, 'modern(ized), Iorms oI the numerals are
used in MSS Elorence, BN II.III.25 (16c.); Vat. Palat. 1343 (end 13c.?); Milan, I.
72 (15c.?). It thus appears evident that the numerals in the Ieonardo manuscripts
Iollow the Iorms current in the known Iatin arithmetical texts. Contrary to what is
sometimes assumed, they do not show the intrusion oI newArabic infuence result-
ing Iro Ieonardo`s oriental travels and his personal contacts with trade centers in
the Arab world.
Pos+scrir+
Eor the Maghribi manuscript Ait Ayache, Hamzawya 80, quoted in this article, it is
now established that it was copied shortly aIter AD 1600; see the detailed descrip-
tion by Ahmad Alkuwaif and Monica Rius, 'Descripcion del Ms. 80 de Al-Zwiya
al-Hamzawya, Al-Qantara 19 (1998), 445463. ThereIore the manuscript can no
longer serve as a testimony to early Iorms oI Western Arabic numerals.
No+rs
1. See Nau.
2. Sezgin V, 211.
3. See al-Brn, India, ch. 14, apud Woepcke 1863, 475I. (note 1), sub 13, 19 and 24
( repr. II, 407I.).
4. Al-Uqldis: Saidan 1973 and 1978; Kshyr: Ievey-Petruck; al-Baghdd: Saidan
1985.
5. About fIteen such titles up to the middle oI the eleventh century are quoted by Sez-
gin, V.
6. Al-Yaqb I, 93; cI. Kbert 1975, 111.
7. Al-Masd I, p. 85 (152).
8. Al-Khwrizm, 193195.
9. Eischer, 783793.
10. Fihrist, I, 18I.; cI. Kbert 1978.
11. Eischer, 792; Kbert 1975, 111.
12. Pellat, 466b.
13. Grohmann, 453I., no. 12, and Plate IXV, 12.
14. ProI. W. Diem, Cologne, who has studied and edited such papyri Ior many years,
inIorms me (in a letter dated 6 August 1996) that the understanding oI the symbols as
a year number is not Iree Irom doubt, because an expression like f sanat ('in the year
. . .), which is usually added to such datings, is here missing. Eurthermore, he con-
frmed that a second dating oI that type in another papyrus, understood by Karabacek,
13 (no. 8), as the Hindu numerals 275 (888/9), is not Iormed by Hindu numerals, but
rather by (cursive) Greek numeral letters. This document, thereIore, must no longer
be regarded as the second oldest occurrence oI Hindu-Arabic numerals in an Arabic
document.
15. See the quotation by Woepcke 1863, 275I. ( repr. II, 358I.).
16. Translated by Woepcke 1863, 496 ( repr. II, 428).
17. Saidan 1985, 33.
18. CI. on this also Woepcke 1863, 482I. ( repr. II, 414I.).
19. Gandz 1927 and 1931.
20. It is in 2 oI the treatise. I owe this inIormation to Richard Iorch. Dr. Iorch is pre-
paring an edition oI al-Sijz`s text.
21. Woepcke 1859, 27, note *** ( repr. II, 191).
22. Eirst cited by Joseph Reinaud in an Addition to his 'Memoire sur lÌnde, 565, Irom
one oI the Iour Hebrew translations that were made Irom Dunas`s original Arabic text,
which itselI has survived only in part.
23. Ibn al-Ysamn, 232I.; a German translation was given by Kbert 1975, 109111.
24. Sid al-Andalus, 58.
25. Ibn al-QiIt, 266, ult.267,3.
26. Woepcke 1854, 359, sub 3 ( repr. I, 456).
27. Woepcke 1859, 67 ( repr. II, 231).
28. Woepcke 186566, 365 ( repr. II, 541).
29. Sid al-Andalus, 169.
30. Edited by Kunitzsch-Iorch.
31. CI. Grundriss, 176II., 181I., 182I.
32. Eor Selden and Pal. lat., cI. the table in Allard, 252; Ior Clm 18927, cI. Iemay 1977,
fgure 1a.
33. See the reproduction in van der Waerden-Eolkerts, 54.
34. Reproduced also in van der Waerden-Eolkerts, 55.
35. Eor reproductions, see, inter alios, van der Waerden-Eolkerts, 58; TropIke, 67; Eolk-
erts 1970, plates 121.
36. See the photographs in Eolkerts 1997, plate 1. etc., Irom the newly Iound and so
Iar oldest known manuscript oI a re-working oI the Iatin translation oI al-Khwrizm`s
Arithmetic.
37. I owe the knowledge oI this manuscript to the kind help oI Dr. S. Brentjes, Berlin,
which is grateIully acknowledged. A detailed description oI the manuscript was given
by Sabra 1977.
38. Rabat, al-Khizna al-mma, MS 321, p. 45. The preceding text, ending on p. 44,
is dated in the colophon to 683/1284. P. 45 was leIt blank by the original writer; a later
hand added in the upper part an alchemical prescription and at the bottom a magic
square with directions Ior its use. I am grateIul to ProI. R. Degen, Munich, Ior bring-
ing this page to my attention, and to ProI. B. A. Alaoui, Ees, and M. A. Essaouri, Rabat,
Ior procuring copies oI the relevant pages Irom the manuscript.Morocco, Ait Ayache,
MS Hamzawya 80. On p. 201 oI the manuscript, in an excerpt Irom the Zf oI Ibn
Azzz al-Qusantn, there is a calculated example Ior JulyAugust 1344 (cI. Kunitzsch
1994, p. 161; 1997, p. 180).
39. It should be added that in the table oI ghubr numerals given by Souissi, 468, the
numerals in the frst two lines (said to date Irom the 10th century and ca. 950, respec-
tively) are not (Arabic) ghubr numerals, but rather Indian numerals (cI. Sanchez Perez,
the table on p. 76, lines 89). It should also be noted that the date given by Sanchez
Perez, 121, table 1, Ior the specimen in line 9 ('Ao 1020) is the Hijra year ( AD
1611/12); the author there mentioned, Ibn al-Qd, died in Ees 1025/1616. Similarly,
the specimen in line 12, ibid., Irom MS Escorial 1952, must belong to the 11th century
Hijra; the manuscript contains a commentary by Abu `l-Abbs ibn SaIwn on the sum-
mary oI Mlik ibn Anas` al-Muwatta oI Abu `l-Qsim al-Qurash.
40. Woepcke 1863, 239 ( repr. II, 322).
41. Eolkerts 1968, 1970.
42. Iemay 1977 and 1982.
43. Eolkerts 1997, 28: MS N, lines 3438.
44. Woepcke 1863, 60I. and 64I. ( repr. II, 297I. and 301I.).
Biiiior:rn.
Allard, A. 1992. Al-Khwri:m, Le calcul indien. Paris and Namur.
Eischer, A. 1903. 'Zur Berichtigung einer Etymologie K. Vollers`. Zeitschrift der
Deutschen Morgenlandischen Gesellschaft 57, 783793.
Eolkerts, M. 1968. 'Das Problem der pseudo-boethischen Geometrie. Sudhoffs Archiv
52, 152161.
Eolkerts, M. 1970. Boethius Geometrie. Ein mathematisches Lehrbuch des Mittel-
alters. Wiesbaden.
Eolkerts, M. 1997. Die alteste lateinische Schrift uber das indische Rechnen nach al-
(wri:m. Munich.
Gandz, S. 1927. 'Did the Arabs know the abacus? American Mathematical Monthly
34, 308316.
Gandz, S. 1931. 'The Origin oI the Ghubr Numerals or The Arabian Abacus and the
Articuli. Isis 16, 393424.
Grohmann, A. 1935. 'Texte zur WirtschaItsgeschichte gyptens in arabischer Zeit.
Archiv Orientalni 7, 437472.
Grundriss. 1982. Grundriss der arabischen Philologie, ed. W. Eischer, I: Sprachwissen-
schaft. Wiesbaden.
Ibn al-Nadm. 18711872. Kitb al-hhrist, ed. G. Elgel, III. Ieipzig.
Ibn al-QiIt. 1903. 1ar al-hukam, ed. A. Mller and J. Iippert. Ieipzig.
Ibn al-Ysamn, Ab Eris. 1973. 'Dall jadd al urbat al-arqm al-mustamala I
l-maghrib al-arab. Al-Lisn al-Arab 10, 231233.
Karabacek, J. 1897. 'Aegyptische Urkunden aus den kniglichen Museen zu Berlin.
Wiener Zeitschrift fur die Kunde des Morgenlandes 11, 121.
al-Khwrizm, Muhammad ibn Ahmad. 1895. Liber Mafatih al-olum, ed. G. van
Vloten. Ieiden.
Kbert, R. 1975. 'Zum Prinzip der gurb-Zahlen |sic pro gubr| und damit unseres
Zahlensystems. Orientalia 44, 108112.
Kbert, R. 1978. 'Ein Kuriosum in Ibn an-Nadm`s berhmtem Fihrist. Orientalia 47,
112I.
Kunitzsch, P. 1994/1997. 'Abd al-Malik ibn Habb`s Book on the Stars. Zeitschrift fur
Geschichte der Arabisch-Islamischen Wissenschaften 9, 161194; 11, 179188.
Kunitzsch, P., and Iorch, R. 1994. Maslamas Notes on Ptolemys Planisphaerium and
Related 1exts. Munich.
Iabarta, A, and Barcelo, C. 1988. Numeros y cifras en los documentos arabigohispanos.
Cordoba.
Iemay, R. 1977. 'The Hispanic Origin oI Our Present Numeral Eorms, Jiator 8, 435
462 (fgures 1a11).
Iemay, R. 1982. 'Arabic Numerals. Dictionary of the Middle Ages, ed. J. R. Strayer,
vol. 1. NewYork, 382398.
Ievey, M., and Petruck, M. 1965. Kshyr ibn Labbn, Principles of Hindu Reckoning.
Madison and Milwaukee.
al-Masd. 1965II. Murf al-dhahab, ed. C. Pellat, III. Beirut.
Nau, E. 1910. 'Notes dàstronomie syrienne. III: 'Ia plus ancienne mentionne orien-
tale des chiIIres indiens. Journal asiatique, ser. 10, 16, 225227.
Pellat, C. 1979. 'Hisb al-Aqd. Encyclopaedia of Islam, new ed., III. Ieiden, 466
468.
Reinaud, J. 1855. 'Addition au Memoire sur lÌnde. Memoires de lInstitut Imperial de
France, Academie des Inscriptions et Belles-Lettres 18, 565I.
Sabra, A. I. 1977. 'A Note on Codex Biblioteca Medicea-Iaurenziana Or. 152. Journal
for the History of Arabic Science 1, 276283.
Sabra, A. I. 1979. 'Ilm al-Hisb. Encyclopaedia of Islam, new ed., III. Ieiden, 1138
1141.
Sid al-Andalus. 1985. Tabaqt al-umam, ed. H. B-Alwn. Beirut.
Saidan, A. S. 1973. al-Uqldis, al-Fusl f l-hisb al-hind. Amman.
Saidan, A. S. 1978. 1he Arithmetic of al-Uqldis. Dordrecht and Boston.
Saidan, A. S. 1985. Abd al-Qhir ibn Thir al-Baghdd, al-1akmila f l-hisb. Kuwait.
Sanchez Perez, J. A. 1949. La aritmetica en Roma, en India y en Arabia. Madrid and
Granada.
Sezgin, E. 1974. Geschichte des arabischen Schrifttums, V: Mathematik, bis ca. 430 H.
Ieiden.
Souissi, M. 1979. 'Hisb al-Ghubr. Encyclopaedia of Islam, new ed., III. Ieiden,
468I.
TropIke, J. 1980. Geschichte der Elementarmathematik, vol. 1, 4th ed. Berlin and New
York.
van der Waerden, B. I., and Eolkerts, M. 1976. Written Numbers. The Open University
Press, Walton Hall, Milton Keynes, GB.
Woepcke, E. 1854. 'Recherches sur l`histoire des sciences mathematiques chez les ori-
entaux. . . . Journal asiatique, 5
e
serie, 4, 348384 ( repr. I, 445481).
Woepcke, E. 1859, Sur lintroduction de larithmetique indienne en Occident. . . . Rome
( repr. II, 166236).
Woepcke, E. 1863. 'Memoire sur la propagation des chiIIres indiens. Journal asi-
atique, 6
e
serie, 1, 2779; 234 290; 442529 ( repr. II, 264461).
Woepcke, E. 18651866. 'Introduction au calcul gobr et haw. Atti dellAccademia
deNuovi Lincei 19, 365383 ( repr. II, 541559).
Woepcke, E. 1986 (repr.). Etudes sur les mathematiques arabo-islamiques, Nachdruck
von Schriften aus den Jahren 18421874, ed. E. Sezgin, III, ErankIurt am Main.
al-Yaqb. 1883. Historiae, ed. M. Th. Houtsma, III. Ieiden.
This chapter considers a group oI Iatin astronomical texts translated IromAra-
bic, or based on Arabic material, which share the same technical language and
have the same systems oI numeration, and explores the possibility that, unlike
the majority oI Arabic scientifc works, which entered Europe Irom Spain,
these came directly Irom the East and were brought to the West by scholars
working in Antioch and Pisa. The texts under consideration here are a transla-
tion oI Ptolemy`s Almagest, known as the 'Dresden Almagest, a Iatin cosmol-
ogy describing the Ptolemaic system, called the Liber Mamonis, and a version
oI the astronomical tables oI al-SI. All these texts appear to have been written
in the second quarter oI the twelIth century.
I 1nr Drrsirx LMLl :xi LILLP MMCÎ
It was Charles Homer Haskins, whose intuitions usually prove remarkably
accurate, who frst pointed out a Ieature which could link the Iatin version oI
the Almagest surviving uniquely in MS Dresden, Iandesbibliothek, Db. 87,
and the Liber Mamonis, a cosmology oI which the only known copy is the
incomplete text in the twelIth-century MS Cambrai, Bibliotheque municipale,
930.
1
The frst section oI this article confrms the link between the two works,
and suggests that the Dresden Almagest represents the frst attempt to translate
the Almagest into Iatin in the Middle Ages, whilst the Liber Mamonis, in turn,
is an early attempt at replacing a cosmology based on Iatin sources with the
Ptolemaic system.
2
MS Dresden, Db. 87 was written in ca. 1300;
3
it once belonged to Ber-
thold oI Moosberg, and became the property oI the Dominicans oI Cologne.
It consists oI several texts related to Ptolemy`s Almagest: Geminus oI Rhodes`s
Introduction to the Phenomena (oIten called 'Introductio Ptolemei in Alma-
gestum), the Parvum Almagestum, and Jbir ibn Afah`s Correction of the
Almagest. The frst item, however, is a unique copy oI a translation oI the frst
Iour books oI Ptolemy`s Almagest. It was frst brought to the notice oI schol-
ars by J. I. Heiberg, who quoted the incipits and explicits oI the Iour books.
4
2
1nr 1r:xsxissiox or Ar:iic As+roxox.
vi: Ax+iocn :xi Pis: ix +nr Srcoxi Qu:r+rr
or +nr 1vrir+n Crx+ur.
Charles Burnett
Cn:rirs Burxr++ 24 1nr 1r:xsxissiox or Ar:iic As+roxox. vi: Ax+iocn :xi Pis: 23
Haskins, in turn, transcribed the list oI titles Ior the frst book and the prologue.
Both scholars considered that the translation was made Irom Greek, on the
grounds oI its vocabulary. However, the Arabic name that appears in the explic-
its to each book already makes one uneasy about a Greek origin: Ior example,
'Here ends the frst book oI the mathematical treatise oI Ptolemy which is
called the Megali Xintaxis oI astronomy, in the translation and> dictation by
the love oI languages (?) oI Wittomensis Ebdelmessie.`
5
The interpretation oI the name, and the precise role oI this 'Abd al-
Mash oI Winchester is unclear. But a comparison oI the text with the Greek
and Arabic versions oI the Almagest reveals clearly that an Arabic text lies at
the base oI this version:
6
1. The word order and terminology oI the frst phrase corresponds to the Ishq/
Thbit version oI the Arabic Almagest.
7
2. Most oI the diagrams are reversed in respect to the Greek; this may have arisen
out oI a mistaken notion oI a translator IromArabic that, since he had to reverse the
direction oI the script, he also had to reverse the diagrams.
3. The terminology is based on Arabic rather than Greek.
8
4. Several turns oI phrase are reminiscent oI other translations IromArabic.
9
Heiberg and Haskins were misled by the translator`s total avoidance oI
transcriptions IromArabic, and by his addition oI a veneer oI Greek or Pseudo-
Greek terms. That this is a veneer is immediately obvious when one looks Ior
these Greek terms in the Greek text oI the Almagest; Ior in most cases they
are simply not there. Where the Dresden Almagest has praxis` (Iol.1r), the
Greek text has apoxtixov` (Heiberg, 4.9), where it has phisialoica` (Iol.1r),
the Greek has uoixov` (Heiberg, 5.9), where it has organum` (Iol. 3r), the
Greek has xotooxruo=` (Heiberg, 13.12), and so on. To the category oI a
Greek veneer may also be ascribed the word occurring in the explicits oI each
book: philophonia` (love oI languages`?), which is meant, no doubt, to sound
stylish, but makes little sense. Nevertheless, it is possible that the translator
attempted to look at a Greek manuscript, or at least consulted someone who
knew Greek. Otherwise it is diIfcult to explain how he came to use the nonce-
word aretius` where the Greek text has oprtmv.`
10
Moreover, he transcribed
quite accurately Greek proper names, and the names oI Egyptian months,
11
as well as the Greek title oI Ptolemy`s work: megali xintaxis` (i.e., ryop
uvtoi=). It is clear that the translator wished to give the impression that he
had taken the whole text Irom the Greek.
Whether the translation is literal can only be ascertained through a close
comparison with the extant Arabic versions oI the Almagest, which (except Ior
the star-tables oI Books 7 and 8) are not yet edited. When one compares the
text with the Greek Almagest one fnds that
1. Although there are reIerences to tables,
12
none oI Ptolemy`s tables has been
copied into the manuscript. It is diIfcult to tell whether they were originally
included.
13
2. In the theorems on trigonometry and spherical astronomy in Books 1 and 2, reI-
erences to the relevant theorems in Euclid`s Elements have been added. That this is
an addition in the original Arabic text is suggested by the presence oI exactly the
same Iorms oI reIerence to the Elements in the text oI al-Nasaw (see item 6 below).
But it is worth noting that here again the translator reIers to Euclid`s work either
with a transcription oI the Greek titlee.g., liber estichie ( otoirio) euclitis,`
Iol. 7vor a literal translation oI that title: liber elementorum euclithis` (Iol. 6v),
14
and not as liber geometrie`/liber institutionis artis geometrie` vel sim. which is
Iound in most oI the translations oI the Elements made IromArabic.
3. OI similar status are the cross-reIerences to other theorems in the Almagest: e.g.,
Iol. 17r (Almagest, II.3): demonstracione .xiiii. fgure primi sermonis huius libri`;
this cross-reIerence is not in Ptolemy`s text, but such cross-reIerences also occur
in al-Nasaw`s work.
4. Sometimes only the geometrical elements oI a theoremare given, and the numeri-
cal values have been omitted; e.g., in Book II, chapters 2 and 3.
5. Sometimes a more precise value replaces Ptolemy`s rounded` value: e.g., Iol.
15r: 11; 39, 59`` where Ptolemy has approximately 11; 40`` (I, 14, Heiberg,
78.12).
6. The only substantial addition vis-a-vis the Greek text is that oI several theorems
on the sector-fgure appended to the end oI Almagest, I, chapter 13 (Dresden MS,
Iols 13v14v). As has been pointed out by Richard Iorch, this passage corresponds
to a section oI chapter 2 oI al-Ishb f sharh al-shakl al-qatt by the eleventh-
century Arabic mathematician, al-Nasaw, who served the Byid amrs in Bagh-
dad.
15
The verbal equivalence between the Dresden Almagest and this passage oI
al-Nasaw suggests that the translator was translating literally Irom an Arabic
text.
16
On the whole, aside Irom the tables, little appears to be missing in the
frst Iour books oI the Almagest.
17
There is nothing, either, to indicate that the
translation stopped aIter the Iourth book, rather than that the scribe oI the Dres-
den manuscript decided at this point not to copy any more oI the translation.
There are two characteristic Ieatures which separate this translation Irom
other Iatin translations oI the Almagest (whether Irom the Arabic or Irom the
Greek): the terminology, and the notation Ior numerals. It was the notation Ior
numerals that attracted the attention oI Haskins, and this needs more careIul
analysis.
The translator
18
began by using roman numerals only. Then, on Iol. 15v
(in Almagest, I, chapter 16), he makes his frst attempt to use alphanumeri-
cal notation. That alphanumerical notation is being introduced at this point is
indicated by the Iact that here, and here only, the numerical values oI the letters
oI the alphabetic are spelt out:
quinquaginta I sex partium et unius a sexagenarie et viginti e quinque
secund-arium ad centum t et viginti k partes (the ratio oI> fIty f (six)
degrees and a (one) minute and twenty e (fve) seconds to t (one hundred)
and k (twenty)
19
degrees`)
ThereaIter, alphanumerical notation is used, sometimes on its own,
sometimes in combination with roman numerals, and sometimes alternating
with them. One can construct the Iollowing key Ior the numerical values oI
the letters:
1 a 10 k 100 t 1000 a mille
2 b 20 (l) 200 u 2000 b milia
3 c 30 (m) 300 x 3000 c milia
4 d 40 (n) 400 y 4000 d milia
5 e 50 (o) 500 z etc.
6 I 60 (p) 600 O
7 g 70 (q) 700
8 h 80 (r) 800 .n.
20
9 i 90 (s) 900 |O|
21
On the whole, in comparison with the Greek text, the numbers in the
Dresden Almagest are accurate, whether they are written in roman numerals or
alphanumerical notation. There is, however, one Iundamental faw in this copy,
namely that, Ior some reason or other, the scribe completely omits any digit in
the 10s or 10000s unless it is Iollowed by a 0`, in which case it is invariably
written as k` Thus:
10 k 20 k 30 k 40 k 50 k etc.
11 a 21 a 31 a 41 a 51 a etc.
12 b 22 b 32 b 42 b 52 b etc.
13 c 23 c 33 c 43 c 53 c etc.
etc.
10000 k milia 20000 k milia etc.
11000 a milia 21000 a milia etc.
Thus, the only numbers which are correctly written in the 10s and 10000s
are 10 and 10000 themselves. As examples one may take the Iollowing:
Iol. 54r: c partes et k sex. et d secunde et h tercie 13;10,34,58`
Iol. 66r: i dies et b dies et k recte hore 122 days and 10 hours
Iol. 64v: x milia et a milia et et c dies 311,783 days (see fgure 1)
That the 10s (and consequently the 10000s) were originally represented
is clear Irom the Iact that the missing letters oI the alphabet, l` to s` (placed
in round brackets in the table above), exactly ft between 10 ( k) and 100
( t). We must also presume that, given the accuracy oI the numbers when they
are written in roman numerals, the original translator also used the alphanu-
merical notation in an accurate way. The parallel examples oI Greek andArabic
alphanumerical notation illustrate how the letters, used in their Semitic order,
progressively represent the units, 10s and 100s, and these are the parallels
the translator oI the Dresden Almagest would have been Iollowing. It seems,
however, that, in one respect, his system diIIered Irom the normal Greek and
Arabic systems: he appears to have added 'k when a 10 or 10000 was not Iol-
lowed by a unit, giving something like:
10 k 20 lk 30 mk 40 nk 50 ok etc.
11 ka 21 la 31 ma 41 na 51 oa etc
12 kb 22 lb 32 mb 42 nb 52 ob etc.
In each case (excepting always '10) the frst letter was dropped at some
time in the copying process, perhaps because it was originally written (or
intended to be written) in rubric, and the rubrics were not flled in.
22
When one turns to the Liber Mamonis one fnds a very similar alpha-
numerical system, but this time, the expected letters Ior the 10s are used:
1 a 10 k 100 t
2 b 20 l 200 u
3 c 30 m 300 x
4 d 40 n
5 e 50 o
6 I 60 p
7 g 70 q
8 h 80 r
9 i 90 s
In the Liber Mamonis the author has no occasion to use any number
between 360 and 1000. ThereIore, it is impossible to know what symbols he
would have used Ior the 100s between 400 and 900. Moreover, he uses 'k only
Ior 10` and not in combination with the letters Ior 20 to 90.
Eor the thousands and above, the author oI the Liber Mamonis usually
writes out the numbers in a way which is also Iound in the Dresden Almagest:
Eigure 2.1
Dresden, Iandesbibliothek, Db. 87, Iol. 64v. A passage Irom the translation oI
Ptolemy`s Almagest, Book IV, ch .7, showing the use oI alphanumerical notation.
In lines 1314 the number 311,783 is written as x milia et a milia et et c.`
Liber Mamonis, Iol. 27v: mille septingenti .l.h. | 1728|.
Dresden Almagest, Iol. 51r: quatuor milia et z et c | 4573|
However, on two Iolios (27v28r), he experiments with using a diIIer-
ent system oI numerals Ior high numbers, namely Hindu-Arabic numerals. He
is writing at a time when Hindu-Arabic numerals were only just beginning to
be used by Iatin scholars, and it is signifcant that he (or perhaps, rather, the
scribe) assimilates their shapes to Iatin letters.
23
Moreover, he uses the Eastern
Iorms oI the numerals, as can be seen Irom the Iollowing table:
Eastern Iorms
1 i or t
2 p
3 +
4 (possibly the abbreviation oI quia`)
5 g
6 7 (the tyronian et`)
7 u
8 a or o
9 q
This active use oI the Eastern Iorms oI the Hindu-Arabic numerals is
quite remarkable. The numerals in current use in Western scripts nowadays,
which are usually reIerred to as 'Arabic, were known in Castile by the late
tenth century,
24
and appear to have become the standard system used by trans-
lators and scholars working in Toledo in the later twelIth century: one variant
oI these Western Iorms is described specifcally as hgure toletane ('Tole-
dan symbols).
25
The Eastern Iorms oI the Hindu-Arabic numerals, on the
other hand, are closer to the original Sanskrit shapes, and developed into the
numeral Iorms used nowadays in Arabic. They are called hgure indice (Indian
symbols`) in the same Iatin manuscript that called the Western Iorms hgure
toletane.
26
The Eastern Iorms are Iound only in copies oI a small number oI
interrelated Iatin works, as will be discussed later.
It would be attractive to think that the author oI the Liber Mamonis was
infuenced by the example oI the Dresden Almagest and refned the alpha-
numerical notation he Iound there, whilst also experimenting with using Hindu-
Arabic numerals Ior the higher numbers. II the similarity between the two texts
stopped here, then this would remain only a weak hypothesis. However, oI even
greater signifcance is the Iact that the majority oI the astronomical terms in the
Liber Mamonis are the same as those oI the Dresden Almagest.
27
Much oI this
terminology is unique to these two works and quite remarkable; e.g., the use oI
the terms sexagenaria` Ior minute,`
28
synodos` Ior conjunction,` and (circu-
lus) rotunditatis` Ior epicycle. The two works must, thereIore, be related.
The question arises, however, as to whether the author oI the Liber
Mamonis was using the contents as well as the terminology oI the Dresden
Almagest. He reIers to Ptolemy`s work or its author on several occasions:
1. 'Ptolemy and the other authorities> who are more sensible in their thinking
concerning climes> . . . .
29
2. 'We know Irom the discovery oI the best men in astronomyPtolemy and the
othersthat this parallax is greater in the case oI> Mercury than in Venus, and
that Venus has a greater parallax> than the Sun.
30
3. 'These two equinoctial> points are fxed and do not move, according to what
Ptolemy stated in the Syntaxis. However, the opinion oI others who came aIter
Ptolemy> and who took it upon themselves to investigate this and other things
more intimately, is that they move eastwards with a slow motionnamely one
degree every 106 years. In this matter it is amazing that the intelligence oI Ptolemy
was deceived. Eor they do move. But what caused him to stumble has been proved
to be a certain astronomer who preceded him, who made a Ialse observation when
seeking the position oI the height oI the Sun. Eor he had said that the apogee oI
the Sun was in the same place in his time as Ptolemy, who dealt with and investi-
gated all the secrets oI the stars more perspicaciously, correctly discovered it was
in, in his own time. Hence Ptolemy is said to have been led into error, although he
said nothing more than that the point> was not moving. He made no mistake at all
in fnding the place; rather, his predecessor made a mistake. ThereIore, no doubts
should be cast on the accuracy oI Ptolemy, but one should blame the ignorance oI
the man who, by committing to writing something that he did not know, made a
wise man stumble. Now that we have said enough in support oI Ptolemy`s> dili-
gence, let us return to the subject.
31
4. 'The movement by which this sphere moves and carries the other spheres>
with it is a slow one Irom west to east, completingas Ptolemy said in his Megale
Syntasisthe path oI one degree in the period oI 100 years; or 106 years in the
opinion oI others.
32
5. 'The spheres oI the three planets, Saturn, Jupiter and Mars, are very similar to
each other in the number and divisions oI their spheres, their movement and their
circles, as Ptolemy is proved to have said in his Syntaxis.
33
6. 'The measurement oI the maximum latitude oI the planets and the Moon to
the north and the south oI the zodiac is: Moon: 4 45 (when the Great Circle is
divided into 360); Saturn 3; Jupiter 2; Mars, 4 20 to the north, 7 to the south;
Venus, as Ptolemy says in his Syntaxis, 6 2, but, as other astronomers have said,
9; Mercury 4.
34
7. 'This lowest sphere oI Saturn moves eastwards round two fxed poles with a
slow motion which is one degree in 100 years. These are the two poles oI the zodiac
which the centre oI the Sun circles, according to what Ptolemy says has been dis-
covered by himselI and his predecessors.
35
The author oI the Liber Mamonis gives the impression (or would like to
give the impression) that he knows the Almagest directly, and there are certain
passages which Iollow the Almagest closely in their argumentation and/or in
the values that they give:
1. The sequence oI the argument Ior working out the relative distances oI the plan-
ets Irom parallax on Iol. 27v (item 2 above) corresponds to that in Almagest, IX,
chapter 1.
2. The description oI fnding the solar anomaly on Iol. 31v corresponds to
Almagest, III, chapters 56.
3. The limits oI the movement in latitude quoted in item 6 above are rounded-oII
fgures Irom Almagest, XIII, 5.
II we are to suppose, on the basis oI this evidence, that the author oI the
Liber Mamonis knew the Almagest directly, then we have to conclude that he
was Iamiliar as much with the later books as with the frst Iour books which
alone are Iound in the Dresden Almagest.
However, the Almagest is not the principal model Ior the Liber Mamo-
nis, nor, perhaps, is it the principal source Ior the numerical values given in
the Iatin work. Eor the Liber Mamonis is a cosmology (in the Arabic haya
tradition),
36
not a mathematical work, and in some respects, the work is closer
to Ptolemy`s Planetary Hypotheses (e.g., in talking in terms oI contingent
spheres on Iols. 28v29r). Moreover, the author reIers specifcally (though not
by name) to Ptolemy`s successors and to astronomers who have criticized or
corrected Ptolemy. One oI these astronomers is, presumably, the unnamedArab
whom he describes, in the preIace to the Iourth book, as being his principal
guide throughout the work:
But since in other books> we Iollowed Ior the most part a certain Arab, in
this also we will Iollow him> through much, although we have Iound cer-
tain things concerning the number oI the spheres and their epicycles, and he
has touched upon the truths about the circles and the obliquities oI the plan-
ets with which the number oI spheres is dissonant.
37
It is presumably Irom these successors oI Ptolemy that he gets the values
which diIIer Irom those in the Almagest, such as 23` 35 Ior the obliquity oI the
ecliptic (Iol. 29r; Ptolemy`s value is 23` 5120), and a precession oI 106 years
per degree.
38
Also, not in the Almagest is his list oI values Ior the 'completion
Eigure 2.2
Cambrai, Bibliotheque municipale, 930, Iol. 27v28r, showing both the alpha-
numerical notation and the Eastern Iorms oI the Hindu-Arabic numerals.
oI movement in latitude ('Completur enim Saturni lati motus . . .) oI each oI
the planets on Iol. 48v:
Saturn: 29 years, e.d. (5.4. ?) days, 15 hours, 24 minutes.
Jupiter: 11 years, 315 days, 14 hours, 29 minutes.
Mars: 1 year, 322 days, 24 hours.
Venus and Mercury: 1 year, 5 hours, 49 minutes.
There do not appear to be any direct quotations Irom the Almagest.
A clue to the direct source oI the Liber Mamonis may be hidden in its
title itselI. Eor it is tempting to see in the 'Mamon, implied by the Iatin geni-
tive 'Mamonis, a reIerence to the caliph al-Mamn (ruled 813833 A.D.), on
whose command the earliest Arabic translations oI the Almagest were made. It
is unlikely that the Liber Mamonis is an allusion to the Almagest itselI. How-
ever, it is possible that 'the book oI al-Mamn is, rather, the correction oI
the Almagest, commissioned by the same al-Mamn, and resulting in al-:f
al-mumtahan (reIerred to as the tabule probationum or tabule probate in Iatin
translations). These astronomical tables were specifcally associated with the
name oI al-Mamn, and at least one oI the values that the Liber Mamonis gives
corresponds to a value attributed to al-Mamn`s correctors,` over against Ptol-
emy`s value: i.e., 23` 35 Ior the value oI the obliquity oI the ecliptic. One may
compare the relevant passage oI the Liber Mamonis with the reIerence to al-
Mamn`s correction in John oI Seville`s translation oI al-Earghn`s Rudimenta,
c. 5 (translated in Iimia` on March 11, 1135):
Liber Mamonis: 'The movement oI this sphere oI the spherical circling (i.e.,
the ecliptic) is Irom the west to the east over two fxed poles and a fxed
radius to the poles. These poles are neither the radius oI the universe nor the
radius (?), but each oI them is distant Irom the pole oI the universe closest to
it by 23 degrees and 35 minutes.
39
Al-Earghn: 'The obliquity oI the ecliptic> is, according to what
Ptolemy discovered, 23` 51 (when the circle is 360`), but, Irom the most
accurate prooI by which Almemon proved it, and with the agreement oI the
majority oI wise men, it is 23` 35.
40
II the Liber Mamonis gives the value oI al-Mamn`s correctors` Ior
the obliquity oI the ecliptic without comment, it is possible that the other val-
ues that he gives and approves are Irom the same source. This remains to be
checked. However, the title given in the Cambrai manuscript is not simply
'Iiber Mamonis, but rather 'Iiber Mamonis in astronomia a Stephano phi-
losopho translatus. This would imply that a certain philosopher called 'Ste-
phen translated a work associated with al-Mamn IromArabic into Iatin. Yet
the Liber Mamonis is not a translation, but rather a treatise written in a liter-
ary style, which has as its leitmotiI a criticism oI the current doctrines in Iatin
cosmology, epitomized in the theories oI Macrobius, and the need Ior their
replacement by the Ptolemaic system. The word translatus` in the title, then,
would mean not translated` (especially since no source or target language is
mentioned) but rather transmitted`, a meaning supported by classical usage,
and, more important, in another work by 'Stephen the Philosopher.
41
Eor, there
is little doubt that the 'Stephanus philosophus oI the Liber Mamonis is the
'Stephanus philosophie discipulus who translated the comprehensive medical
work, al-kitb al-malak (Regalis dispositio) oI Al ibn al-Abbs al-Majs.
The reasons Ior this identifcation have been explained in detail else-
where,
42
but can be summarized here. We fnd the same style oI literary Iatin,
including whole phrases, in both works; we fnd consultation oI Greek as well
as Arabic sources; but, above all, we fnd the same system oI alpha-numerical
notation. Moreover, we fnd a place and a date, or rather, several dates, attached
to diIIerent books oI the translation oI the Regalis dispositio. The place is
Antioch, and the dates all Iall within the year 1127. This makes it very likely
that Stephen the Philosopher is 'Stephanus thesaurarius Antiochie Ior whom
a copy oI the Rhetorica ad Herennium (now MS Milan, Ambrosiana, Cod. E.
7 sup.) was written in 1121.
43
Eor, this manuscript also uses the alphanumeri-
cal notation in the same way as in the Liber Mamonis. Richard Hunt pointed
out that there was a treasurer called Stephen at the Benedictine monastery oI St
Paul, one oI the principal religious Ioundations in Antioch, who had been given
a house in the city between 1126 and 1130.
44
This is probably our 'Stephen.
Eurther biographical details are given by the medical writer, Matheus Eerrar-
ius, who states that 'Stephen, a certain Pisan, went to those parts (meaning
the Orient?), and, learning that language, translated the whole oI the Practica
(i.e., the practical portion oI the Regalis dispositio),
45
and Irom an unknown
twelIth-century supporter oI the medical school oI Montpellier who calls Ste-
phen the 'nephew oI the Patriarch oI Antioch.
46
The correspondence in language and the Iorm oI the alphanumerical
notation in the Liber Mamonis, the Regalis dispositio and the Milan copy oI the
Rhetorica ad Herennium (alphanumerical notation only) corroborates that the
'Stephen mentioned in all three works is the same man, and we can be reason-
ably sure that he was working in Antioch in the 1120s. He had a strong interest
and competence in astronomy; he translated Irom Arabic, but also had some
knowledge oI Greek; he had the help oI Arabic-speaking colleagues. The Dres-
den Almagest is a little diIIerent Irom these three works: the name Stephen is
not attached to it; all Arabic transliterations are avoided whereas, in the Rega-
lis dispositio at least, Stephen deliberately transcribes the Arabic terms when
he does not know the Iatin equivalent; the alphanumerical notation is not as
advanced, and no Hindu-Arabic numerals are used. Nevertheless, the similari-
ties between the Dresden Almagest and the Liber Mamonis are such that the
work must, at least, have arisen in the same milieu, iI it was not directly used
by Stephen the Philosopher.
47
That milieu could have been Pisa, which was not
only an important centre Ior Greek-Iatin translations, but also had close links
with the Arabic world through its quarters in Antioch, Iaodicea, Acre, and
elsewhere.
48
But it was more likely Antioch itselI, where Stephen was actively
engaged in translating works IromArabic.
One last clue associating the Dresden Almagest and the Liber Mamo-
nis must be taken into account. At the beginning oI the Liber Mamonis the
author states: 'Since we have Iulflled our promise, having written a treatise on
the rules we had proposed Ior the canon oI astronomy, I approach this second
task. . . .
49
He has, then, already written rules Ior astronomical tables. These
he appears to reIer to in the body oI the Liber Mamonis,
50
and he makes Ire-
quent reIerences to the tables themselves.
51
Altogether, Iour tables are men-
tioned, which are to be used in succession Ior ascertaining the position in
longitude oI the planets; a Iurther table is used Ior their latitudes. These tables
and their rules have not been identifed,
52
but the late thirteenth-century astro-
nomer, Henry Bate, in listing a number oI authorities that agree in measuring
the movements oI the planets in respect to the 'ninth sphere, mentions among
the tables written by the Iollowers oI the magistri probationum, tabule pisane
wintonienses` (the tables oI Pisa and> Winchester).
53
It is true that one oI
the sets oI instructions Ior using the tables oI Pisa mentions the longitude oI
Winchester.
54
But it is intriguing that this title should nicely join the names oI
Stephen (oI Pisa and Antioch), the author oI the Liber Mamonis and Abd al-
Mash oI Winchester, the translator oI the Dresden Almagest.
55
II As+roxoxic:i 1:iirs Coxxrc+ri vi+n Pis: :xi Iucc:
There is a very strong reason to connect the Tables oI Pisa and the Liber
Mamonis: namely, that the earliest copy oI the Tables and the instructions Ior
their useBerlin, Staatsbibliothek, lat.Iol. 307 (Rose, no. 956)are written
entirely in the Eastern Iorms oI the Hindu-Arabic numerals. The Tables oI Pisa
appear to be based on the lost Arabic tables oI Abd-al-Rahmn ibn Umar
al-SI (d. 986),
56
and the Berlin version gives 1149 completed years as their
starting point (i.e., they were written in 1150 A.D.). The Eastern Iorms are
actively used
57
only in a very restricted range oI works. Most oI these are Iatin
texts attributed to, or based on the work oI the Jewish polymath, Abraham ibn
Ezra (1089/92aIter 1160).
Eigure 2.3
Berlin, Staatsbibliothek, lat. Iol. 307, Iol. 32r. A page Irom the Pisan Tables showing
the use oI the Eastern Iorms oI Hindu-Arabic numerals.
A text very similar to the anonymous instructions Ior the use oI the Pisan
Tables in the Berlin manuscript occurs in Iondon, British Iibrary, Arundel
377, Iols 56v63r. This text is entitled 'Tractatus magistri Habrahe de tabulis
planetarum, and is immediately Iollowed by a text on the astrolabe according
to the words oI 'Abraham, outstanding among the philosophers oI his time,
and our master, on whose dictation we wrote this account oI the astrolabe.
58
This copy oI these two texts employs the Western Iorms oI the Arabic numer-
als, which (as we have seen) became the norm, but the presence oI a key (Iol.
56r) Ior converting the Eastern Iorms to the Western Iorms implies that an ear-
lier copy had been written using the Eastern Iorms, and this is what we fnd
in another manuscript oI the astrolabe text: MS British Iibrary, Cotton Vespa-
sian A.II, Iols 37v40v. Yet another text was written to accompany the Tables
oI Pisa: this text, known nowadays as the Fundamenta tabularum, is a schol-
arly discussion oI the theory oI astronomical tables, drawing on many Arabic
sources. In certain manuscripts it is called 'Abrahismus and Millas Vallicrosa
has shown that it was written byAbraham ibn Ezra in Dreux in 1154. In the ear-
liest manuscripts oI the work the Eastern Iorms oI the numerals are used.
59
Abraham Ibn Ezra was born in Tudela in the Muslim kingdom oI Sara-
gossa between 1089 and 1092 and spent the earlier part oI his liIe in North-
ern Spain, though he also visited other parts oI Spain and North AIrica. In the
early 1140s he started to visit Jewish communities in Christian Europe, frst in
Rome, Iollowed soon aIter by Iucca (11421145), which is only some fIteen
kilometers Irom Pisa.
60
We are Iortunate in possessing a manuscript to which
have been added some notes which mention the year 1160 and were presumably
written then, or soon aIter: MS British Iibrary, Harley 5402, Iols 69rv.
61
These
notes include a handy way oI calculating the position oI the Moon, in which
the writer uses Hindu-Arabic numerals in their Eastern Iorm, and instructions
Ior tables Ior the meridian oI Iucca which are similar to those accompanying
the tables oI Pisa, as we read them in the Berlin manuscript.
62
What is strik-
ing is that these notes are written in a mixture between Italian and ungram-
matical Iatin. Examples oI written Italian beIore the end oI the twelIth century
are rare; the most substantial document, in Iact, is a religious poem written
by a Jew in Hebrew script.
63
We may, thereIore, be in the presence oI a Jew-
ish scholar writing in Iucca and using the Eastern Iorms oI the Hindu-Arabic
numerals: i.e., someone just like Abraham ibn Ezra.
But the Harley manuscript tells us more. The original scribe, who pre-
sumably wrote the text beIore 1160, did not use Hindu-Arabic numerals. He
transcribed a copy oI Pseudo-Ptolemy`s Iudicia (an astrological work) and the
Iatin corpus oI Sahl ibn Bishr`s astrological texts, using, where necessary,
roman numerals. But between these two texts he added an astronomical table
in which the numerical values (aside Irom those oI the frst column which are
in roman numerals) are written entirely in alphanumerical notation. The nota-
tion is that oI the Liber Mamonis, Regalis dispositio, and the Milan Rhetorica
ad Herennium, but in this case, and in this case only, a key is provided (see
fgure 2.4).
64
We see, then, in the Harley manuscript the two kinds oI notation Ior
numerals which appear in the Liber Mamonis: the Iatin alphanumerical sys-
tem and the Eastern Iorms oI the Hindu-Arabic numerals. The writer oI the
Liber Mamonis used alphanumerical notation Ior lower numbers, and the East-
ern Iorms Ior numbers consisting oI more than three digits. This mixed sys-
tem occurs regularly in Islamic astronomical tables, and occasionally in Greek
tables, in which alphanumerical notation is used Ior all numbers up to 360 (the
number oI degrees in the circle), but Hindu numerals are used where the num-
bers do, or in principle could, reach the higher hundreds or exceed 1000.
65
It is
quite possible that the tables to which Stephen, the author oI the Liber Mamo-
nis, reIers, were written in this mixed system just as the Liber Mamonis itselI
is. Eor some reason or other, however, neither the alphanumerical notation nor
the mixed system caught on. Rather, as it seems, the Eastern Iorms, as used in
the mixed system, were used Ior all numerals, as we see in the Berlin manu-
script oI the Pisan Tables.
It is unlikely that Ibn Ezra himselI introduced the Eastern Iorms oI the
Hindu-Arabic numerals. At the beginning oI his Hebrew work on arithmetic,
Sefer ha-Mispar, he mentions Hindu-Arabic numerals and substitutes Ior them
the letters oI the Hebrew alphabet. OI the several manuscripts oI this text that
I have seen, only one gives the Hindu-Arabic numerals in their oriental Iorm,
as an alternative to the Western Iorms which Ibn Ezra could have been Iamiliar
with in Spain.
66
It is only in the Iatin versions oI Ibn Ezra`s works that the East-
ern Iorms are preIerred, and this preIerence must have been due to Ibn Ezra`s
Iatin collaborators or students. It could also be questioned whether Ibn Ezra
was responsible Ior drawing up the Pisan Tables, which is inIerred by modern
scholars Irom the Iact that he bases his instructions on how to use astronomical
tables on the Pisan Tables. The Berlin manuscript does not mention his name,
and Henry Bate diIIerentiates between 'the tables oI Abraham and the 'Pisan
Tables.
67
Since Henry adds 'oI Winchester to the to the Tables oI Pisa, it is at
least worth considering whether 'Abd al-Mash oI Winchester had anything
to do with the transmission oI al-SIs tables to the West. An Antiochene origin
would also solve the problem that, apparently, al-SI`s tables were not known
to Andalusi astronomers.
68
The use oI the Eastern Iorms oI the Hindu-Arabic numerals, even in
the small number oI works I have mentioned, implies at least a community
oI scholars who understood them. The anonymous Iatin scholar who colla-
borated with Abraham ibn Ezra on writing his text on the astrolabe, and the
other Iatin scholars who presumably put into good Iatin his instructions Ior
the use oI the Pisan Tables, belong to this group. It is possible that Hermann
oI Carinthia, Robert oI Ketton, and Hugo oI Santalla who were working 'on
the banks oI the Ebro and in Pamplona, Tarazona and Tudela, and were all
involved in compiling compendia on astrological judgments (the Liber trium
iudicum and the Liber novem iudicum) also belonged to this group.
69
Abraham
ibn Ezra, as we have seen, came IromTudela, and could well have known these
scholars. Subsequently both Hermann (in 1143) and Ibn Ezra (in 1148) are
attested in Beziers, where Hermann`s student Rudolph oI Bruges wrote a text
on the astrolabe (in 1144), which is sandwiched between Abraham`s Funda-
menta tabularum and his treatise on the astrolabe in MS Iondon, Cotton, Ves-
pasian A.II (all three texts, and these only, have been written in the same hand),
though it uses only roman numerals.
70
There is some evidence that these scholars also knew, and may have used
the Eastern Iorms oI the numerals. Three early copies oI works by Hugo oI
Santalla preserve these Iorms, in two cases in a table only,
71
in the third case in
the text.
72
It is clear that the scribes oI these three manuscripts were not Iamil-
iar with the Eastern Iorms, and either copied them wrongly (in the table), or
abandoned them (in the text). But in one copy oI the Liber trium iudicum the
scribe uses the Eastern Iorms, confdently and consistently, not only Ior writing
the Liber trium iudicum itselI, but also Ior the copy oI Hyginus`s Astronomi-
con which accompanies that text. This is a twelIth-century manuscript incorpo-
rated into MS Iondon, Arundel 268, which is remarkable because it is one oI
the earliest examples we have oI a Western manuscript made oI paper. The
Liber trium iudicum is an astrological compendium addressed in this manu-
script to 'karissime R., who is presumably Robert oI Ketton, which would
make the compiler Hermann oI Carinthia; in another manuscript the same
Liber trium iudicum is dedicated to Michael, bishop oI Tarazona, who was
Hugo oI Santalla`s patron.
73
The use oI paper at this date suggests a close con-
nection with the Islamic world. The hand-writing suggests that it was written
in Italy, as does that oI Hugo`s text in MS Digby 50.
Hermann oI Carinthia was a student oI Thierry oI Chartres, and some oI
his translations arrived in Chartres; the Berlin manuscript oI the Pisan tables
is said to have originated Irom Chartres, though the instructions Ior the Pisan
tables mention the longitudes oI Angers and Toledo, and another text in the
manuscript implies that Paris is place oI writing.
74
Ibn Ezra himselI traveled
North, to the Jewish communities in Rouen, Dreux (11531156), and Evreux,
and eventually to Iondon (1158 and 1160), where, according to some sources,
Eigure 2.4
British Iibrary, Harley 5402, Iol. 16r, written in alphanumerical notation.
he died. Thus, it is not surprising that the majority oI the manuscripts oI Ibn
Ezra`s works (and hence including the Eastern Iorms oI Hindu-Arabic numer-
als) are Norman and English.
75
Much more work needs to be done on this subject. That the works oI Ibn
Ezra, Robert oI Ketton, Hermann oI Carinthia, Rudolph oI Bruges, Hugo oI
Santalla, and Stephen the Philosopher and Abd al-Mash oI Winchester, and
the Eastern Iorms oI Hindu-Arabic numerals associated with these works, did
not remain in currency Ior very long may be due to the Iact that, aIter 1150,
Toledo became the centre oI the translating activity IromArabic into Iatin, and
this industry` in Toledo was so successIul that it replaced or drove underground
much oI what went beIore. It is becoming increasingly obvious that, presum-
ably under the supervision oI Gerard oI Cremona, there was a concerted pro-
gram oI revising and translating aIresh those works that were perceived to be
most important in astronomy and astrology. The mechanics oI the transmission
oI the Toledan translations and the reasons Ior their success are still not clear.
At frst sight it would seem that the translations oI Greek astronomical works
directly Irom Greek should have been preIerred to the translation oI their Arabic
versions: but in the case oI Ptolemy, Theodosius, Menelaus, Archimedes, and
so many other authors, although Greek-Iatin translations were made, and oIten
made beIore the Toledan Arabic-Iatin translations, nevertheless it was the latter
which succeeded in becoming established in the scholarly community.
The same conclusion can be reached in regard to the works oI Stephen the
Philosopher and his colleagues. The Dresden Almagest and the Liber Mamonis
both survive incomplete in one manuscript each, and both the alphanumerical
notation and the Eastern Iorms oI the Hindu-Arabic numerals Iailed to catch
on. Nevertheless, some credit must be given to Stephen and his Iriends. II the
arguments that are put Iorward in this chapter are convincing, then the Dresden
Almagest is likely to predate the earliest hitherto known translationthat made
Irom the Greek in Sicily in ca. 1160 A.D.
76
by up to Iorty years, and the Liber
Mamonis shows a remarkably advanced understanding oI Ptolemaic astronomy
and its Arabic developments at a correspondingly early date. Both works are
indicative oI a considerably higher level oI astronomical learning among cer-
tain Iatin scholars in the second quarter oI the twelIth century than has hitherto
been recognized, and they invite us to look to the Eastern Mediterranean, rather
than to Spain, as a source Ior this learning.
Ackxovirixrx+
I am very grateIul Ior the help oI Richard Iorch, Paul Kunitzsch, Eritz Saaby-
Pedersen, Tzvi Iangermann, and Tony Ievy.
No+rs
1. C. H. Haskins, Studies in the History of Mediaeval Science, 2nd ed. (New York,
1927), p. 109, n. 155: 'There is also a conIusing Iorm oI numerals |in the Dresden
Almagest|: b 2, etc. CI. supra, n. 128 |discussion oI the numerals in the Liber
Mamonis|.
2. The Iullest account oI the Liber Mamonis up to now is that oI Haskins, Studies, pp.
98103. An edition oI the work has been promised by Richard Iemay.
3. I am very grateIul to Menso Eolkerts Ior lending me a microflm oI Dresden, Db. 87,
and to Richard Iorch Ior inIormation on the text. Eor a description oI the manuscript
see Bertoldo di Moosburg, Expositio super elementationem theologicam Procli, 184
211, De animabus, ed. I. Sturlese (Rome, 1974), pp. xlviixlix. I am also indebted to
Gerhard Brey Ior sending me a copy oI the on-line description oI the same manuscript
Irom the ICCMSM database.
4. J. I. Heiberg, 'Noch einmal die Mittelalterliche Ptolemaios-bersetzung, in
Hermes, 46, 1911, pp. 207216 (see pp. 215216).
5. MS Dresden, Db. 87, Iol. 15v: 'Explicit primus sermo libri mathemathice Ptolomei,
qui nominatur megali xintaxis astronomie translacione dictamine (elsewhere dictami-
nis`) philophonia wittomensis ebdelmessie (with variants in the other explicits win-
tomiensis` and wuttomensis`). The interpretation oI these explicits and oI the name oI
the translator/dictator has been explored in C. Burnett, 'Abd al-Mash oI Winchester,
in Between Demonstration and Imagination. Essays on the History of Science and
Philosophy Presented to John D. North, ed. I. Nauta and A. Vanderjagt (Ieiden, 1999),
pp. 159169. 'Abd al-Mash ('Servant oI the Messiah) is sometimes a generic name
Ior a Christian in a Muslim society.
6. This Iact was frst pointed out to me by Richard Iorch.
7. 'Preclare Iecerunt qui corrigentes scienciam philosophie, o Syre, diviserunt theori-
cam partem philosophie a practica nima m faala fm ar lladhna staqsaw ilm
al-falsafa y Srus f ifrdihim fu: al-falsafa an-nazar an al-amal (P. Kunitzsch,
Der Almagest, Wiesbaden, 1974, p. 133). Contrast the Greek: Hovu xom= oi yvpoim=
iooopoovtr= m Eupr, ooxouoi oi xrmpixrvoi to 0rmpptixov tp= ioooio=
oao tou apoxtixou (Ptolemy, Syntaxis mathematica, ed. J. I. Heiberg |Ieipzig, 1898|,
p. 4, lines 78) The true philosophers, Syrus, were I think, quite right to distinguish
the theoretical part oI philosophy Irom the practical` (trans. G. J. Toomer, Ptolemy`s
Almagest |Iondon, 1984|, p. 35).
8. discordia` (Iol. 36r; anomaly`) corresponds to ikhtilf (diIIerence`), rather than to
ovmoio; discordia visus` (Iol. 50r; parallax`) ikhtilf al-manzar (diIIerence oI
vision`), rather than aopooi=; circulus rotunditatis` (Iol. 36v; epicycle`) falak
al-tadwr (circle oI roundness`), rather than raixuxo=; longinqua longinquitas` (Iol.
37r; apogee`) al-bud al-abad (the most distant distance`), rather than oaoyriov,
etc.
9. E.g., Iol. 2r: 'rememorat qualitatem communem, which is Iollowed immediately by
another translation oI the same phrase: 'vel commemoratio communis qualitatis and
is Iollowed soon aIter by another double translation: 'rememorabimus vel commemora-
bimus ergo. The words 'rememoratio and 'rememorare are particularly common in
the Arabic-Iatin translations oI John oI Seville and Gerard oI Cremona, corresponding
to Iorms oI the Arabic root dh-k-r.
10. Eol. 1r: 'propter quod aretius morum anime possit esse in pluribus sine doctrina;
see Heiberg, p. 4, line 12: oio to tmv rv p0ixmv oprtmv rvio= uaopoi ouvoo0oi
aooi= xoi mpi= o0porm=. The Iatin word-order and sense Iollow the Arabic text
here.
11. Hipparchus is always 'ypparcus, Kalippus is 'caliphi (Iol. 70r), Phanostratos,
the Athenian archon, appears as 'Phanastrato Athenarum custode and 'Phanastri pri-
matis Athanarum (both Iol. 69r; cI. trans. Toomer, pp. 211212), the Egyptian months
appear as 'Thuth (Thoth), 'Phamenoth (Phamenoth), 'Messurem (Mesore), and
'Mechir (Mechir) on Iols 69r70v. The Greek months, however, either betray the
infuence oI Arabic transcription (Iol. 69r 'Bussiothos Ior Poseideon) or are omitted
altogether. A list oI month names, including the Egyptian and the Greek, preceded the
Almagest text in WolIenbttel, Gud. lat. 147, which added the preIace oI the Sicilian
Greek-Iatin translation oI the Almagest to a text oI the Toledan Arabic-Iatin translation
(see Kunitzsch, Almagest, p. 95).
12. E.g., Iol. 9v: Ponemus igitur istud in tabulis` ( Almagest, I.11); Iol. 20v: in
tabulis que hoc verbum secuntur` ( Almagest, II.6, tables Ior the third parallel up to
the North Pole); Iol. 22v: 'posuimus autem tabulas orientalium .x. et decem graduum
( Almagest, II.8); Iol. 30v: 'ponentes eorum tabulas ( Almagest, II.13); Iol. 55r: 'et
posuimus in prima tabularum tempora collectorum annorum et displicatorum (sic) et
mensium et dierum et horarum et post istud motum in longo et post illum in discor-
dia et post illum in lato et post illum in longinquitatem que est inter Solem et Iunam
( Almagest, IV.4).
13. It is possible to imagine that the tables were leIt in the original Iorm, in Greek or
Arabic (or merely with their rubrics glossed in Iatin), and that the reader was expected
to be able to interpret the numeral values: this would provide a possible reason Ior the
adoption oI alphanumerical notation (the notation oI the Greek and Arabic tables) in the
Iatin text; see below.
14. The combination oI both Iorms occurs in 'liber estichie elementorum (Iol. 7r) and
the translator also uses the abbreviations 'l.e.e. and 'l.he.e. It is unlikely that a scholar
who knew Ptolemy`s Almagest did not have access to Euclid`s Elements, and it is worth
considering whether our translator also translated the Elements and iI any extant version
shows the characteristics oI his style.
15. MS Istanbul, Ahmet III, 3464, Iols 216r217r. The correspondence between the
two works begins with the theorem which precedes the addition, which is the last the-
orem in Almagest, I, chapter 13. Eor an edition oI the relevant section oI the Dresden
manuscript see R. Iorch, 1hbit ibn Qurra On the Sector-Figure and Related 1exts,
ErankIurt 2001, pp. 362375.
16. The relationship between the Dresden Almagest and the work oI al-Nasaw needs
Iurther investigation. Both texts reIer to the same earlier theorems oI Ptolemy (e.g.,
Dresden Almagest, Iol. 13v: 'quod est ostensum in .xii. fgura istius sermonis), and
the style oI writing in the Dresden Almagest gives the impression that the 'addition
is integral to the work: the same system oI cross-reIerences and appeals to the relevant
theorems oI Euclid`s Elements are Iound throughout the section oI the Almagest on trig-
onometry and spherical astronomy.
17. I have noticed only that fgure 4.5 (the rejected eccentric model Ior the Moon`s
motion) is missing.
18. 'Scribe and 'translator need not be distinguished here in that the scribe oI this
unique manuscript appears to have tried to copy what was beIore him without altera-
tion. They need only be distinguished when (as sometimes happens) it is clear that the
scribe misunderstood what was in Iront oI him.
19. The use oI '.k. Ior twenty here is due to the faw in the system mentioned below,
and not to conIusion with the Greek and Semitic alphanumerical notations in which
'k is 20.
20. The scribe uses the abbreviation Ior 'enim, but in two Iorms: the capital H with the
middle bar extended in both directions (oIten described as an English practice) on Iol.
54r, and .n. on Iol. 66v. The frst Iorm, which is a distinctive symbol, may have been the
original one. It is also possible that the Greek capital eta (H) was originally intended,
and that, in copying the text on Iol. 66v, the scribe mistook this sign Ior the abbrevia-
tion Ior 'enim.
21. The isolated use oI O where the Greek text has '900 on Iol. 63v is apparently a mis-
take, since there are several examples oI O being used Ior '600. In other cases where
the Greek text has '900 the number is written out in the Iatin translation as 'nongenti.
Whether the scribe`s letters are closer to Greek majuscules or miniscules can only be
determined by comparison with Greek letters occurring both in Greek and Iatin manu-
scripts oI the period. Iatin authors tended to use the Greek majuscule, and we have sev-
eral examples oI the knowledge among themoI the numerical values oI the Greek letters;
see W. Berschin, Greek Letters and the Latin Middle Ages, revised and expanded edition,
trans. J. C. Erakes, Washington, D.C., 1988, pp. 2930 and 289, n. 40.
22. The alphanumerical notation in Liber Mamonis (see below) is generally rubri-
cated.
23. On Hindu-Arabic numerals in Arabic manuscripts see Kunitzsch`s chapter in this
volume. The Eastern Iorms in this manuscript are more closely assimilated to letters oI
the Iatin alphabet than in other manuscripts. The Iact that diIIerent letters are used Ior
the same numerals in the case oI '1 and '9 in the Cambrai manuscript oI the Liber
Mamonis might suggest that the assimilationwas made bythe scribe, whoexpectedtofnd
letter-Iorms, rather than by the author, Irom whom one would expect a one-to-one
equivalence oI symbols.
24. Richard Iemay, in 'The Hispanic Origin oI Our Present Numeral Eorms (Jiator,
8, 1977, p. 43562) suggests that this 'Western Iorm arose out oI a mixture oI the
original Sanskrit numerals and Visigothic Iorms, but the question oI origin is not yet
settled.
25. MS Munich, Bayerische Staatsbibliothek, clm 18927, I. 1r; see Eigure 1a in R.
Iemay, 'The Hispanic Origin . . .
26. The two Iorms are also distinguished in Arabic in the twelIth-century Maghrib by
Ibn al-Ysamn: see Ab Eris, 'Dall jadd al urbat al-arqm al-mustamala I al-
maghrib al-arab, Al-lisn al-arab, 10 (1392/1973), pp. 232234; the Iull text oI Ibn
al-Ysamn has been edited in a dissertation by a student oI Ahmed Djebbar, see p. 345.
27. OI the terms already mentioned (n. 8 above), the Liber Mamonis (M), gives 'visus
discordia (Iol. 27v) Ior 'parallax, 'rotunditas or 'rotundus circulus (Iol. 39r et pas-
sim) Ior 'epicycle (cI. D|resden Almagest|, 'circulus rotunditatis), and 'longinqua
longinquitas and 'propinqua longitudo (M, Iol. 29v) Ior 'apogee and 'perigee.
Other parallels in terminology are 'speralis (rather than 'spericus) Ior 'spherical (M
and D passim), 'austrum (rather than 'meridies) Ior 'south (M, Iol. 49v, D, Iol. 10v),
'circulus recti diei Ior 'equator (M, Iol. 8v; cI. D, Iol. 6r, 'circulus rectitudinis diei),
'punctum capitis Ior 'zenith (M, Iol. 18v, D, Iol. 16r), 'synodos Ior 'conjunction
(oI planets) and 'new moon (M, Iol. 39r, D, Iols 55r, 65r and 68r), 'pansilini Ior 'Iull
moon (M, Iol. 34r, D, Iols 55r and 65r), 'sexagenaria Ior 'minute (M and D passim),
'latum Ior 'latitude (M, Iol. 49v, D, Iol. 51r), 'circulus signorum Ior 'zodiac (M,
Iol. 49v, D, Iol. 10r), and, fnally, the word Ior Ptolemy`s work itselI: 'megali sintasi
(ablative; M, Iol. 39v), 'megali xintaxis (nominative; D, Iol. 15v etc.). Occasionally,
the author oI the Liber Mamonis uses diIIerent terminology, perhaps because oI his
greater Iamiliarity with the Iatin astronomical and geometrical tradition: e.g., 'liber
geometrie (M, Iol. 7v), rather than 'liber elementorum Ior Euclid`s work; 'circu-
lus extra centrum (M, Iol. 39r), rather than 'circulus Iorinseci centri (D, Iol. 36v)
Ior 'eccentric; 'caput draconis and 'cauda draconis rather than 'ligans capitis and
'ligans caude (D, Iol. 67r) Ior the ascending and descending nodes.
28. To call a 'minute a 'sixtieth part (oI a degree) would seem a natural thing to do,
but such terminology does not appear to have been used either in the Greek or in the
Arabic versions oI the Almagest: see Kunitzsch, Almagest, pp. 156160.
29. Eol. 22v: 'Tholomeus autem et ceteri quibus sanior est intellectus. . . .
30. Eol. 27v: 'Deprehensum autem a probatissimis in astronomia viris Tptolomeo (sic)
et reliquis cognovimus maiorem hanc esse visus discordiam Mercurio quam in Venere,
Veneremque maiorem habere Sole.
31. Eol. 29v30r: 'Sunt autem hii duo puncti fxi nec moventur sicut Ptolomeus posuit
in Sintaxi. Aliorum autem qui secuti sunt deinceps astronomicorum quibus Iuit etiam
et hec et alia secretius rimari, sententia est in orientem moveri tardo motu, in .t. scili-
cet et .I. annis uno gradu. Qua in re mirum est Ptolomei deceptam Iuisse sollertiam.
Moventur enim. Set illi casus causa extitisse probatus est> quidam sui temporis pre-
cessor astronomus cuius inquirendo altitudinis Solis loco Ialsa Iuit inspectio. Illo enim
in loco longuinquam Solis longinquitatem suo tempore dixerat, quo Ptholomeus perspi-
catius omnia astrorum secreta discutiens et investigans suo tempore illam esse veracis-
sime comperuit. Hinc ergo deceptus Ptolomeus Iuisse dicitur, tametsi non ipse nisi quod
non moveri dixit. In loco enim inveniendo nichil peccavit, sed qui precesserat ipsum
deceptus Iuit. Non ergo Ptholomei sagacitas cecidisse arguenda est, sed illius depra-
vanda ignorantia qui quod nescivit scripture tradens sapienti viro cadendi causa Iuit.
Nunc satis quod pro illius deIendenda industria diximus, ad nostra redeamus.
32. Eol. 39v: 'Motus autem is quo hec spera movetur et alias secum movet tardus est ab
occidente in orientem, complens, sicut dixit Ptolomeus in Megali sintasi sua in .t. anno-
rum curriculo unius cursum gradus. Aliorum sententia est in .t.I. annis.
33. Eol. 44v: 'Trium autem planetarum, Saturni scilicet Iovis et Martis, spere simil-
lime sunt adinvicem in numero et divisionibus sperarum, motu quoque et circulis Iere,
sicut dixisse Ptolomeus in sua probatur Sintaxi.
34. Eol. 49v: 'Mensura autem maioris lati planetarum et Iune in septentrionem et aus-
trum a circulo signorum est Iune quidem .d. graduum .n.e. sex., per mensuram qua
dividitur magnus circulus in .x.p. partes, Saturni trium, Iovis duorum, Martis in septen-
trione .d. graduum .l. sex., in austrum .g. graduum, Veneris, sicut dicit Ptolomeus in sua
Sintaxi, .I. graduum .b. sex., sicut autem alii dixerunt astronomi, novem, Mercurii .d..
35. Eol. 49v: 'Ima vero spera Saturni hec movetur in orientem tardo motu qui est in .t.
annis uno gradu super duos fxos polos. Hii sunt duo poli circuli signorum quem Solis
centrum circinat, sicut dicit Ptolomeus sua et precedentium ratione inventum.
36. Eor an account oI the haya tradition seeY. T. Iangermann`s introduction to his edi-
tion and translation oI Ibn al-Haytham, On the Conhguration of the World, New York
and Iondon, 1990, especially pp. 2534, and A. I. Sabra, 'Confguring the Universe:
Aporetic, Problem Solving, and Kinematic Modeling as Themes oI Arabic Astronomy,
Perspectives on Science, 1998, pp. 288330.
37. Eol. 38r: 'Verum cum in aliis Arabem quendam plurimum secuti sumus, in hoc
quoque per multum sequemur, licet quedam de sperarum numero et rotunditatum inve-
nerimus et de circulis quidem et inclinationibus planetarum vera perstrinxit a quibus
sperarum numerus dissonat.
38. See items 3 and 4 above. This is possibly a mistake Ior 66 years, al-Battn`s value,
unless the author has conIused the precession rate with the small Iraction oI a day
(1/106) by which a year is less than 365 1/4 days long, according to al-SI (see J.-M.
Millas-Vallicrosa, 'El magisterio astronomica, n. 59 below, p. 314).
39. Eol. 29r: 'Motus autem huius spere speralis circuitus est ad orientem ab occidente
super duos fxos polos et fxum polorum radium. Non sunt autem hii poli aut radius poli
mundi aut radius, set horum uterque a sibi proximo mundi polo .l.c. gradibus .m.e. sex.
distat.
40. John oI Seville`s translation, printed Paris, 1546, p. 16 and in MS Elorence, Con.
soppr. J.II.10, Iol. 154r: 'Et est secundum quod invenit Ptholomeus .xxiii. graduum et
.li. minutorum, cum Iuerit circulus .ccclx. graduum, probatione autem certissima qua
probavit Almenon . . . et convenerunt in ea plures sapientes esse .xxiii. graduum et .xxx.
minutorum. The errors in this text can be corrected Irom the translation by Gerard oI
Cremona (ed. R. Campani, Il libro dellaggrega:ione delle stelle, Citta di Castello,
1910, p. 229): ' . . . secundum considerationem vero consideratam et expertam quam
Johannes flius Almansoris (1his is Yahy ibn Ab Mansr, who was one of al-Mamns
correctors) consideravit in diebus Maimonis et convenit in ea numerus sapientum est
23 gradus et 35 minuta, and by the Iatin translation accompanying J. Golius`s edi-
tion (Muhammedis hl. Ketiri . . . qui vulgo Alfragani dicitur, Elementa astronomica,
Amsterdam, 1669, p. 18): 'et vero juxta dimensionem quam probatum (al-mumtahan)
vocant, quamque piae memoriae Almamon institui jussit, adhibitis et eam rem viris
doctis compluribus, ea declinatio continet gradus 23, et minuta 35.
41. See the preIace to the second part oI the Regalis dispositio, in which Stephen apol-
ogizes Ior leaving certain terms in Arabic: 'Malui igitur paulo infrmus videri quam
scientiam non transIerre ('I preIerred, thereIore, to seem a little infrm than not to
transmit knowledge). The preIaces to the Regalis dispositio are edited in the article
mentioned in the Iollowing note.
42. The identity oI the two Stephens is explored in detail in my 'Antioch as a Iink
between Arabic and Iatin Culture in the TwelIth and Thirteenth Centuries, in Occident
et Proche-Orient. contacts scientihques au temps des Croisades, ed. I. Draelants, A.
Tihon, and B. van den Abeele, Iouvain-la-Neuve, 2000, pp. 178.
43. That both this date and the dates oI the translation oI the Regalis dispositio are
given in terms oI 'a passione domini, constitutes another link between the two works.
It does not imply that the 33 years oI Christ`s liIe have to be added to the dates; the argu-
ments Ior this are given in my 'Antioch as a link . . ., Appendix III.
44. R. W. Hunt, 'Stephen oI Antioch, in Medieval and Renaissance Studies, 6, 1950,
p. 172173.
45. 'Stephanon autem quidam Pisanus ad illas partes ivit et linguam illam addiscens
eam |Practica| ex toto transtulit (MS ErIurt, WissenschaItliche Bibliothek, Amplon.
O 62, I. 50r).
46. MS British Iibrary, Sloane 2426, Iol. 8r (within a 1heorica attributed by a later
scribe to Cophon): 'Stephanus nepos patriarche Antiochensis. This preIace is edited
in C. Singer, 'A Iegend oI Salerno: How Constantine the AIrican Brought the Art oI
Medicine to the Christians, 1he Johns Hopkins Hospital Bulletin, 28, 1917, pp. 6469
(see p. 67a).
47. A comparison between the translating techniques and the terminology used in the
Dresden Almagest and the Regalis dispositio might reveal Iurther links. SuIfce to say
here that both works give elaborate details concerning the author and translator, and use
the word 'sermo, in preIerence to 'liber, Ior the constituent books ( Arabic maqla).
In the explicit to the Iourth book oI the Dresden Almagest the word 'liber has been
expunged, and 'sermo has been written in its place.
48. See Burnett, Antioch as a link . . . `, passim.
49. Eol. 2r: 'Quoniam in canonem astronomiHA quas proposueramus regularum exsequ-
to tractatu promissum exsolvimus, secundum hoc opus . . . aggredior.
50. Eol. 44v: 'Hence it is that, in fnding their latitude in the rules for the canon, we
have realized that a certain latitude is taken . . . ('Inde est quod in eorum inveniendo
lato in canonis regulis quoddam latum sumi precepimus). Italics mine.
51. E.g., Iol. 46r: In quarte tabula postquam rectifcata Iuerit . . . `; Iol. 46v: Quare
autem quarte tabule nec (?) tercia auIeratur sepe quinta iungitur, evidens hoc modo fet
ratio. Secunde tabule numerus ab obliqua longinquitate circuli extra centrum usque .s.
| 90| Iere decrescit, exinde usque .t.r. | 180| recrescens.`
52. They are unlikely to be the Regulae in Canonem Astronomiae in Elorence, Bibl.
Naz., Con. Soppr. J.II.10, I. 235ra239va (227r231v), as maintained by R. Iemay
in De la Scolastique a lhistoire par le truchement de la philologie. itineraire dun
medieviste entre Europe et Islam, in La diffusione delle scien:e islamiche nel medio evo
europeo, Convegno internazionale dellÀccademia Nazionale dei Iincei, Rome, 1987,
pp. 399535 (see p. 472), not only because these rules are not accompanied by tables,
but also because the astronomical terminology is completely diIIerent. It is true that
both texts criticize the opinion among the Iatins that retrogradation is caused by the
attraction oI the rays oI the Sun (Elorence MS, Iol. 238/230rava, Iiber Mamonis, Iol.
39r, preIace to the Iourth book), but the language oI the criticisms is diIIerent and the
criticised opinion was widespread.
53. Henry Bate, Descriptio instrumenti pro equatione planetarum, which Iollows his
Magistralis compositio astrolabii (1274), printed with Liber Abraham iudei de nativ-
itatibus, Venice, 1485, sig. d4 rectoverso: Ptholomeus vero: et Geber: Albategni:
Abrahamque Iudeus: et Aophius ceteri quoque magistri probationum et maxime orien-
tales astronomi motus planetarum secundum nonam speram considerantes radices suas
super hoc Iundauerunt: et hoc patet in tabulis Ptholomei: Albategni et Abrahe: in tabulis
Pisanis vuintoniensibus et aliis.` The punctuation oI the Renaissance printing has been
careIully observed, and corresponds to that oI MS OxIord, Digby 48, Iols 152v155v.
54. British Iibrary, Arundel 377, Iol. 56v; see Burnett, 1he Introduction of Arabic
Learning into England, Iondon, 1997, pp. 4849. The tables are not given in this copy.
55. The implications oI this coincidence are discussed in my '!Abd al-Mash oI
Winchester.`
56. Eor the arguments Ior believing that the Iatin tables preserve those oI al-SI see
R. Mercier, The Iost Zj oI al-SI in the TwelIth Century Tables Ior Iondon and Pisa,`
in Lectures from the Conference on al-Sf and Ibn al-Nafs, Beirut and Damascus,
1991, pp. 3872. Al-SI served several members oI the Byid dynasty and is attested
in Dinawar and IsIahan.
57. I discount (1) the Iossilized` Iorms in which a Iatin translator or scribe seems to
be copying Eastern Iorms Irom Arabic manuscripts without understanding their signi-
fcance (though these are valuable as giving evidence Ior which Iorms were used in
Arabic manuscripts), and (2) the instances in which a Iatin scribe gives the series oI
Eastern Iorms as an alternative to the Western Iorms, the latter oI which he uses exclu-
sively. Eor a Iuller account oI the use oI Eastern Iorms in Iatin manuscripts and their
Arabic exemplars, see C. Burnett, 'Indian Numerals in the Mediterranean Basin in the
TwelIth Century, in From China to Paris. 2000 Years 1ransmission of Mathematical
Ideas, ed. Y. Dold-Samplonius et al. Stuttgart, 2002, pp.237288.
58. British Iibrary, Cotton Vespasian A.II, Iol. 40v: Ut ait philosophorum sibi con-
temporaneorum Habraham magister noster egregius quo dictante et hanc dispositionem
astrolabii conscripsimus.`
59. See J.-M. Millas-Vallicrosa, El magisterio astronomica de Abraham ibn Ezra
en la Europa latina,` in id., Estudios sobre historia de la ciencia espaola, Barcelona,
1949, pp. 289347 and the same author`s edition oI the Fundamenta tabularum. El
libro de los fundamentos de las 1ablas astronomicas de R. Abraham Ibn E:ra, Madrid
and Barcelona, 1947. This publication includes illustrations oI the numeral Iorms in the
manuscripts.
60. Shlomo Sela, in a valuable article on Ibn Ezra`s versions oI astronomical tables and
instructions Ior their use, adduces arguments Ior Abraham being in Pisa itselI: see his
Contactos cientifcos entre judios y cristianos en el siglo XII: el caso del Libro de las
tablas astronomicas de Abraham ibn Ezra en su version latina y hebrea,` Miscelanea de
estudios arabes y hebraicos, seccion Hebreo, 45, 1996, pp. 185222 (see pp. 212222).
Eurther inIormation is given in id., Algunos puntos de contacto entre el Libro de las
tablas astronomicas en su version latina y las obras literarias hebreas de Abraham ibn
Ezra,` Miscelanea de estudios arabes y hebraicos, seccion Hebreo, 46, 1997, pp. 37
56 and Astrology and Biblical Exegesis in Abraham ibn E:ras 1hought (in Hebrew),
Ramat-Gan, 1999.
61. Eor a study oI this manuscript see C. Burnett, Iatin Alphanumerical Notation
and Annotation in Italian in the TwelIth Century: MS Iondon, British Iibrary, Harley
5402`, in Sic Itur ad Astra, EestschriIt Paul Kunitzsch, ed. M. Eolkerts and R. Iorch,
Wiesbaden, 2000, pp. 7990; this includes editions oI the mixed Italian-Iatin texts and
the astrological table mentioned in n. 64 below.
62. One may compare the opening oI these instructionsSciatis quod tabule iste Iacte
sunt super ciuitas luce. Et da angle daòccidente usque ad illam abet (i.e., habet) G
radus> .xxxiv;with the statement by Jose Bonfls that Ibn Ezra in the tables which
he composed in Iucca in the land oI Iombardy, aIfrmed that the longitude oI Iucca,
that is, the distance Irom the West, is 34 degrees`: see Sela, Contactos cientifcos,` p.
206.
63. B. Migliorini, in Storia della lingua italiana, Elorence, 1985, p. 111. I am grateIul
to Erancesca Ziino and Iaura Iepschi Ior advice on early examples oI Italian.
64. This table occurs elsewhere only in a later manuscript, PommersIelden 66, Iol.
84r.
65. See Antioch as a Iink . . .`, Appendix II.
66. Two sets oI numeral Iorms are given in the margin oI I. 1v oI Paris, BNE, hebreu
1052 (fIteenth century); in the text oI this manuscript, however, and in the copies oI
Sefer ha-Mispar in MSS Paris, BNE, hebreu 1049, 1050 and 1051, only the Western
Iorms are used. I am grateIul to Tzvi Iangermann Ior alerting my attention to these
manuscripts.
67. See quotation in n. 53 above. On the other hand, MS Cambridge, University
Iibrary, Kk.I.1, Iol. 145v mentions the tabule mediorum cursuum Solis ad meridiem
Winton. ab Abrahamo condite` (I owe this reIerence to Eritz Saaby Pedersen).
68. The tables are not mentioned, as Iar as I know, in any Arabic text, though there
remains the problems oI how, in the Fundamenta tabularum, they are described as
secundum Azof compositum, et Azarchelis probatio Iuit in anno 482 ab helligera` (El
libro de los fundamentos, ed. J. M. Millas Vallicrosa, p. 87). With this should be com-
pared a passage in the Sefer ha-Olam in which Abraham says that al-Zarqlluh`s value
Ior the amount by which the year Ialls short oI 365.25 days is the same as that oI al-SI,
without explicitly saying that the later astronomer was indebted to the earlier; I owe this
inIormation, once again, to Tzvi Iangermann.
69. See C. Burnett, A Group oI Arabic-Iatin Translators Working in Northern Spain in
the mid-twelIth Century,` Journal of the Royal Asiatic Society, year 1977, pp. 62108.
70. See R. Iorch, 'The Treatise on the Astrolabe by RudolI oI Bruges, in Between
Demonstration and Imagination. Essays in the History of Science and Philosophy
Presented to John D. North, ed. I. Nauta and A. Vanderjagt (Ieiden, 1999), pp. 55
100. (This article includes an edition oI the text.)
71. That is, in two copies oI Hugo`s translation oI Ibn al-Muthann`s commentary on
al-Khwrizm`s astronomical tables: Cambridge, Gonville and Caius, 456/394, Iol. 73r
and OxIord, Bodleian Iibrary, Selden, Arch. B, 34, Iols 32v33r; the third copy, ibid.,
Savile, 15, does not include the tables.
72. Hugo`s translation oI a work on geomancy, in OxIord, Bodleian, Digby 50, Iols
94r101r.
73. See Burnett, A Group oI Arabic-Iatin Translators.`
74. Eor a thorough description oI the manuscripts, see V. Rose, Jer:eichniss der latein-
ischen Handschriften . . . der Koniglichen Bibliothek :u Berlin, II, 3, Berlin, 1905, pp.
11771185. The Chartres provenance is suggested in a modern note on the cover oI the
manuscript itselI (Rose gives 'Erance as the provenance). Eor the inhabitants oI Pisa
and Angers, see Iol. 27r: 'Et scias quod iste tabule conposite sunt secundum meridiem
Pissanorum quorum longitudo occidentalis 33 graduum est, Andegavensium vero lon-
gitudo occidentalis 24 graduum; on Iol. 30r there is a 'Tabula diversitatis aspectus lune
ad longitudinemToleti 39 graduum 54 minutorum et eius hore 14 minuta horarum 51;
Paris is mentioned on Iol. 1r: 'nos sumus Parisius existentes. I am grateIul to Raymond
Mercier Ior some readings oI this manuscript.
75. Eor the diIIusion oI these manuscripts in Normandy and England see C. Burnett,
1he Introduction of Arabic Learning into England, pp. 4660.
76. Haskins, Studies, pp. 157163, J. E. Murdoch, 'Euclides Graeco-Iatin, Harvard
Studies in Classical Philology, 71, 1966, pp. 249302 (see pp. 263270; the date 1165
is proposed on p. 269).
II
1ranstormations ot Greek Optics
The section on 'the science oI optics (ilm al-manzir) in al-Erb`s (d. 339/
950) Catalogue of the Sciences (Ihs al-ulm)a work representing the state
oI many felds up to about the mid-4th/10th centurycovers fve distinct sub-
jects within the Iull text oI its Arabic edition (hereaIter I):
1
the need Ior such an
autonomous science and its distinctions Irom geometry (I.1), its Iunction and
methods (I.2), its applications and instruments (I.3), its assumptions and mech-
anisms (I.4), and its domains oI inquiry (I.5).
2
The respective passages, on the
other hand, cover their intended subjects all with reIerence to some aspect oI
'appearances (their veracity in I.1, justifability in I.2, Iallibility in I.3, and
elements and eIIects in I.4I.5), the appearances involved being in most cases
those other than commonly experienced, either in terms oI perceived shapes or
sizes (I.1 and I.2) or in the Iorm oI appearance through mediums other than air
(I.4 and I.5).
Erom the standpoint oI the history oI optics beIore 339/950, the long pas-
sage is an indispensable source Ior providing a most ftting outline and time-
Irame Ior discussion: on the one hand, it contains evidence on the state oI early
Arabic optics: its sources and problems, traditions and orientations, and bound-
aries and methods,
3
along with the terminological aspects oI the 'appearances
that it covers; on the other hand, its treatment oI optics Irom the perspectives
originally intended (its classifcation as a science and a mathematical science)
deserves close attention, iI only Ior the partial, and potentially misleading,
nature oI 'evidence provided by it, Irom a historical perspective.
Take the case oI the opening passage (I.1), which addresses the problem
oI the veracity oI appearances with reIerence to the age-old example oI the cir-
cular appearance oI Iar rectangular objects (see below). It is true that what that
passage contains provides valuable historical evidence Ior the identifcation oI
one oI the most central problems oI early optics (the most central, iI one goes
by the prominent place it occupies in that passage alone). But it is also true that
what the passage does not contain (in this case, reIerence to other central prob-
lems oI optics at this early period), is, at best, incomplete in the light oI what
is available through other early sources. Whether al-Erb is not Iully aware
3
1nr M:x. Asrrc+s or Arrr:r:xcrs:
Ar:iic Or+ics +o ,o :i
Llaheh Kheirandish
Li:nrn Knrir:xiisn 36 1nr M:x. Asrrc+s or Arrr:r:xcrs: Ar:iic Or+ics +o ,o :i 37
oI the optical problems oI his time (including problems as central as clarity oI
vision, on how clear something is seen, rather than how real the nature oI its
appearance is), or it is the case that he is not reIerring to these in the context oI
his intended discussion (in this case, distinguishing between the two mathemat-
ical sciences oI optics and geometry, with the same objects but diIIerent Iunc-
tions), one may argue that the fve passages, marked here I.1I.5, are all notable
Ior both what they contain and what they do not. Indeed, what is contained in
the texts oI the middle and closing passages too is quite refective oI many
aspects oI 'the science oI aspects(ilm al-manzir): in the case oI the second
and third passages (I.2 and I.3), its methodological and practical aspects, and in
the case oI the Iourth and fIth passages (I.4 and I.5), its theoretical and physi-
cal aspects respectively. It is the purpose oI the present essay to include a dis-
cussion oI what is not contained in all the respective passages (I.1I.5) as well,
with a historical analysis oI their Iull texts in the light oI early optical sources
previously treated by the present author Irom other perspectives.
4
What Iollows is an individual treatment oI the passages Iorming the
optics chapter oI al-Erb`s Catalogue in historical context, as the basis Ior
the argument that its coverage oI the 'science oI aspects (ilm al-manzir) in
general and oI 'appearances (m yazhar) in particular, while extensive and
inIormative, is Iar Irom exhaustive oI the Iull range oI problems and concepts
present within the manzir (optics), mary (catoptrics), and misha (survey-
ing) core oI the early Arabic optical tradition.
The 'manzir tradition, which in its earliest stages was primarily
Arabic,
5
is a rich tradition where the Euclidean concept oI 'appearance (Irom
oi vro0oi),
6
takes on many aspects oI its own, Irom what 'appears to sight
(yazhar f al-basar), 'seen or 'seen as (yur), or else 'viewed (yunzar.
Irom the same verbal root as manzir)
7
in al-Erb (I.1I.2), to what appears
(zahara), 'seen (yubsar), 'thought to be seen (yuzannu annahu yur), 'imag-
ined to be seen (qad yutawahhamu an yubsar), 'more accurately seen (asdaqu
ruyatan)and outside oI the Euclidean text, also what sight or the eye (basar)
'perceives (yudrik).
8
The manzir tradition is also a versatile and complex tra-
dition, one that even iI the particular 'agenda oI al-Erb`s Catalogue allowed
Ior the Iull coverage oI its versatility (I.3I.5), it could not possibly be expected
to refect the many shades oI its complexities.
The misha (surveying) and mary (catoptrics) components oI the
manzir tradition are similarly illuminated by al-Erb`s discussions, but
understandably limited in their scopes oI representation: The case oI misha is,
despite the involvement oI a wide range oI methods (determination oI heights,
widths, and lengths) and objects (Irom trees, walls, valleys, and rivers to moun-
tains and the heavenly bodies) in the corresponding al-Erb passage (I.3),
still short oI a range oI problems posed, not only by the one condition men-
tioned in passingthat oI 'sight (basar) Ialling on (yaqau al) objectsbut
especially by the most common (and as we shall see, problematic) instrument
Ior surveying it Iails to mentiona plane mirror. As Ior mary, treated by al-
Erb at once as an integral part and separate branch oI optics (I.5), the rather
unusual categorization oI 'indirect (ghayr mustaqm) rays, frst as 'defected
(munatif), 'reversed (munakis) and 'bent (munkasir) next to 'direct
(mustaqm) rays, and then, to the exclusion oI refracted rays,
9
is both wanting
and conIusing, this time to be examined in the light oI the largely unsettled,
and particularly problematic state oI Arabic optics during the earliest phases
oI its development.
I Vrr:ci+. :xi Accur:c. or Visiox
I.1: 'The science oI optics (ilm al-manzir) investigates the same things as
does the science oI geometry, such as fgures, magnitudes, order, position,
equality and inequality, but not in so Iar as these exist in abstract lines, sur-
Iaces and solids, whereas geometry investigates them insoIar as they exist
in abstract lines, surIaces and solids. Thus geometrical investigation is more
general. But there was a need for a separate science of optics, although |its
objects| are included among the objects oI geometry, because many of the
things which are proved to be of a certain shape or position or order or
the like, acquire opposite properties when they become obfects of vision.
thus obfects which are really square are seen (r--y) as circular when seen
(n-z-r) from a distance and equal objects appear to be unequal and unequal
ones appear equal, and many objects which are placed in the same plane
appear to be some lower and some higher, and many Ioreground objects
appear to be Iarther back. And such things are many. . . .
The opening passage (I.1) poses the question oI the need Ior a science
oI optics (ilm al-manzir) in terms oI the problem oI the truth correspondence
(al-haqqa) oI 'what appears (m yazhar) a certain way. In discussing the
problem oI the veracity oI appearances, the present passage addresses a histori-
cally important aspect oI appearances, while leaving out another pro-minent
aspect: that oI the accuracy oI vision (sidq al-ruya).
In the selection above, the discussion oI the appearance oI visible pro-
perties other than what they really are, begins with the case oI shapes, and the
circular appearance oI Iar rectangular objects, a problem that is, by itselI, nota-
ble Ior the extensive and diverse explanations it had received beginning with
the Greek sources. The author oI the passage in the corpus Problemata Physica
attributed to Aristotle, treats the case oI a 'square (trtpoymvov) |literally, a
fgure with Iour angles| that appears (oivrtoi) to have sundry angles, but
Irom a distance, 'looks like a circle,
10
an eIIect that is explained in terms oI
the 'cut-oII shape oI the angles owing to the uneven strength and distribu-
tion oI rays within the base oI the visual cone. There is also the proposition
in Euclid`s Optics that orthogonal magnitudes (op0oymvio ryr0p) appear
(oivrtoi) circular Irom aIar,
11
with a 'geometrical prooI based on the dis-
appearance oI the object`s angles beyond a certain distance (i.e., where visual
rays lose contact with objects).
In the Arabic tradition, various treatments oI the problem beIore al-
Erb include diIIerent versions oI the same Euclidean proposition, starting
with its supposed translation that 'sight (basar) by moving (yantaqil) Irom
one point on the object`s outline to another, skips some points in between (m
bayn).
12
Among other curious Iormulations, some emphasizing the temporal
aspects oI appearance and disappearance by pointing to angles as the hrst part
oI a fgure to disappear,
13
there is an apparently early treatment by Ahmad ibn
s
14
in his book on optics (H in Appendix), that is oI interest on many levels.
On one level, it is a 'critique oI the Euclidean proposition on both physi-
cal and epistemological grounds, the assumption oI the 'leap (tafra) oI sight
and the inability to perceive the fgure`s real shape (al-shakl al-haqq) being
called 'the most amazing oI amazements (min afab al-afab);
15
on another
level, geometrical demonstrations are oIIered Ior a variety oI 'angular fgures
appearing as circles Irom a distance, demonstrations in which the 'circle itselI
is no longer conceived as a fgure without angles, but rather, as a fgure equi-
distant Irom its own center.
16
The treatment is also oI interest beyond what was or was not available
to an apparently early optical author such as Ahmad ibn s.
17
In a book, bear-
ing the striking title oI Kitb al-1arb wa al-tadwr (Book of Rectangularity
and Circularity), al-Jhiz (d. 255/868869), the Iamous ninth-century literary
fgure and mutakallim, makes a number oI relevant and revealing remarks,
including the statement that: 'iI we say a quadrilateral structure is seen Irom
a distance as circular, then perhaps the sun is polygon-shaped (mudallaa)
|musallaba (cross-shaped) in the published edition)| and the stars are quadri-
lateral (murabbaa).
18
The addressee, a certain Ahmad ibn Abd al-Wahhb,
whom al-Jhiz describes as 'quadrilateral (murabba) in Iorm, but estimated
as round (mudawwar) due to his fgure!, may be compared to Ahmad ibn s
beyond an overlap oI names and apparent intellectual orientations.
19
II it is no
incident that Ahmad Ibn s devotes exceptionally long and thorough discus-
sions in his Optics to the Euclidean proposition on the circular appearance oI
Iar rectangular objects, then, the problem oI 'the circle and the square would
take us much beyond questions oI identity and onto the exciting milieu oI the
3rd/9th century itselI.
20
But Ior early Arabic optics, the problem oI the appearance oI shapes
other than what they actually are, was not limited to the case oI the circle and
the square, just as the problem oI 'appearance itselI was not limited to the
determination oI its truth correspondence. Al-Kind (d. ca. 257/870), himselI
the subject oI attack by none other than al-Jhiz, Iocused, Ior example, on the
case oI the rectilinear appearance oI circular shapes as part oI his arguments
against intromission in the De aspectibus (A in Appendix),
21
and spoke, not so
much oI veracity oI appearances, but rather, oI clarity and accuracy oI vision,
using two distinct expressions: 'clearer (abyan) and 'more accurate (asdaq)
vision (ruya), in his later 1aqwm (Rectihcation), extant in Arabic
22
(Q in
Appendix).
'Clear and 'accurate vision were, in Iact, subjects treated extensively
and variously by other authors on optics: in the period beIore al-Erb, by Ibn
s (beIore 250/846?), Hunayn ibn Is-hq (d. 264/877) and Qust ibn Iq
(d. 300/912) in addition to al-Kind (d. ca. 257/870,), and in the centuries Iol-
lowing the composition oI Ihs al-ulm, by Ibn al-Haytham (d. ca. 432/1040),
Nasr al-Dn al-Ts (d. 672/1274) and Kaml al-Dn al-Eris (d. ca. 718/
1318). And yet, oI the two aspects oI appearances that were prominent in the
early optical tradition, namely veracity oI appearances (as presented through
the passages oI Ibn s and al-Jhiz), and clarity and accuracy oI vision (as
developed through the transmission oI another Euclidean proposition), it is the
latter that is leIt out oI al-Erb`s account altogether, while the Iormer occu-
pies a prominent place, not just in al-Erb`s opening passage, but also in the
one immediately Iollowing it.
II us+iri:iiii+. :xi V:rir+. or Drxoxs+r:+ioxs
I.2: 'By means of this science discrimination (m-y-:) is made between what
is seen (z-h-r) as different from what it truly is and what is seen (z-h-r) as it
truly is, and the reasons (s-b-b) why all this should be so are established by
certain (y-q-n) demonstrations. And with regard to all that can be subject to
visual error (gh-l-t) this science explains various devices (h-y-l) Ior avoiding
error and apprehending what the seen thing truly is in respect to size, shape,
position, order, and all that can be mistaken by sight (b-s-r) . . .
While the opening passage (I.1) is about the justifcation Ior having a
science oI optics as a separate branch Irom geometry (with which it shares
its subjects: fgures, magnitudes, order, position, equality and inequality), the
second passage (I.2) is about the justifcations oIIered by the science oI optics
itselI: not only is this science one that discriminates (yumayyi:) between what
literally 'appears to sight (yazhar f al-basar) according to its real (haqq)
properties and what does not; it is the science by which one establishes the
reasons (asbb) behind appearances (whether these are consistent with or at odds
with real properties), and why they are as such (lima hiya) by means oI 'cer-
tain demonstrations (barhn yaqniyya). The explanation oI various 'devices
(hiyal), this time Ior avoiding 'errors (ghalat) oI sight (basar), with respect to
size, shape, position, order, and all that may be subject to error
is, once again,

expressed in terms oI 'coincidence with the truth (yusdif al-haqqa).
In speciIying demonstrations as distinct Iorms oI explanation with reIer-
ence to appearances, the present passage well refects aspects oI the feld that
were central to the early optical tradition. Demonstrations were indeed among
the key Ieatures oI the science oI optics; but these were expressed in terms
that were much more specifc than al-Erb`s 'certain (yaqn) demonstra-
tions:
23
Demonstrations (barhn)oIten meaning mathematical prooIs to an
early fgure such as al-Erbcould be geometrical demonstrations (barhn
handasiyya demonstrationes geometricae), as reIerred to by al-Kind in
distinction Irom 'philosophical demonstrations (barhn falsahyya); there
were also such variations as 'demonstrations by lines (barhn khuttiyya),
in the words oI Qust ibn Iq with specifc reIerence to 'the science oI rays
(ilm al-shut), as there was the genre, 'illustration (mithl), itselI used by
someone like Ibn s, in the Iorm oI both 'geometrical illustration (mithl
handas) and 'sensible illustration (mithl hiss). The latter pair, correspond-
ing to Ibn Iq`s 'illustration by lines (mithl khutt) and 'sensible explana-
tion (bayn hiss) respectively, were to act as justifcation by means oI textual
prooIs (in the frst case), and experience-based setups (in the second case).
But what best reveals the orientation oI a wide range oI explanations in early
Arabic optics, is neither in the Galenic language oI 'demonstration by lines
(as Iound in book X oI De usu partium), nor in the Euclidean language oI geo-
metrical illustrations (as a Iormal division oI a Euclidean prooI), but rather,
in the Aristotelian language oI scientifc reasoning (as set out in Aristotle`s
Posterior Analytics).
24
With the Aristotelian distinction between knowledge oI Iact (to oti) and
oI the reason why (to oioti) in a well-known passage in the Posterior Analytics
(I.13: 78a3435), and especially, the reservation oI the privileged knowledge
oI the cause (or reasoned Iact) Ior the mathematician, and the optician (with
respect to inIerior sciences), a method oI providing 'demonstrations had
Iound its way quickly into optical texts. The terminology oI illa aIter the Aris-
totelian expression Ior 'knowledge by reasoning to oioti (ilm bi al-illa),
25
was accordingly more common in optics than the 'demonstration (burhn)
oI the Arabic title oI Posterior Analytics (Kitb al-Burhn), the text on which
many authors on optics wrote commentaries, including al-Erb.
26
So central
was the process oI reasoning to optics Irom early on, that the term illa was
even included in the Iull Arabic titles oI a Iew early optical texts.
27
Discussions on demonstrative sciences in general, and optics in partic-
ular, were in Iact transmitted through two Aristotelian works: The Posterior
Analytics and Physics were both widely circulated in Islamic lands, through
fgures such as al-Kind, Ibn Iq, and among al-Erb`s (d. 339/950) own
contemporaries, Abu `l-Hasan al-mir (d. 382/992), who all wrote on optics
and on methodology in diIIerent Iorms and capacities:
28
Al-Kind, charac-
terized optics as a 'science in which geometrical demonstrations proceed
in accordance with the requirements oI physical things, and supplemented
geometrical demonstrations throughout his text with others Irom the world oI
experience.
29
Qust ibn Iq, on the other hand, spoke explicitly about the
'cooperation (ishtirk) between natural philosophy, Irom which we acquire
sense perception (idrk hiss), and geometry and its demonstrations through
lines (barhn khuttiyya), and about the incomparable 'excellence oI this
coming together in the science oI rays (ilm al-shut). Ibn s similarly
used, in addition to Euclidean illustration (mithl handas), 'sensible illustra-
tion (mithl hiss), with reIerence to devices like tubes (unbb).
30
In this way,
aspects oI appearances that became subject to demonstrations oIten involved
repeated demonstrations, typically supporting mathematically-based argu-
ments with experience-based set ups. The subjects oI demonstrations them-
selves were Iocused more Irequently on problems such as clarity or accuracy
oI vision, than the reality or veracity oI appearances.
The dominance oI the subject oI visual accuracy in the Arabic optical
tradition may be measured by the abundance oI demonstrations that involve
quantifed treatments oI visual clarity in terms oI the amount oI radiation
involved. Al-Kind, who treats the subject in more than one place, demon-
strates geometrically that what determines the eIIectiveness oI the central
region oI the visual feld is the Ialling oI the greatest amount oI radiation on
the central region.
31
Other cases include demonstrations by Ibn Iq, where
the amount oI radiation Ialling on a region determines how many objects are
seen, rather than how clear they are seen.
32
In yet another variation by Ibn
s, alternative demonstrations Ior the privileged position oI the visual axis
Ior 'more accurate vision (asdaqu ruyatan) is based on there being 'more oI
the ray Ialling (m waqaa min al-shu akthar).
33
All these demonstrations
(including those involving repetitions) are somehow related to the problematic
transmission oI Euclidean optics,
34
which in this case amounts to a missing
causal premise in the proposition demonstrating that 'that on which more oI
the ray Ialls is seen more accurately (m waqaa alayhi al-shuu akthara fa
ruyatuhu asdaqu).
III Vrrs:+iii+. :xi l:iiiiiii+. or Arriic:+ioxs
I.3: By means of this art (sina) too, one can determine the si:e of distant
and inaccessible bodies, the magnitudes of their distances from us and their
distances from one another. Examples are. the heights of tall trees and walls
and the widths of valleys and rivers, even the heights of mountains and the
depths of valleys, provided that sight (b-s-r) can reach (w-q-) their limits, the
distances of clouds and other obfects from our location and above any place
on the earth . . . In general, every visible magnitude oI which the size or dis-
tance Irom something else we seek to know, |can be determined| sometimes
by means oI instruments which are made Ior guiding the passage oI sight
(b-s-r) so that it may not err, and sometimes without such instruments.
The passage on the applications oI optics as an art (sina) rather than
its explanations as a science (ilm) (I.3) points to an important, and oIten over-
looked, aspect oI the discipline: The determination oI the size and distance
oI objects was, just as the explanation oI their appearance, an integral part
oI optics. In including among the applications oI the manzir tradition (i.e.,
optics), the misha problems, based on a set oI Iour propositions in Euclid`s
Optics (i.e., the 'surveying problems), al-Erb acknowledges an aspect oI
the discipline that was Iar Irom marginal to the practices oI early Arabic optics.
But the absence oI the one Euclidean proposition in that set where the deter-
mination oI a magnitude involves the use oI a plane mirror, gives al-Erb no
occasion to move Irom the discussion oI instruments and errors to the more
problematic aspects oI each case, the mirror and the principle oI refection
respectively.
The omission is all the more surprising in the light oI the thoroughness
oI al-Erb`s account regarding the many applications oI 'the art (sina).
The inclusion oI applications such as the determination oI unknown heights,
depths and lengths Irom known values through measurements and calculations
is indeed close to the common practices oI the discipline, both beIore and aIter
al-Erb, as is the extension oI the objects and distances involved to include
natural objects and far distances. A good example oI an early work that com-
bined the methods oI magnitude determination with principles in optics is a
short treatise by Sinn ibn al-Eath (ca. 4th/10th cent.) with the selI-explana-
tory title oI al-Misht al-manziriyya (Z in Appendix).
35
Another text that
extended its methods oI magnitude determination to the case oI natural objects
and Iar distances, is a treatise by al-Kind, entitled On Clarihcation of Finding
Distances between an Observer and the Centers of Mountain Heights (Risla f
/dh wifdn abd m bayn al-nzir wa marki: amadat al-fibl (W in Appen-
dix).
36
Comparable works oI a period slightly later than al-Erb include two
short tracts bearing the name oI Ibn al-Haytham (d. ca. 432/1040), an author
much better known Ior his infuential Optics (Kitb al-Manzir):
37
one is On
the Determination of the Height of Upright Bodies, Mountains and Clouds
(Maqla f Marifat irtif al-ashkhs al-qima wa amidat al-fibl wa irtif
al-ghuym), and the other, On the Extraction of the Elevation of Mountains
(Qawl f Istikhrf amidat al-fibl).
38
But the misha tradition had close associations, not just with the manzir
tradition, most directly through the visual and solar rays involved in magnitude
determinations; it also had direct links with the mary tradition, as it involved
indirect Iorms oI radiation, Ior example, through the Euclidean proposition
involving a plane mirror Ior height determination. The close association is
clear Irom a title such as Mirtiyya (Related to Mirrors), an alternative title
Ior a treatise by Badr al-Dn al-Tabar (ca. 824/1421) called Height (Irtif),
39
in this case involving the determination oI the height oI lower objects, once
by means oI a 'rod (amd), and once by a plane mirror. The late author oI
the Mirtiyya reveals a useIul piece oI inIormation about the transmission oI
the methods involved, when he says that 'regarding the determination oI the
height oI tall objects by means oI a mirror, what the people oI the art (ahl-i-
sinat) may have oIIered on the subject, has not reached him or has not been
seen by him anywhere. The curious omission oI this aspect oI the misha
tradition Irom al-Erb`s extensive account oI the 'art, may be related to the
reportedly poor transmission oI surveying techniques using mirrors; but this is
something that must also be viewed in the light oI other problematic cases in
transmission.
The transmission oI the Euclidean propositions on height determina-
tion by means oI refecting visual rays is a case that had already taken a mis-
directed course as a result oI the Arabic terminology used Ior the principle oI
refection Irom a plane mirror. In the Arabic version oI that proposition, the
term refection was not translated into Arabic in the later standard Iorm iniks,
but rather, as initf, a Iorm also employed by al-Erb with reIerence to the
defection oI rays at a polished surIace such as mirrors, just as in both the early
and late Arabic versions oI Euclid`s Optics (3rd/9th and 7th/13th centuries
respectively).
40
The terminological twist is particularly surprising, because
the more standard term Ior refection (iniks) was used as early as in Hunayn
ibn Is-hq`s 1en 1reatises on the Eye, in the Arabic versions oI the Pseudo-
Euclidean De speculis and Aetius` Placita philosophorum,
41
as well as in late
works by Ts and his commentator, Tabar.
42
The exact source oI al-Erb`s
combined Iormulation where 'defected (munatifa) (i.e., refected) rays
become 'reversed (munakisa) (i.e., on their return paths Irom a mirror), is
not quite clear. Neither does the extensive literature on the subject contain
Irequent reIerences to 'bent (munkasira) rays or to 'seeing behind oneselI, as
we fnd in al-Erb`s next passage.
IV Lirxrx+s :xi Mrcn:xisxs or Visiox
I.4: Now all that can be looked at (n-z-r) and seen (r--y) is seen (r--y)
by means of a ray (sh--) that penetrates (n-f-dh) the air or any transpar-
ent body in contact (m-s-s) with our eyes (b-s-r) until it reaches (w-q-) the
obfect seen (n-z-r). And rays that pass through transparent bodies to a visible
obfect are either straight (mustaqma) or deected (munatifa) or reversed
(munakisa) or bent (munkasira). Straight rays are those that, having issued
(kh-r-f) Irom the eye, extend rectilinearly on the line (s-m-t) oI sight (b-s-r)
until they weaken and come to an end. . . . Defected rays are those that, hav-
ing passed out oI the eye, meet on their way, and beIore they weaken, a mirror
that precludes them Irom passing through in a straight line, thereby causing
them to be defected (-t-f) and turned (h-r-f) to one side oI the mirror. They
then extend in the direction into which they have turned towards the beholder
(n-z-r). Reversed rays are those that return Irom the mirror on the path they
traversed at frst, until they Iall (y-q-) on the body oI the beholder (n-z-r) Irom
whose eyes they have issued (kh-r-f), and it is by means oI this |kind oI| ray
that the beholder (n-z-r) sees (r--y) himselI. Bent rays are those that return
Irom the mirror towards the beholder (n-z-r) Irom whose eyes (b-s-r) they
have issued (kh-r-f), but extend obliquely beside him until they Iall (y-q-)
on something else behind the beholder or on his right or his leIt or above
him, and it is thus that we see (r--y) what lies behind or beside us.
Al-Erb`s detailed account oI the elements and mechanisms oI vision
points to complex problems in transmission: the elements oI vision, being in
this case, the Iour cases oI visual radiation, and the mechanisms, their respec-
tive Ieatures and Iunctions. By containing explanations that clearly represent a
combination oI the various Iormulations oI visual radiation and their distinct
Ieatures and Iunctions, the account well refects the problematic nature oI the
theoretical aspects oI the discipline. But by excluding some common Iormula-
tions, while including uncommon ones, the detailed account is still a Iaint
refection oI the range and complexity oI the problems characterizing the early
Arabic optical tradition.
Al-Erb`s account may be considered primarily Euclidean, though it is
neither identical with any oI the two most common Iormulations oI the Euclid-
ean visual-theory, nor comparable to any one particular variation to reveal the
exact source oI its own 'mixed Iormulations. Transmitted through a textual
tradition, noted Ior having been not just problematic, but physically deIec-
tive in both Greek and Arabic,
43
the Euclidean tradition itselI has a complex
history, beginning with the statement oI the Euclidean visual-ray hypothesis.
OI the two most distinct Iormulations oI that statement passing under the name
oI Euclid,
44
none are Iully represented in al-Erb`s version: 'a ray (shu)
passing through (yanfudh) a transparent body in contact (yumss) with our
eyes (basir) until it Ialls (waqaa) on the viewed object (manzr ilayh),
does not correspond exactly, with the elements and mechanisms oI vision as
described in the opening lines oI the Arabic versions oI Euclid`s Optics (Kitb
al-Manzir li-Uqldis, M in Appendix), and largely Iollowed in the late Arabic
versions oI the same text (i.e., Tss 1ahrr al-Manzir and Ibn Ab Jarda`s
1afrd al-Manzir); neither does it match the version Iound in the pseudo-
Euclidean De speculis (Kitb al-Manzir li-Uqldis, S in Appendix), the latter
corresponding almost word by word, and only in this part, to the Iormulations
oI al-Kind`s 1aqwm, and Ibn s`s al-Manzir wa al-mary al-muhriqa (H
and Q in Appendix, respectively).
45
A critical part Ior the discussion oI appearances, including al-Erb`s
distinct Iormulations, is the many Iorms oI the third Euclidean assumption
in the Optics that only those things upon which rays Iall are seen (opotoi).
It is not insignifcant that the corresponding Arabic verb Ior 'seeing
is,
in the Optics (M), yubsar, in the De speculis (S) yudrak and in al-Erb`s
account, yunzar and yur: 'All that can be looked at (yunzar ilayh) and seen
(yur), states al-Erb, 'is seen (yur) by means oI a ray (shu) that . . .
Ialls (waqaa) on the viewed object (manzr ilayh). It is clearly the case that
al-Erb`s Iormulation is Iree Irom the occurrence oI the terminology oI per-
ception (idrk), a conception that appears in similar terms in the statements oI
predecessors like al-Kind, Ibn s, and Ibn Iq (as part oI their reIormula-
tion oI the third Euclidean defnition), as well as the subsequent elaborations
oI Avicenna and Ibn al-Haytham.
46
In the case oI direct vision, thereIore, al-
Erb`s account oI 'appearances remains partial by virtue oI being strictly
Euclidean. This is a tradition in which various expressions occur Ior both the
passive and active modes that are involved in vision |what is viewed (yunzar),
looked at/seen (yubsar), or appears (yazhar); and what is seen (yur), thought
to be seen (yuzannu annahu yur), or imagined to be seen (qad yutawahhamu
an yubsar)|, with the similarly striking exception oI the term Ior perception
(idrk).
The Euclidean character oI al-Erb`s account oI indirect vision repre-
sents restrictions in other theoretical directions. This time, what is striking is
the absence oI any reIerence to vision through reIraction.
47
In the part oI the
passage on appearances through mediums (i.e., indirect vision) rather than
those through air (i.e., direct vision), everythingincluding the term 'initf
(later, standard Ior reIraction)is a reIerence to refection: visual radiation
passes out oI the eye (nhdh min al-basar) . . . meets a mirror (al-mirh), is
defected (tanatif) and turned (inharafat) to one side oI the mirror; and then,
the reversed (munakisa) rays return (tarfi) Irom the mirror on the path they
frst traversed. Clearly, al-Erb`s account, where al-shut al-munatifa
stand Ior defected rays (as in the Arabic Euclidean tradition), next to non-
Euclidean cases such as reversed (munakisa) or bent (munkasira) rays, lacks
mention oI reIracted rays in its most standard Iorm. And when we read in the
closing passage that ilm al-mary is the division within ilm al-manzir that
investigates what is visible through indirect (ghayr al-mustaqma) rays, this
does not include reIracted rays as the reIerence to transparent mediums such as
water or glass would have us believe.
V 1nr Moirs :xi Mriiuxs or Orrr:+iox
I.5: 'The medium that lies between the eye (b-s-r) and what is looked at
(n-z-r) is, in general, a transparent body, whether air, or water, or celestial
body or an earthly composite body such as glass and the like. And mir-
rors, which send back the rays and prevent them Irom rectilinearly passing
through, are either those made by us oI iron or the like, or they consist oI a
thick moist vapour, or water, or some other body similar to these. 1he sci-
ence of optics, then, inquires into all that is looked at (n-z-r) and seen (r--y)
by means of these four rays {straight (mustaqma), defected (munatifa),
reversed (munakisa), and bent (munkasira)} and into every kind of mirror
and all that pertains to the obfect of vision (n-z-r). It is divided into two
parts, the hrst of which investigates what is visible (n-z-r) through rectilinear
rays, and the second is visible through non-rectilinear rays, and this [latter]
is specially called the science of mirrors (ilm al-mary).
The closing passage (I.5) concludes with optics` domains oI inquiry
(fahs), as distinct Irom the subjects oI investigation oI the opening lines (I.1)
that were meant to make optics itselI distinct Irom the 'more general feld
oI geometry. In extending the discussion oI indirect or mediated appearances
(introduced in the previous passage), to cases involving refecting surIaces
other than mirrors (thick moist, vapor, or water) on the one hand, and medi-
ation oI transparent bodies other than air ('water, celestial sphere and earthly
composite bodies like glass and the like) on the other, the fnal passage pro-
vides a IaithIul account oI mediated appearances, insoIar as it distinguishes
between mediation through opaque bodies (such as mirrors) and transparent
bodies (water or glass). But insoIar as the passage includes one type oI medi-
ted appearance (through refection) to the exclusion oI the other (through
reIraction), the account does not Iully represent the stage reached by optical
writings oI the period beIore or during al-Erb`s compositions.
The two-Iold and hierarchical division oI the discipline, the frst (ilm
al-manzir) investigating what is viewed (yunzar ilayh) through direct
(mustaqma) rays, and the second, through indirect (ghayr mustaqm) rays,
must itselI be understood in terms oI the mediumsand not just modesoI
propagation. On the one hand, mustaqma and ghayr mustaqma represent
direct ( unmediated) and indirect ( mediated), rather than rectilinear and
non-rectilinear (i.e., in terms oI a medium oI propagation other than air, rather
than the rays being in the mode oI rectilinearity), simply because the rays
involved in appearances through a mirror, Ior example, are both rectilinear (
not bent or curved) and indirect ( changing course), this being true oI both
defected (munatifa) and reversed (munakisa) rays. On the other hand, the
part, mary in ilm al-mary, itselI commonly translated as 'the science oI
mirrors, can neither be reduced to mirrors, nor understood to exclude surIaces
now commonly considered reIractive mediums, especially since al-Erb`s
mary stands Ior 'defecting surIaces such as vapor or water thick enough to
produce such an eIIect.
The limited knowledge oI optical reIraction in general and oI the
treatment oI enlarged objects in water in particular has long been noted by
A. I. Sabra with reIerence to specifc works by al-Kind, Ahmad ibn s, even
early Ibn al-Haytham (d. ca. 432/1040), Ior their lack oI understanding oI the
phenomenon oI reIraction as Iound already in Ptolemy`s Optics.
48
Ironically
enough, the so-called 'pre-Ptolemaic stage represented by the astronomer`s
explanations oI the enlarged appearance oI bodies through mediums in terms
oI visual angles, is, in some sense, more advanced than the stage represented
by the optical tradition itselI, a tradition where the appearance oI the principle
oI refection in a single proposition oI Euclidean optics came and circulated
with the vocabulary oI initf to 'conIuse the phenomena oI refection and
reIraction in the works oI Nasr al-Dn al-Ts (d. 672/1274) and Qutb al-Dn
al-Shrz (d. 711/1311), and all the way up to Kaml al-Dn al-Eris (d. 718/
1318) who noted the puzzles involved.
49
The nonstandard terminology oI 'initf as 'refection (or refection)
has already been mentioned with reIerence to the part oI al-Erb`s previous
passage (I.4), where defected (munatifa) rays appear alongside the standard
Iormmunakisa, itselI used to mean reversed, rather than refected rays (i.e., the
returned rays on the same path oI incidence). That the applications oI the two
principles remained inconsistent is immediately clear Irom the relevant writ-
ings oI a late author such as Ts: in his short treatise, Iniks al-shut wa
initfuh, where the two terms are used jointly (and in some manuscript tran-
scriptions, also interchangeably), as well as in the Persian work, Shu, the word
'iniks is used Ior refection, while in Ts`s 1ahrr al-Manzir (Recension
oI Euclid`s Optics), the Iorm initf (by then, standard Ior reIraction) is still
the term used Ior refection Irom a mirror; this is all the more curious, because
initf appears in the frst oI these works as a Iorm oI reIraction a cone oI rays
undergoes at the surIace oI transparent bodies like still water, so that ':wiyat
al-initf is no longer Ts`s 'angle oI defection at the surIace oI polished
bodies such as mirrors (as in the 1ahrr), but rather, the angle that the reIracted
end oI the visual cone makes with the reIracting surIace, and in such a way that
it is still equal with the angle oI incidence (:wiyat al-shu).
50
With the treatment oI reIraction, in particular, the conIusion between
the principles oI refection and oI reIraction, and especially their nonstandard,
inconsistent, and orthographically comparable terminology, are Iactors to be
considered, in addition to the apparently poor transmission and circulation oI
Ptolemy`s relevant treatments. The important statement oI A. I. Sabra about
all the historical evidence pointing to the limited use oI Ptolemy`s Optics in
both Antiquity and the Islamic Middle Ages,
51
must then be combined with
no less qualifed statements that would also take into account, not just what
was transmitted, but also how whatever did get transmitted was transmitted.
The diIfculty is that historical evidence may successIully reveal the frst (the
what oI transmission), but not all historical evidence would reveal the second
(its how). In the case oI early Arabic optics, we are Iortunate to have a good
number oI texts, including al-Erb`s passage, that may still act as histori-
cal documents, to determine what sources or concepts, were transmitted up to
about the year AD 950. But it takes a close examination oI the available sources
Irom the perspective oI the transmitted terms and expressions, in addition to
the sources and concepts, and these through extant manuscripts in addition to
published editions, to determine the exact nature and manner oI the eIIect that
all oI these have had on the state oI Arabic optics during a critically important
stage in its development.
To conclude with remarks that take into account the entire passage and
overall plan oI the optics section in al-Erb`s Catalogue, it should be remem-
bered how optics is presented throughout that text: as an established scientifc
discipline (ilm) within the mathematical sciences (talm), supplied not just
with 'reasons (asbb) but 'certain (yaqn) demonstrations, and not just
with 'explanations (marifa) but also 'devices (hiyal), a discipline in search
oI 'conIormity with reality (yusdif al-haqqa), one that at once demystifes
and justifes 'what appears to sight (m yazhar f al-basar) both within and
beyond the ordinary realms oI vision. All this leaves little doubt about how al-
Erb conceived oI, or at least presented, the 'program oI the early optical
tradition.
But how close is such a 'program to the character oI the early opti-
cal tradition itselI ? There is no question that al-Erb`s account is extremely
valuable as a historical document, especially in the light oI the rarity oI such
general accounts on the early history oI any scientifc discipline; but with the
Iocus oI the text on disciplinary and pedagogic concerns, rather than histori-
cal, or even scientifc ones, the coverage oI the 'science oI aspects, remains
inexhaustive, in regards to the many aspects oI the feld, in terms oI both the
orientations and associations oI the discipline, and the Iorms and expressions
oI its concepts. Appearances are treated, in the opening passage, in terms oI
problems oI veracity to the exclusion oI the slightly diIIerent, and more com-
mon, themes oI clarity and accuracy; in the second passage, in terms oI dem-
onstrations to the exclusion oI their sense-perceptible dimensions; in the third
passage, in terms oI applications to the exclusion oI their more problematic
extensions; in the Iourth passage, in terms oI the elements and mechanisms oI
vision to the exclusion oI their multiple variations; and fnally in the closing
passage, in terms oI modes oI investigation to the exclusion oI all the modes
and mediums oI operation. The terminological aspects oI appearances are also
at once refective and restrictive, as the concept oI 'appearance itselI emerges
Irom the active involvement oI an observer (nzir) viewing (n-z-r) or seeing
(r--y), to the passive presence oI an object appearing (z-h-r)all to the exclu-
sion oI other Iorms, including the Iorm al-manzir, meaning appearances (and
not just visual rays), as in the 'science oI aspects (ilm al-manzir) itselI.
Einally, the Iew extant early Arabic texts examined in the present study,
themselves act as important historical documents in demonstrating that it is not
only the case that al-Erb`s coverage oI 'aspects and 'appearances, is not
Iully representative oI the concepts and problems oI the earlyArabic texts oIten
covering the very same items; it is also the case that such a coverage, leaves
out, in eIIect obscures, the extremely complex character oI an early tradition,
to whose 'unsettled aspects, as well as rich dimensions, al-Erb`s account
itselI is suIfcient testimony.
Arrrxiix
Primary Sources vithin the Early Arabic Optical 1radition
M ,1..! ..- _. .,|.i ,..:
Kitb Uqldis f Ikhtilf al-manzir Euclid`s Optics (Arabic version)
|Krause (1974); Sezgin (1974) GAS V; Kheirandish (1991, 1996). ed.
Kheirandish (1999): vol. 1: pp. 1225; Rashed (1997)|.
S .,|.,. ,! ,..:
Kitb al-Mirh li-Uqldis |Pseudo-| Euclidean De speculis (Arabic ver-
sion) |Sabra (1979); Kheirandish (1991, 1999); Rashed (1997); Iatin text:
Bjrnbo and Vogl (1900), pp. 97173; Theisen (1972)|.
H ..,-! .,,!, ,1..! ,..:
,.,i |. _. .,|.i ,:.. _|. _.,. , ..-i ,,i..:
Ahmad ibn s, Kitb al-Manzir wa al-mary al-muhriqa talf Ahmad
ibn /s al madhhab Uqldis f ilal al-basar (Arabic text) |Krause
(1936); Kheirandish (1991, 1996, 1999; Sabra (1989); Rashed (1997), in-
cludes edition oI section on Burning Mirror, Sabra and Kheirandish, edi-
tion in preparation|.
Q ..,-[ ., _i[ ..i _.-.[ , ,,.., ,.,, _,i ,..:
,1..!., ,,.,! .,..: _. .,|.,. _.i :.!, .a-i ,,,.. _.
Al-Kind, Kitb Ab Ysuf Yaqb ibn Ishq al-Kind il bad ikhwnihi f
1aqwm al-khata wa al-mushkilt allat li-Uqldis f Kitbihi al-mawsm
bi al-Nzir |al-Manzir| (Arabic text) |Marash (v. 19); Rashed (1997),
ed. pp. 162335|.
A :.,,|. .,...,i ,:,,i, ,1..! ..- |. _. ,..:
Al-Kind, De aspectibus ( Ikhtilf al-manzir?) (Arabic text extant
in Iatin) Kitb f Ilal ikhtilf al-manzir wa al-barhn al-handasiyya
alayh? De causis diversitatum aspectus et dandis demonstrationibus
geometricis super eas). |Bjrnbo-Vogl (1912); ed. Hugonnard-Roche, tr.
Jolivet, Sinaceur: Rashed (1997)|.
I ,1..! ..- . .,,! _. ,., .. |. _. ,..:
Qust ibn Iq, Kitb f Ilal m yaridu f al-mary min ikhtilf al-
manzir allafahu . . . Qust ibn Lq al-Ynn (Arabic text) |Gulchn
Man (13501971), vol. 8; Toomer (1976); Kheirandish (1991, 1996,
1999); Rashed (1997), ed. pp. 572646|.
W ,1..i ,, .. ...,i .-, _.., _. ..i -. , ,,.., .i..
.i.,,., .,.. ,,., .,. .. ,|., .,....i ,|., .,-i ....i ,:,.,
Al-Kind, F /dh wifdn abd m bayn al-nzir wa marki: amidat al-
fibl wa uluww amidatih . . . wa huwa yusamm Mrstus. (Arabic text)
|Ritter and Plessner (1932); Krause (1936); Brockelmann (1937) GAI
S I; Kheirandish (1991, 1999)|.
Z .,,1..! :.-..!
Sinn ibn al-Eath, al-Misht al-manziriyya (Arabic text) |King (1986a)
Cairo Catalogue, v. 2, p. 1030; King (1986b) Cairo Survey, p. 39;
Sezgin, (1979), GAS VII; Kheirandish (1991, 1999)|.
NOTES
1. Al-Erb, Ab Nasr, Ihs al-ulm, ed. Uthmn Amn | Osman Amine|, Cairo:
Iibrairie Anglo-Egyptienne, 1968 |earlier editions, Cairo, 1931, 1949; Arabic edi-
tion Irom Escorial manuscript by A. Gonzalez Palencia, Alfarabi Catalogo de las cien-
cias (ACIS), includes two medieval Iatin and a modern Spanish translation (Madrid,
1932)|.
2. The English translation oI the Iull passage is quoted Irom A. I. Sabra`s 1he Optics
of Ibn al-Haytham. Books IIII On Direct Jision, 1ranslated with Introduction and
Commentary, 2 vols., Warburg Institute, 1989, vol. 2, pp. lvilvii. The translation is
described by Sabra as 'made Irom a composite text constructed Irom two editions oI
the Arabic text and the Iatin version . . . in the absence oI a single satisIactory edi-
tion (emphasis and verbal root indications are added to passages quoted in this chapter
by the present author). An earlier English translation oI a large part oI the same pas-
sage based on the Arabic edition in ACIS (and using the Cairo edition) is published as
'The Science oI Aspects in Marshall Clagett, 'Some General Aspects oI Physics in the
Middle Ages, Isis, 1948, 39: 2944, pp. 3235.
3. The section 'Aimand Scope oI The Optics in Sabra`s 1he Optics of Ibn al-Haytham,
vol. 2, pp. liiilxiii, contains discussion oI this and other key passages.
4. Kheirandish, Elaheh, 1he Arabic Jersion of Euclids Optics. Kitb Uqldis f Ikhtilf
al-manzir, Edited and Translated with Historical Introduction and Commentary, 2
volumes, Springer-Verlag: Sources in the History of Mathematics and Physical
Sciences, no. 16, 1999; also, 'TheArabic Version` oI Euclidean Optics: TransIormations
as Iinguistic Problems in Transmission, 1radition, 1ransmission, 1ransformation.
Proceedings of 1wo Conferences on Pre-modern Science Held at the University of
Oklahoma, ed. E. Jamil Ragep and Sally P. Ragep with Steven Iivesey, Ieiden: Brill,
1996: 227243.
5. Kheirandish, Elaheh, 1he Arabic Jersion of Euclids Optics, 2 volumes, Springer-
Verlag, 1999 (see above); also, 'The Manzir` Tradition through Persian Sources, Les
sciences dans le monde iranien, ed. Z. Vesel, H. Beikbaghban et B. Thierry de Crussol
des Epesse, Tehran: Institut Eranais de Recherche en Iran (IERI), 1998: pp. 125145.
6. Some aspects oI appearances are discussed with reIerence to the Greek and Iatin
traditions: appearance versus visual perception, by C. D. Brownson in 'Euclid`s Optics
and Its Compatibility with Iinear Perspective; vision versus reality, by Vasco Ronchi
in 'Classical Optics is a Mathematical Science; image reception versus perception,
by Richard Tobin in 'Ancient Perspective and Euclid`s Optic; objective and subjective
elements in vision, by Gerard Simon in 'The Notion oI the Visual-Ray, by WilIred R.
Theisen in 1he Mediaeval 1radition of Euclids Optics, and by Kim Veltman in Optics
and Perspective. A Study in the Problems of Si:e and Distance, see Bibliography.
7. On the general and specifc senses oI the plural manzir as well as the singular Iorms
manzar and manzara with examples Irom relevant literature; see Sabra, 'Manzir, or
Ilm al-manzir, EI
2
6, p. 376 and Sabra, 'Ibn al-Haytham, DSB, 4, p. 203, n. 9;
see also, Kheirandish, 1he Arabic Jersion of Euclids Optics, vol. 2, Index oI Arabic
Terms.
8. See ed. Amn, pp. 7983, tr., Sabra, 1he Optics of Ibn al-Haytham, v. 2, pp. lvilvii
and Kheirandish, 1he Arabic Jersion of Euclids Optic: Index oI Arabic Terms.
9. The exclusion is noted by Sabra, 1he Optics of Ibn al-Haytham, vol. 2, p. lviii.
10. The Iull passage in the Iacing translation oI the Ioeb edition is as Iollows: 'Similar
to this is the phenomenon that a square appears to have sundry angles, but iI we stand
Iarther oII it looks like a circle. Eor as the Iall oI the rays is in the Iorm oI a cone, when
the fgure is removed to a distance, those rays that are at the angles are cut oII and do not
see anything because they are weak and Iew, when the distance grows greater, but those
that Iall on the center persist because they are collected together and strong. When the
fgure is near they can see also the parts at the angles, but when the distance becomes
greater they cannot, Problems, bk. XV, 911b1921, Ioeb edn., p. 335. The Arabic ver-
sion is now available in the edition oI I. S. Eilius: 1he Problemata Physica Attributed to
Aristotle, 1he Arabic Jersion of Hunain ibn Ishq and the HebrewJersion of Moses ibn
1ibbon, Aristoteles Semitico-Iatinus, Ieiden: Brill, 1999, pp. 658659.
11. Heiberg, J. I., Euclidis Optica, Euclidis Opera Omnia, ediderunt J. I. Heiberg et H.
Menge, vol. VII, Ieipzig: Teubner, 1895, pp. 16 and 166.
12. Kheirandish, 1he Arabic Jersion of Euclids Optics, vol. 1, pp. 3034, vol. 2, pp.
4448.
13. There is a statement in the margin oI one variant oI the early Arabic version,
where a circle is also defned as 'a fgure Ior which there are no angles, as well as in
Ts`s recension oI that proposition, that act as alternative causal premises that this is
'because (li-anna) the smallest parts oI an object (namely angles) are the frst to disap-
pear (yaghb) Irom sight (an al-basar) at a Iar distance. Ibn al-Haytham`s explanation
oI 'why a polygonal (mudalla) fgure is perceived to be circular (mustadr), treats the
problem with the terminology oI the concealment (khaf) oI the angles owing to their
relative smallness (sighar) at a distance (bud): Kitb al-Manzir, bk. III, sec. 9, ed.
Sabra, p. 416; tr. Sabra, vol. 1, p. 281. The Arabic version oI Ptolemy`s Optics, which
contains such a problem, has not reached us Ior specifc or linguistic comparisons; Ior
the Iatin version based on Arabic, see Iejeune, LOptique de Claude Ptolemee, 1956,
1989.
14. Max Krause, who frst reported the two Arabic manuscripts oI Ibn s`s Optics
and Burning Mirrors (Kitb al-Manzir wa al-mary al-muhriqa) |H in Appendix|,
dates it as 'beIore 250H ( 864 A.D.) without Iurther specifcation: 'Stambuler
HandschriIten islamischer Mathematiker, pp. 513514; A. I. Sabra, elaborates on
Krause`s 'conjecture by noting the 'peculiar vocabulary oI the text and its lack oI
any mention oI Arabic authors alongside Greek authors: 1he Optics of Ibn al-Haytham,
vol. 2, p. xxxvii, and n. 39; Roshdi Rashed, who includes part oI the text in a recent
publication (Cuvres Philosophiques et Scientihques dAl-Kind, vol. I: LOptique et la
Catoptrique in Islamic Philosophy 1heology and Science, Texts and Studies edited by
H. Daiber and D. Pingree, vol. xxix), insists that the text is a 'relatively late compila-
tion preserving some works by al-Kind, and that the 'discretion with regard to naming
Arabic authors is 'deliberate, Cuvres Philosophiques et Scientihques dAl-Kind, vol.
1, pp. 5760, see also note 17 below.
15. The Iull passage is as Iollows: 'Euclid said in his book Ikhtilf al-manzir that fg-
ures having angles, like a quadrilateral (murabba), are seen Irom a certain distance as
circular (mustadr), so iI they become distant Irom the eye they are seen (yur) as round
(mudawwar). . . this is 'the most amazing oI amazements (afab al-afab) because iI
it is in the nature (tab) oI sight (basar) to make a leap (tafra) and see a quadrilateral
object Irom a distance as round, then as a result oI that leap the object`s real (haqq)
shape is not seen . . . : H |see Appendix|. The reasoning (illa) oIIered Iollows in a much
clearer text (note that the term tafra is not in the Arabic proposition oI Euclid).
16. The geometrical demonstration shows that iI Irom the center oI the quadrilateral
fgure a line is drawn perpendicularly such that there is a point Irom which the excess
(fadl) oI lines connecting the fgure`s center to its Iar corner, and to the middle oI its
sides is not a sensible magnitude (qadr mahss), then Irom that point the fgure is seen
as circular. In contrast to other explanations oI this visual eIIect based on the charac-
terization oI a circle as a fgure with no angles, this prooI is based on the conception oI
a circle as a fgure having all its points equidistant Irom a center. The defnition oI circle
in Euclid`s Elements is 'a plane fgure contained by one line such that all the straight
lines Ialling upon it Irom one point among those lying within the fgure are equal to one
another: see Heath, 1he 1hirteen Books of Euclids Elements, book 1, vol. 1, p. 153.
17. At the present state oI research, it is diIfcult to determine the exact dates oI Ahmad
Ibn s, a name with no Iew occurrences in historical records (12 in the list oI al-
SaIad`s alone, see Kitb al-Wf bi al-wafayt, v. 7, ed. Ihsn Abbs, Wiesbaden, 1389
1969, pp. 271275). There is no concrete evidence Ior considering Ibn s`s Optics
as 'pre 250/864 with Krause, though the content certainly points to an early date oI
composition; nor is there conclusive evidence Ior considering the text as a post al-Kind
'compilation with Rashed, since with corresponding passages in particular, the chron-
ological arrow may go either way; see note 14 above.
18. Mudallaa (..|..) makes more sense than musallaba (.,|.. cross-shaped) in
Pellat`s edition (treated as such also in Adad`s Erench translation), see Kitb al-1arb
wa al-tadwr, ed. Pellat, p. 91 (tr. Adad, p. 308).
19. On a character whose name, dates and intellectual orientations are comparable to
both the author oI the optical text and the target oI al-Jhiz`s text, see the article under
Ahmad ibn s (d. 247/861) by Madelung in the Encyclopaedia of Islam. EI
2
, Suppl.,
pp. 4849. The article is about a sh scholar and leader associated with the early
Abbasid court.
20. Eor a favor oI the early periods oI intellectual activity and the rivalries involved, see
Gerhard Endress, 'The Circle oI Al-Kind: Early Arabic Translations Irom the Greek
and the Rise oI Islamic Philosophy, 1he Ancient 1radition in Christian and Islamic
Hellenism (Gerhard Endress and Remke Kruk, eds.), Ieiden: Research School CNWS,
1997.
21. Alkindi, 1ideus und Pseudo-Euklid. Drei optischeWerke, herausgegeben und erklrt
von A. Bjrnbo und Sebastian Vogl, Abhandlungen :ur Geschichte der mathematischen
Wissenschaften, Ieipzig/Berlin, 1912, 26, 3: 141; In Iindberg, 1heories of Jision, reI-
erence is made to al-Kind`s use oI this proposition Ior his reIutation oI the intromission
theory: 'II sight occurred through intromission oI the Iorms oI visible things, he |i.e.,
al-Kind| argues, a circle situated edgewise beIore the eye would impress its Iorm in the
eye and consequently would be perceived in its Iull circularity; see p. 23; Ior the con-
text oI discussion and the relevant reIerences, see pp. 2224, pp. 223224, n. 2327.
22. 1aqwm al-khata wa al-mushkilt allat li-Uqldis f Kitbihi al-mawsm bi al-
Nzir |al-Manzir|, Qum MS.: Marashi-yi NajaI 7580, 69b-102b, 960H (unique?), ed.
Rashed, Cuvres Philosophiques et Scientihque dAl-Kind, vol. 1, pp. 162335.
23. On the methodological aspects oI optics, see Kheirandish, 'The Mixed Mathema-
tical Sciences oI the Islamic Middle Ages, 1he Cambridge History of Science, 8 vols.
ed. David C. Iindberg and Ronald Numbers; vol. 2: 1he Middle Ages, Iorthcoming.
24. See respectively: Galen, De usu Partium, English translation. On the Usefulness of
the Parts of the Body. 1ranslated from the Greek with an Introduction and Commentary,
by Margaret T.May, 2 vols., Ithaca: Cornell University Press, 19681969, Arabic trans-
lation: Kitb al-Manh al-ad, Bibliotheque Nationale: MS ar. 2583; Euclid`s
Elements of Geometry, English translation: 1he 1hirteen Books of Euclids Elements,
second edition (revised with additions) Thomas, I., Heath, 3 vols, Cambridge: Cam-
bridge University Press, 1926 (Dover, 1956), especially, 'The Eormal Divisions oI a
Proposition, vol. 1, pp.117131; Aristotle, Posterior Analytics, edited by G. P. Goold,
with an English translation by Hugh Tredennick (Cambridge: Harvard University Press,
Iondon: William Heinemann Itd., 1976); Arabic version, Kitb al-Burhn min Mantiq
Arist, ed. A. Badawi (Cairo: Dr al-Kutub al-Misriyya, 1949), Islamica VII, part 2,
309465.
25. Aristotle, Kitb al-Burhn, ed. A. Badawi, Organon Aristotelis in version Arabica
Antiqua, Part 2, pp. 349353; on the distinction, see also Ragep, Nasr al-Dn al-Tss
Memoir on Astronomy, p. 386, and Crombie, Robert Grosseteste and the Origins of
Experimental Science, pp. 2526, and pp. 5354.
26. Al-Erb, Kitb al-Burhn, ed. M. Eakhry, Beirut: Dr al-Mashriq, 1986. Eor
the commentaries, see Peters, E. E., Aristoteles Arabus. 1he Oriental 1ranslations and
Commentaries on the Aristotelian Corpus (Ieiden: E. J. Brill, 1968), pp. 1719.
27. Kitb fhi al-Manzir wa al-mary al-muhriqa talf Ahmad ibn /s ala mad-
hhab Uqldis f ilal al-basar, Kitb f Ilal m yaridu f al-mary min ikhtilf
al-manzir alifahu (Qust ibn Iq al-Ynn), Arabic texts and Erench translations in
Roshdi Rashed, Cuvres Philosophiques et Scientihques d Al-Kind, vol. 1, p. 649 (in-
cludes only the section on Burning Mirrors), and pp. 572646 respectively (H and I in
Appendix); and Kitb f Ilal ikhtilf al-manzir maa al-barhn al-handasiyya lah (a
likely Iorm Ior the original title oI al-Kind De causis diuersitatum aspectus et dandis
demonstrationibus geometricis super eas, better known as De aspectibus), ed. Bjrnbo,
in A. Bjrnbo and S. Vogl, 'Alkindi, Tideus und Pseudo-Euklid: Drei optische Werke,
Abhandlungen :ur Geschichte der mathematischen Wissenschaften, Ieipzig/Berlin, 26
(1912), 3, pp. 141, tr. J. Jolivet, H. Sinaceur, H. Hugonnard-Roche, in Rashed, Cuvres
Philosophiques et Scientihques dAl-Kind, vol. I, p. 437.
28. On the commentaries, see Peters, Aristoteles Arabus, pp. 1719 and 3031 respec-
tively. On al-mir`s al-Qawl f al-Ibsr wa al-mubsar (Discourse on Jision and Jisual
Obfects), see KhalIt (ed.), Rasil-i Abu al-Hasan-i mir b muqaddamih va tashh-
i Sahbn Khalft, tarfumih-i muqqadamih, Mihd 1adayyun, Tehran: Markaz-i Nashr-i
Dnishgh (University Publications), 13751996 |earlier edition, Rasil Ab al-Hasan
al-mir wa-shadhartuhu al-falsafyah. dirsah wa-nuss, Amman: al-Jmiah al-
Urdunyah, 1988|; Ior reIerences relevant to optics, see Kheirandish, 1he ArabicJersion
of Euclids Optics, vol. 1, pp. xlvi, lviii, vol. 2, p. 13, p. 17.
29. Al-Kind, De aspectibus, ed. Bjrnbo, p. 1.
30. Kheirandish, 1he Arabic Jersion of Euclids Optics and 'The Mixed Mathematical
Sciences oI the Islamic Middle Ages; Sabra, 1he Optics of Ibn al-Haytham, vol. 2,
pp. 2526.
31. De aspectibus, sec. 12, 14, 22, ed. Bjrnbo, pp. 1719, p. 24; sec. 22, and pp. 37
39. The clarity oI a close object is discussed in terms oI the strength oI its illumination,
which is enhanced by proximity to the visual axis, in the wording oI De aspectibus:
'illumination in many parts (plures partes) and Irom all sides; and in 1aqwm, ed.
Rashed, p. 171, as 'whatever is under more intense illumination (nr al-shadd) is seen
more clearly (tur abyan) and so more accurately (asdaq).
32. Qust ibn Iq, F Ilal (I), p. 6: 'II one ray Ialls (yaqau) upon an object, it is seen as
one, iI two rays Iall, it |the object| is seen as two (ra ithnayn), and iI more than two rays
(aktharu min shuayn) Iall, it is seen as more than two (ra akthara min ithnayn).
33. Eor the relevant passage in Ibn s`s H |see Appendix|, see Sabra`s related discus-
sions in 1he Optics of Ibn al-Haytham, v. 2, pp. 2526. Another passage in H contains
the interesting combination that 'whatever is seen by a large angle is seen as larger
(azam) and its vision is more accurate (asdaqa ruyatan).
34. In the second proposition oI the Optics, Ior example, the Arabic translation oI the
Euclidean vocabulary oI visual clarity (oxpirotrpov) as visual accuracy (sidq al-
ruya), along with the problematic Iorm kathra used Ior the concept oI clarity in the
seventh Euclidean defnition, gives rise to a range oI treatments, see Heiberg, Euclidis
Optica, p. 4 and p. 156; Kheirandish, 1he Arabic Jersion of Euclids Optics, vol. 1, p.
and p. 156, and vol. 2, pp. 3034: see also deI. 14, and Prop. 1, 3, 9, 23.
35. The apparently unique manuscript copy is in Dr al-Kutub; see King, Cairo
Catalogue, v. 2, p. 1030; Cairo Survey, p. 39. On the author, see Kitb al-Fihrist, ed.
Elgel, p. 281, tr. Dodge, p. 665; Sezgin GAS: V, p. 301; VI, p. 207, and VII, p. 406.
36. One oI the two extant manuscripts has an additional title with reIerence to Muristus:
wa uluww amidatih wa ilm umq al-br wa urd al-anhr wa ghayr dhlika wa
huwa yusamm Mrstus (and the Elevation of its Height, and the Science of the Depth
of Wells and the Width of Rivers and other 1hings, and he? is called Mrstus |there is
an entry under this Greek author in EI
2
|. The short Arabic treatise is not among the list
oI al-Kind`s works reported by Ibn al-Nadm; see Kitb al-Fihrist, ed. Elgel, pp. 257
261; tr. Dodge, pp. 618620. Atiyeh, Al-Kindi. 1he Philosopher of the Arabs, 1966, p.
200, lists it as no. 230 (citing Brockelmann, GALS I, p. 374); it is also cited in Ritter and
Plessner, 'SchriIten Jaqb ibn Ishq al-Kind`s in Stambuler Bibliotheken, p. 370.
37. See Sabra, 1he Optics of Ibn al-Haytham, Books IIII: On Direct Vision, 2 vols.,
1989 |Includes an English translation and commentary based on an earlier critical edi-
tion; see Bibliography. To be Iollowed by a similar study oI books IVVII|.
38. The Iormer (attributed to Shaykh Ab Al Ibn al-Haytham), seems to be more
directly in the Arabic Euclidean tradition, than the latter (bearing the more commonly
encountered name, al-Hasan ibn al-Hasan ibn al-Haytham) Ior including reIerences to
the eye (basar) and its rays (shu) in the course oI discussions on height determina-
tion.
39. The author oI Height is reported as a commentator on two works, the Persian S
Fasl (1hirty Chapters) attributed to Ts and Euclid`s Elements; see Munzav, Persian
Manuscripts, vol. 1, p. 132.
40. On the corresponding Greek and Arabic terms Ior the bending oI rays, see Kheiran-
dish, 1he Arabic Jersion of Euclids Optics, vol. 2, p. xliv. ReIerences to sources con-
taining the more common Arabic Iorm appear on pp. 5758, n. 205206.
41. On the frst oI these, see Max MeyerhoI `s edition, p. 109, lines 78 |reIerence Irom
Sabra, 1he Optics of Ibn al-Haytham, v. 2, p. lviii, n. 80|, where iniks is used together
with inkisr, in the sense oI the 'turning back (ruf) oI the visual rays (manzir); in
the Arabic manuscript oI De speculis, the verb yanakis is used |Iol. 104b, the Iatin has
convertitur|; see 1ractatus |pseudo-| Euclidis De speculis, Bjrnbo and Vogl, Alkindi,
1ideus und Pseudo-Euklid, p. 100 cited by Theisen, 1he Mediaeval 1radition, p. 294, n.
54. Eor iniks in Ibn Iq`s Arabic translation oI Aetius` Placita philosophorum, see
Daiber, Aetius Arabus, Die Jorsokratiker in arabischer berlieferung, p. 204.
42. In Ts`s Shu and Iniks al-shut wa initfuh, the expression is used Ior the
defection oI the cones oI ray at equal angles when encountered by polished surIaces. In
the case oI Ts`s commentators, there is Badr al-Dn al-Tabar`s short Persian treatise,
Irtif, with two chapters devoted to the problem oI height determination by means oI a
plane mirror. The author, 'recalling another method by which the height oI tall objects
can be made known (malm) by means oI a mirror placed at diIIerent locations on the
ground, states that what the people oI the art (ahl-i sinat) may have oIIered on this
subject 'has not reached him, nor has it been seen by him anywhere. The commentator
oI S fasl (1hirty Chapters) attributed to Ts, and oI Euclid`s Elements expresses the
equality oI visual angle (:wiyat al-shu) and the angle oI refection (:wiyat al-
iniks) in the more common Iorm and explicitly associates the principle with the sci-
ence oI optics (ilm al-manzir), see Kheirandish, 1he Arabic Jersion of Euclids Optics,
vol. 1, p. l.
43. On the evidence Irom the Arabic tradition, see Kheirandish, 1he Arabic Jersion
of Euclids Optics, vol. 1, pp. xxvixxvii, and p. xxix, and vol. 2, p. 6; on the case oI
the Greek tradition, see Jones, 'Peripatetic and Euclidean Theories oI the Visual Ray,
p. 52, and Knorr, 'Pseudo-Euclidean Refections in Ancient Optics, p. 29, n. 4849.
44. Kheirandish, Elaheh, 'What Euclid Said to his Arabic Readers: The Case oI the
Optics, Proceedings oI the XXth International Congress oI History oI Science, Iiege,
1997, Published in Optics and Astronomy (Simon, G. and Debarbat, S., eds.), 2001,
pp. 1728.
45. In the frst Iormulation, namely Euclid Kitb al-Manzir (Optics), the indicators
are the ray (al-shu), issuing (yakhruf), the eye (al-ayn), paths (sumt), infnte (l
nihya) multitude (kathra), cone (makhrt), apex (ras), and object (mubsar), while in
the second Iormulation, that is, Euclid`s Kitb al-Mirh (De speculis) they are, lumi-
nous power (quwwa nriyya), spreading (yanbathth), pupil (nzir), sanawbar (pine-
shaped), :uff, mustahadd (pointed). Note that the latter term (:uff), a non-standard
Iorm Ior the cone`s apex, seems to be intended in the Arabic De speculis (S), Ahmad
ibn s`s Manzir wa al-marya al-muhriqa (H), and in al-Kind`s 1aqwm (Q), rather
than the similarly transcribed terms 'rahb and 'wa bihi in Rashed`s edition oI Q and
S respectively: see Cuvres Philosophiques et Scientihques d Al-Kind, vol. 1, p. 163
and p. 338.
46. Some aspects treated by Ibn al-Haytham are perception by glancing (idrk bi al-
badha), by contemplation (idrk bi al-taammul) by recognition (idrk bi al-marifa),
and ascertained (muhaqqaq) perception, see Sabra, 1he Optics of Ibn al-Haytham, vol.
2, p. 241; Ior a general discussion, see H. WolIson, 'The Internal Senses in Iatin, Arabic
and Hebrew Philosophical Texts, Harvard 1heological Review, 1935, 28: 69133.
47. Sabra, A. I., 1he Optics of Ibn al-Haytham, vol. 2, pp. lviiilix.
48. Sabra, 1he Optics of Ibn al-Haytham, vol. 2, pp. lviiilix (on the limited circulation
oI Ptolemy`s Optics, with specifc reIerence to relevant works including the exceptional
cases oI Ibn Sahl and Ibn al-Haytham); see also Sabra, 'Psychology vs. Mathematics:
Ptolemy and Alhazen on the Moon Illusion, Mathematics and its Applications to
Science and Natural Philosophy in the Middle Ages. Essays in the Honor of Marshall
Clagett, ed. Edward Grant and John E. Murdoch, Cambridge: Cambridge University
Press, 1987: 217247, pp. 219221.
49. Eris`s reIerence to 'the shortcomings oI Euclid`s book |i.e., Optics| and his rea-
sons Ior undertaking his own optical researches appear in the introduction oI his impor-
tant commentary on Ibn al-Haytham`s Kitb al-Manzir, entitled 1anqh al-Manzir,
ed. Hyderabad, p. 16: 'I saw in the statements oI some leading philosophers, and in
more than one oI them, that light shines Irom a luminous object in straight lines, and
when it encounters a surIace such as the surIace oI water, it is refected Irom it at
angles equal to their opposite |side| and penetrates (yanfdh) into it on the extension oI
illumination, and is reIracted (inatafat), in the extension oI refection (iniks), and
Irom this Iour equal angles are produced, angles oI direct radiation, refection, pen-
etration, and reIraction (:awy al-istiqmah, al-iniks, al-nufdh, wa al-initf). So
I became puzzled (tahayyartu) by these rules (ahkm). Ts`s treatment in 'Risla f
iniks al-shut wa initahh |see note below| is cited by A. I. Sabra in this connec-
tion: 1he Optics of Ibn al-Haytham, vol. 2, pp. lxixlxxi, n. 112.
50. See English translation, Winter and AraIat, 'A Statement on Optical Refection
and ReIraction` Attributed to Nasr ud-dn at-Ts, p. 141; partial German trans-
lation, Wiedemann, 'ber die Refection und Umbiegung des Iichtes von Nasr al Dn
al-Ts.
51. Sabra, 1he Optics of Ibn al-Haytham, vol. 2, p. lix: 'all the historical evidence we
have points to the Iact that Ptolemy`s book |i.e., Optics| was little used both in Antiquity
and in the Islamic Middle Ages almost up to I.H.`s time.
Biiiior:rn.
Aristotle, Problems, v. 1: Books IXXI, edited by T. E. Page, E. Capps, and W. H. D.
Rouse, with an English translation by W. S. Hett, Ioeb Classical Iibrary, Cambridge:
Harvard University Press, and Iondon: William Heinemann Itd., 1936.
Aristotle, Posterior Analytics, edited by G. P. Goold, with an English translation by
Hugh Tredennick (Cambridge: Harvard University Press, Iondon: William Heinemann
Itd., 1976; Arabic version, Kitb al-Burhn min Mantiq Arist, ed. A. Badawi (Cairo:
Dr al-Kutub al-Misriyya, 1949), Islamica VII, part 2, 309465.
Atiyeh, G. N., Al-Kind. 1he Philosopher of the Arabs, Rawalpindi: Islamic Research
Institute, Publication No. 6, 1966.
Bjrnbo and Vogl, Alkindi, 1ideus und Pseudo-Euklid Alkindi, 1ideus und Pseudo-
Euklid. Drei optische Werke, herausgegeben und erklrt von A. Bjrnbo und Sebastian
Vogl, Abhandlungen :ur Geschichte der mathematischen Wissenschaften, Ieipzig/
Berlin, 1912, 26, 3: 1176.
Brockelmann, GAL Brockelmann, Carl, Geschichte der arabischen Litteratur, 2nd
edition, Ieiden: E. J. Brill, 2 vols. and 3 suppl., 19371949 (GAL I vol. I, 1943; GAL
II vol. II, 1949; GALS I Suppl. I, 1937; GALS II Suppl. II, 1938; GALS III Suppl.
III, 1943).
Brownson, C. D., 'Euclid`s Optics and Its Compatibility with Iinear Perspective,
Archive for History of Exact Sciences, 1981, 24, 3: 165194.
Burton, H. E., 'The Optics oI Euclid, Journal of the Optical Society of America, 1945,
35: 357372.
Clagett, Marshall, 'Some General Aspects oI Physics in the Middle Ages, Isis, 1948,
39: 2944.
Crombie, Alistair C., Robert Grosseteste and the Origins of Experimental Science.
11001700, OxIord: At the Clarendon Press, 1953.
Daiber, Hans, Aetius Arabus. DieJorsokratiker in arabischer berlieferung, Wiesbaden,
1980.
DSB Dictionary of Scientihc Biography, ed. C. C. Gillispie, 16 vols. (including
Supplement and Index), NewYork: Charles Scribner`s Sons, 19701980.
EI
2
Encyclopaedia of Islam, 2nd edition, Ieiden: E. J. Brill, 1960.
Euclid, Elements, see Heath, 1he 1hirteen Books of Euclids Elements.
Euclid, Optics: see Heiberg, Euclidis Optica; Burton, 'The Optics oI Euclid; Ver
Eecke, Euclide. lOptique et la Catoptrique.
Endress, Gerhard, 'The Circle oI Al-Kind: Early Arabic Translations Irom the Greek
and the Rise oI Islamic Philosophy, in 1he Ancient 1radition in Christian and Islamic
Hellenism (G. Endress and R. Kruk, eds.), Ieiden: Research School CNWS, 1997.
al-Erb, Ab Nasr, Ihs al-ulm, ed. Uthmn Amn | Osman Amine|, Cairo:
Iibrairie Anglo-Egyptienne, 1968 |1st and 2nd edn., Cairo, 1931, 1949|; A. Gonzalez
Palencia, Alfarabi Catalogo de las ciencias: two medieval Iatin translation, modern
edition and Spanish translation, Madrid, 1932.
al-Erb, Kitb al-Burhn, ed. M. Eakhry, Beirut: Dr al-Mashriq, 1986.
al-Eris, Kaml al-Dn, 1anqh al-Manzir li-dhaw al-absr wa al-basir, 2 vols.,
Hyderabad, 13471348 (19281930).
Eilius, I. S., 1he Problemata Physica Attributed to Aristotle, 1he Arabic Jersion of
Hunain ibn Ishq and the Hebrew Jersion of Moses ibn 1ibbon, Aristoteles Semitico-
Iatinus, Ieiden: Brill, 1999.
Galen, De usu Partium, Arabic translation: Kitb Manh al-ad, Bibliotheque
Nationale: MS ar. 2583; English translation, see Margaret May.
Gulchn Man, stn-i Quds A Catalogue of Manuscripts in the Astan Quds Ra:avi
Library (Fihrist-i Kutub-i khatt-yi Kitabkhnih-i stn-i Quds-i Radav) by Ahmad
Golchin Maani, v. 8, Mashhad: A publication oI the Cultural and Iibrary AIIairs, no.
6, 13501971.
Heath, Thomas, I. 1he 1hirteen Books of Euclids Elements, second edition (revised
with additions), 3 vols, Cambridge: Cambridge University Press, 1926 (Dover, 1956)
Heiberg, J. I., Euclidis Optica Heiberg, J. I., Euclidis Optica Opticorum recensio
1heonis, Catoptrica, cum scholiis antiquis, edidit Heiberg, Euclidis Opera Omnia, edi-
derunt J. I. Heiberg et H. Menge, vol. VII, Ieipzig: Teubner, 1895: 1121.
Heiberg, Opticorumrecensio1heonis Heiberg, J. I., Euclidis Optica, Opticorumrecen-
sio 1heonis, Catoptrica, cum scholiis antiquis, edidit Heiberg, Euclidis Opera Omnia,
ediderunt J. I. Heiberg et H. Menge, vol. VII, Ieipzig: Teubner, 1895: 143247.
Hunayn Ibn Ishq, Kitb al-Ashar maqlt f al-ayn, see MeyerhoI, 1he Book of the
1en 1reatises of the Eye Ascribed to Hunain ibn Is-haq (809877 AD).
Husayn and Marash, Marash Husayn, S. A., and Marash, S. M., Fihrist-i
Nuskhihh-yi khatt-yi Kitbkhnih-i Hadrat-i yatullh al-Uzm Naff-yi Marash .
. . , v. 11, 1364 (1985); v. 12, 1365 (1986).
Ibn al-Haytham, Kitb al-Manzir, ed. Sabra, A. I., Al-Hasan Ibn al-Haytham, Kitb
al-Manzir. Books IIIIII On Direct Jision~. Edited with Introduction, Arabic-Latin
Glossaries and Concordance 1ables by Abdelhamid I. Sabra, Kuwait: The National
Council Ior Culture, Arts and Ietters, 1983. See also, Sabra, A. I. 1he Optics of Ibn al-
Haytham. Books IIII On Direct Jision, Iondon, 1989.
Ibn al-Nadm, Kitb al-Fihrist, ed. Elgel Ibn al-Nadm, Muhammad ibn Ishq, Kitb
al-Fihrist, edited by Gustav Elgel, 2 vols., Ieipzig: Verlag von E. C. W. Vogel, 1871
1872; reprinted, Beirut: Khayyt, 1964.
Ibn al-Nadm, Kitb al-Fihrist. 1he Fihrist of al-Nadm. A 1enth-Century Survey of
Muslim Culture, Bayard Dodge (editor-translator), 2 vols. (Records oI Civilization:
Sources and Studies no. 83), NewYork and Iondon: Columbia University, 1970.
al-Jhiz, Kitb al-1arb wa al-tadwr: edition, Pellat, Le Kitb al-tarb wa-t-tadwr de
Ghiz.
Jones, Alexander, 'Peripatetic and Euclidean Theories oI the Visual Ray, Physis, 1994,
31, 1: 4776.
KhalIt (ed.) mir, Rasil Rasil-i Abu al-Hasan-i mir b muqaddamih va
tashh-i Sahbn Khalft, tarfumih-i muqqadamih, Mihd 1adayyun, Tehran: Markaz-i
Nashr-i Dnishgh (University Publications), 13751996 |earlier edition, Rasil Ab
al-Hasan al-mir wa-shadhartuhu al-falsafyah. dirsah wa-nuss, Amman: al-
Jmiah al-Urdunyah, 1988|.
Kheirandish, Elaheh, 'The Arabic Version` oI Euclidean Optics: TransIormations as
Iinguistic Problems in Transmission, 1radition, 1ransmission, 1ransformation: 227
247, Ieiden: Brill, 1996.
Kheirandish, Elaheh, 'The Manzir` Tradition through Persian Sources, La sciences
dans la monde iranien, ed. Z. Vesel, H. Beikbaghban et B. Thierry de Crussol des
Epesse, Tehran: Institut Eranais de Recherche en Iran (IERI), 1998: pp. 125145.
Kheirandish, Elaheh, 1he Arabic Jersion of Euclids Optics. Kitb Uqldis f Ikhtilf al-
manzir. Edited and translated with historical introduction and commentary (revised
dissertation, Harvard University, 1991), 2 vols., Springer-Verlag: Sources in the History
of Mathematics and Physical Sciences, no. 16, 1999.
Kheirandish, Elaheh, 'What Euclid Said to His Arabic Readers: The Case oI the
Optics, Proceedings oI the XXth International Congress oI History oI Science, Iiege,
1997, Published in Optics and Astronomy (Simon, G. and Debarbat, S., eds.), 2001,
pp. 1728.
Kheirandish, Elaheh, 'The Mixed Mathematical Sciences oI the Islamic Middle Ages,
1he Cambridge History of Science, 8 vols. ed. David C. Iindberg and Ronald Numbers;
vol. 2: The Middle Ages, Cambridge University Press, Iorthcoming.
al-Kind, De aspectibus: see Bjrnbo and Vogl, Alkindi, 1ideus und Pseudo-Euklid.
Drei optische Werke. See also Rashed, Cuvres Philosophiques et Scientihque d Al-
Kind, 1997.
King, David A., A Catalogue of the Scientihc Manuscripts in the Egyptian National
Library, Cairo: General Egyptian Book Organization in collaboration with The American
Research Center in Egypt, and the Smithsonan Institution, Part II, 1986 |Part I, 1981|.
King, DavidA., A Survey of the Scientihc Manuscripts in the Egyptian National Library,
published Ior The American Research Center in Egypt, Catalogs, vol. 5, Winona Iake,
Indiana: Eisenbrauns, 1986.
Knorr, Wilbur R., 'Pseudo-Euclidean Refections in Ancient Optics: A Re-Examination
oI Textual Issues Pertaining to the Euclidean Optica and Catoptrica, Physis, 1994, 31,
1: 145.
Krause, Max, 'Stambuler HandschriIten islamischer Mathematiker, Quellen und
Studien :ur Geschichte der Mathematik, Astronomie und Physik, Berlin, 1936, B, Band
3, HeIt 4: 437532.
Iejeune, Albert, LOptique de Claude Ptolemee dans la version latine dapres larabe
de lemir Eugene de Sicile. Edition critique et exegetique augmentee dune traduction
franaise et de complement, Ieiden: E. J. Brill, Collection de 1ravaux de lAcademie
Internationale dHistoire des Sciences, no. 31, 1989. |earlier edn.: Iouvain: Universite
de Iouvain, Recueil de travaux dhistoire et de philologie, 4. ser., Iasc. 8, 1956|.
Iindberg, David C., 1heories of Jision from Al-Kindi to Kepler, Chicago and Iondon:
The University oI Chicago Press, 1976.
Madelung, WilIerd, 'Ahmad ibn s, EI
2
, Supplement, pp. 4849.
May, Margaret T., Galen On the Usefulness of the Parts of the Body. 1ranslated from
the Greek with an Introduction and Commentary, 2 vols., Ithaca: Cornell University
Press, 19681969.
MeyerhoI, Max, 1he Book of the 1en 1reatises on the Eye Ascribed to Hunain ibn Is-haq
(809877 A.D.), Cairo: Government Press, 1928.
Munzav, A Catalogue of Persian Manuscripts (Fihrist-i Nuskhih'h-yi khatt-yi frs)
by Ahmad Monzavi, Tehran: Regional Cultural Institute, publication, v. 1, no. 14, 1348
(1969).
Pellat, Charles, Le Kitb al-tarb wa-t-tadwr de Ghiz, Damascus: Institut Eranais
de Damas, 1955.
Pellat, Charles, 1he Life and Works of Jhiz. 1ranslations of selected texts, Berkeley and
Ios Angeles: University oI CaliIornia Press, 1969.
Peters, E. E., Aristoteles Arabus. 1he Oriental 1ranslations and Commentaries on the
Aristotelian Corpus, Ieiden: E. J. Brill, 1968.
Ptolemy, Optica: see Iejeune, LOptique de Claude Ptolemee, and Smith, 'Ptolemy`s
Theory oI Visual Perception: An English Translation oI the Optics with Introduction
and Commentary.
Ragep, E. J., Nasr al-Dn al-Tss Memoir on Astronomy. al-1adhkira f ilm al-haya,
Edition, 1ranslation and Commentary, 2 vols., New York: Springer-Verlag, Sources in
the History of Mathematics and Physical Sciences, no. 12 (ed. G. J. Toomer), 1993.
Rashed, R., Cuvres Philosophiques et Scientihque d Al-Kind. Vol. I: LOptique et la
Catoptrique in Islamic Philosophy 1heology and Science, Texts and Studies edited by
H. Daiber and D. Pingree, vol. xxix, Cuvres Philosophiques et Scientihques dAl-Kind
editees par Jean Jolivet et Roshdi Rashed, Ieiden: E. J. Brill, 1997.
Ritter, H. and Plessner, M., 'SchriIten Jaqb ibn Ishq al-Kind`s in Stambuler Biblio-
theken, Archiv Orientalni, 1932, 4: 363372.
Ronchi, Vasco, 'Classical Optics is a Mathematical Science, Archive for History of
Exact Sciences, 1961, 1, 2: 160171.
Sabra, A. I., 'Ibn al-Haytham, DSB, 1972, 6: 189210 |reprinted in Sabra, Optics,
Astronomy and Logic, 1994, II: 189210|.
Sabra, A. I., 'A Note on Codex Biblioteca Medicea Iaurenziana Or. 152, Journal for
the History of Arabic Science, 1977, 1, 2: 276283.
Sabra, A. I., 'Manzir, or Ilm al-Manzir, EI
2
, 1987, 6, Iasc. 103104: 376377.
Sabra, A. I. 'Psychology vs. Mathematics: Ptolemy and Alhazen on the Moon Illusion,
Mathematics and its Applications to Science and Natural Philosophy in the Middle
Ages. Essays in the Honor of Marshall Clagett, ed. Edward Grant and John E. Murdoch,
Cambridge: Cambridge University Press, 1987: 217247.
Sabra, A. I., 1he Optics of Ibn al-Haytham. Books IIII On Direct Jision, 1ranslated
with Introduction and Commentary by A. I. Sabra, 2 vols. (Studies oI the Warburg Insti-
tute edited by J. B. Trapp, vols. 40 iii), Iondon: The Warburg Institute, University oI
Iondon, 1989.
Sabra, A. I., and Heinen, Anton, 'On Seeing the Stars: Edition and Translation oI Ibn
al-Haytham`s Risla f Ruyat al-kawkib, Zeitschrift fur Geschichte der arabisch-isla-
mischen Wissenschaften, 1991/92, 7: 3172.
Sabra, A. I., 'On Seeing the Stars II: Ibn al-Haytham`s Answer` to the Doubts` Raised
by Ibn Madn. Zeitschrift fur Geschichte der arabisch-islamischen Wissenschaften,
1995/96, 10: 159.
al-SaIad, Kitb al-Wf bi al-wafayt, v. 7, ed. Ihsn Abbs, Wiesbaden, 13891969.
Sezgin, Euat, Geschichte des arabischen Schrifttums, Ieiden: E. J. Brill, 19701979
(GASJ: Mathematik, 1974; GASJI: Astronomie, 1978; GASJII: Astrologie, Meteorol-
ogie, und Verwandtes, 1979).
Simon, Gerard, 'The Notion oI the Visual-Ray, presented in a conIerence on optics
at the Dibner Institute, May, 1990, 'Ia notion de rayon visuel et ses consequences sur
lòptique geometrique grecque, published in Physis, Rivista Interna:ionale di Storia
della Scien:a (Olschki, I. S., ed.), 94, pp. 77112.
Theisen, 1he Mediaeval 1radition Theisen, WilIred R., 1he Mediaeval 1radition of
Euclids Optics, Ph.D. Thesis, University oI Wisconsin, 1972, Iacsimile, Ann Arbor:
University Microflms International, 1984.
Tobin, Richard, 'Ancient Perspective and Euclid`s Optics, Journal of the Warburg and
Courtauld Institutes, 1990, 53: 1442.
Toomer, G. J., Diocles on Burning Mirrors. 1he Arabic 1ranslation of the Lost Greek
Original, Edited with English Translation and Commentary by G. J. Toomer, NewYork:
Springer-Verlag, Sources in the History of Mathematics and Physical Sciences, no. 1
(ed. M. J. Klein and G. J. Toomer), 1976.
al-Ts, Nasr al-Dn, 1ahrr al-Manzir, Mafm al-rasil, Hyderabad: Uthmniyya,
part 1, 1358 (1939), 5: 224.
Veltman, Kim H., Optics and Perspective. A Study in the Problems of Si:e and Distance,
Ph.D. Thesis, Iondon: The Warburg Institute, 1975.
Ver Eecke, Euclide. lOptique et la Catoptrique Euclide. lOptique et la Catoptrique,
uuvres traduites pour la premiere Iois du grec en Iranais, avec une Introduction et des
Notes par Paul Ver Eecke, Nouveau Tirage, Paris: Iibrairie Scientifque et Technique
Albert Blanchard, 1959.
Wiedemann, Eilhard, 'ber die Refexion und Umbiegung des Iichtes von Nasr al
Dn al Ts, Jahrbuch fur Photographie und Reproduktionstechnik, Halle, 1907, 21:
3844.
Winter, H. J. J. and AraIat, W., 'A Statement on Optical Refection and ReIraction`
Attributed to Nasr ud-Dn at-Ts, Isis, 1951, 42: 138142.
WolIson, H. A. 'The Internal Senses in Iatin, Arabic and Hebrew Philosophical Texts,
Harvard 1heological Review, 1935, 28: 69133.
al-Yaqb, Tarkh: Ibn-Wdhih qui dicitur Al-Jaqb, Historiae, pars prior historiam
ante-Islamicam continens, edidit indicesque adjecit M. Th. Houtsma, 2 vols., Ieiden:
E. J. Brill, 1883.
|T|he nature oI our subject being conIused, in addition to the continued dis-
agreement through the ages among investigators who have undertaken to
examine it, and because the manner oI vision is not ascertained, we have
thought it appropriate that we direct our attention to this subject as much as
we can, and seriously apply ourselves to it, and examine it, and diligently
inquire into its nature.We should, that is, recommence the inquiry into
its principles and premises, beginning our investigation with an inspection
oI the things that exist and a survey oI the conditions oI visible objects. We
should distinguish the properties oI particulars, and gather by induction what
pertains to the eye when vision occurs and what is Iound in the manner oI
|visual| sensation to be uniIorm, unchanging, maniIest and not subject to
doubt.AIter which, we should ascend in our inquiry and reasonings, grad-
ually and orderly, criticizing premises and exercising caution in regard to con-
clusionsour aim in all that we make subject to inspection and review being
to employ justice, not to Iollow prejudice, and to take care in all that we judge
and criticize that we seek the truth and not to be swayed by opinion.
Ibn al-Haytham, Optics, PreIace |6|
|A|ll that sight perceives it perceives by reIraction.
Ibn al-Haytham, Optics, Bk VII, Ch. 6.
I Ix+roiuc+iox: A P:r:iox or Ar:iic Or+ics
The history oI Arabic science is Iull oI puzzles, one oI which is a paradox
revealed by comparing the history oI optics with that oI astronomy. Arabic
astronomy was launched in the eighth century AD with a series oI transla-
tions that included Ptolemy`s Almagest, Euclid`s Elements, and other Greek
mathematical works deemed necessary Ior pursuing a serious study oI the
subject. This clearly concerted eIIort was part oI a spectacular cultural move-
ment actively supported by the Abbsid caliphate in Baghdad. The ensuing
scientifc endeavor, which continued with renewed surges oI energy under
various dynastic rules in various parts oI the Islamic world Ior more than seven
4
Iix :i-H:.+n:x`s Prvoiu+iox:r. ProJrc+ ix Or+ics:
1nr Acnirvrxrx+ :xi +nr Ois+:cir
A. I. Sabra
Ior ir Lrust om|ri.h, 19092001
A. I. S:ir: 86 Iix :i-H:.+n:x`s Prvoiu+iox:r. ProJrc+ ix Or+ics 87
hundred years, always gave pride oI place to astronomy as the discipline at the
very top oI the Greek mathematical sciences and, sometimes, even at the top
oI the entire hierarchy comprising the sum total oI Greek theoretical knowl-
edge, as defned in Greek antiquity and somewhat modifed by later traditions
represented in particular by Ptolemy in the second century AD. In most or per-
haps all oI the patronized scientifc activities Irom the eighth to the fIteenth
centurywhether in Baghdad, Cairo, Muslim Spain, Ghazna, Margha, or
Samarkandastronomy tended to be Iavored as the pursuit most worthy oI
the attention oI both the patron rulers and the patronized mathematicians who
were keen on mastering and exploiting the Greek legacy. This must have been
due in large part to the practical benefts widely expected Irom astronomy, in
particular those promised by astrology, usually conceived as applied astron-
omy. But along with this practical motivation, astronomy came to be viewed,
almost Irom the start, as a supreme Iorm oI scientifc knowledge that pos-
sessed exact methods oI observation and calculation and prooI, and as a com-
pelling or at least highly persuasive way oI maniIesting God`s wisdom and
the perIection oI His handiwork. Many Muslim astronomers (probably most
oI them at frst) embraced a pre-Islamic, Hellenistic view oI the world and oI
man`s place in it, but almost all oI them were also willing to put their skills in
observation and computation at the service oI Muslim religious practice. Eor
example, they developed exact methods Ior determining prayer times and the
direction oI Muslim prayer toward Mecca at all localities within the Islamic
world. As a result oI all these perceptions, expectations, and assorted claims,
which were in part practical, in part theoretical or spiritual or religious, the
amount oI resources and intellectual energy that were spent on promoting and
perIecting the science oI astronomy Iar outweighed what was available Ior any
oI the 'ancient sciences, al-ulm al-qadma or ulm al-awil, with the pos-
sible exception oI medicine, which itselI Irequently received more than cur-
sory attention Irom scholars deeply engaged in the mathematical/astronomical
sciences.
1
And yet, despite the many refnements, corrections, and innovations,
oIten motivated by keen intellectual interest and real investigative and criti-
cal attitudes, Arabic astronomy never managed to break out oI the Ptolemaic
paradigm. The planets and the orbs that carried them around continued to
move in circles, with defnite preIerence given to traditional uniIorm speeds;
Ibn al-Shtir in Iourteenth-century Damascus argued Ior discarding Ptolemy`s
eccenters, but the epicycles remained, still serving the cause oI circular uni-
Iorm motion; and, despite repeated ventures into theoretical aporetic (raising
doubts/shukk/aporiai and attempting solutions), the earth stayed frmly put
and unrotating at the center oI a universe largely structured by basic tenets
oI Aristotelian natural philosophy. The twelIth-century 'Andalusian Revolt
(itselI inspired by strictly Aristotelian cosmology) against the Ptolemaic eccen-
ters and epicycles, proved a non-starter that Iailed to induce later developments
in the Islamic world (Sabra, 1984; Samso, 1994). Arabic astronomy, whose
impressive accomplishments have yet to be suIfciently appreciated by the gen-
eral historian oI science, never achieved what can meaningIully be called 'rev-
olution (Sabra, 1998).
Now compare the above account with the story oI Arabic optics. As in
the case oI astronomy, Arabic 'optics/ilm al-manzir, by which term I shall
always reIer to the mathematical study of vision (the meaning this termhad in the
Greek and Arabic traditions), also started with the translation oI ancient works
into Arabic. And while the early translations in astronomy included Indian and
Persian, as well as Greek, materials, Arabic research in optics can defnitely be
said to have been originally based entirely on Greek sources. OI the two surviv-
ing Greek works bearing the title OP1IKA, namely those oI Euclid and Ptolemy,
the frst is known to have circulated in Baghdad in Arabic version(s) already
in the ninth century (Kheirandish, 1999). Apart Irom the text or texts purport-
ing to be translations oI Euclid`s treatise, the Arabic optical writings that have
reached us Irom that period (they include those oI Hunayn ibn Ishq, Ahmad
ibn s, al-Kind, and Qust ibn Iq),
2
exhibit a mixture, or rather mixtures oI
doctrines that can be traced back to Galen and, through him, to Plato and to the
Stoics, as well as, independently oI these authorities, to Euclid`s Optics itselI.
What all oI these ninth-century writers on optics had in common was their
adherence to one version or another oI what has been called the extramission
hypothesis, according to which vision oI an external object was mediated by a
visual emanation Irom the eye that extended in the shape oI a cone (makhrt,
sanawbara/konos) all the way to the object seen. That emanation was either the
sensitive material pneuma itselI that had frst descended Irom the brain into
the eyes through the optic nerves or, in most cases, and almost certainly under
the infuence oI Galen`s arguments (especially in his De placitis Hippocratis
et Platonis, see below), a sensitive power (quwwa/dynamis) conIerred by the
pneuma upon the surrounding air when it struck the air as it emerged Irom the
pupil.
3
Besides the pneuma`s sensitive capacity indicated by calling it 'visual/
basar or, literally, 'seeing spirit/breath (al-rh al-bsir), the pneuma was
given a Iurther property indicated by also calling it 'luminous breath (al-rh
al-nr or al-nayyir) or, simply, 'the light (al-nr).
4
Since everybody knew
that vision does not take place in the dark, the Iurther assertion was made that
illumination oI the air by a shining body (the sun, the stars, or a torch) was nor-
mally a necessary condition Ior the instrument oI sight, that is, the pneuma or
the qualitatively altered air, to perIorm its action. Vision happened as a result
oI the 'contact or 'coalescence or 'cooperation oI the visual light occu-
pying the visual cone with the external light, and it occurred along the radial
lines (khutt al-shu, also called 'visual rays, manzir/opseis) constituting
the cone which spread outwardly Irom its vertex at the eye to its base at the
object.
5
Apart Irom a passage in al-Kind`s De aspectibus (Proposition 11) that
S. Vogl pointed out in 1912,
6
and a brieI, ambiguous reIerence in al-Kind`s
1aqwm al-khata wa al-mushkilt allat li-Uqldis f Kitbihi al-mawsm bi-
al-Manzir/'Rectifcation oI Euclid`s Optics to Ptolemy and Theon oI Alex-
andria,
7
Ptolemy`s Optics is conspicuously absent Irom the Arabic discussions
oI the ninth century. In these discussions it is always Euclid`s Optics that is
cited as an authority, or sometimes, as in the case oI al-Kind, taken to task Ior
some Iailing or anotherbut never in conjunction with an explicit mention oI
Ptolemy`s treatise.It would thus seem that the continuous radiation oI visual
light, which is the view defnitely Iavored (but not necessarily invented) by
Ptolemy (d. ca. 170), had reached the ninth-century Arabic writers, not directly
through acquaintance with Ptolemy`s treatise, but through Galen (d. ca. 214),
or through sources not all oI which are presently known to us.
One is particularly struck by two obseravations that strongly argue
against direct acquaintance on the part oI those writers with Ptolemy`s Optics,
and against any correct understanding or appreciation oI his distinctive contri-
butions. The frst observation is the absence Irom their known compositions
oI the certainly non-Euclidean and importantly (though not exclusively) Ptol-
emaic emphasis on the primacy oI color as a precondition Ior gaining percep-
tion oI all other visible properties as qualifcations oI color, and oI coloration
as the primary 'eIIect (passio/inhl or tathr/pathos) produced in the organ
oI sight (visus/basar/opsis).
8
(See below, Section II.)
The second observation is that when some oI the same ninth-century
writers attempted a reasoned account oI optical reIraction in terms oI lines
and angles (we have two such accountsone by Ahmad ibn s and the other
by al-Kind), they only produced disastrously wrong arguments that could not
possibly have been conceived by someone who had read and understood Book
V oI Ptolemy`s Optics.
9
It is also signifcant that the somewhat extended expo-
sition by al-Erb (d. AD 950) oI the 'mathematical science oI optics counts
three modes oI optical refection (all oI which cases oI bending oI the visual
ray back in the direction oI the viewer, and to which he assigns three diIIer-
ent names), but Iails to mention reIraction (Ibn al-Haytham, 1989, II, Intro-
duction, pp. lvilviii). Add to this the Iact that the same erroneous account,
Iound in Ibn s and in the Iragment ascribed to al-Kind, reappears, without
correction, in an early astronomical work by Ibn al-Haytham ('Commentary
on Almagest, MS Saray, Ahmet III 3329),
10
a work which may have helped
to spread misunderstanding oI the reIraction phenomenon among mathemati-
cians as late as the thirteenth century, such as Nasr al-Dn al-Ts (d. 1274)
and Qutb al-Dn al-Shrz (d. 1311).
11
That the Arabic tradition oI Ptolemy`s
Optics diIIered greatly Irom that oI his Almagest should be suIfciently clear
Irom the additional Iact that the optical part in the widely-used collection oI
'the middle books (al-Mutawassitt), prepared by al-Ts Ior the use oI stu-
dents oI astronomy, represented not the advanced state oI knowledge Iound in
Ptolemy`s treatise, but the stage reached Iour-and-a-halI centuries earlier in
Euclid`s Optics.
12
The frst clear evidence we have oI a correct understanding oI Ptolemy`s
theory oI reIraction does not appear in the Arabic sources available to us until
the second halI oI the tenth century, when the Persian mathematician al-Al
ibn Sahl was able to put Ptolemy`s ideas to use in Iormulating entirely orig-
inal geometrical arguments Ior the construction oI burning instruments by
means oI reIraction.
13
That was a landmark achievement, and one oI the Iruits
oI Buwayhid patronage in Iraq, at a time remarkable Ior its intensive scientifc,
especially mathematical activity. But none oI the extant works oI Ibn Sahl is
concerned, either wholly or in part, with problems oI vision, a subject that in
Iact is never mentioned in them: his interests, it appears, lay elsewhere, prob-
ably with an eye to practical application.
It is, thereIore, still true to say today that, Ior the frst substantial treat-
ment oI vision that was directly inspired by Ptolemy`s contribution, we have
to turn to the works oI Ibn al-Haytham in the frst halI oI the eleventh century,
that is just short oI nine hundred years aIter Ptolemy and some two hundred
and fIty years aIter the Abbsid rulers lent their support to the Greco-Arabic
translation enterprise.
Ibn al-Haytham was one oI the most prolifc and most competent geo-
metricians in the Arabic tradition. In his younger years he wrote commentaries
and/or summaries oI Euclid`s Elements and Apollonius`s Conics.
14
A 'Com-
pletion oI the Conics (a reconstruction oI the lost Book Eight) possibly also
belongs to an early period in his liIe.
15
A large proportion oI his extant writings
are devoted to problems oI elementary and advanced geometry, including geo-
metrical methodology.
16
In the Optics, which belongs to a later period in his
career, Ibn al-Haytham solves the problem that has become known since the
seventeenth century as 'Alhazen`s problem: to fnd the point(s) oI refection
on the surIace oI a spherical mirror, convex or concave, given the positions oI
the eye and the visible object-pointa problem equivalent to a Iourth-degree
equation and thereIore not solvable by ruler and compass (Sabra, 1982;
Hogendijk, 1996). Armed with this solution he is able in Optics, Book V to
undertake the frst systematic investigation oI images produced by mirrors oI
various shapes: spherical, cylindrical, and conical, convex and concave.
The titles and/or descriptions oI works mentioned in lists oI Ibn al-Hay-
tham`s writings also reveal a sustained interest in natural philosophy, which he
called physics (al-ulm al-tabiyya). In his early career, sometime beIore 417/
1027, he wrote a combined 'summary oI the Optics oI Euclid and oI Ptolemy,
in which he attempted a 'completion oI Ptolemy`s work by oIIering a recon-
struction oI Book I that, apparently, had been missing Irom the Greek text.
This 'summary is not extant, and we can only speculate about the contents oI
that Arabic addition to Ptolemy`s truncated treatise. But it must have been con-
cerned at least in part with what Ibn al-Haytham considered to be the appro-
priate Ioundation Ior Ptolemy`s experimental and psychological treatment oI
a subject which Euclid had dealt with in predominantly geometrical terms.
Also non-extant is a treatise on the 'Nature oI Sight, and another on 'Optics
According to the Method oI Ptolemy.
17
These three compositions, then, and
possibly a Iourth whose title we do not know (Optics, PreIace, para. |8|) were
preliminary exercises, so to speak, that eventually led to the large 'Book oI
Optics (Kitb al-Manzir), in seven treatises, which Ibn al-Haytham com-
posed in mature age (at any rate, aIter AD 1028). That major work, announcing
a decisive break with the basic assumptions oI earlier mathematicians, includ-
ing Euclid and Ptolemy, had the declared ambition oI not simply proposing
an alternative view, but oI building up a new and complete system oI optics
based on new Ioundations deliberately planned to combine both 'physical
and mathematical doctrines and modes oI argument.
18
As is now well known,
Ibn al-Haytham`s Book of Optics, once rendered into Iatin (almost certainly in
Spain), not long aIter a Iatin version oI Ptolemy`s treatise had also been made
in Sicily Irom the Arabic, quickly established itselI as the chieI authority on its
subject, a status which it maintained among philosophers and mathematicians
in Europe up to the time oI Kepler.It is quite remarkable (perhaps the single
most remarkable thing about the history oI Arabic optics) that the frst math-
ematician to have Iallen under the infuence oI Ptolemy`s Optics as the most
developed mathematical treatment oI vision in antiquity was also the one who
wrote the frst treatise that superseded it.This chapter has the immediate aim
oI identiIying and bringing together the main ingredients that went into the
making oI Ibn al-Haytham`s revolutionary project.
II 1nr Suis+r:+ux: A Pnrxoxrx:iis+ Pn.sics or Iin+
It is clear that Ior Ibn al-Haytham, as Ior Euclid and Ptolemy, 'optics/ilm al-
manzir/h optik techn, was a study essentially concerned with visual per-
ception. Everybody had accepted that the presence oI external light (however
understood) was a condition oI vision; and, apparently, Ptolemy, in the lost
Book I oI his Optics, had included a discussion that reaIfrmed the Platonic
doctrine oI synaugeia (iftim, ishtirk), meeting or cooperation oI external
light with visual emanation in the production oI actual vision (Iejeune,
1948, premiere partie).
19
In the remaining Iour Books oI Ptolemy`s treatise,
however, external light appears as a Iactor previously stated but no longer
investigated as an independent agency: the visual ray or fux (visus, which
must have translated basar, itselI corresponding to opsis), spreading out Irom
the eye in the Iorm oI a continuous cone, constantly fgures in Ptolemy`s work
as the indispensable instrument oI vision in all three oI its known modes:
rectilinear, refected, and reIracted; while illuminated color is asserted to be
the primary cause aIIecting the organ oI vision, in such a way as to convey to
the perceiver 'accidents or qualifcations oI colorIor example, shape as the
outline oI a colored area, size as the magnitude oI such an area, motion as the
changing spatial relations between adjacent colored areas. True vision (vere
videre) happens when the direct ray is blocked by a dense and shining object
(lucidum spissum) which thus signals its presence where it is actually located
with respect to the viewer (Ptol. Opt. II, |4, 26|; see below, sec. IV). Untrue
vision occurs when the ray is defected by a refecting or reIracting surIace,
thereby giving rise to the perception oI an imaginary object lying behind the
surIace. This account was to be entirely abandoned or drastically reinterpreted
when Ibn al-Haytham decided to reject the visual-ray hypothesis altogether,
proposing instead a theory oI vision based on a coherently articulated theory
oI light as an independent physical propertya property whose behavior in
rectilinear propagation, refection, and reIraction was subject to experimen-
tally verifed rules, and whose characteristic physical eIIect on physiological
systems oI vision initiated specihed processes ultimately ending in the visual
perception oI external objects and oI all their visible qualities or properties
(man/intentiones) through mental operations oI 'inIerence (qiys, istidll)
or interpretation.
In the PreIace to his Optics, Ibn al-Haytham announced a new starting
point Ior his investigation: the mathematical treatment oI vision was to be pre-
served, as he declared, and thus the traditional conception oI optics as a math-
ematical science remained, but it had to be based on a correct understanding
oI light as the sole agent oI vision. As viewed by him, the crisis oI the sci-
ence oI vision ('conIusion was his term Ior the crisis) consisted in what he
regarded as an unsatisIactory separation between a rigorous, and to that extent
commendable, mathematical approach, and a not-so-thorough approach, to be
Iound in the works oI natural philosophers, which nevertheless had physical
truth on its side. There was, thereIore, need Ior a new 'synthesis (tarkb) that
combined the advantages oI these two opposed methods.
It must be emphasized that Ibn al-Haytham was contrasting, not two
authorities, but rather two methodologies, and that what he consciously looked
Ior, was not a synthesis by means oI juxtaposition, but a new system oI expla-
nations to be discovered only by a Iresh inquiry based on newly established
'principles and premisses. And, as the PreIace also announced, the proposed
inquiry consisted not in simply conceding what was due to existing doctrines,
but frst in subjecting the relevant 'particulars and 'properties (oI vision and
oI light) to 'inspection and then, and only then, in 'gathering by induction
(istiqr) what is Iound in the manner oI vision to be uniIorm, unchanging and
not subject to doubt (Optics, I, 1|6|).
20
Consistently with the declared aim oI
the book, and in agreement with the proposed procedure outlined in the PreI-
ace, the chapter Iollowing the PreIace, on the 'properties oI sight, immedi-
ately presents the reader with a general, orderly description oI the 'conditions
oI vision as revealed by a series oI detailed observations supported by care-
Iully described experiments: the existence oI distance between eye and object;
the existence oI unobstructed straight lines between points on the surIace oI
the eye and points on the object`s perceived surIace; luminosity oI the object;
a minimum size oI the object that varies with the strength oI the viewer`s eye-
sight; opacity and hence color oI the seen object; and ascertainable variability
oI distance with the size oI the object, the degree oI the object`s luminosity and
color, and with the strength oI eyesight. These were the general, empirically
established data oI vision to be explained in terms oI equally general 'prop-
erties oI light to which Ibn al-Haytham turns in the next chapter in Book I,
with more details to be added in Books IV and VII on refection and reIraction,
respectively.
The theory describes how the physical property oI light shines or radi-
ates (ashraqa) Irom a selI-luminous object (the sun, a star, a fame) in straight
lines Irom every point on the surIace oI the object in all directions; how it
rectilinearly extends (imtadda/extendere) through a transparent medium (air,
water, glass); how it is refected Irom smooth surIaces at a certain angle and in
a given plane; and how it is reIracted as it passes through surIaces separating
media oI diIIerent transparencies. The theory also takes care to describes how,
once 'fxed (thabata) in the surIaces oI illuminated opaque (nontransparent)
objects, or in the body oI transparent media considered as always endowed with
a certain degree oI opacity, the light will shine Iorth in exactly the way it does
Irom selI-luminous objects: that is, Irom every individual point where the light
is fxed in the surIace oI the opaque object or in the incompletely transparent
medium, on all straight lines that can be drawn Irom that point. In all oI these
statements the theory makes no appeal to any metaphysical entities or doc-
trines and no reIerence to anything other than the observable phenomena. The
resulting classifcation oI all 'lights is consequently a classifcation oI these
phenomena: 'essential light is the observable light in permanently luminous
bodies; 'accidental light is that observed in bodies when they are illuminated
by external sources; 'primary light is what radiates Irom essential light; and
'secondary light that which radiates Irom accidental light. All oI these 'spe-
cies or kinds oI light are 'Iound to behave in exactly the same manner; and
what are classifed as opaque, smooth, or reIracting surIaces are all 'Iound to
behave in identical ways with respect to all kinds oI light; and, fnally, all lights
are said to weaken in strength or intensity as they recede Irom their respective
sources. Thus we are led to characterize the 'physical theory oI light under-
lying Ibn al-Haytham`s Optics as a phenomenalist theory: the theory invokes
no hidden entities or properties, and it is merely concerned to establish regular
Ieatures oI the behavior oI light by reIerence to direct observation and experi-
mentation.
Color, in Ibn al-Haytham`s theory, is a distinct property oI material,
opaque bodiesdistinct, that is, Irom light. SelI-luminous objectsIor
instance the sunare said to have 'something that behaves like color, or 'oI
the nature oI color, which means that besides their intrinsic luminosity they
are also opaque. Colors exist in other opaque bodies whether these are illumi-
nated or not. We do not know (we cannot know) whether colors extend them-
selves into the adjacent medium in the absence oI light. But when illuminated
they are 'Iound to 'radiate in the company oI the illuminating light, with
which they 'mingle; we can Iurther veriIy that colors behave exactly as light
does in rectilinear transmission, refection, and reIraction; and we know that
they are perceived only when they enter the eye mixed with light. Thus all the
experimental statements about light in the preceding paragraphs also hold Ior
color, and Ibn al-Haytham`s theory oI light is at the same time a theory oI the
distinct property called 'color.
In his theory oI light and color, as well as in his theory oI visual percep-
tion, Ibn al-Haytham employs the Aristotelian-Peripatetic term 'Iorm (eidos/
sra) to reIer to the physical properties involved. His experimental, phenom-
enalist account (in Optics, I, 3) oI light and color as objective properties gains
nothing in explanatory power by substituting 'Iorm oI light and 'Iorm oI
color Ior 'light and 'color. The term 'Iorm does not in Iact occur until
somewhat late in this account (Optics, I, 3|113|), when the discussion turns to
color, and it is not always adhered to aIterward. But the use oI these expres-
sions was oI course in keeping with Ibn al-Haytham`s declared aim to inject the
mathematical study oI vision with what he believed to be objective, physical
truth. He does, however, introduce a crucial refnement oI the concept oI Iorm
which is essential to the Iulflment oI his program Ior a physical theory oI light
that is also amenable to mathematical treatment. (And as will be seen pres-
ently, 'Iorm was the term inextricably linked to the natural-philosophical doc-
trine oI minima naturalia which Ibn al-Haytham accepted.)
21
This refnement
appears in his concept oI point-forms oI light and color. The expression itselI is
not used by Ibn al-Haytham, but he repeatedly speaks oI the 'Iorms oI points
oI light and oI color in the surIaces oI visible objects, and it is these points oI
light and color that extend themselves on straight lines Irom illuminated points
on the objects (whether these are selI-luminous or not), and Irom points within
an illuminated transparent medium like air or water.
The idea or, iI we like, the assumption, is that, Ior any material body or
part oI a body, opaque or transparent, to carry, receive, repel or transmit the
'Iorm or property light (or color) it has to be oI a certain minimal size. II the
body is 'divided Iurther (say, by narrowing the naturally luminous (or illumi-
nated) part on the body`s surIace, or by narrowing the aperture through which
the light passes in a transparent body), the light will vanish. Thus the 'point
Irom which light shines in all directions must have a minimal, fnite size; and
the 'ray proceeding Irom this 'point in the adjacent medium, what Ibn al-
Haytham calls 'the least light (aqall al-qall min al-daw) or 'the small-
est light (adaqq al-daqq min al-daw, asghar al-saghr min al-daw), must
occupy a minimal fnite width through the middle oI which the mathematical
ray can be imagined to pass.
22
It will be remembered that Newton`s Opticks (1704), which had also
deliberately presented a phenomenalist/experimentalist theory oI light claim-
ing independence Irom any hypothesis about the 'nature oI light, operated
with what he regarded as a neutral concept oI the 'light ray as 'the least Iight
or part oI Iight passing through a suitably small aperture. He, too, assumed
that a minimal width oI the passing light beam could be isolated by suIfciently
dividing or narrowing the beam`s path, while additionally assuming that, by
simultaneously allowing a 'least part oI light to pass alone in the direction oI
propagation (e.g., by successively chopping oII the beamperpendicularly to the
direction oI propagationsay, by means oI a Iastly rotating shutter), a single
minimal part could be isolated |Opticks, pp. 12|. By substituting the concept
oI ray as 'least part oI light Ior the covert reality oI a light corpuscle, Newton
thought he could claim to be presenting a purely experimental and nonhypo-
thetical, or phenomenalist theory which he hoped would accommodate a host
oI light properties including, not only rectilinear propagation, refection and
reIraction, but also diIIerential reIrangibilities oI color, diIIraction/infexion,
fts oI easy refection and oI easy transmission, and polarityall oI which he
believed to be inseparable and indeed 'connate dispositions oI the light 'ray.
Ibn al-Haytham`s position was made much easier by the severely limited reper-
toire oI light phenomena that he had to deal with; and thus the simple concept
or assumption oI a least part, already supported by a widely accepted natural
philosophy, was enough Ior endowing the mathematical ray, strictly an imag-
inary sraight line or a straight direction oI activity, with solid physical real-
ity, that is, the phenomenal reality oI a sensible, diIIused quality oI a material
object. And, perhaps even more important, such a simple concept was all that
he Ielt was needed Ior attacking the specifc and basic problem of vision he set
out to solve in his Optics.
A theory oI light and color as properties oI material objects is not itselI
a theory oI vision. But it is clear that Ibn al-Haytham`s theory, whatever the
range oI its intrinsic possibilities, is utilized in his Optics only to the extent
required Ior establishing the new theory oI vision. It has been stated above
that the Ioundational theory oI light is Iree Irom metaphysical presupposi-
tions, Ior instance such as those associated with the so-called thirteenth-
century Iatin 'perspectivists, such as Roger Bacon, John Pecham, and Witelo,
whose enthusiastic appropriation oI Ibn al-Haytham`s ideas should not be
allowed to obscure the Iact that their basic commitment was to a very diI-
Ierent project. Would it be equally true to assert that the theory also Iailed to
envisage a mechanical structure oI matter necessary Ior or at least enhancing
the understanding oI the behavior oI light? In Books IV and VII oI the Optics
Ibn al-Haytham contemplates, and indeed explores at some length, 'explana-
tions (he calls them ilal, causes or reasons) oI optical refection and reIraction
in terms oI concepts oI motion, speed, density, impact, resistance, repulsion,
and the concept oI itimd (endeavor, pressure), current in earlier and contem-
porary Mutazilite kalm discussions; and all oI these concepts are directly
applied to small, solid, spheres projected against hard or yielding surIaces.
23
What should we make oI these analogies (Ior that appears to be their Iunction
in the Optics) which, we know, later attracted IruitIul attention in Europe, Ior
example Irom Kepler, and especially Irom Descartes who Iound some oI their
mathematical Ieatures well suited to his own mechanistic view oI the behavior
oI light? I pose this important question here although it is not essentially rel-
evant to my present concern, since these analogical explanations (or 'compari-
sons, as Descartes called them) do not play a part in Ibn al-Haytham`s account
oI vision. Ibn al-Haytham`s own opinion seems to be in Iact that, strictly speak-
ing, considerations such as those involved in the study oI mechanical collision
oI bodies or the penetration oI yielding media did not properly belong in a trea-
tise on 'optics in the restricted sense oI a theory oI vision.
24
III Ax Ix:r-orirx+ri 1nror. or Visiox
Ibn al-Haytham`s theory oI vision was the only one circulating in Europe, up
to the time oI the Renaissance, that interposed between the center oI vision
and the seen object a surIace on which a confguration oI illuminated points
oI color directly corresponded to their arrangement in the feld oI vision. That
interposed surIace was the slightly fattened spherical surIace oI the crystal-
line humor, and the visually relevant class oI points oI light and color existing
in it marked intersections oI the surIace with the straight rays proceeding Irom
points in the feld toward the center oI the eye, or vertex oI the geometrically
dehned 'visual cone. The theory maintained that 'perception/idrk/ compre-
hensio oI any object in the feld, and oI all its visual properties (size, shape, dis-
tance, and the rest), consisted in a mental reading oI this color mosaic (which,
alone, is said to be frst 'sensed or registered on the crystalline`s surIace)
aIter it has been transIerred as a coherent whole through the humors oI the eye
and through the optic nerves, and aIter being ultimately presented to the brain
where the fnal reading process was perIormed by a sense-Iaculty understood
as a Iaculty oI discrimination and judgment (tamy:).
The analogy between the interposed surIace in Ibn al-Haytham`s theory
and the 'picture plane envisaged by the Renaissance perspectivists is worth
noting: the analogy would not hold iI Ibn al-Haytham`s eye Iunctioned as a
pin-hole camera (as in Ieonardo) or as a lens-camera (as in Kepler). But the
diIIerence between a spherical surIace in the one case and a plane surIace in
the other was oI signifcance to the Renaissance perspectivists Ior whom there
also remained the unique task oI artifcially constructing on a fat surIace a pic-
ture which an external eye would read as a representation oI objects deployed
in three-dimensional space. In the present section oI this chapter it will be our
concern to investigate the status oI what Ibn al-Haytham regarded, in Book
I oI the Optics, as a 'Iorm oI the object having the same order or arrange-
ment oI parts on the surIace oI the crystalline as on the surIace oI the external
object, and to consider the Iate oI that concept oI ordered form, and indeed oI
the whole theory as we fnd it in Book I, in the light oI experiments reported
only in Book VII.
Ibn al-Haytham`s was not oI course the frst image- or picture-oriented
theory oI vision; such had been Ior example the theory oI the ancient atom-
ists in terms oI coherent likenesses, or flms or idols (eidla), and the theory
Iavored by Aristotle and by the Peripatetic tradition throughout the Middle
Ages, in terms oI Iorms received in the eye on the analogy oI impressions made
by a signet-ring in wax.
25
But the theory proposed by the eleventh-century
mathematician was the frst to attempt a mathematical way oI constructing a
physically produced 'Iorm (sra) inside the eye that could serve as the imme-
diate basis Ior a Iull mental representation oI the seen object. HenceIorth, and
until the publication oI Kepler`s Ad Jitellionem Paralipomena in 1604, much
oI the history oI optics as a mathematical theory oI vision was concerned with
problems suggested by Ibn al-Haytham`s attempted construction, and by prob-
lems arising Irom the new experiments Iully preserved in the Arabic text and
in the medieval Iatin translation oI Book VII. Am I ultimately contemplating
a success story? Yes, but not one that is simple, straightIorward, or historically
uninstructive. AIter all, Ibn al-Haytham wrote some nine hundred years aIter
Ptolemy; and Iour hundred years oI rather active interest in vision were to pass
between the transmission oI Ibn al-Haytham`s Optics to Western Europe and
the publication oI Kepler`s treatise.
In Optics, I, 6 ('On the manner oI vision) Ibn al-Haytham clearly Ior-
mulates his problem as one oI identiIying the conditions, necessary and suI-
fcient, Ior obtaining the normal visual perception oI the external world as a
world oI distinct objects and distinctly diIIerentiated colors and shapes (I, 6|5
11|)having previously argued Irom observation and experiment that we see
as a result oI an eIIect (athar, tathr, inhl/ablatio, operatio, passio) produced
in the eye by the agency oI external light (Optics, I, 2 & 4) and having experi-
mentally determined the general characteristics oI the behavior oI light (I, 3)
and having provided a general description oI the structure oI the eye adapted
Irom the current anatomical literature (I, 5). Ibn al-Haytham agreed with the
physicists` view that visual perception was a matter oI receiving Iorms, but he
insisted that asserting this was not enough to explain 'distinct vision, given
what he had just established about how Iorms oI light and color arrive at the
surIace oI the eye Irom points on the Iacing objects (in accordance with rules
oI emission and propagation) and what the Iorms must undergo as they pass
through the eye`s layers (in accordance with the known rules oI reIraction).
A Iull explanation, he argued, must provide means oI somehow isolating the
point-Iorms which originate at distinct points on the objects, but which must
necessarily mingle together on the surIace oI the eye, and inside the eye itselI,
where they must make their frst impressions.This was not only Ibn al-Hay-
tham`s new problem; it was the real problem oI distinct vision.
Now it is important to realize that, both in the way he understood his
problem and in the way he went about fnding a solution in Book I (and later
in Book VII), Ibn al-Haytham simply accepted, on trust, the empirical evi-
dence claimed by Galen and by the dominant medical tradition up to Ibn al-
Haytham`s timeto the eIIect that the crystalline humor, alone among the
parts oI the eye, was the 'frst or 'principal seat oI visual 'sensation. The
evidence consisted in the reportedly observed Iact that vision ceased when the
crystalline alone was damaged or obstructed (Optics, I, 6|14, 15|; II, 2|10|).
26
I say 'important because this generally received doctrine proved to be a cru-
cial Iactor, and, as it turned out, a major 'obstacle, that seemed persistently
to push Ibn al-Haytham`s thought into a certain direction. Since Iorms oI light
and color will come Irom all points on the Iacing objects to all points on the
surIace oI the eye, they will all mingle together as they individually spread over
the whole oI that surIace, and most oI them will have been reIracted beIore
they have reached the crystalline`s surIace, thus giving rise to more mingling
among them; and Iurther conIusion oI Iorms will take place again as a result
oI Iurther reIraction at the crystalline`s surIace. '1herefore |as Ibn al-Haytham
concludes in one oI many passages to the same eIIect in Book I|, the crystal-
line cannot perceive the visible object as it is |i.e., truly and distinctly| unless it
perceives the color and light oI each point on the object by means oI the Iorm
reaching it through one point only on the surIace oI the eye (I, 6|16|, emphasis
added). In other words, Ior a distinct perception/idrk (read: sensation/ihss or
sense impression) oI a given object to be at all possible at the crystalline`s sur-
Iace and within the crystalline`s body, only point-Iorms that have entered both
the surIace oI the eye and the crystalline`s Iront surIace at right angles can be
considered to be visually eIIective. And Ior this necessary condition to be real-
ized, these two surIaces must be considered concentric at the region cut oII by
the 'visual cone defned by the width oI the pupil, namely the cone with ver-
tex at the center oI the eye, and base at the object. As Ior the Iorms that strike
the crystalline`s surIace aIter having been reIracted at the eye`s surIace, they
are (in Books I & II) rendered ineIIective by virtue oI two assumptions. The
frst, expressed briefy in Book I, invokes the superior eIIect oI perpendicular
action (I, 6|24|). The other, much more signifcant and oI greater consequence
Ior the theory as a whole, postulates the crystalline`s natural disposition to
'sense only Iorms that go through it perpendicularly, along the lines oI the
visual cone.
27
As Ibn al-Haytham spells out this second assumption, the crys-
talline, as a transparent body, will reIract the Iorms (or rays) that reach it on
lines intersecting the radial lines, and this reIraction will take place in accor-
dance with the rules oI reIraction; but, as a sensitive body, it will only take
notice oI the unreIracted Iorms that strike its surIace at right angles and that
subsequently pass through its body along the radial lines. ThereIore, what the
crystalline senses in its surIace and throughout its body is a total 'Iorm oI the
object consisting oI a confguration oI point-Iorms oI illuminated color that
correspond one-to-one with all their distinct points oI origin on the object seen;
and it is this sensed Iorm that will eventually reach the common nerve where
two total Iorms Irom both eyes will be united and, fnally, perceived by the ulti-
mate sentient Iaculty (al-hss al-akhr/ultimum sentiens) that resides in the
Iront oI the brain.
Ibn al-Haytham has an answer to one who might object to his argument
as an example oI ad hoc conjecturing or hypothesizing, that is, as something
perhaps unworthy oI the mathematizing natural scientist he claimed to be.
Such an objection, he replies, would not be really justifable, considering that
other privileged lines oI activity are known to be attested in the natural world:
heavy bodies Ireely Iall only in lines normal to the earth`s circumIerence, heav-
enly bodies move only in circles, and light itselI Ireely travels only in straight
lines (Optics, I, 6|43|).He has a point. It is clear, however, that in choosing to
privilege one natural mode oI receptivity in the crystalline over another he was
undoubtedly swayed by the widely received doctrine that assigned sensitivity
in the eye to the crystalline.
We should note that while Ibn al-Haytham is certainly concerned to
explain distinct vision, he does not speak oI distinct forms (or images) in the
eye actually portraying the array oI light and color that distinctly appear on the
object. These Iorms are real, being the eIIect physically produced in the eye
by the incoming Iorms oI light and color. And, in accordance with his physics
oI light and color, he characterizes that eIIect in the eye as actual illumination
and coloration inside the eye. But as the Iorms arrive on top oI one another
at the surIace oI the eye, and subsequently on the surIace oI the crystalline,
they become conIused or 'mixed (mumta:ifa/permixtae) and thereIore do not
by their combined physical eIIect inside the eye represent their original dis-
tinct distribution on the object. Ibn al-Haytham repeatedly speaks, however, oI
Iorms in the eye that have the same arrangement oI parts as the Iorms (or con-
fgurations oI light and color) existing in the objects (Optics, I, 6 |20, 27, 33,
38, 40, 63, 66|). These, as we have just seen, are the Iorms geometrically dis-
tinguished by means oI the perpendiculars drawn Irom the object-points to the
center oI the eye or vertex oI the visual cone. But Ior their distinct perception,
Ibn al-Haytham simply invokes the power oI perception itselI, thus by-pass-
ing the need Ior a segregating optical apparatus (such as a pin-hole arrange-
ment or a Iocusing lens), and immediately and precipitously embarking on a
psychological explanation oI vision.So, here again, we recognize the extent
to which Ibn al-Haytham`s theory in Book I oI the Optics depended on the sec-
ond, physiological assumption reIerred to above (p. 98)an assumption that,
we shall now see, Ibn al-Haytham was led to modiIy in an important passage at
the end oI chapter 6 in Book VII that explicitly and emphaticaly makes room
Ior reIraction as a constant Iactor in vision.
28
That we see objects outside the cone oI vision is a Iact that Ibn al-Haytham
demonstrates in that passage by a simple experiment (fgure 4.1). He places a
slender object (a surgeon`s 'probe/ml/Gr. ml) close to the outer corner oI
one eye while the other eye is closed: the probe will be visible even when it is
defnitely located outside the limits oI the geometrical cone as defned by the
narrow pupil. Thus the light reaching the crystalline Irom points on the probe
will have arrived at the crystalline`s surIace on lines all of which intersect the
lines oI the cone, and, thereIore, all are inclined to the eye`s surIace (i.e., the
portion opposite the pupil) where they will have been reIracted to points on the
crystalline`s surIace where, we are told, sensation will frst take place. In these
circumstances, then, as Ibn al-Haytham concludes, objects are seen only by
reIracted Iorms or rays.
29
Other experiments demonstrate that all objects within the cone are also
seen by reIraction as well as being visible by the agency oI the light or Iorms
that arrive unreIracted at the crystalline`s surIace in the manner explained in
Book I (fgure 4.2). A needle MN is held close to one eye against a white wall
AB. Iet E be the center oI the eye, GH the surIace oI the eye, and TV the
parallel surIace oI the crystalline; G, E, H thus defne the visual cone. CD,
the area on the white background that is concealed by the opaque needle MN,
cannot be seen by rays perpendicular to the eye`s surIace, such as PY, pro-
ceeding rectilinearly Irom points on CD toward E. To be visible, the light
Irom points on CD would have to arrive at the eye`s surIace on non-per-
pendicular lines, such as PMS and PNS`, and must thereIore be reIracted
through GH beIore it strikes the crystalline`s surIace (where it will be reIracted
again). However, as Ibn al-Haytham observes, the eye will actually per-
Eigure 4.1 (constructed Irom the text)
Probe experiment. Only reIracted Iorms Irom E, placed outside the geometrical cone
defned by GEH, will reach the crystalline`s surIace KI. All will be 'sensed by the
crystalline, and all will be 'perceived by the 'sensitive Iaculty as located on the per-
pendicular EE, itselI considered a 'radial line proceeding Irom E though not one oI the
lines constituting the cone that is limited by the pupil`s width.
ceive the screened area CD, though not as clearly as it does the neighbor-
ing areas on AB, but will perceive it as iI it were visible through a transparent
body much broader than the needle`s diameter. The experiment thus shows
that while the eye sees the near side oI the needle by means oI Iorms reach-
ing it along the radial lines, it can only see CD by lights or Iorms that have
been reIracted at the eye`s surIace. Hence the impression oI seeing both the
needle and the screened area at the same time, as iI the much closer needle
had the eIIect oI an interposed transparent bodya conclusion which Ibn al-
Haythamconfrms Iurther byreplacingthe thinneedle byanopaque object broad
enough to prevent the Iorms/rays Irom reaching the crystalline`s surIace Irom
points on the screened area by reIraction. He fnally concludes that what has
been shown with regard to points on CD must be true oI all visible points
on the white wall and oI all points included in the visual cone as previously
defned.
30
Aware that he has just said and done something new, Ibn al-Haytham
proudly declares, at the end oI chapter 6 in Bk VII, that no one beIore him, in
ancient or in recent times, had recognized the Iact that 'all that sight perceives
it perceives by reIraction.
31
Thus we are fnally led to ask: How does the introduction oI an essential
role Ior reIraction as a 'universal cause or explanation (illa kulliyya/causa
universalis)
32
in Book VII aIIect the theory oI vision presented in Book I,
Eigure 4.2 (constructed Irom the text)
Needle Experiment
which the later account obviously makes more problematic? The description oI
the above experiments in Iavor oI this later addition is preceded by a 'general
statement in the course oI which we read the Iollowing:
|a| We have shown in the Eirst Book that iI the sensitive organ |i.e., crystal-
line| were to sense, through (min) every point on its surIace, every Iorm that
comes to it, then it would sense the Iorms oI visible objects as mixed with
one another and not distinctly. This indicates that the sentient organ does not
sense the Iorms except through (min) the perpendiculars to its surIace only,
and that iI it senses the Iorms through the perpendiculars to its surIace then
the visible objects will be distinguished and none oI their Iorms will (appear)
to it mixed. And we have shown in the present Book [JII] that refracted
forms can be perceived by sight only on (al) the perpendiculars drawn from
the obfects to the surfaces of transparent bodies. All that being so, the Iorms
reIracted through the layers oI the eye must be perceived only on the perpen-
diculars drawn Irom the objects to the surIaces oI the eye`s layers.
|b| Now, these perpendiculars are the lines that issue Irom the center oI the
eye. ThereIore sight will perceive all Iorms reIracted through the eye`s lay-
ers and sense them, but will perceive them on the straight lines drawn Irom
the eye`s center. The Iorms oI all objects Iacing the portion oI the eye`s sur-
Iace that lies opposite the uvea`s aperture will thus occur in this portion oI
the eye`s surIace, and will be reIracted through the transparency oI the eye`s
layers, and reach the sentient organ, viz. the crystalline, and the crystalline
will sense them, and the sensitive Iaculty will perceive them on the straight
lines joining the eye`s center and those objects.
|c| I mean that the Iorm oI every point on every object Iacing the portion oI
the eye`s surIace that lies opposite the uvea`s aperture will occur in the whole
surIace oI this portion, and will be reIracted Irom the whole oI this portion
and reach the crystalline humor, and the crystalline humor will sense all that
will reach it oI the Iorm oI every one oI these points, and the sensitive Iaculty
will perceive all that will reach the crystalline oI the Iorm oI each point on
the single line that joins the eye`s center and that point.That is the manner
in which sight perceives all visible objects.
33
In the paradigmatic representation oI optical reIraction here cited in the
underlined sentence in paragraph |a|a representation transmitted by Ptol-
emy`s Optics (in terms oI visual rays)
34
and adopted by Ibn al-Haytham and
Iully stated earlier (in terms oI light rays) in the same chapter Irom which the
above passage is quotedthe eye sees an object by reIraction on the extension
oI the reIracted ray (Ptolemy`s incident ray) at the point where it intersects the
perpendicular Irom the object to the reIracting surIace. In that representation,
both the direction oI the visible 'image (called khayl by Ibn al-Haytham)
and its distance Irom the eye are determined. Note, however, that Ibn al-
Haytham`s citation here oI the Ptolemaic rule specifes the direction but not the
perceived distance oI the Iorm/sra: the Iorm arriving at the crystalline`s sur-
Iace and reIracted through the crystalline`s body is said to be perceived 'on
(Arabic al/Iatin in) the perpendicular Irom the object to the crystalline`s sur-
Iaceperiod. Thus, neither in Ibn al-Haytham`s 'general statement, nor in his
subsequent experimental reports, is anything said about the meeting or inter-
section, oI reIracted rays or their prolongations, with the perpendicular Irom
the object to the crystalline, this perpendicular being one oI the radial lines
drawn Irom the eye`s center. And there is no mention in these reports oI an
'image/khayl, the word consistently used elsewhere Ior a Iorm/sra that is
seen behind a refecting or reIracting surIace.
What, then, do we fnally learn in Book VII with regard to the perception
oI objects whose light has reached the surIace oI the eye, by way oI rectilinear
propagation or by refection or reIraction?Answer: (1) that these objects are
in all cases 'perceived by reIraction; (2) that some oI these objects (namely
those inside the radial cone defned by the size oI the pupil) will generally,
though not always, also make their eIIect by perpendicular Iorms or rays; but
(3) that all Iorms, no matter how they have arrived at the eye, will be 'sensed
by the crystalline and 'perceived by the 'sensitive Iaculty as Iorms oI objects
situated on the perpendicular Irom the object to the crystalline`s surIace.
As a consequence oI (1)(3), we learn Iurther, and contrary to Book
I, that (4) it is no longer a condition oI distinct vision that a given point on a
visible object be perceived Irom a Iorm reaching the crystalline Irom a sin-
gle point only on the eye`s surIace (see above, p. 98); rather, the light Irom an
object-point, which always shines in the Iorm oI a cone whose base covers
the portion on the eye`s surIace opposite the pupil, will always stimulate the
crystalline`s sensitivity by means oI all the Iorms that may reach it Irom any
point on that portion. And, fnally, we learn (5) that the concept oI 'visual
cone has been defnitely though quietly re-defned to include lines, such as EE
in fgure 4.1, which is now called a radial line 'by way oI extension (al tarq
al-istira) 'because it resembles |the traditional radial lines| in that it issues
Irom the center oI the eye.
35
The 'obstacle created by the role traditionally assigned to the crystal-
line humor as the sensitive part oI the eye thus remains. But the experiments
described in the Optics, Bk VII, Ch. 6, along with their stated conclusions,
now point to new problems (including oI course the question as to where on
the perpendicular is the object seen) and to possibilities which some oI Ibn al-
Haytham`s readers in the Muslim world (e.g., Kaml al-Dn al-Eris) and in
Europe (especially Kepler) were led to grapple with.
36
IV A Nrv Poir ror Ps.cnoio.:1nr Lx:xrir or Drr+n Prrcrr+iox
In Book II oI the Optics, having already argued against the visual-ray doctrine
as Iutile (abath wa fadl/superuus et otiosus, Optics, I, 6|54|Risner 14:26)
and even 'impossible/mustahl (Optics, I, 6|59|), Ibn al-Haytham goes on to
expose the argument levelled by 'mathematicians |and by Galen| against the
view preIerred by the 'proponents oI physical science: '. . . iI |so the argu-
ment went| vision takes place by means oI a Iorm which passes Irom the vis-
ible object to the eye, and iI the Iorm occurs within the eye, then why does sight
perceive the object in its own place outside the eye while its Iorm exists inside
the eye? He answers: 'Those people have ignored the Iact that vision is not
achieved by pure sensation alone, and that it is accomplished only by means
oI discernment and prior knowledge (bi-al-tamy: wa taqaddum al-marifa/per
cognitionem et distinctionem antecedentemRisner 39:34), and that with-
out discernment and prior knowledge sight would achieve no vision whatever
(lam yatimma li-l-basari shayun min al-ibsr/ non compleretur in visu visio),
nor would there be perception oI what the visible object is at the moment oI
seeing it (Optics, II, 3|71|, emphasis added). This passage, occurring at the
beginning oI a long discussion oI depth perception, points out, clearly and suc-
cinctly, the aim and scope oI the whole project oI the Optics. Ibn al-Haytham
is saying that no theory oI visual perception in terms oI the reception oI Iorms
or images (which according to him are analysable in terms oI the eIIect oI light
radiation in the eye) can do without the support oI a theory oI vision which
is essentially psychological in character. It is thereIore no accident that psy-
chological investigations have come to occupy such a considerable part oI the
Optics: Book II, which is wholly devoted to a general theory oI the psychology
oI vision; Books III and VI, on the errors oI rectilinear and oI refected vision
respectively; and the parts oI Book VII that deal with the errors oI vision by
reIraction, including Ibn al-Haytham`s new explanation oI the 'moon illusion
as a psychological phenomenon (Sabra, 1987, 1991/92, 1995/96).
There is, thereIore, a necessary connection between Ibn al-Haytham`s
theory oI vision as an image-oriented theory and the particular emphasis he
lays on psychology. More than that, this connection can easily be shown to
have substantially infuenced the very character oI his psychological explana-
tions. II, as Ibn al-Haytham believed, distinct vision required a distinct image
as the only accessible message Irom the object, then the mental activities oI
'recognition, 'comparison, 'discernment, 'judgment, and 'inIerence,
as operations accompanying every act oI visual perception beyond the mere
reception oI sense impressions, must be concerned in the frst place with the
image itselI as the sole optical datum. The signifcance oI this remark can per-
haps be best appreciated when we compare Ibn al-Haytham`s psychological
explanation oI depth perception with visual-ray accounts such as we fnd in
Ptolemy and Galen.
BeIore we do this, let us note briefy that the Iorm/image fnally pre-
sented in the common nerve to the sense Iaculty, though still diIIerent Irom
the image produced by a pin hole or a lens, can be said to be distinct in a sense
that is not applicable to Iorms/images in the crystalline. The reason is that the
Iorm/image in the common nerve has already been disengaged (selected and
isolated) Irom the countless Iorms with which it was mixed beIore it was sent
oII, alone and undistorted, to the common nerve. Ibn al-Haytham thus speaks,
in Optics, II, 3|47|, oI the eIIect exerted by the arriving Iorm on the pneuma
that flls the common nerve as one oI luminosity and coloration, but now these
qualities are understood to represent, distinctly and accurately, their distribu-
tion on the object: 'The Iorm then reaches the cavity oI the common nerve,
whereupon that part oI the sentient body | pneuma| in that cavity where the
Iorm oI the object has arrived becomes colored by the color oI that object and
illuminated by its light. II the object is oI one color, then that part oI the sen-
tient body will be oI one color; iI the parts oI the object have diIIerent colors,
then the colors oI the portion oI the sentient body that is in the cavity oI the
common nerve will be diIIerent.
As A. Iejeune has already demonstrated in 1948, Ptolemy`s explana-
tion oI the perception oI the distance oI a visible object along the stretch oI the
visual ray, this distance being a Iactor in discerning the object`s localization in
space, amounted to nothing more than the assertion that estimation oI the dis-
tance was simply due to our consciousness oI the length oI the ray we have sent
out to grasp the object (Iejeune, 1948, IV.1, pp. 8695; Ptol. Opt., II, |26|). Not
much oI an explanation, perhaps, to a modern mind, but it must have counted
as one oI the points actually supporting the widely received visual-ray hypoth-
esis, both in antiquity and in the early Arabic tradition, and even later in Islam
and in Europe. It will thus be interesting to reIer here to one oI the most reveal-
ing passages in Galen`s De placitis Hippocratis et Platonis where he deIends
the view adopted by the 'geometers, in words reminiscent oI the position and
language Iound in Ptolemy`s treatise. Galen states frst that 'the proper object
oI sight, which I also call its primary sense-object, is the class oI color. Eor col-
ors are the frst thing it perceives, and it perceives them by itselI, and it alone
oI all the sense organs can perceive them, . . . sight alone can discern (syndia-
gignskein) along with the color oI the thing seen its size and shape, . . . There
is a detailed discussion oI such things in the fIth (book) On Demonstration. In
any case, sight can discern, in addition to other things, the position (thesin) and
distance (diastma) oI the colored body. . . . I demonstrated in what was said
there |in the fIth (book) On Demonstration| that everything supports the view
that the observed body is seen in the place where it actually is. 1his is clearly
revealed also through sense-perception itself, and for that reason the geome-
ters give no proof of it but posit it as self-evident,. . . , emphasis added.
37
Reading Ibn al-Haytham on the question oI how an object is seen in its
own place (mawdi/locus) we fnd ourselves in an entirely diIIerent Iramework
oI explanation. The question is split into others concerning the perception oI
distance as such or remoteness/bud/ex remotione, direction/fiha/ex parte uni-
versi, and the magnitude or measure oI distance/miqdr or kammiyyat al-bud/
ex quantitate remotionis (Optics, II, 3|67|Risner 38:4243). The same dis-
tinctions (or some oI them) are in Ptolemy (Ptol. Opt., II, |26|; Iejeune, 1948,
pp. 87II). But now we are treated to an incomparably rich discussion the like
oI which in scope and sophistication is not Iound in earlier writers in Greek
or Arabic. Most important in this discussion is the Iact that it is dominated
throughout by strict adherence to the various elements oI an integrated theory,
including especially the all-important element represented by the newly argued
concept oI an optical image in the common nerve. The passage quoted earlier
(p. 104) Irom Optics, II, 3|71|, continues as Iollows: 'Eor what the object is is
not perceived by pure sensation, but |either| by recognition or by resuming the
|original| discernment and inIerence at the |current| moment oI vision. Thus
iI vision were eIIected by pure sensation alone, and iI all perceptible proper-
ties |size, shape, etc.| in the visible objects were perceived only by pure sensa-
tion, then the object would not be perceived where it is unless it was reached
by something which touched and sensed it. But iI vision is not eIIected by pure
sensation alone; and iI all perceptible properties oI visible objects are not per-
ceived by pure sensation; and iI vision is not accomplished without discern-
ment, inIerence, and recognition; and iI many oI the visible properties are
perceived only by discerment; then to perceive a visible object in its own place
there is no need Ior a sentient |thing| to extend to it and touch it.
In this context, the sentient |thing|/al-hss is oI course the pneuma or
visual fux, and it should be clear against whom this argument is addressed.
Other passages in the Optics also make it clear that by pure sensation/mufarrad
al-hiss Ibn al-Haytham means nothing more than the light and color excita-
tion in the crystalline and in the common nervenot even the awareness oI
light as strong or weak, or the awareness oI color as red or blue, these modes
oI awareness being actually 'judgments involving the various mental opera-
tions mentioned (II, 3|4966|). But the crucial point to be borne in mind (espe-
cially in comparisons with Aristotle and Ptolemy) is that these operations do
not only presuppose 'pure sensation as 'aIIection (inhl/passio) suIIered by
the eye. They are applied in the frst place to Ieatures oI the 'ordered Iorm/
image that has been transported to the common nerve. Eor although the imme-
diate eIIect oI light, as Ibn al-Haytham tells us, is a sensation/excitation 'oI the
nature oI pain and can, as such, share the manner oI nervous transmission oI
painIul and tactile excitations (I, 6|67|),
38
it is also a sensation oI luminosity
and colors and their confguration (I, 6|81|). And, as such, the sensation must
be mediated through the visual apparatus by a succession oI physiological pro-
cesses according to defnite modes oI transmission: that is, along segments oI
the radial lines within the crystalline, then on suitably reIracted lines within
the vitreous, and fnally along the fbers stretching through the curving optic
nerveall oI these diIIerentiated modes oI transmission being designed to pre-
serve the integrity oI the object`s Iorm/image (I, 6|8081|).
39
Ibn al-Haytham calls those Ieatures in the image 'signs or 'cues,
amrt/signa (Optics, II, 3|22, 24, 45|). Thus the size oI the image on the crys-
talline`s surIace (equivalent to the visual angle) can be a cue to the object`s size,
the image`s shape a cue to the object`s shape, the image`s brightness or obscu-
rity a cue to the object`s nearness or remoteness, and so on. Hence the char-
acteristic complexity we fnd in the psychological explanation oI estimating
distance in the Optics. Eirst, we are told that, as with all properties oI visible
objects other than their luminosity and color, distance is perceived separately
only by discernment, but as a result oI accumulated experience it is perceived
by 'prior knowledge |72|. Certain experiences (opening and closing the eye-
lids, turning our eyes toward or away Irom the object seen) lead to the judgment
that the object is spatially separate Irom us |73|. With time, this discernment,
necessary at frst, becomes automatic or, as Ibn al-Haytham puts it, 'estab-
lished in the soul |74|. While objects are always seen to be at some distance
Irom the eye, the magnitude oI their distance is not always 'ascertained. II an
object`s distance lies along a sequence oI contiguous bodies, then the distance`s
magnitude will be perceived by estimating the size oI the intermediate bodies
|76|, and this judgement will be correct (or nearly so) when the distance is a
moderate one. A correct judgement oI distance will thus depend on an accurate
(or nearly accurate) estimation oI si:e |7677|. This is supported by observa-
tions oI clouds below the tops oI not very high mountains, and is Iurther con-
frmed by experiments on the ground |7980|. Such 'inIerences, however, are
not possible in the absence oI intermediate bodies, as when we look at a star
high up in the sky |84|, and the general conclusion is drawn that distances can
be ascertained in magnitude only when they are moderate in size, and only
when they stretch along a series oI visible objects whose sizes are ascertain-
able |86|.
What happens then when the distances oI objects cannot be 'ascer-
tained? The answer is that 'the Iaculty oI judgement immediately conjectures
their magnitudes by comparing their distances with those oI |similar| objects
which sight has previously perceived and |the magnitudes oI whose distances|
it has ascertained |87|. Strictly speaking, this is a comparison between cur-
rently perceived 'Iorms and 'similar Iorms previously perceived and ascer-
tained. 'That |says Ibn al-Haytham| is the limit oI what the discerning Iaculty
is capable oI in the process oI attaining perception oI the magnitudes oI the
distances oI visible objects |88|. Accompanying the immediate perception oI
remoteness (distance as such) there is, thereIore, also a perception oI the mag-
nitude oI that remoteness, which is either ascertained or conjectured |90|. And,
with regard to Iamiliar objects 'recognized Irom Iamiliar distances, there will
be 'no great discrepancy between the conjectured magnitude oI their dis-
tances and the real magnitude |93|.
It remains to show how the size oI an object is accurately estimated by
sight, this being a condition Ior ascertained perception oI distance. In his dis-
cussion oI size (Optics, II, 3|135171|) Ibn al-Haytham begins by accepting
Ptolemy`s arguments (against 'the majority oI the mathematicians) in support
oI the view that size is determined not by the angle oI vision alone but also
by the object`s perceived distance and orientation (wad/situs), citing, among
other observations, what is now known as size-distance constancy |135140|.
The angle, or equivalently, the area cut oII on the eye`s surIace by the visual
cone, remains a basic criterion/asl/radix Ior judging size, which itselI varies in
magnitude with the object`s distance |142143|. How, then, is the length oI the
cone determined? We are back to the question simply answered in the visual-
ray theory by postulating immediate consciousness |144148|. The question
is now crucial especially with regard to 'ascertained, as distinguished Irom
'conjectural, distances. And it is at this point that Ibn al-Haytham`s consid-
erations rise to a still higher level oI elaborateness by correlating angles oI
vision, or Iorms oI objects in the eye, with experiences that Iurnish non-optical
measures oI magnitude. He proposes, in eIIect, a sort oI unconscious inIerence
that leads step by step Irom estimating distances oI objects close to our Ieet to
judging distances oI Iarther and Iarther objects on the ground which subtend
progressively decreasing angles at the center oI the eye |149156|. The frst
distance-length, that oI a point on the ground next to my Ioot, is measured by
my height; the si:e oI an immediately adjacent area on the ground that has been
measured (and correlated with corresponding angles) many times by my step-
ping on it or by stretching my arm to it. This gives me an estimate oI the length
oI the ray reaching my eye Irom a point displaced by a given amount Irom the
frst. And so on, by continually comparing ray-lengths with angles (or corre-
sponding sizes oI their images in the crystalline), we get to Iorm notions oI
the size oI terrains separating us Irom the objects whose distances we come to
judge, frst by calculation, and subsequently by recognition based on repeated
experiences and on repeated and unintended verifcations. Ibn al-Haytham
fnally adds that by 'estimation oI distances he does not mean anything quan-
titatively precise, but simply the process oI Iorming a standard in our imagina-
tion Ior making Iurther estimations.
V Coxciusiox
The seven books oI Ibn al-Haytham`s Optics constitute a unifed project with a
single, continuous argument running through them. In the working out oI this
project the Iundamental concept oI Iorm is Aristotelian, and so is the basic
idea oI vision as something initiated in the perceiver by the impression made
in the eye by a 'Iorm combining the qualities oI luminosity and color. But the
punctiform analysis oI Iorms (Vasco Ronchi`s apt term) as a means oI bringing
points oI the object`s light and color into the eye, and as a guide leading them
through the eye`s tunics and humors, is certainly not. Ptolemy is believed to
have presented in the lost Book I oI his Optics a theory describing the role oI
external light, as distinguished Irom the radial emission oI visual 'fux Irom
the eye, but we do not have a precise idea oI the details oI that theory (though
possibly more can be learned about it by excavating the Arabic sources). At any
rate, Ptolemy`s account oI the role oI external light (whatever its precise nature)
was clearly intended to work in company with, and not as a substitute Ior, the
explanations in terms oI the visual-ray doctrine that actually dominates the rest
oI his treatise. Ibn al-Haytham proposed, in contrast, a theory oI light radiation
whose purpose was to do away with visual radiation altogether by relegating
to itselI all Iunctions previously assigned to the visual ray as a physical entity,
while retaining its use as a geometrical abstraction. His theory oI vision Iurther
employed the concept oI an optical image in the eye, the construction oI which
according to the established behavior oI light had not been part oI either the
Aristotelian or the Ptolemaic program. And an immediate consequence oI his
attempts in this direction was his implementation oI a persistent psychological
approach to the whole problem oI visual perception.It is evident that the syn-
thesis worked out in Ibn al-Haytham`s Optics incorporates elements oI various
sorts which it has derived Irom earlier treatments oI vision. It is also evident
that those elements are now made to serve the aims oI an entirely new project
in which they perIorm new Iunctions.
One inherited element in particular, the role assigned to the crystalline
humor as the sensitive part oI the eye, had two distinct consequences. The frst
was that it encouraged (perhaps even pushed) Ibn al-Haytham to postulate a
geometrical structure oI the eye that retained the basic geometry oI the visual
cone whose lines alone defned the course oI eIfcacious rays (or point-Iorms)
within the crystalline. Thus he lost sight oI a crucial Iact oI ocular physiol-
ogywhich is not to say that the 'Iact was lying out there Ior all to see! The
experiments described by Ibn al-Haytham in Optics, Book VII, chapter 6, some
oI whose striking conclusions he boldly articulated and accepted, were strictly
incompatible with the hypotheses developed in Book I in terms oI perpen-
dicular rays. These experiments appear, however, to have come late to Ibn al-
Haytham`s notice. And although they eventually led him, explicitly, to widen
the visual cone in order to admit the experimentally proven eIfcacy oI reIracted
rays within the eye, he thought he could accommodate them by appealing once
more to psychologywhich brings us to the second consequence, namely the
new emphasis on psychological explanations.
In and oI itselI the new level oI emphasis on psychology in Ibn al-
Haytham`s Optics was historically an important development that has recently
attracted the attention oI psychologists and historians oI psychology. The
emphasis allowed him to open a new chapter in the history oI the systematic
study oI vision. But the impressive psychological observations and explora-
tions that we fnd displayed throughout the book were also Irequently overbur-
dened. As an element in a complete explanation oI the act oI visual perception,
the psychological approach was oI course appropriate, indeed indispensable;
but, in Ibn al-Haytham`s treatise, psychology was required to accomplish more
than it could possibly achieve by itselI. This requirement had a lot to do with
the 'obstacle we noted in Section III, and the diIfculties involved are already
evident in Ibn al-Haytham`s theory oI direct vision, Ior example in his treat-
ment oI depth perception. But the inherent problems become especially acute
in Ibn al-Haytham`s treatment oI specular images, where we fnd him Iollowing
an admirably consistent but ultimately doomed course. I believe that Kepler
at one point became aware oI these problems which, as I believe, he correctly
diagnosed, and which he was able to overcome by removing the 'obstacle. It is
my intention to pursue these last observations in a sequel to this discussion.
No+rs
1. Prominent examples oI mathematicians/astronomers who contributed to medical lit-
erature are Thbit ibn Qurra and Ibrhm ibn Sinn in the ninth and the tenth century
AD, Ibn al-Haytham and al-Brn in the eleventh, Nasr al-Dn al-Ts and Qutb al-Dn
al-Shrz in the thirteenth. Thbit (whose medical writings were considerable) is also
known to have practiced medicine, as did other members oI the Ibn Qurra Iamily.
Eor Thbit see Ibn al-QiIt, 1arkh al-hukam, ed. J. Iippert, Ieipzig, 1903, esp. pp.
11619; E. Sezgin, GAS, III (Medizin, etc.), pp. 26063.Work number 44 in the auto-
graph list oI writings by Muhammad ibn al-Hasan ibn al-Haytham in the felds oI 'nat-
ural philosophy and metaphysics, all composed beIore 10 Eebruary 1027, consisted oI
summaries/epitomes (fumal wa fawmi!) oI 30 works oI Galen; see IAU/Mller, 1882,
II, p. 95.Eor al-Brn see C. Brockelmann, Geschichte der Arabischen Litteratur, I
(Ieiden: E.J. Brill, 1943), p. 627 (no. 27), S II(Ieiden, 1937), pp. 87475; S.H. Nasr,
al-Brn. An Annotated Bibliography, Tehran, 1973, pp. 10103.The medical activi-
ties oI Sinn ibn Thbit ibn Qurra and Ibrhm ibn Sinn are mentioned in IAU/Mller,
1882, I, pp. 224, 226.Eor Ts and Shrz see Brockelmann, GAL, S II(1938), pp.
932-33, and II(1949), pp. 27475, S II, pp. 29697.
2. Eor Hunayn`s 1he Book of the 1en 1reatises on the Eye, see MeyerhoI, 1928. The
integral text oI Ahmad ibn !s remains in manuscript (an edition is being prepared by
E. Kheirandish and A.I. Sabra). Two Arabic MSS are: Istanbul, Ialeli 2759(2), and
Ragip Pasha 934; Ior dates and discussions oI Ibn !s, see the Introduction to Ibn al-
Haytham, 1989, vol. II; Rashed, 1997; Kheirandish, 1999, vol. I. The surviving optical
writings by al-Kind are edited, translated into Erench and annotated in Rashed, 1997,
which includes a revision by J. Jolivet, H. Sinaceur, and H. Hugonnard-Roche oI Liber
Jakob Alkindi De causis diuertitatum aspectus, frst edited by A. Bjrnbo and S. Vogl
see Bjrnbo & Vogl, 1912. Also included in the 1997 volume is the work by Qust ibn
Iq on specular images, F !Ilal m ya!ridu f al-mary min ikhtilf al-manzir.
3. Galen/De Iacy, 1980, Second Part, Bk vii, esp. pp. 453II. See n. 37 below.
4. See, Ior example, Hunayn, 1en 1reatises, in MeyerhoI, 1928, Arabic Text: pp. 91;
77, 80; 79, 81.
5. Hunayn, 1en 1reatises, in MeyerhoI, 1928, p. 95 (where manzir is translated as
'glances); Kind, 'Rectifcation, in Rashed, 1997, p. 167, line 11, and passim; and
Qust ibn Iq, !Ilal, in Rashed, 1997, p. 583, lines 1718.
6. A. Bjrnbo & S. Vogl, 1912, pp. 42II.
7. Rashed, 1997, p. 173, lines 1925.
8. Iejeune, 1948, pp. 2224.The work oI the tenth-century philosopher, Abu `l-
Hasan al-mir (d. 382/992) (see al-mir/KhalIt, 1988), Qawl f al-basar wa al-
mubsar ('On Vision and the Object oI Vision), betrays certain interesting similarities
oI ideas and terms with Ptolemy`s Optics which are not easily detectable in the ninth-
century writers reIerred to above. He mentions Ptolemy`s name (but not his Optics) side
by side with Euclid`s as two 'philosopher-geometricians (hukam al-muhandisn)
who upheld a theory according to which |visual| 'light goes out oI the eye in the shape
oI a cone to the 'sensible color; this light is said to be 'strengthened by the external
light beIore it is 'obstructed (yaquhu) by the 'colored body, thereby causing the
'Iorm oI the color to be Iastened (tanaqidu) to the end or base oI the cone, whereupon
the 'light turns back (yankusu munakisan) to the source/origin (yanb) oI the cone.
He then attributes this account oI ittisl (contact between the organ oI sight and the
object oI vision) also to 'Galen and his Iollowers among physicians p. 435.He oIIers
(p. 419) a list oI seven 'visible properties (al-man al-mariyya), in this order: (1)
color, (2) size, (3) shape/fgure (shakl), (4) number, (5) distance, (6) the state oI being at
rest or in motion (hayat al-sukn wa al-haraka), and (7) substance (fawhar). Three oI
these (here numbered 4, 5, 7) are not included in Ptolemy`s Iormal list oI res videndae in
his Optica (II, |2|: 'Dicimus ergo quod uisus cognoscit corpus, magnitudinem, colorem,
hguram, situm, motum et quietem (Iejeune, 1989, p. 12, and n. 2), although 4 and 5 are
actually discussed by Ptolemy. On the other hand, two visible properties in Ptolemy`s
list, namely corpus/fism/corporeity (which should not be conIused with mir`s no. 7:
fawhar), and situs/wad/local position, are missing Irom mir`s list.
Also according to mir, some oI the visible properties are perceived by them-
selves/ bi-al-dht (these are: color, size, shape, being at rest or in motion), while others
(substance, number, and distance) are perceived 'incidentally/bi-al-arad; only color
is perceived 'frstly (mudrak
un
idrk
an
awwaliyy
an
), whereas size, shape, and the state
oI being at rest or in motion are each perceived 'by second intention/ bi-al-qasd al-
thn. Some oI these expressions correspond to similar ones in Ptolemy: compare, Ior
example, mir`s hrstly and by second intention, with Ptolemy`s primo, and sequenter,
respectively (Ptol. Opt., II, |3, 6|.
The above similarities and discrepancies immediately pose the question as to what
might have been mir`s source(s). I hope to discuss this question in more detail in
another publication. Here I am inclined to postulate a Peripatetic origin Ior these terms
and concepts in both Ptolemy and mir.
9. Eor Ibn s, see Ibn al-Haytham, 1989, vol. II, Intro., pp. xxxvixxxvii, and lix, n.
85; Ior al-Kind`s text see Rashed, 1997, pp. 424427.
10. The relevant text is quoted in Ibn al-Haytham, 1971, pp. 7477. See also Ibn al-
Haytham, 1989, vol. II, Intro., pp. xxxivxxxvii.
11. Ibn al-Haytham, 1989, vol. II, Intro., pp. lxixlxx, n. 110, and p. lxxi, n. 112.
12. Incidentally, this is a Iact to be borne in mind when considering the role oI optics
in Arabic astronomical investigation up to the time when Ibn al-Haytham`s Optics
became generally known to astronomers, that is, aIter the composition oI Kaml al-
Dn al-Eris`s 'Commentary on it at the end oI the 13th century.On the astronomer
Muayyad al-Dn al-Urd`s knowledge oI the Optics see Sabra, 1998/99, p. 307, n. 24.
13. Ibn al-Haytham, 1989, II, Intro., pp. lixlx; Rashed, 1993; Sabra, 1994c.
14. IAU/Mller, 1882, II, pp. 9394.
15. Hogendijk, 1985, pp. 6263.
16. Eor the latter topic see Sabra, 1998, pp. 2730, and reIerences cited there.
17. IAU/Mller, 1882, II, p. 87 (no. 20), and p. 98 (no. 27); Ibn al-Haytham, 1989, vol.
II, Intro., p. xxxiii.
18. Ibn al-Haytham, 1989: Optics, I, 1|2|; vol. II, pp. 47.
19. Eor the Arabic translation oI synaugeia in Plato`s account see Daiber, 1980, pp.
202, 550, 590. I suspect that ishtirk in the sense oI Plato`s term is what underlies com-
municant in Ptolemy`s Optica, in the Iatin translator`s PreIace, para. |2|see Iejeune,
1989, p. 5, and Iacing translation. Eor translating this verb in the same paragraph,
Iejeune has 'sassocient; Smith, 1996, has 'interact.
20. The reader should note that the PreIace and the two Iollowing chapters oI Book I
are missing Irom the Iatin version.
21. See Ibn al-Haytham`s Qawl f al-Daw ('Discourse on Iight) in Rasil, pp. 17
18.
22. Optics, IV, MS Eatih 3215, Iols 62bII.
23. Sabra, 1981, pp. 7278, 9398.
24. At the end oI his interesting mechanical discussion (in Book IV) oI the refection oI
vertical Iree and Iorced motion, Ibn al-Haytham says that 'this not being the place Ior
a thorough discussion oI this matter, it will be enough to make do with what we have
determined regarding the examination oI the motion oI refection (Eatih MS 3215, Iol.
70b).
25. Aristotle, De anima II.12: 424a15. A detailed account oI what might be described
as the most developed 'image-oriented theory oI vision up to the time oI Ibn al-
Haytham can be Iound in Book III (al-maqla al-thlitha) oI the De anima com-
posed by his contemporary Avicennasee Avicenna/Rahman, 1960; Avicenna/Van
Riet, 1972. The theory is basically Aristotelian, but it combines Aristotle`s view oI vi-
sion as the reception oI visible Iorms in the organ oI sight with physiological and pneu-
matic doctrines transmitted through Galen`s works: The 'Iorm oI an object is 'con-
veyed by an actually transparent medium (illuminated air) to the eye where it makes
an impression on the crystalline humor. But this is not perception yet. Visual 'percep-
tion takes place only when the two Iorms in the two crystallines reach the Iront oI
the brain, aIter their uniting together in the optic chiasma where the optic nerves cross
each other beIore diverging again. Only in the Iront ventricle (fargh) oI the brain,
which is flled by the pneuma that carries 'the power oI common sense, will the Iorm
fnally make the eIIective impression received by the common sense itselI, and vision
will be completed.Avicenna reIers to the proponents oI this theory as 'the upholders
oI the doctrine oI simulacra/ashb al-ashbh/auctores sententiae de simulacris
apparently using shabah in this context interchangeably with sra/eidos. (The Medieval
Iatin translation alternates between simulacrum and forma, but also uses other terms
including imagosee Lexique Arabo-Latin in Van Riet`s edition.) Avicenna repeatedly
contrasts this doctrine with that oI ashb al-shu/auctores sententiae de radiis, namely
the doctrine Iavored by the mathematicians, which he rejects. But rather than suggest
any possible way oI reconciling these two positions, Avicenna seems consistently to
present them as mutually exclusive. And there is no hint oI concern Ior a mechanism by
means oI which the Iorm is 'conveyed to the crystalline or Irom the crystalline to the
brain.There is as yet no satisIactory modern translation oI Avicenna`s rich discussion;
see however the section devoted to Avicenna in Iindberg, 1976, pp. 4352.
26. See also Ibn al-Haytham, 1989, vol. II, pp. 5354; and Galen/May, 1968, pp. 463
464, where Galen asserts as a matter oI empirical truth 'that the crystalline humor is
the principal instrument oI vision, a Iact clearly proved by what physicians call cata-
racts, which lie between the crystalline humor and the cornea and interIere with vision
until they are couched. See also the Index in May`s edition Ior related statements on
the crystalline. Ibn al-Haytham does not mention Galen (or any other medical author-
ity) by name, but aIter describing the experiments which we fnd reported throughout
the medical tradition up to his time and later, remarks: 'All this is attested by the art oI
medicine Optics, I, 6|14|.
27. Optics, I, 6|65|Risner 15:7-9; I, 6|90|Risner 17:1-4; II, 2|11|Risner 26:1-9;
Sabra, 1978, pp. 165166.
28. There are reasons (which cannot be discussed here Ior lack oI space) that lead me
to conclude that the passage in question must represent a later development oI Ibn
al-Haytham`s thought, that is, an expansion oI his theory made necessary in light oI
compelling experiments the Iorce oI which he either was not aware oI or had not Iully
realized when he frst conceived the arguments expounded in Book I.
29. MS Eatih 3216, Iols. 102b-103bRisner 269:2646.
30. Ibid., Iols 103b106aRisner 269:46270:29.
31. Ibid., Iol. 106b:89Risner 270:3940; emphasis added.
32. Ibid., Iol. 106bRisner 270:37.
33. Ibid., Iols. 101a:2102a:1Kprl MS 952, Iol. 73a:17-31Risner 268:61269:
16.
34. Iejeune, 1957, pp. 167II.
35. MS Eatih 3216, Iol. 102bRisner 279:24.
36. I may briefy mention here two modern comments on Ibn al-Haytham`s treatment
oI reIraction in the eye. The frst is by MustaI NazI, who gives a lucid account oI the
experiments in Optics VII.6, points out the serious departure Iromthe theory expounded
in Book I, notes howclose Ibn al-Haythamcame in the fnal account to the breakthrough
which we associate with Kepler, and (in connection with this last observation) ends with
the counterIactual: had Ibn al-Haytham proposed that 'the image (khayl) oI the visi-
ble point was located at the intersection oI the perpendicular with the reIracted ray itselI
inside the eye (rather than with its prolongation outside the eye), he would have placed
himselI 'centuries ahead oI his own time (NazI, 1942, pp. 233239).True, but the
Iact that he did not is itselI interesting and historically signifcant.
The second comment is by David Iindberg. He too points out the discrepancy
between Books I and VII, and the impossibility (in some cases) oI intersection between
the perpendicular and the prolongation oI the reIracted ray (Iindberg, 1976, pp. 76
78).But, as indicated above, such an intersection was not intended. Both comments
seem to me to underestimate the eIIect oI Ibn al-Haytham`s initial suppositions and the
implications oI his precipitous though understandable appeal to psychology (see Sec-
tion IV below).
My own analysis above is meant to answer two questions: one is about what drove
Ibn al-Haytham in the direction he actually took, and the other concerns the nature of
the problem he handed down to his successors. I would however agree with Iindberg`s
remark that 'Kepler`s principal innovations were a response to precisely this problem oI
nonperpendicular rays |i.e., those that must be reIracted within the eye| and the neces-
sity oI establishing a one-to-one correspondence |between points on the object seen and
points on the image| (p. 78).
37. Galen/De Iacy, 1980, vii.5: 3240, emphasis added. On the ninth-century Ara-
bic translations oI Galen`s De placitis and oI parts oI his On Demonstration, see
Bergstrsser, 1925, Hunayn`s Arabic text and German translation, pp. 2627/2122 and
4748/3839; Sezgin, GAS, III, pp. 105106, no. 37; see Ptol. Opt., II, |22, 23, 26|.
38. See also Ibn al-Haytham, 1989, vol II, Commentary, p. 56, and Optics, II, 2|14
15|Risner 26:39II.
39. Ibn al-Haytham, 1989, vol. II, Commentary, pp. 7374; Sabra, 1989.
Prrrrrxcrs
Note on abbreviations.The Optics oI Ibn al-Haythamis reIerred to by Book (maqla),
chapter, and Paragraph numbers, in this Iorm: I, 2|3|. These numbers are the same Ior
the Arabic text (Ibn al-Haytham, 1983), and the English translation (Ibn al-Haytham,
1989, vol. I).
The abbreviation Ptol(emaei). Opt(ica). reIers to the second edition and Erench
translation in Iejeune, 1989, by Book and Paragraph: e.g., II, |3|.
ReIerences to Risner`s edition oI Ibn al-Haytham`s Optics in Opticae thesaurus,
1572 are by page and line numbers, e.g., 38:4243.
al-mir/KhalIt, 1988: Abu `l-Hasan al-mir, Rasil. The Arabic texts, edited by
Sahbn KhalIt, and with an Introduction translated into Persian by Mahd Tadayyun.
Tehran: Markaz Nashr Dnishgh.
Avicenna/Rahman, 1960: Avicennas De Anima (Arabic 1ext). Being the psychological
part oI Kitb al-ShiI. Edited by E. Rahman. Iondon: OxIord University Press. Reprint
oI 1959 edition by the University oI Durham.
Avicenna/Van Riet, 1972: Avicenna Latinus. Liber de Anima IIIIII. Edition critique
de la traduction latine medievale par Simone Van Riet. Iouvain-Ieiden: E. Peeters-E.J.
Brill.
Bergstrsser, 1925: GotthelI B., 'Hunain ibn Ishq ber die syrischen und arabischen
Galen-bersetzungen, Abhandlungen fur die Kunde des Morgenlandes, XVII(2),
Ieipzig.
Bjrnbo & Vogl, 1912: A. Bjrnbo, and S. Vogl (eds), Alkindi, 1ideus und Pseudo-
Euclid. Drei optische Werke. Abhandlungen zur Geschichte der mathematischen Wis-
senschaIten mit Einschluss ihrer Anwendungen, XXVI.3, Ieipzig-Berlin.
Daiber, H., 1980: Aetius Arabus. Die Sokratiker im arabischer berlieferung. Wies-
baden: Eranz Steiner Verlag.
Galen/May, 1968: Galen, On the Usefulness of the Parts of the Body (De usu partium).
Translated Irom the Greek with an Introduction and Commentary by Margaret Tall-
madge May. 2 vols. Ithaca, N.Y.: Cornell University Press.
Galen/De Iacy, 1980: Galen, De placitis Hippocratis et Platonis. Edition, Translation
and Commentary, by Phillip de Iacy. In three parts. Corpus Medicorum GraecorumV4,
1, 2. Berlin Akademie Verlag.
Hogendijk, 1985: Jan P. H., Ibn al-Haythams Completion of the Conics of Apollonius,
NewYork: Springer.
, 1996: 'Al-Mutaman`s Simplifed Iemmas Ior Solving Alhazen`s Problem,` in
From Baghdad to Barcelona. Studies in Honour of Prof. JuanJernet, ed. Josep Casulle-
ras and Julio Samso. 2 vols. Barcelona: Instituto Millas Vallicrosa de Historia de la
Ciencia Arabe. Vol I, pp. 61101.
Hunayn ibn Ishq, see Bergstrsser; MeyerhoI.
IAUIbn Ab Usaybia/Mller, 188284, Uyn al-anb f Tabaqt al-atibb, ed.
August Mller. 2 vols. Cairo-Knigsberg. (Reprinted with additions as Islamic Med-
icine, vols 15, ErankIurt: Institut Ir Geschichte der Arabisch-Islamischen Wissen-
schaIten, 1995.)
Ibn al-Haytham, 1971: al-Hasan ibn al-Hasan ibn al-H., al-Shukk al Batlamys, ed.
A.I. Sabra and N. Shehaby, Cairo: Dr al-Kutub.
, 1983: Kitb al-Manzir, Bks IIII. On Direct Jision, edition oI the Arabic text
by A.I. Sabra. Kuwait: al-Majlis al-Watan li-l-ThaqIa wa al-db wa al-Eunn.
, 1989: 1he Optics of, Bks IIII. On Direct Jision. English Translation, Commen-
tary, etc. by A.I. Sabra. 2 vols. Iondon: The Warburg Institute.
, 1357 A.H.: Rasil, Hyderabad, Dn.: Dirat al-MariI al-Uthmniyya.
, see Risner.
Kheirandish, 1999: Elaheh K., 1he Arabic Jersion of Euclids Optics (Kitb Uqldis f
Ikhtilf al-manzir). Edited and translated with Historical Introduction and Commen-
tary. 2 vols. NewYork: Springer.
Iejeune, 1948: Albert I., Euclide et Ptolemee. Deux stades de loptique geometrique
grecque. Iouvain: Bibliotheque de lÙniversite de Iouvain.
, 1957: Recherches sur la catoptrique grecque dapres les sources antiques et
medievales. Brussels: Academie Royale de Belgique.
, 1989: LOptique de Claude Ptolemee dans la version latine dapres larabe
de lemir Eugene de Sicile. Edition critique et exegetique augmentee dune traduction
franaise et complements. Ieiden: E.J. Brill. (Eirst published without the Erench trans-
lation in 1956.)
Iindberg, 1976: David I., 1heories of Jision fromAl-Kind to Kepler. Chicago & Ion-
don: University oI Chicago Press.
MeyerhoI, 1928: Max M., 1he Book of the 1en 1reatises on the Eye Ascribed to Hun-
ain ibn Is-Hq (809877 AD). The Arabic Text, edited Irom the only two known manu-
scripts, with an English Translation and Glossary. Cairo: Government Press.
NazI, 194243: MustaI N., al-Hasan ibn al-Haytham. Buhthuhu wa kushfuhu al-
basariyya. 2 volumes with continuous pagination. Cairo: Eu d I University.
Newton, 1952: Isaac N., Opticks. 4th ed., Iondon 1730. Reprint oI Iondon 1931, New
York: Dover.
Ptolemy: Ptolemaei Optica, see Iejeune, 1989.
Rashed, 1993: Roshdi R. (ed. & transl.), Geometrie et dioptrique au x
e
siecle. Ibn Sahl,
al-Qh et Ibn al-Haytham. Paris: Belles Iettres.
, 1997: LOptique et la catoptrique. Vol. I oI Cuvres philosophiques et scienti-
hques dal-Kind. Ieiden: E. J. Brill.
Risner, 1572: Alha:eni arabis libri septem |Opticae|. In Opticae thesaurus, ed. Eried-
rich Risner, Basel. (Reprint, with Introduction by D. C. Iindberg, NewYork, 1972.)
Sabra, 1972: A.I. S., 'Ibn al-Haytham, in Dictionary of Scientihc Biography, ed. C.C.
Gillispie, NewYork: Scribner, VI, pp. 189210.
, 1978: 'Sensation and InIerence in Alhazen`s Theory oI Visual Perception, in
Peter K. Machamer and Robert G. Turnbull (eds), Studies in Perception. Interrelations
in the History of Philosophy and Science, Columbus, Ohio: Ohio State University Press,
pp. 160185.
, 1981: 1heories of Light from Descartes to Newton. 2nd ed., Cambridge: Cam-
brigde University Press. (1st ed., Iondon: Oldbourne, 1967).
, 1984: 'The Andalusian Revolt Against Ptolemaic Astronomy, in Everett Men-
delsohn (ed.), 1ransformation and 1radition in the Sciences. Essays in Honor of I.
Bernard Cohen, Cambridge: Cambridge University Press, pp. 133153. Reprinted as
chapter XV in Sabra, 1994a.
, 1987: 'Psychology versus Mathematics: Ptolemy and Alhazen on the Moon
Illusion, in Edward Grant and John Murdoch (eds), Mathematics and Its Application
to Science and Natural Philosophy in the Middle Ages. Essays in Honor of Marshall
Clagett. Cambridge: Cambridge University Press, pp. 217247.
, 1989: 'Eorm in Ibn al-Haytham`s Theory oI Vision. Zeitschrift fur Geschichte
der Arabisch-Islamischen Wissenschaften 5, pp. 115140.
, 1991/92: 'On Seeing the Stars: Edition and Translation oI Ibn al-Haytham`s
Risla f Ruyat al-kawkib, Zeitschrift fur Geschichte der Arabisch-Islamischen
Wissenschaften, 7(199192), pp. 3172. In collaboration with Anton Heinen.
, 1994a: Optics, Astronomy and Logic. Studies in Arabic Science and Philosophy.
Aldershot: Ashgate Publishing, Variorum.
, 1994b: 'The Physical and the Mathematical in Ibn al-Haytham`s Theory oI
Iight and Vision. Paper frst delivered in Jyvskl, Einland, in 1973, frst published in
Tehran in 1976, and reprinted with corrections as chapter VII (20 pages) in Sabra,
1994a.
, 1994c: Review oI Rashed, 1993. Isis, 85 (1994), pp. 685686.
, 1995/96: 'On Seeing the Stars, II. Ibn al-Haytham`s Answers` to the Doubts`
Raised by Ibn Madn, Zeitschrift fur Geschichte der Arabisch-Islamischen Wissen-
schaften 10, pp. 159.
A. I. S:ir: 118
, 1998: 'One Ibn al-Haytham or Two? An Exercise in Reading the Bio-Biblio-
graphical Sources, Zeitschrift fur Geschichte der Arabisch-Islamischen Wissenschaften
12, pp. 150.
, 1998/99: 'Confguring the Universe. Aporetic, Problem Solving, and Kinematic
Modeling as Themes oI Arabic Astronomy, Perspectives on Science (MIT Press), 1998,
vol. 6, no. 3, published in 1999.
, see Ibn al-Haytham.
Samso, 1994: Julio S., 'On al-Bitrj and the hay a Tradition in al-Andalus, in idem,
Islamic Astronomy and Medieval Spain. Aldershot: Ashgate Publishing, Variorum.
Chapter XII, pp. 113, frst presented at a Symposium held in Granada, 1992.
Sezgin, GAS. Euat S., Geschichte des arabischen Schrifttums, Ieiden: E.J. Brill, 1967.
Smith, 1996: A. Mark S., Ptolemys 1heory of Jisual Perception. An English 1ransla-
tion with Introduction and Commentary. Transactions oI the American Philosophical
Society, Philadelphia, vol. 86, pt. 2.
III
Mathematics. Philosophy and Practice
The Andalusian jurist, philosopher, and scientist Ibn Rushd, the great Averroes
who died eight hundred years ago in Marrakesh, put his liIe`s work in the ser-
vice oI one project, perceived dimly in his youth, defned ever more clearly in
the course oI a prolonged struggle with the epistemic paradigm oI the religious
community, and brought to Iruition in his years oI maturity in a series oI Iong
Commentaries on fve principal works oI Aristotle: establishing demonstrative
science, the law oI reason, as the basis oI thought and action in human society.
An ever increasing sense oI urgency is pervading all oI his approaches
to this task, seen as an ultimate duty. The recurring prayer Ior a last delay that
God should grant him to achieve his goal accompanies his praise oI Aristotle as
a guide and guarantor. It is a goal most diIfcult to attain where the principles
oI 'natural philosophy (al-ilm al-tab) were at variance with the observa-
tions described and calculated, precisely and predictably, by the mathematical
proIessions. In his Commentarium Magnum on Aristotle`s De Caelo, the solu-
tion oI some remaining doubts is reIerred to the discussion oI Metaphysics:
'Perhaps, iI we shall see the last oI our term in liIe, we may explain this point
when devoting a literal commentary to Aristotle`s discourse on this science
|i.e., Metaphysics|; indeed, this is one oI my highest hopes, and perhaps God,
in his grace and compassion, will help us to live and see this time and to attain
this goalHe is benefcent and generous.
1
But when he fnally achieved his
Great Commentary on 'this science, that is, the Metaphysics or Eirst Philoso-
phy oI Aristotle, toward the end oI his liIe, he despaired oI his task: the task
he had set himselI to explain and justiIy an astronomy true to the principles oI
Aristotle`s cosmology: a valid model Ior calculation as well as a true represen-
tation oI reality:
'We must examine this ancient astronomy Irom the beginning. It is the true
astronomical scheme which is valid in accordance with the natural principles
|al-hayatu l-qadmatu llat tasihhu al l-usli l-tabiyya|. That is, according to
my conviction, an astronomy based on |the assumption oI| the movement oI one
and the same sphere around one and the same center, revolving on two diIIerent
3
M:+nrx:+ics :xi Pniiosorn. ix Mriirv:i Isi:x
Gerhard Lndress
Grrn:ri Lxirrss 122 M:+nrx:+ics :xi Pniiosorn. ix Mriirv:i Isi:x 123
poles and more according to what is ftting the appearances. Indeed, motions
like these can make a star go Iaster and slower, Iorward and backward, and have
all the motions Ior which Ptolemy was unable to fnd a model (an yasnaa lah
l-hayata). . . . In my youth, I hoped to make a complete study oI this, but now
that I have grown old, I have given up this idea because oI the obstacles I Iound
in my way beIore. But this explanation will perhaps induce somebody to study
these things later. In our time, astronomy is not about something real (laysa
minhu shayun mawjdun); the |model oI the| sphere existing in our time is a
model conIorming to calculation, not to reality (muwhqatun li-l-husbni l li-
l-wujdi).
2
This brings us to the heart oI the matter at hand: The relation oI philoso-
phy and mathematics, and the relation oI both to reality. In the fnal analysis,
this concerns the question oI what the true and frst reality is: the ultimate
object oI study Ior the philosopher-scientist.
Knowledge is power. In assuming the prerogative oI defnition, philoso-
phers and scientists, using diIIerent paradigms oI concept and method, com-
peted Ior authority. This authority is based on the belieI that true understanding
oI the being, order and movement oI the world will warrant proper action,
and will constitute the ultimate good and the Ielicity oI man. This conviction
inspired the philosophy oI science Irom Antiquity until the early modern age:
Irom Plato`s Academypori= oyrmrtppto= rioitm
3
until Kepler, who
illumined the frst book oI his Harmonice Mundi libri \
4
with a quotation,
in Greek, Irom Proclus` Commentary on Euclid`s Elements: 'Mathematics
makes contributions oI the very greatest value to physical science. It reveals
the orderliness oI the ratios according to which the universe is constructed, and
the proportion that binds together all that is in the cosmos.
5
1 1nr Pniiosornic:i 1r:ii+iox ix Grrrk Scirxcr
Philosophy as a transmitted text and as a system oI instruction entered Arabic-
Islamic society in the baggage oI specifc social and proIessional groups: oI
scientists and physicians.
It is true that premodern societies did not know the narrow proIessional-
ism typical oI the modern division oI labor. It is nevertheless true that since
early Hellenism, philosophy itselI competed with the individual sciences Ior
recognition oI a proIessional status in society, and sought to Iound its claim
on the unconditioned knowledge (raiotpp ovuao0rto=) oI the principles.
6
On the one hand, the philosophical schools assumed competence, and took
charge oI education, in the mathematical sciences. The conception oI philoso-
phy in Aristotle and the old Peripatetic school had embraced, ideally at least,
the applied sciencesthese in turn being regarded as elements oI a paideia in
the sense oI propaedeutics to philosophy: a stage in philosophical education
leading the way to the advanced level oI dialectic, the 'science oI sciences.
7
Neither in the Hellenistic nor in the Roman period, on the other hand, did
mathematical studies Iorm part oI a general education. Outside the enkyklios
paideia,
8
the 'comprehensive education oI philosophy, such studies were
linked up with, and restricted to, the proIessional training oI engineers, archi-
tects, geometers, and musicians.
9
But here, even in the individual and practical
sciences, the teaching oI the leading authorities and their basic texts maintained
the intimate connection between applied mathematics and its epistemological
and metaphysical background. Beyond the decline oI the philosophical schools
in the civilization oI late Hellenism, the philosophical doctrine oI the principles
and oI the cosmos survived in the gnostic Platonism oI the natural sciences,
in the Neoplatonism oI the mathematicians, in the Peripatetic cosmology oI
Ptolemy, as alsobut this is a matter diIIerent and apartin the elementary
logic reading oI the Christian schola (uskl).
10
This is how Greek philosophy
entered the urban and courtly society oI Islam: as methodology and ideology
oI the proIessional sciences, notably oI mathematics and astronomy on the one
hand, oI medicine on the other. It is a philosophy neither pagan nor Christian
nor Islamic, but universal: a rational religion oI the intellectuals oI Greek eru-
dition, giving an ulterior sense to their activity.
Each scientifc tradition carried its own philosophical discourse: a choice
oI authorities, a methodology, a classifcation and hierarchy oI the sciences,
and a general orientation oI cosmology and ethics. With the physicians we fnd
Galen`s platonism as also Galen`s own logic, anthropology and ethics, com-
peting with philosophy in pretending to teach an ars vitae. (In consequence,
the philosophic or non-philosophic character oI medicine, being techn or
epistm, was under dispute in apology and polemic Irom both sides.) The
mathematician and astronomer, and the proIessional astrologer or geometer,
pretended to a universal competence no less than the physician, but on a diI-
Ierent scale: on the authority oI a time-honored tradition, and oI an eminent
ancestry, in the history oI philosophy itselI. The mathematicians were Pla-
tonists and Pythagoreans in the tradition oI Nicomachus, Proclus and Iambli-
chus. But the astronomers cherished the Aristotelian propaedeutic and, above
all, the Aristotelian cosmology conjoined with the authority oI Ptolemy. Hence
it was Aristotle who came to dominate the system oI the physical world, and it
was a Peripatetic structure which, since being adopted by Ptolemy, prevailed
in the method and epistemology oI proIessional science.
11
1he \orld Is Number. 1he Platonic Heritage
The philosophy oI mathematics owes to Pythagoras two signifcant contribu-
tions, perennial legacies to the history oI thought: frstly, oI the attitude and
the word oI philosophos, and secondly, oI the concept oI mathmata as being
everything that can be precisely known and learned. Even Aristotle will use
mathmata in this basic sense, and there is a long way leading Irom those
essentials oI knowledge to what is being learned as mathematics in today`s
schools. But although the mathmata oI Pythagoras were diIIerent in subject,
method, and the state oI conscious approach Irom the mathematical discipline
oI the later schools, mathematical principles in the proper sense oI number
12
and oI mathematical harmony were Irom the beginning regarded as the very
center oI Pythagorean thought. Everything knowable has number.
13
II we may
trust Aristotle, the Pythagoreans held that numbers have an essential likeness
with things oI the world, sometimes that things are numbers, and that number
is 'the essence oI all.
14
This is, oI course, reIuted by Aristotle, being assimi-
lated to Plato`s doctrine oI eternal Ideas.
15
Through Plato, the mathematical
disciplines oI the Pythagorean canon were introduced as Iorming the basis oI
intellectual education. It was the program oI the Academy: arithmetic, geome-
try, astronomy, harmony, and dialectic. This canon is presupposed as a matter
oI course byAristotle in the opening oI the Posterior Analytics. But it was Aris-
totle who, in the second book oI the Analytica posteriora, extended the math-
mata to include the universe oI knowledge, leaving behind the more narrowly
mathematical paradigm oI the Academy and oI its Pythagorean model.
16
The programme oI the Academy is expounded by Plato as a programme oI
educating the Guardians oI the Republic (Resp. VII, 526II.): The subjects to be
taught are arithmetic, geometry (to which is added stereometry as a special sub-
ject), astronomy, and music (i.e., the science oI harmonical proportions); these
are the only disciplines recognized as sciences in the proper sense, yielding a
priori knowledge oI immutable and eternal reality. They are described in Resp.
book VII with respect to their power oI turning the soul`s eye Irom the material
world to objects oI pure thought. Eirst comes arithmetic, the science oI number,
the numbersideal unitsoI mathematics being considered 'by thought, in
the abstract, only;
17
hence 'this study is really indispensable Ior our purpose,
since it Iorces the mind to arrive at pure truth by the exercise oI pure thought.
18
Geometry, coming second, is equally 'knowledge oI the eternally existent, and
'will tend to draw the soul toward truth and to direct upwards the philosophic
intelligence.
19
The rise oI arithmology on the one hand, and oI musical theory on the
other, in later Platonism and Neoplatonism is closely connected with the re-
emergence oI Pythagoreanism, the mathematical oIIshoot, as it were, oI Middle
Platonism.
20
And as Plato had built upon Pythagoras, Neoplatonism relied on
Neopythagoreanism. Not the metaphysics oI the One and Intellect oI Plotinus,
but the gnostic and occult tendencies oI the later Neoplatonists drew on this
source, leaning toward asceticism, supported by magical practice, contriving
the perIection oI the soul in view oI its ultimate ascent to the world above.
The central philosophic message oI Pythagoreanism, echoed and supported
by many a statement oI Plato as well, made it a religion oI mathematicians:
The world is number; through mathematicals, the transcendental and the
divine can be perceived. As mathematicians, some oI the leading fgures oI the
Pythagorean school, such as Apollonius oI Tyana, were Ioremost in their age.
The 'Introduction to Arithmetic oI Nicomachus oI Gerasa (frst century C.E.),
on the other hand, is the work oI a philosopher rather than a mathematician,
intended as a guide to the late works oI Plato and to the Pythagorean treatises,
frst read by philosophers rather than mathematicians, and still popular at a
time when there were no mathematicians leIt, but only philosophers who inci-
dentally took an interest in arithmology.
21
As a true Pythagorean, Nicomachus makes arithmeticthe science oI number
the primary object oI philosophy, the name oI philosophia being ascribed to
Pythagoras (as was common in the school). This 'wisdom, sophia, is defned as
knowledge oI the truth in 'real things, things immaterial, unchanging and eter-
nal, among which the subject oI arithmetic is Ioremost, because 'it existed beIore
all the others in the mind oI the creating God like some universal and exemplary
plan (/ntrod. arithm., IV.2).
22
Iamblichus, the 4th century disciple oI Porphyry, Iollower oI Plotinus,
put Iorward a program to pythagoreanize Platonic philosophy. As philosophers,
the Iollowers oI Pythagoras made mathematics, starting with arithmetic, not
only the leading propaedeutic art, but also the Ioremost object oI philosophical
study in its own right. 'One signifcant result oI this is the mathematization oI
all areas oI philosophy that is so striking a Ieature oI later Greek philosophy.
23
In the writings oI Proclus, again, Plato supplants Pythagoras as the central
authority, and while accepting the pivotal role played by mathematics in the
philosophical sciences, Proclus chooses geometry rather than arithmetic as the
pre-eminently mediatory mathematical science (as is evident Irom his com-
mentary on Euclid`s Elements). Mathematicals are projections by the soul oI
innate intelligible principles; and it is particularly in geometry, according to
Proclus` teacher Syrianus, that the soul projects such innate principles into
imagination because in its weakness soul is better able to grasp these principles
in the extended Iorms given in the fgures oI geometry.
24
1he 1ransmission ot Philosophy through Science
While the principal works oI these authors may not have been available to
the frst generation oI Arabic scientists, as al-Kind and his contemporaries,
the general attitude to the mathematical sciences clearly goes back to such
authorities, transmitted through the basic manuals and their commentators and
translators. Even Ior astronomy, where the Peripatetic doctrine and attitude
determined the method and many oI the basic cosmological assumptions,
Plato had sanctioned the pursuit oI mathematics as a philosophical assign-
ment. Going beyond the mere calculus oI an auxiliary model, mathematics
established the sympathetic rapport between the higher and lower worlds, man
and the universe. The designation as philosophos oI the savant who in his spe-
cial feld oI application evinced this quality, underlined this selI-image. Hence
the Platonic concept and rle oI the mathematica prevails in the system oI the
mathematical sciences and in the tradition oI number theory and oI the doctrine
oI musical harmony.
But long since, philosophy and the sciences had driIted apart with regard
to their social status and their rle in education and intellectual liIe; philosophy
had ceased to pretend to the status oI a proIession. Even beIore the decline
oI pagan Hellenism, philosophy had lost the remarkable role it had played in
education and intellectual liIe as against the applied sciences; it had also lost its
social statusphilosophy remaining but small Iare in the provisions oI proIes-
sional physicians and astrologists. The last Alexandrian commentators earned
their living not as proIessors oI philosophy, butas indicated by explicit hints
and by the implicit evidence oI their metaphors and their examples given to
illustrate a pointas doctors, grammarians, rhetoricians, and astrologers.
25
Instead oI the philosopher oI universal competence, the authority oI defnition
among the intellectual elite is assumed by the 'philosopher (in Arabic, hakm)
specialist oI the applied arts.
Apart Irom particular proIessional Ieatures, regional traditions persisted
locally Irom pre-Islamic time beIore uniting in Baghdad: On the side oI the
Nestorians working in Sasanid Iran, we fnd b oI Kashgar (c. 600), Iamiliar
with astronomical as well as medical sources; on the side oI the Monophysites
oI upper Mesopotamia, Sergius oI Reshayna translated not only Galen, and
books on astrology, but also works oI the Christian Neoplatonist known under
the name oI Dionysius Areopagita, and the translations Irom the Persian, made
by Severus Sebokt oI Qinnasrin, provided a remarkable range oI astronomical
and mathematical works.
26
Most important Ior our subject is the tradition oI
the philosopher-scientists Irom Harrn, the ancient Carrhae, where worship
oI the heathen star-gods survived until the tenth century: the Sbia, claiming
the protection due to the ahl al-kitb and rising to high stations in the Abbasid
administration. Even though the activity oI Thbit ibn Qurra and his descen-
dants, both as translators and as original mathematicians, does not suIfce to
attribute to these 'Sabian sources every text oI Plato and every Neoplatonic
or gnostic interpretation oI Aristotle leaving traces in the Arabic tradition,
27
the
Platonic-Neoplatonic heritage oI the mathematicians Irom Harrn may go a
long way to explain the knowledge oI Plato, and the infuence oI Platonism, in
the philosophical orientation oI early Arabic science. But Iar beyond this spe-
cifc transmission oI mathematical science, the arithmology oI neo-Pythago-
rean origin pervaded the multiple strands oI the popular and practical traditions
oI Hellenism, Irom the occult sciences to gnomological wisdom literature; the
fnal triumph oI Peripatetic jalsaja as a school oI demonstrative science is due
to its reception and adaptation in a new milieu.
Al-Kindi and the Platonic 1radition
The translations which were commissioned by one oI the leading philosopher-
scientists oI the early ninth century, and infuenced his own writings, cover
a wide range: Ab YsuI Yaqb ibn Ishq al-Kind, astronomer, astrologer,
versed in mathematics and optics, medicine and pharmacy, a polymath oI
his agethe age oI the caliph al-Mamn (813833) and his sons and suc-
cessorswho died in or shortly aIter 866. Taken altogether, the works oI al-
Kind, and the sources made available through his eIIorts, and translated on his
demand, are the most impressive witness to the triumph oI Hellenism aIter an
earlier period where the import oI Iranian traditions had been prominent both
in the political and religious community and in the reception oI science. But
the diIIerent strands, proIessional and doctrinal, are yet unconnected, even in
confict, diIIerent in style and approach, in his vast ouvre.
Al-Kind`s Plato, where he is named, stands Ior the Platonism oI the
gnostic, 'Hermetic subculture oI popular Hellenism, a religion Ior intel-
lectuals like the one upheld by the Sabians, oI mathematicians and astrolo-
gers. An expose oI Neoplatonism, transmitted on al-Kind`s authority by Ibn
al-Nadm,
28
is put Iorward as a doxography oI the Sbia. It is true that more
oI Plato`s authentic works were available in al-Kind`s generation than were
preserved beyond the next century (mainly through the philosophical tradition
oI medical authorsthe tradition oI Galen the Platonist). Among the Platonic
dialogues available was the Meno, the frst exposition oI the doctrine oI recol-
lection (anamnesis). But al-Kind`s own treatise on 'What the soul remembers
oI what it had in the world oI the intellect, while invoking the Platonic con-
cept, is based on the Neoplatonic tradition oI the Arabic Plotinus source and
related texts.
29
What is stressed most Iorcibly here is the existence and nature
oI a priori knowledge: a knowledge oI the eternal principles, but brought
Iorth (tukhrijuh) in the individual soulaIter its exposure to the secondary
intelligibles, the Iorms-in-matter oI the material worldby the autonomous
activity oI reason only. Not everybody is qualifed to attain, to re-collect, this
knowledge in its primordial splendor. By stressing the incorporeal substance
oI the intelligibles and oI the rational soul, al-Kind drives home the ultimate
value oI his science: Only he who purifes his soul will gain true happiness and
the ultimate vision oI truth.
Our main testimony Ior the Arabic Meno is Iound in the work oI a mathemati-
cian: Thbit ibn Qurra`s Epistle 'On the argument ascribed to Socrates on the
rectangle and its diameter.
30
This is the well-known problem used in the Meno
in order to demonstrate how even a mind not trained in mathematics can be
guided toward mathematical insight, because the human mind can be made to
'remember what it obtained in its preexistence while viewing the ideas in the
world above. But the mathematician Thbit does not touch on anamnesis, clos-
ing his mathematical analysis with a very general remark on the goal and value
oI mathematical science. Some oI these matters, he says, are more elevated than
others; whosoever confnes himselI to the basic matters oI geometry, does this
either Irom incompetence, 'like some people in our era, or because he wants
to guide learners through gradual stages according to their capacity. This latter
was the intention oI Socrates, using this problem as a paradigm oI his intention,
Ior 'mathematics (talm, o0poto) is Ior the soul what nourishment is Ior the
bodyas one oI the Ancients said.
31
The Platonism oI Proclus (who had written a long commentary on book
I oI Euclid`s Elements),
32
and the Pythagorean attitude oI Nicomachus oI
Gerasa and oI the Hellenistic theory oI music are obvious in al-Kind`s extant
works and in the titles oI some which have been lost. As a mathematician, al-
Kind was Iamiliar with the tradition oI Iamblichus` treatise 'On the Common
Mathematical Science, which presented mathematics as the absolute object
oI contemplation; he knew the Neoplatonic reading oI Euclid`s Elements,
and he reworked Ior his own use the Neoplatonic philosophers` vademecum
oI number theory, the Introduction to Arithmetic by Nicomachus oI Gerasa,
available in a contemporary Arabic version.
33
Also available was Nicomachus`
'Great Book on Music,
34
and hence, the interest in the actualisation oI perIect
mathematical relations in musical harmony can be Iollowed up in this same
school oI thought.
But this is only one side oI al-Kind`s philosophical program. On the other
hand, he was an heir to, and a conscious continuator oI the Academic tradition
surviving at the hands oI commentators with a Neoplatonic orientation, notably
in the school oI Alexandria and its Byzantine oIIshoots. Both sides compete in
this early period oI translation and adaptation, at times in a striking contrast
oI style. Al-Kind`s Aristotle is not yet the master oI logic and oI demonstra-
tive science, styled the Eirst Teacher by al-Erb`s school in the next century,
emancipating philosophy Irom the applied arts, relegating Plato to an inIerior
rank restricted to sharing out practical, political wisdom. But already, Aristotle
had taken on the rle oI super-philosopher, the 'Ioremost (mubarri:) oI the
Ancients who, in al-Kind`s words, 'Ior luminous, harmonious souls will lead
the way toward the highest spiritual rank.
35
al-Kind`s Aristotle, albeit Platonic
in matters oI theology and cosmology, is representing the encyclopaedia oI the
rational sciences; he is the undisputed authority on the physical world.
The systemoI philosophical (including mathematical) studies is modelled
on the Alexandrian curriculum, well-known through the Neoplatonic commen-
tators oI Aristotle Irom the school oI Ammonius; and the victory oI Hellenism
over the Iranian tradition in astronomy and astrology prepared the way Ior Aris-
totle to become the Eirst Teacher oI Arabic Islamic philosophy. This is evident
in the work oI al-Kind`s rival, the astrologer Ab Mashar (m. 272/886)his
Great Introduction to astrology is the frst Iull-grown handbook to be written in
Arabic Ior any oI the ancient sciencesas well as al-Kind`s own.
36
The result oI al-Kind`s reading is, in more than one respect, a compro-
mise between the obvious contradictions apparent in the Corpus Aristotelicum
itselI, between the 'Platonic and the 'Peripatetic Aristotle, and between the
tendencies oI the Greek commentators. The synthesis arrived at in al-Kind`s
division oI the sciences is a case in point, and refecting a long discussion. The
Platonic tripartition oI the sciences and oI being had been kept by Aristotle
in his division oI the sciences into physical, mathematical and theological
(Metaph. K 7, 1064b1), but was in confict with the ontology oI the later Aris-
totle whoin the fnal analysisdenied the subsistence oI the mathematical
entities as well as oI soul. In the Iurther attempts oI the Academy to reconcile
these positions, Irom Speusippus to Proclus, soul was coordinated ontologi-
cally with the mathematicals.
37
In his introduction to the study oI Aristotle,
38
al-Kind gives the frst
place to mathematics (al-riydiyyt) as a preparation to the study oI philoso-
phy proper (plausible in the contextthere were scarcely any mathematical
writings ascribed to Aristotle
39
and in accordance with the inIormation given
by some oI the commentators, reIerring to the Platonic curriculum).
40
al-Kind
wrote a treatise confrming this attitude, 'That philosophy can be acquired
through the science oI mathematics only.``
41
In the Iollowing classifcation
oI Aristotelian philosophy, we do not fnd mathematics as an 'intermediate
science ranged between physics (on the motion oI sensible substances) and
metaphysics (on intelligible being), but psychology: an intermediate discipline
concerning objects which are independent oI bodily substance, subsistent, but
perceived by the senses in conjunction with bodies. These are the topics oI
psychology, in one instance.But, in another context, the topic oI this 'inter-
mediate science is mathematics. The quadrivium oI mathematics is interme-
diate between the natural sciences on the lower end, andIollowing the way
Irom the multiplicity oI sensual phenomena to the universal simplicity oI the
principles through abstractionmetaphysics and theology. The soul belongs to
the same intermediate realm 'in between: the soul as subject, and the math-
ematicals as object, belong both to the world oI the eternal intelligibles and the
corruptible sensibles. This realm 'in between, to metaxy, is the realm oI the
recollection in the soul oI its prenatal view oI the universals. Iogic, the Iourth
and propaedeutical discipline beside the Iormer three, is being treated as part
oI the catalogue, but the fnal apotheosis oI the science oI demonstration, based
on the Analytica Posteriora, was achieved by the Arab logicians oI the tenth
century. Al-Kind`s starting point Ior the study oI philosophy, as in the Platonic
Academy, is mathematics.
Al-Kind`s double esteem oI mathematics as a propaedeutic to philoso-
phy, and as a subject worthy oI philosophical study in itselI, unIolds in great
detail, and in the best rhetorical tradition oI the Platonic mathematicians, in
his treatise 'On the string instruments producing sound (K. al-Musawwitt
al-watariyya). Introducing musik as a discipline oI mathematics, he expounds
the position oI the mathematical sciences as intermediate between physics and
metaphysics:
42
'It is a custom with the philosophers to practice the middle science, ranged
between a science beneath it and a science above it. The one beneath is the sci-
ence oI nature and what is moulded Irom nature; the one above is called the sci-
ence oI what is not oI nature, albeit its impact is observed in nature. This interme-
diate science, which leads the way both to the science oI what is above and what
is below it, is divided into Iour sections: viz. the sciences oI arithmetic, musical
harmony (ilm al-talf wa-huwa l-msq), geometry, and astronomy
in this order, music being given the second place beIore the other disciplines
according to the 'Pythagorean system oI Nicomachus.
43
AIter providing some remarks on the precedence oI knowledge (ilm)
over action (amal), al-Kind goes on to expound the philosophers` teaching oI
mathematics in general as a theoretical basis oI rational practice:
'So it was a habit with the philosophers to present the secrets oI the science oI
nature and its maniIestations in many oI the subjects they treated in books, as in
those on arithmetic and the amicable and hostile numbers, on the proportional
lines, and on the fve polyhedra (mufassamt bodies`) ftting into the sphere.
AIter demonstrating that there is no sensible thing the matter oI which is not
constituted Irom the Iour elements and the fIth nature, viz. fre, air, water, earth,
and the sphere, their acumen, intelligence and refection guided them to establish
the stringed instruments oI sound, and thus to mediate between the soul on the
one hand, and the composition oI the elements and the fIth nature on the other, by
means oI such instruments. They designed many stringed instruments in accor-
dance with the composition oI the animal bodies, and brought Iorth Irom them the
sounds corresponding to the human composition, demonstrating thereby to intel-
ligent minds how noble and excellent this wisdom (hikma, philosophy) is.
44
The skilled philosopher-musician will be able to adapt his music to any
given situation, creating harmony between the soul and the universe, like the
physician diagnosing the humors oI his patient and prescribing a treatment
inducing the equilibrium oI health.
45
The musical instruments, and the string
instruments in particular, are constructed so as to present the cosmic structure
oI the physical and intelligible world. Al-Kind describes the ethnic and his-
torical varieties oI the lute as models oI the universe, allocating the number oI
strings in each case to ontological classes. The Iour strings oI the Greek, as also
the earlier Arab lute, are oI course in correspondence with the Iamiliar series oI
cosmic tetrads, the elements, the senses, and the humors, and many others, not
Iorgetting the Iour cardinal virtues, and the Iour primary questions put Iorward
by Aristotle in his introduction to the Analytica Posteriora.
46
The fve-stringed
lute, more Iamiliar Irom the later period oI Arab musical practice, receives a
similar treatment.
47
In an analogous procedure, the particular characteristics
oI the Iour individual strings are associated with the elements, the humors,
ecliptic arcs, sections oI the zodiac, Iaculties oI the soul and other aspects oI
physiology and astronomy, the macrocosm and the microcosmassociations
which are used to explain the specifc reactions and aIIections evoked by the
sound oI each.
48
A diIIerent use, but going back to the same tradition oI Neopythagorean arith-
metic, is made oI mathematical proportions in the composition oI drugs. Since
al-Kind`s Kitb f Marifat quw l-adwiya al-murakkaba
49
and its critique by
Ibn Rushd is being discussed by Tzvi Iangermann (in another contribution to
this volume),
50
I will dispose oI it briefy. What al-Kind does is to apply Galen`s
doctrine oI the 'grades oI action in the simple drugs to the compound drugs,
and he extends the Galenic model by calculating the eIIects oI the compounds on
the basis oI geometric proportions and progressions. The arithmology underly-
ing this speculation can be traced back, again, to the Arithmetic oI Nicomachus
oI Gerasa, and it is the continuing infuence, in twelIth-century Andalusia, oI a
theory based on 'the art oI number and the art oI music which exasperates Ibn
Rushd and drives him to a torrent oI abuse.
51
1he Gnostic 1radition and Mathematics
It is in these various domains that we encounter concurrent oIIshoots oI a
tradition which in the work oI al-Kind Iound its most versatile, serious and
infuential exponent, and which lived on in his school, but gradually sank to
the lower strata oI occult practice and heretical obscurantism against the pure
rationalism oI Peripatetic philosophy.
Cognate traditions oI arithmology can be traced in the philosophical
encyclopaedia oI the Isml Ikhwn al-SaI on the one hand, and in the
corpus oI alchemy and the occult sciences ascribed to Jbir ibn Hayyn on the
other. I cannot go into the discussion oI age, authenticity and unity oI the Jbir
corpus oI writings; suIfce it to say that the transmission oI the corpus was car-
ried on in Isml circles, where the Kitb Ikhwn al-Saf was read as well,
but goes backin part at leastto an earlier period oI the Arabic reception oI
Hellenistic thought. It is a tradition oI the philosophical and scientifc 'sub-
culture oI late Hellenism. Al-Kind knew this tradition and drew upon it, but
achieved a frst, though incomplete integration with the Peripatetic paradigm.
In the Jbir writings, as also in the Kitb Ikhwn al-Saf, this tradition was
carried on independently, and in a diIIerent context.
Eor Jbir the role oI arithmology is signifcant. It pervades the 'science
oI the balance, ilm al-m:n, a central concept governing the philosophy oI
nature.
52
This is meant to reduce all domains oI human knowledge to a system
oI quantity and measure, conIerring on these the character oI an exact sci-
ence. In particular, it is Jbir`s intention to submit nature to measure, and to
determine the proportions oI the elemental qualities or Iorcesthe hot and the
cold, the moist and the dryrepresented in the bodies and in their interactions.
As in al-Kind`s treatise on the compound drugs, the Galenic physiology, and
in particular the Galenic theory oI the Iour degrees oI intensity or potency in
regard oI the elemental qualities, determining the eIIects oI a specifc medical
or nutritive substance, is at the basis oI this theory.
53
Jbir Iollows closely the
Galenic classifcation. At the same time, the values oI the Iour degrees and
their subdivisions are calculated on the basis oI arithmetical progressions,
Iamiliar Irom Greek number theory as it was Iound in Nicomachus, and on the
other hand, in the literature on the harmonical proportions.
54
On a larger scale,
the rapport between the musical harmony governing the celestial spheres and
the harmony Iound in the physical world is investigated in the Jbirian Kitb
al-Bahth, a philosophical justifcation oI theurgy (the ilm al-tilasmt). Here
again, the author makes reIerence to Plato`s 1imaeus, or rather the Platonic-
Pythagorean tradition elaborated Irom the cosmology oI the Platonic work.
55
Also based on Greek sources, and put into a systematic Iramework struc-
tured on the lines oI the Alexandrian curriculum and system, are the Rasil
Ikhwn al-Saf, written in second halI oI the 10th century.
56
One remarkable
trait which springs to the eye is the priority given to mathematics. The book
opens with a treatise on arithmetic. The introduction, explaining falsafa, phi-
losophia, as mahabbat al-hikma, is closely dependent on the Pythagorean tra-
dition
57
as transmitted through the Eisagge oI Nicomachus,
58
leading on to the
classifcation oI the sciences.
'The aim and scope oI this Epistle is training (riyda) the souls oI the disciples
oI philosophy, those who choose wisdom,` and who study the real (nature oI the)
things, and search aIter the causes oI all things; in it, there is explained that the
Iorm oI number in the souls corresponds to the Iorms oI the beings in matter,
being the models Irom the upper world. Through their knowledge, the novice
is led on toward the other propaedeutical |riydiyyt, i.e., mathematical| and
physical disciplines. Indeed, the science oI number is the root oI the sciences, the
essence oI wisdom, the Ioundation oI knowledge and the (principal) element oI
all things |al-man, objects oI the mind|.
59
Arithmetic is Iollowed by geometry, astronomy, and music. Apart Irom
the general concept oI mathematics, a number oI closer parallels to al-Kind`s
treatment oI musical harmony is Iound in the Epistle on Music, the Iourth part
oI the quadrivium.
60
The author oI the Risla is convinced (Iollowing the Pythagorean exam-
ple)
61
that the science oI music is the principal wisdom leading to philosophical
thought:
'Musical harmony in its most exalted and perIect Iorm is embodied in the heav-
enly spheres and the music that they make, and earthly harmony, including that in
the music made by man, is only a pale refection oI that same loIty universal har-
mony. . . . Since an ordinary mortal cannot hear this music beIore he is cleansed
and purifed, he will aspire to be redeemed Irom the prison oI this earthly liIe, to
be prepared Ior the contemplation oI an eternal harmony which is the most real
and truthIul.
62
The harmony that governs all celestial and earthly phenomena is
explained by means oI number. Iike al-Kind`s treatise mentioned beIore, the
Epistle abounds oI arithmetical speculations that spread into many and varied
domains oI cosmology, physiology and philosophy. The section on the aIflia-
tion oI the Iour strings oI the lute with various cosmic and physical and physi-
ological tetrads agrees almost verbatim with the third section oI al-Kind`s
treatise on string instruments.
The general attitude is expressed in nuce in one oI the aphorisms col-
lected at the end oI the Risla, recalling the Iamiliar analogy between soul and
number:
'Since the substance oI the soul is oI the same nature as that oI harmonic
numbers (al-add al-talhyya) and corresponds to them, when the beats oI the
rhythms presented by the musicians are measured, when in these rhythms the
period oI beats and silences are proportionate, human nature takes delight in
them, the spirit rejoices and the soul experiences happiness. All this is because oI
the resemblance, the relation and the kinship which exists between the soul and
musical harmony.
63
1he School ot al-Kindi
The school oI al-Kind was brought to the East by Ab Zayd al-Balkh (d. 332/
934), who spent in his youth eight years oI studies in Baghdad, better known
through what has survived oI his geographical ouvre, but also bent on a sys-
tematical treatment oI the sciences (his 1artb al-ulm is now lost): a man
who combined competence in the rational sciences with a conservative piety
praised by his contemporaries.
64
In Balkh, the meeting place oI trade routes
Irom Central Asia, Transoxania and Iran, the tradition oI the Faylasf al-Arab
was passed on to Abu `l-Hasan al-mir who spent some time in Baghdad and
at the Buyid court in Rayy beIore he returned to Nishapur (where he died in
381/992). The most detailed attempt to determine the relation oI the religious
and the philosophic disciplines in a harmonious symmetry is al-mir`s Ilm
f manqib al-Islm.
65
The very title signals an apologetic program: the rational
sciences are put into the service oI Islam, the absolute religion, and oI the reli-
gious sciences. Both spheres are based on tenets which agree with pure reason
and are supported by valid demonstration.
Here again, we fnd al-Kind`s attitude toward the mathematical sciences:
the 'science oI number (ilm al-adad) will 'immerse the mind into the intel-
lectual pleasures;
66
and the science oI harmony (sinat al-talf) will give
demonstrations oI the harmonious relations, measures and Iorces in the ter-
restrial and celestial world, and beyond this, in the corporeal and the spiritual
world (al-lam al-rhn) in general; without this, the astronomers were not
able to veriIy the states oI the celestial bodies. Astronomy, in its turn, will alert
the mind to the 'doors oI Ielicity.
67
The ethical component oI this hikma, the
autonomous ethics oI the philosopher who fnds in the encyclopaedia oI sci-
ences the instruction Ior educating his soul toward purity and ultimate bliss, is
Iound again in the 1ahdhb al-akhlq oI Ab Al Miskawayh (m. 421/1030).
It was al-Kind`s concept oI philosophy as an autonomous way oI thought and
way oI liIealbeit in the service oI the Muslim community and compatible
with the Koranic revelationwhich stayed alive in the circles oI the hukam:
oI scientists, oI learned courtiers, and oI physicians who in the spirit oI Galen`s
platonism revered in philosophy the healing art oI the soul. It was the concept
integrated by Ibn Sn with al-Erb`s concept oI philosophy as demonstrative
science: a universal encyclopedia which in the ranks oI scientists and physi-
cians, but also, and increasingly, among the elite oI Muslim administration,
Iound an eager readership.
2 Aris+o+ir :xi +nr Lxivrrs:i Ci:ix or Pniiosorn. :s
irxoxs+r:+ivr scirxcr
Erom the rise oI philosophical schools in late Hellenism to their reception
by the Christian transmitters and the Muslim heirs to this tradition, Aristo-
tle was venerated as Iounder oI the paradigms oI rational discourse, and oI
a coherent system oI the world. In the course oI the hellenization oI Arabic
science and philosophy, aIter the initial dominance oI Iranian traditions in
medicine and astrology (transmitting, it is true, their own brand oI hellenism),
Aristotle was elevated to the rank oI absolute philosopher, al-hakm or al-
faylasf.
68
At the same time, Plato was shoved gradually aside into the domain
oI popular wisdom and vulgar gnosticism. Not the philosophers, but the learned
doctors oI medicine, disciples oI Galen, continued to cherish the Platonism oI
Galen`s school, and the little that is extant oI the texts still available to the frst
generation oI translators has been preserved not by philosophers but by the dil-
ettanti oI philosophy, most oI them physicians. The mathematicians, and above
all the astrologers, Iollowers oI Ptolemy, equally made Aristotle the supreme
guide to the 'science oI sciences,
69
according to the traditional defnition oI
philosophy.
Plato`s dialectic oI ideas was replaced by Aristotle`s alternative dialectic
oI discourse: a deductive epistemology. This was frst put Iorward in a radical
Iorm in the Posterior Analytics: 'Eor we can say goodbye to the eide, Ior they
are nonny-noes, and iI there are any, they are nothing to the argument.
70
Even
here, Aristotle`s closeness to the Platonic model he is replacing, and to the
study course oI the Academy, is evident in his allusions to Platonist vocabulary
and concepts: 'All teaching and all intellectual knowledge come about Irom
already existing knowledge. This is evident iI we consider it in every case; Ior
the mathematical sciences are acquired in this Iashion, and so is each oI the
other arts.
71
With Aristotle as with Plato, mathematics is the science par excel-
lence, providing both examples and the general problematic. A passing shot at
the Meno is making it plain, iI only implicitly, that Plato`s theory oI recollec-
tion (anamnesis), which explained the preexistence oI mathematical universals
through reIerence to eternal ideas, has been discarded. In the 'aporia oI the
Meno, the partner oI Socrates, baIfed in his search Ior virtue, asks: 'And in
what Iashion, Socrates, will you seek that oI which you do not even know iI it
exists? (Meno 80d). Aristotle Iollows Plato in maintaining that the seeker does
in a sense already knowwhat he is looking Ior. Having explained that the learner
must already know the premises, Aristotle adds (An. Post. I.1.71a24I.) that in
a sense he also knows the conclusion. But only in the end (An. Post. II.19), he
returns to the implications oI his own theory oI preexisting knowledge, and pro-
ceeds to account Ior the acquisition oI the frst, lowest universals by induction
(epagge). Indeed, we cannot demonstrate the principles. Aristotle`s primary
contention is to expound the universal structure in the acquisition oI knowledge
based on what is already available to human knowledge Iromsuch principles, by
expounding the universal system oI demonstration, apodeixis.
The role oI axiomatic mathematics as a background to demonstrative
method in philosophy is evident Irom its very conceptualization and termino-
logy. As early as in the fIth century B.C.E., Greek mathematics had taken the
step Irom simple demonstration, oao-orii=, Irom visual evidence, to demon-
stration Irom principles: defnitions and axioms. Iike the science oI geometry,
logical demonstration had 'to rely on principles, which, though unprovable, are
nonetheless true and indisputable.
72
In this, Aristotle continued an intellectual
tradition which recognized a Iundamental aIfnity between mathematics and
dialectic. Even though the mathematical and physical sciences apprehend their
principles in a diIIerent way, Aristotle regards mathematical procedureaxi-
omatization, and the use oI hypothesesparticularly helpIul Ior the acquisi-
tion oI all scientifc knowledge. Mathematics provided to him a model oI
deductive-demonstrative science parting Irom principles (opoi ).
73
But his noetic concerns are not separable Irom ontology, and especially
Irom the basic ontological aporia oI relating individuals qua individuals to
individuals as being exemplars oI universals. 'The newly declared ouoio, the
individual substance, had as individual substance become unknowable except
in universal terms, and the abstracted essence took on the detached charac-
ter oI the rejected Iorms oI Plato.
74
Eor Aristotle himselI, this remained the
'greatest aporia (Metaph. 1087a13, 999a2425). It was here that the ways oI
late Hellenistic metaphysics, oI philosophical theology, in Christianity and in
Islam as well, parted with the master oI demonstrative science: returning over
and again to the assumption oI preexisting, eternal, hypostatized objects oI
knowledge. According to Aristotle`s theory, presented as a general epistemol-
ogy, the sciences are to deduce the properties oI substances Irom their essences
through syllogisms. Still, in expounding the sciences in a Iormal axiomatized
system, Aristotle proposed Ior every branch oI human knowledge what early
Greek mathematics had done Ior mathematicals (and what Euclid consum-
mated Ior geometry later on
75
infuencing, in his turn, an axiomatic approach
to ontology and cosmology in Neopythagorean and Neoplatonic metaphysics).
In Iollowing Aristotle in this overall orientation, all subsequent philosophical
systems, notably those oI Islam, are essentially Aristotelianwhatever their
Platonic or Neoplatonic paradigm in the allegories oI the World Above.
The place accorded to mathematics has to be seen in this context.
In c. 2 oI the Metaphysics, book B, Aristotle discusses the question under
which science, or sciences, iI any, the 'principles oI demonstration will come,
that is to say, the 'common opinions or axioms (aximata), which are the
starting-point oI all demonstrations. The science oI these principles cannot be
any one oI the special sciences, as e. g. geometry. In a later passage (Metaph. I.
3, 1005a19b1, cI. K. 1, 1059b1421), Aristotle gives his solution concerning
the position oI mathematics:
76
'We have now to consider whether it belongs to one science or to diIIerent sci-
ences to inquire into what mathematicians call axioms, and into substances. It is
maniIest that the inquiry into these axioms belongs to one science and that the
science oI the philosopher; Ior they hold good oI all existing things, and not Ior
some one genus in particular to the exclusion oI others. Everyone makes use oI
them because they belong to being qua being, and each genus is (part oI) being.
. . . This is why none oI those who study the special sciences tries to enunciate
anything about them, their truth or Ialsehood; neither the geometer, Ior instance,
nor the mathematician does so, though it is true that some oI the physicists have
made the attempt, and not unnaturally seeing that they supposed that the inquiry
into the whole oI nature and into being belonged to them alone. But since there
is a class oI inquirer above the physicist (nature being only one particular genus
oI being) it is Ior the thinker whose inquiry is universal and who investigates pri-
mary substance to inquire into these axioms as well. Again, since the mathema-
tician, too, uses the common axioms in a particular application, it must be the
business oI frst philosophy to investigate the principles oI mathematics also.
As Iar as the axioms oI mathematics hold oI all being, they belong to
philosophy, investigating all being so Iar as each oI them is.
The physical part oI philosophy and mathematics study the same
objects, but there is a diIIerenceespecially with regard to optics, harmonics,
and astronomy, which among the mathematicals are 'nearest to the study oI
nature.
77
In physics, both matter and Iorm are studied: the substances oI the
physical bodies as well as their shapesbounded by planes, lines, and points.
Mathematics studies these geometrical attributes only, not as attributes oI
physical bodies, but in abstraction:
78
separate in virtue oI cognitive abstraction,
notpace Platoqua being ontologically separate.
Al-Iarabi and Demonstrative Science
By the end oI the ninth century, philosophers had gained a readership which
had spread considerably beyond the circles oI the scientifc proIessions: among
litterateurs, among the elite oI the secretarial class (the kuttb), and in circles
attending the courtly maflis oI learned and literary exchange. These may also
have shared the philosophers` distrust oI the inIection oI Kalm by the growing
tide oI traditionist orthodoxy. But then, addressing the same readership, a diI-
Ierent program was draIted by al-Erb (d. 339/950): based on a wider choice
oI the sources which had become accessible, and envisaging a comprehensive
system oI knowledge, and integrating the Aristotelian theory oI the principles
with Neoplatonic cosmology and a Platonic model oI the political-religious
community.
Encompassing all oI Aristotle`s logic, physics and metaphysics, the early
Erb was the frst oI the falsifa to turn Irom the compromises oI al-Kind`s
creationist Platonism in his 'Integration oI the opinions oI the two Sages,
Plato and Aristotle. The philosopher realizes that the primary subjects oI his
inquiry are the universals, not as hypostatized species subsistent in the 'world
oI the intellect, such as ideasor Ior that matter, mathematical entitiesbut
in rebus, principles oI reality subjected to induction and demonstration. He
discards with the Platonic concept oI anamnesis by re-interpreting it on Aristo-
telian lines: In substance, Aristotle agrees with Plato when he defnes the true
Iunction oI recollection in the beginning oI the Posterior Analytics (the Kitb
al-Burhn).
79
Al-Erb added to earlier concepts oI philosophy in Islam the
radically Aristotelian concept oI philosophy as a demonstrative science (ilm
al-burhn) which proves universally what in the particular sciences is deduced
by particular 'indications or 'signs (dalil), and which perceives absolutely
what in the individual religious-linguistic communities is conveyed individu-
ally. Philosophy as a science is a method oI deduction and demonstration, not
an ideology competing with theology: being an independent way toward know-
lege, it could be proclaimed as a saIeguard Ior the religious community itselI.
It is here that Aristotle as being the author oI exemplary and encyclopaedic
instruction is transIormed to become the authority oI a method leading to abso-
lute knowledge. The philosopher claims rulership, not only inside the scientifc
community, but in the religious community as well.
80
Al-Erb integrated the sciences in the Iramework oI a Iormal axioma-
tized system, the system oI Aristotle`s Posterior Analytics. Philosophy and
religion, the universal, rational sciences and the disciplines specifc to the reli-
gious and linguistic community, are shown to be complementary parts oI the
same hierarchical system oI cognition and interpretation.
Aristotle`s Posterior Analytics, the Kitb al-Burhn, provided al-Erb
with a coherent system oI deduction and demonstration, comprising all levels
oI rational activity, and serving as a guide Ior the division and hierarchical
classifcation oI the sciences, leading up to the Eirst Philosophy, metaphysics.
The basic text is the exordium oI the Analytica posteriora (I.1): 'All teaching
and all learning come about Irom already existing knowledgeby deduction
(Irom the specifc), induction (Irom the particular), and individual 'signs
(dalil), or, in the practical arts, experience, in descending order oI certainty.
Al-Erb`s own summary contains explicit consequences as to the coherence
and ranking oI the sciences:
'OI the theoretical sciences, some are universal |sc. the Eirst Philosophy and the
universal demonstrative sciences, Topica and Sophistica| and some are particular
|sc. mathematics, physics and theology|. The universal sciences have in common
the subjects, the objects and most oI the premises, but diIIer in the conditions
aIorementioned |sc. in the relative status oI the principles, subjects and objects
used as premises in their specifc demonstrations|. The particular sciences are
all below the Eirst Philosophy, participating in it in so Iar as all their subjects
are below the Absolute Existent. This science |sc. Eirst Philosophy| will employ
universal premises which all the particular sciences employ in the way we have
described |i.e., in the mode applied to their particular subject|, while the particu-
lar sciences employ premises which are demonstrated in that science |sc. in Eirst
Philosophy|.
81
The subordinate, particular sciences and the superior, universal sciences
'help each other in that 'the prior sciences provide in the subsequent sciences
the knowledge oI the causes or oI both the causes and the existence, while the
subsequent sciences provide in the prior ones the existence. It Iollows that
'each art (sina) which provides the principles oI another art is governing (rasa
li-) that art. Now the governing science in an absolute manner among the sci-
ences which provide the causes, is that which provides the ultimate causes oI the
beings: and this must be the Eirst Philosophy.
82
The mathematical sciences are posited between the physical and the
metaphysical in being the 'abstractive sciences, abstracting immaterial enti-
ties Irom the material substances: separating intellectually what is inseparable
Irom matter in its actual existence.
83
The position oI the mathematical sciences
(ilm al-talm) in al-Erb`s 'Enumeration oI the sciences (Ihs al-ulm)
Iollows the same Aristotelian premisses: Theoretical arithmetic (arithmology,
ilm al-adad) examines numbers absolutely (bi-itlq), abstracted in the mind
(mufarrada f l-dhihn) Irom the bodies and Irom anything actually numbered,
sensible or not; through this universal application, it enters the realm oI the
sciences
84
(as against the particular, practical technai, like medicine, excluded
Irom the ulm in al-Erb`s Ihs).
85
An analogous statement is made about
theoretical geometry (handasa).
86
This latter is qualifed as being 'more gen-
eral than the one immediately Iollowing, optics (ilm al-manzir), but optics
requires the status oI a science in its own right (an yufrad) in examining the
'aspects oI what 'appears to sight (m yazharu li-l-basar) as distinguished
Irom 'what (a thing) is in reality (m huwa alayhi bi-l-haqqa).
87
Here the scientifc character oI optics
88
is pointed out as being a way
to establish this diIIerence by means oI 'certain demonstrations (barhn
yaqniyya), that is, prooIs yielding certain knowledge.
89
In this, al-Erb Ior-
mally asserts the claim oI a mathematical science to demonstrative method, and
to the quest Ior knowing reality as suchthe objective oI philosophy by defni-
tion. The same claim was raised, and had been raised beIore, by mathemati-
cians who (a) were able to point out Aristotle`s use oI an optical example, in his
Posterior Analytics, Ior elucidating the conclusion Irom sensible existence (to
o ti) to cause (to oio ti) in scientifc demonstration and the discussion oI optical
phenomena in the Ps.-Aristotelian Problemata Physica,
90
(b) took the concept
and method oI geometrical apodeixis Irom Euclid, and (c) claimed the status
oI universals Ior the mathematical 'causes fguring as a middle term in such
demonstration, in accordance with Aristotle`s own procedure (An. Post I.2).
One generation prior to al-Erb, Qust ibn Iq (died c. 300/912
13)mathematician, philosopher, and translator oI Greek scientifc texts
introduces his epistle on catoptrics with a praise oI demonstrative science as
'the fnest oI the humaniora, and then continues to commend his own subject,
optics, as being 'the fnest oI the demonstrative sciences: the one in which the
natural science and the science oI geometry partake, since Irom the natural sci-
ence it takes the sensual perception, and Irom the geometrical, the demonstra-
tion by means oI lines |i.e., linear constructions|such, par excellence, is the
science oI rays (catoptrics).
91
Erom here Ibn al-Haytham was able to go on toward establishing math-
ematical astronomy and optics as the noblest oI sciences about universalia in
rebus.
Ibn Sina. the Nev Encyclopdia
Avicenna united and integrated the early traditions oI falsafa, both in respect to
groups oI readership and proIessional circles, and also in uniting the Platonic
and Peripatetic Iundamentals. Taking up and completing the work oI al-Erb,
he projected the conceptual Iramework oI the Arabic Posterior Analytics onto
all domains oI scientifc and philosophical knowledge, conceiving all strata oI
cognitionincluding the highest degrees oI discursive and intuitive thought
(the latter being the hads, bereIt altogether oI its mystical connotation)as
applications oI the syllogism.
In his 'Division oI the Intellectual Sciences (1aqsm al-ulm al-aqli-
yya),
92
the mathematical quadrivium is dealt with in a basic and straightIorward
manner, dependent upon the manuals oI the Hellenistic tradition, and repeat-
ing the classical topos oI the 'intermediate position oI the mathematicals:
Mathematics, as a part oI the theoretical philosophy (al-hikma al-nazariyya) is
intermediate, al-ilm al-awsat, between physics and metaphysics ('theology);
its objects, regarding their existence, are bound up with matter and motion, but
their conceptstheir 'defnitions (hudd)are not, since they can be under-
stood without reIerence to any bodily substrate.
Mathematics is duly mentioned in his 'Autobiography, which is an idealized cur-
riculumoI the accomplished philosopher: In his youth, preceding systematic stud-
ies, he learned some practical geometry and 'Indian calculation. Then he studied
the Isagoge and elementary logic with his frst teacher in philosophy, al-Ntil,
going on to Euclid and the Almagest, and then to physics and metaphysics. AIter
learning the practical art oI medicine, he took up the systematical study oI theo-
retical philosophy, to be crowned by a deepened understanding oI metaphysics.
93
The question oI the place oI mathematics and its objects in the philosoph-
ical sciences is dealt with in some more detail in the Metaphysics (al-Ilhiyyt)
oI the philosophical encyclopdia al-Shif.
94
The subject oI mathematical sci-
ence is measure (al-miqdr), qua being abstracted, in the mind, Irom matter.
95
Number may be Iound both in sensible and in non-sensible objects; measure,
whether said oI a corporeal dimension or oI a limited quantity taken Irom a
continuous extension, is never separate Irom matter, although in the frst sense
(oI dimension), it is a principle in the existence oI all natural bodies, hence it is
prior in essence to the sensible beings.
96
He goes on to discuss the subject-matter oI metaphysics, that is, 'what
is beyond nature, M bad al-taba.
97
It might be called properly 'the science
oI what is beIore nature, because its objects oI study are essentially and gen-
erally beIore the natural bodies. Now someone might objectsomeone, we
might add, in the tradition oI al-Kind or the Ikhwn al-SaI'that the ques-
tions oI pure mathematics, studied in arithmetic and in geometry, are equally
beIore nature,` and especially number, because this can exist independently
oI a material substrate. As Ior the lines and surIaces treated in geometry they
subsist in bodies. The measure (miqdr) treated in geometry is not an absolute
principle or Iorm, as oI prime matter, but accidental, as oI bodies possessing
the three dimensions width, breadth and height. As Ior arithmetic, it does not
study absolute number. It is true that number can be Iound in the separate
beings, but it is also Iound in the natural things. It may arrive that in the imagi-
nation (f l-wahm) a number abstracted Irom any adventitious substrate be
conceived. But number cannot exist unless it adheres to a thing in existence (l
yumkinu an yakna l-adadu mawfdan ill ridan li-shayin f l-wufd). As
Iar as being among the separate existents (al-ash al-mufriqa), on the other
hand, number cannot be subjected to any quantitative relation, augmentation or
reduction, but subsists as it is. As an object oI quantitative relation, augmenta-
tion or reduction, it must be in matter. This, however, is the subject oI arithme-
tic: studying number in respect oI it being in bodily nature, albeit abstracted
Irom its natural states in the imagination (fa-idhan ilmu l-hisbi min haythu
yanzuru f l-adadi innam yanzuru fhi wa-qad hasala lah l-itibru lladh
innam yaknu lah inda kawnih f l-taba).
98
Arithmetic, in consequence, is
not a study oI the essence oI number, or oI absolute number, but oI number in
its accidental inherence in matter. The essence oI number, however, is an object
oI metaphysics indeed.
Book VIII oI the Shif On unity and multiplicity
99
brings an extended
reIutation oI the Platonic doctrine oI ideas, and especiallyon the lines oI
Aristotle`s Metaphysics, A.5 (985b23II.) oI the Pythagorean notion oI subsis-
tent 'numeric Iorms or 'numeric numbers (313.11 al-adad al-adad, 314.3
al-suwar al-adadiyya), viz. oI 'numeric numbers Irom which they constitute
the Iorms oI the natural existents (319.10 al-qilna bi-l-adadi l-adadiyyi
l-murakkibna minh suwara l-tabiyyiyt).
On the one hand, Ibn Sn insists on the primacy oI philosophyoI the
Eirst Philosophy: on the rank oI its subject matter, and the universal control
oI demonstrative method. On the other hand, he integrates the Iundamentals
oI the quadriviumgeometry, arithmetic (including numerology), spherical
astronomy (the science oI the Almagest), and musical harmonyinto the
encyclopdia oI the sciences established under the aegis oI philosophy. But his
is not a mathematicians` philosophy; and contrary to all oI his predecessors,
he leaves out all aspects oI mathematical science where observational practice
meets demonstrative method.
3 Iix :i-H:.+n:x: M:+nrx:+ics :s irxoxs+r:+ivr scirxcr
The enormous success oI Ibn Sn`s encyclopedia was not only due to his
new metaphysics, which promised to solve the antinomies oI metaphysics and
religious thought, but also to the systematic coherence oI his logic and episte-
mology. Scientists and physicians were his most eager students and readers. It
was not an Avicennian scientist, however, who at frst, and uncompromisingly,
made demonstrative science his very own:
'Truth is sought Ior itselI; and in seeking that which is sought Ior itselI one is
only concerned to fnd it. To fnd the truth is hard and the way to it is rough.
. . . But God has not protected scientists Irom error. . . . The seeker aIter the truth
is, thereIore, not he who studies the writings oI the ancients and, Iollowing his
natural disposition, puts his trust in them, but rather the one who suspects his
Iaith in them and questions what he gathers Irom them, the one who submits
to argument (al-huffa) and demonstration (burhn) and not to the arguments
oI a human being whose nature is Iraught with all kinds oI imperIection and
defciency.
100
The Iamous introduction to the Shukk al Batlamys oI Avicenna`s con-
temporary Ibn al-Haytham (c. 354/965432/1040) is like a radical restatement
oI Aristotle`s Irequent proposal, beIore studying a problem, frst to consider the
opinions oI the Ancients. It echoes in spirit, iI not in its wording, the statement
oI Avicenna in the introduction to his Mantiq al-mashriqiyyn, acknowledging
the merit oI 'the most excellent oI their |the Peripatetics`| predecessors, Aris-
totle, but scoIfng at 'the common philosophasters who are inIatuated with the
Peripatetics and who think that no one else was ever guided by God or attained
to his mercy.
101
But the approach and method oI Ibn al-Haytham, although
evolving Irom a mathematical science embedded in the traditional system oI
ancient cosmology, end up in a radical rejection oI transmitted authority where
it contradicts the results established by the very method oI Aristotle. Aristotle
remains a vague symbol oI the authority oI any philosophy whatsoever, while
mathematical science becomes the Ioremost oI the demonstrative sciences.
Evincing the principles oI his science, Ibn al-Haytham enjoins the true
scientist to be a true philosopher, Iollowing the rules oI demonstration. Remarks
on method are Irequent. Eor the general principles oI physics, Ibn al-Haytham
turns to the opinions oI 'all the philosophers or 'those oI the philosophers
who arrived at the truth (al-muhaqqiqn min al-falsifa).
102
Aristotle 'laid
down the principles Irom which the way to the truth will be Iound, its nature
and substance be attained, and its essence and quiddity be Iound (ahkama
l-usula llat fh yuslaku il l-haqqi fa-yudraku tabatuh wa-fawharuh wa-
tfadu dhtuh wa-mhiyyatuh).
103
Aristotle`s physical philosophy was, as a
matter oI course, his point oI departure, an authority invoked Irequently, and
the subject oI summaries and commentaries listed among his early writings.
But in the end, Ibn al-Haytham remained an Aristotelian only in the sense oI a
general methodological orientation. In an earlier treatise 'On the Confguration
oI the World (f Hayat al-lam),
104
he expounds, in a separate appendix, the
principles oI celestial movement, all oI which can be traced up to Aristotelian
physics.
105
In the later treatise 'On the Iight oI the Moon (f Daw al-qamar),
he spurns all mention oI Aristotle`s celestial physics, such as the nature oI the
fIth body, aither, to be used as premisses Ior his theory. Instead oI metaphysi-
cal doctrines, such general principles as can be observed behind his argument
are specifc theorems, developed Irom physical theory, but closer to the Iacts
under discussion.
106
Aristotlethe only philosopher actually namedremains
but a symbolic authority oI demonstrative method: a virtual text, while his own
writings Iall into oblivion.
The observance oI demonstrative method by itselI had become the pass-
port oI competence Ior the pursuit oI knowledge in the epistemic community.
When in his 'Solution oI the Aporias in Euclid`s Elements, Ibn al-Haytham
raises his own apodeictic method above the time-honored authority oI the
master oI demonstration in geometry, he still reIers to the principles oI science
pronounced by Proclus and Aristotle, but claiming their ultimate perIection:
'The causes in matters scientifc are the premises employed in the geometrical
prooIsthese are the proximate causes; but what we seek in each construction
is the remote and frst causeand this has not been pointed out by any oI the
earlier nor any oI the later authorities.
107
In Ibn al-Haytham`s remarks on his method oI inquiry, the use oI
istiqr (epagge, 'induction)
108
is an explicit pointer to the logical procedure
described in the fnal chapter oI Aristotle`s Posterior Analytics as the way to
detect the principles or universals used as premisses in a valid demonstration.
It is true that the word is used somewhat loosely by Ibn al-Haytham in many
instances.
109
According to al-Erb`s reading oI Aristotle, induction (istiqr)
aims at establishing a universally aIfrmative or negative proposition. As a pro-
cedure, he understands induction as the act oI surveying all or most oI the par-
ticular cases Ialling under a given universal to see whether a certain predicate
applies or does not apply to the particulars surveyed. II complete, the induction
is called 'perIect, iI incomplete, 'imperIect. al-Erb`s understanding oI
induction in terms oI a one by one examination oI the particulars does not cor-
respond to the meaning oI this term in the relevant Aristotelian passages; there,
it is not attending to the particular cases, bur rather the advance Irom these
particular cases to the corresponding universal which is known as induction.
'It seems that al-Erb`s understanding oI the matter is a consequence oI the
Iact that in the Arabic Prior Analytics, epagge was rendered as istiqr, a term
that must in this case be taken to reIer to the act oI 'collecting the individual
cases.
110
The mathematician Ibn al-Haytham goes on Irom here to check the
limits oI the theoretical model by means oI systematic observation (itibr,
'experience).
111
But Ibn al-Haytham, starting Irom the Iamiliar concepts oI Aristotelian
epistemology and Irom the traditional models oI astronomy and optics, trans-
Iormed both. In his hands, the objective oI induction, instead oI a collection oI
universals Irom the particulars oI any observation whatsoever, became Iocused
on the refnement oI complex procedures, apt to provide criteria Ior the validity
oI the models and hypotheses they were to yield. The true progress, owing to a
true revolution in method, was based on a new conception oI the use oI mathe-
matics Ior the description oI those particulars, collected and surveyed in order
to support a perIect inductive inquiry yielding valid results. While mathematical
models are based on the data oI observation, the philosopher-mathematician is
convinced oI the essential coherence between valid models and the planthe
logosoI nature. This conception oI mathematical relations in natural science
is Iounded on the basic assumption that physical theory, in order to be valid,
must deal with real bodies, and not with imaginary hypotheses, and must be
developed through a process oI observation, experiment, and induction. It is
thus possible to conclude that the most simple oI mathematical relationships,
arrived at under the most excellent conditions oI scientifc observation, can be
supposed to correspond to the structure oI the physical world.
112
Through induction Irom the phenomena, the Aristotelian-mathematician
grasps Ior demonstrable evidence oI the absoluteIorms-in-matter, but none-
theless universal. The only medieval dissertation on the aesthetically beautiIul
not bound up with ethical instruction about the morally good is Iound in Ibn
al-Haytham`s Optics: the beautiIul (al-mustahsan, what is regarded as beauti-
Iul) as harmonious proportion.
113
The ultimate object oI contemplation oI the
Pythagorean and Platonic philosophy oI mathematics reappears in demonstra-
tive mathematics: not assimilation to the Eirst Good, but absolute Iorm.
Philosophy and Spherical Astronomy
The rationalism oI Hellenistic philosophy, in Islam as beIore, is made visible
through the reality oI the cosmos. The order oI the spheres, the eternal, circular
motion oI the heavenly bodies, the progression Irom the one and frst cause to
its maniIestations in the celestial hierarchy and to the changeable and corrupt-
ible substances oI the sublunar region evolved into an increasingly diIIerenti-
ated system (in Arabic, the haya, 'shape or 'confguration, oI the world).
Plato, according to a well-known tradition, had enjoined the astronomers to
fnd out 'by the assumption oI what uniIorm and orderly movements the appar-
ent motions oI the planets can be acccounted Iorto 'save the phenomena
through mathematics: both were Iounded on the eternal Ideas.
114
This was achieved, or very nearly so, at an early age oI Greek science.
But Aristotle`s assumption that Eudoxus` geometrical model oI concentric
spheres was a physical reality, a mechanical system in conIormity with pure
mathematics, and obeying the laws oI natural movement established in his
Physics, created new problems. The more precise the astronomical observa-
tion, and the more intricate the mathematical calculus oI the celestial revolu-
tions grew, the more diIfcult it became to reduce all phenomena oI the heavens
to a coherent system oI uniIorm, circular motion: the natural movement oI
aither, the celestial substance. Based on two theorems oI Apollonius, Ptolemy
crowned the achievement oI his predecessors, with an elaborate construction oI
epicyclic and eccentric vectors explaining (in the Almagest) the phenomena oI
the planetary cycles. And indeed, he took this both as a mathematical solution
oI Plato`s assignment and as a true model oI the physical cosmos (a system oI
contiguous nested shells, as sketched in his Hypotheses).
115
But this quantita-
tive conception had to deal with the variations in angular velocity oI the eccen-
tric deIerent, with the variances oI precession and trepidation in the sphere oI
the fxed stars, and related diIfculties, which violated the principle oI uniIorm
circular movement. Consequently, the application oI Aristotle`s physical theory
to the Ptolemaic system required a new planetary theory.
116
The problematic was known, and was seen as a Iundamental aporia,
Irom the times oI late Antiquity. Various doubts are raised and reIuted by
the commentators oI Aristotle with reIerence to early astronomers as well as
the philosophers oI the Academy.
117
While the main bulk oI our literature is
concerned with the confict between the mathematical astronomy oI Ptolemy
and the physical philosophy oI Aristotle, the Platonists had a diIfcult stand as
well. Proclus the philosopher, having to give an appropriate place to astronomy
among the Iour disciplines leading the way to the Good, had to deIend Plato
the astronomer against the 'modern astronomy oI his time, devoting many
pages to the task oI reIuting the assumption oI eccentric spheres and epicyclic
motion. The main Iault oI the astronomers was 'to pass Irom the domain oI the
physical bodies to mathematical considerations, which are imaginary, 'and
to give an account oI the natural movements on the basis oI things which do
not exist in naturerepeating, as it were, the mistake oI the Pythagoreans
criticized by Aristotle in the Metaphysics, in making mathematical realities
account Ior natural processes.
118
This is, in principle, the very same criticism which was raised against
Ptolemy by Ibn al-Haytham, and by the Aristotelian philosopher-scientists
oI Andalusia. The astronomy received by the Arab authors represented the
state oI the art oI proIessional science, and was adopted as well in al-Erb`s
philosophical cosmology. But the Ptolemaic system, while valid as a purely
mathematicalgeometricalmodel to serve as a basis Ior hypothetical cal-
culus, was interpreted as physical reality, and hence, got into confict with the
principles oI celestial physics.
Ibn al-Haytham`s Critique ot Ptolemy
Convinced oI the power oI inductive method, Ibn al-Haytham leads a vehe-
ment critique against those models oI Ptolemaic astronomy which under the
scrutiny oI the mathematician proved insuIfcient to suit both the universal
principles prevalent in the cosmos and the observation oI the celestial motions.
Ptolemy`s celestial model (haya) consists oI magnitudes supposed to move in
epicyclesimaginary vectors 'which cannot have by |themselves| a sensible
movement so as to produce something real in the world (laysa yataharraku bi-
dhtih harakatan mahssatan tuhdithu manan mawfdan f l-lam).
119
What
is more, 'the assumptions made in Ptolemy`s astronomy (haya, i.e., a model oI
celestial mechanics) Ior the movements oI the fve |lower| planets are invalid,
because they are contrary to theory (khrifa an al-qiys, i.e., to the theory
demonstrable Irom valid premisses) and to sound principles.
120
'It is not possible that the movement oI the stars, being eternal, homogeneous,
Iollowing a single order, unchangeable and incorruptible, should be against the
principles oI theory (khrif an al-qiys, aooo :ov !o,ov). It has become evident
Irom all that has been said that the confguration (haya) established by Ptolemy
Ior the motions oI the fve planets is invalid, and that a valid model to be con-
structed Ior the motions oI the planets based (on the assumption) oI bodies in
homogeneous, eternal, and uniIorm movement, implying neither absurdity nor
admitting doubt, will be diIIerent Irom the model established by Ptolemy.
121
While admitting that some oI the contradictions Iound in the Almagest
may be excused, being due to inadvertence, others were admitted consciously;
indeed, Ptolemy admits that at some points, he was 'compelled by the nature
oI our subject to use a procedure not in strict accordance with theory (ashy
khrifa an al-qiys).
122
Once the imaginary circles and lines (dawir wa-
khutt mutakhayyala) posited Ior the celestial model were assumed to apply to
existing bodies, contradiction was bound to Iollow, but Ptolemy accepted this
consciously, being Iully aware oI the methodological implications:
123
'II someone posits a line in his imagination and moves it in his imagination, a
line analogous to this line in the heaven will not move in the same way, and nei-
ther will a star, iI somebody imagines a circle in the heaven and imagines a star
to move along this circle, move on this same imaginary circle. II this is the case,
then the models which Ptolemy imposed on the fve planets are Iutile, and he
posited them although being aware that they are Iutile, because he was not able to
fnd others. But . . . it is not valid to assume a sensible, eternal, orderly movement
unless it conIorms to a valid model to be Iound in existing bodies.
124
The philosopher-scientist takes the mathematician to task, and calls Ior
the repair oI a system branded as being alogon: contrary to the universals oI
'reasoned theory on which demonstrative science rests.
125
But he stays within
the mathematical paradigm oI explaining the physical phenomena: he does not
call Ior the abolishment oI the Ptolemaic system.
126
Contrary to the spokesmen
oI the 'Andalusian revoltwe shall come back to thishe was content, in the
fnal analysis, to replace the Aristotelian doctrine oI the aetherial body with a
physical theory oI solid nested shells obeying the cinematic laws established
and calculated by Ptolemya celestial mechanic in accordance with observa-
tional data. But while in his optics, he evolved a sophisticated methodology
oI experimental control oI the mathematical models, the division between the
celestial and sublunar realms oI the cosmosboth in cosmological theory and
in Iactual experienceIorbade an inductive, experimental control in the light
oI the results obtained.
127
Among mathematicians, criticism oI Aristotle the mathematician was growing,
and was raised explicitely. Abu `l-Euth Ibn al-Sar (Ibn al-Salh, died in 548/
1153) wrote a treatise reIuting Aristotle`s assumption, put Iorth in De Caelo III.
8: 306b38, that there are two regular solids which can fll up three-dimensional
space, the pyramid and the cube, and proving that cubes only can fll a space.
See the text ed. by Mubahat Trker, 'Ibnu`s-Salahìn De Coelo ve onun erhleri
hakkindaki tenkitleri, Aratirma 2 (Ankara, 1964): 179, p. 71I.
4 1nr Prix:c. or ioc+rixr: +nr Axi:iusiox rr:c+iox
'Once such imaginary assumptions were applied to existing bodies, contradic-
tion Iollowed as a consequence.
128
Ibn al-Haytham`s point oI criticism against
Ptolemy was at the basis oI the attacks against Ptolemy rising in the Andalus.
But the philosophers oI MuslimSpain would deny the mathematicians suIfcient
competence to grasp the intelligible reality oI the cosmos. Starting with Ibn
Bjja, they would attack Ibn al-Haytham Ior trespassing into Ioreign territory:
'Al-Zarqlluh did not cease to criticize Ptolemy in most oI his opinions. Others
beIore him have held this opinion, and I do not wonder that Ibn al-Haytham
detected his |Ptolemy`s| evident errors; and iI you want to study what I reIerred
to, read Ibn al-Haytham`s treatise On the Doubts Raised Against Ptolemy. . . .
But iI you should look at this treatise, you will detect that Ibn al-Haytham read
this discipline only in the most elementary manner (innah lam yaqrai l-sinata
ill min ashali l-turuq), and perhaps he could not decide in his time iI he should
give a frm judgement in view oI its reIutation, or iI he should just disregard it.
Nay, he was not one oI true experts oI his science, and much Iarther oII the goal
than al-Zarqlluh.
129
There is more to this attack on Ibn al-Haytham than a subtle point oI
mathematical-physical analysis. It is the start oI a ferce competition Ior the
prerogative oI universal, rational knowledge. BeIore going into the notorious
problem oI the planetary motions, I would like to recapitulate a more specifc
criticism raised by Ibn Rushd in his Commentary-Paraphrase (1alkhs) oI
Aristotle`s Meteorologica, reIerring to Ibn al-Haytham`s explanation oI the
halo surrounding the moon.
In his commentary, he looks down upon the doubts oI the earlier com-
mentators, and exalts Aristotle whom God distinguished among all mankind.
In a comment on Meteorologica III.3, he reIers to Ibn al-Haytham`s explana-
tion oI the shape oI the halo, 'why it is a circle and why it appears round the
sun or the moon or one oI the other stars (Meteor. 372b1213). At the outset,
he makes a distinction between the methods oI 'this science and mathemati-
cal optics:
'Et quia subiectum istorum |sc. signorum e.g. halonis etc.| sunt corpora natura-
lia, et cum hoc ipsa accidunt in situ determinato et in fguris determinatis, neces-
sarium est, ut sit investigatio de eis secundum unum modum naturalis, secundum
alium mathematica. Nos autem consideremus hoc de dispositione istorum de
eis, de quibus considerat naturalis, utendo illis rebus quae declaratae sunt in
mathematicis tanquam suppositionibus et Iundamentis positis, et maxime eis, de
quorum consuetudine est ut accipiantur hic principia directionis.
130
In his exposition, Ibn al-Haytham had made an explicit statement on
the role oI mathematics: the substratum oI the object under consideration is a
physical body, hence the method oI investigation must be physical; but since
these objects have a round shape, they must also be investigated mathemati-
cally.
131
'This is why the inquiry by means oI which the nature oI these two
eIIects is investigated comes to be composed oI a physical and a mathematical
(examination).
132
This is precisely Ibn Rushd`s point oI criticismor rather,
apology oI Aristotle`s apparent omission: Physical science, that is to say: the
philosophical theory oI physical phenomena, and the mathematical science oI
optics are two disciplines oI diIIerent orders. What Ibn al-Haytham explained,
in his 'Iamous treatise Iound in the hands oI everybody, 'does not belong to
this science (laysa min hdh l-ilm). In his own Epitome, Ibn Rushd contin-
ues, he had taken the mathematical principles as postulates. But mathematics
and physical science diIIer with regard to the causes they look into: while the
philosopher studies the proximate causes, essential and evident, oI the phe-
nomena qua physical bodies, the mathematical science oI optics studies the
remote causes: geometrical models accidental to the physical substances.
133
Those who conIound the two sciences will make mistakes such as Ibn
al-Haytham committed. The principles oI optics cannot serve as premises in
a valid demonstration brought Iorward in physical philosophy.
134
'Praise to
the Iord, he continues, 'who distinguished him |Aristotle| with the human
perIection. What he understood easily, is understood by the common man only
aIter prolonged study and many diIfculties, and what others understand easily
is contrary to what is understood |to be correct| by him.
135
Is he singing in the dark? Averroes takes care not to engage in a discus-
sion oI Ibn al-Haytham`s advanced mathematics against Aristotle`s. Indeed,
Aristotle had evidently omitted to give a comprehensive explanation which
would 'save the phenomena in the context oI mathematical optics. Hence
Averroes Ielt the need Ior an apologyan apology oI his own philosophy
which, in order to be universally valid, was dependent on the binding authority
oI Aristotle`s method.
On the authority oI Aristotle, only the physical part oI philosophy stud-
ies real substances; the abstraction oI mathematics is but a tool Iur the purpose
oI analogous description, and cannot penetrate into reality in its own right.
136
Interpreting Plato`s program oI intellectual education, Averroes Ioists upon
him a decisive shiIt in view oI the rank oI the mathematicals:
'The intelligibles |oI the mathematical sciences| are deIective intelligibles since
they are not conceived oI in any particular objects but in what imitates them.
Hence Plato divides the intelligibles oI things into two parts. One oI them he calls
direct; these are the intelligibles oI things that truly are. And the second |he calls
thought|; these are the intelligibles oI the appearances oI existing thingsand
they are the mathematical sciences |Plato, Resp. VII, 533E45|. . . . Plato asserts
oI them that they are not oI the rank oI the other theoretical sciences as regards
human perIection. Hence he says oI them that they are sciences whose begin-
nings are unknown and whose ends are unknown; and |only| what is between
the beginnings and the ends is known |Resp. VII, 533B6C8|. This being so, the
mathematical sciences are not intended |mekhuwwanot| initially and essentially
Ior human perIection, as is the case with physics and metaphsics. Although they
diIIer in this respectand particularly in what these two sciences take Irom them
|sc., the mathematical sciences| by way oI principles Ior the investigation oI the
end (as when the divine science |i.e., metaphysics| accepts the number oI move-
ments Irom astronomy)this diIIerence is not only with respect to their kinds
|i.e., arithmetic, geometry, astronomy, and music|, but also exists with respect to
the parts oI the particular science.
137
The mathematician gives a descriptive model oI the phenomena, apt to
yield correct calculations, but abstracted Irom reality; the physical philosopher
looks into the proximate causes governing the reality oI the world. In cos-
mology, the noblest object oI both physics and mathematical astronomy, the
philosopher proves his competence to grasp the highest, yet remote object oI
intellectual study.
138
1he Andalusian Restoration ot Aristotle`s Cosmos
The relation between the 'physical science (al-ilm al-tab) oI philosophy
and mathematical astronomy, and the relation oI either to the cosmic reality,
was discussed Irom the side oI natural philosophy throughout the Andalusian
school oI falsafa: Ibn Bjja, Ibn TuIayl, Ibn Rushd, al-Bitrj, Maimonides.
Starting Irom the same principles oI cosmological theory, Arabic astronomers
oI the West put Iorward solutions based on the Aristotelian models oI homo-
centric spheres. Al-Zarqlluh, Ibn TuIayl (Ibn Rushd`s predecessor as physi-
cian to the Almohad court) and the latter`s disciple, the astronomer Ab Ishq
al-Bitrj were the most prominent advocats oI such theories. The point oI criti-
cism turned against Ptolemy and his Iollowers remains the same as beIore: The
calculus oI the mathematicians may ft, albeit imperIectly, the observations, but
does not account Ior the actual processes governed by uniIorm principles and
eternal laws; indeed, it is in evident contradiction to the principles established
by Aristotle. Going Iurther than previous critics oI Ptolemy, they made bold
to build a new confguration which should conIorm both to the principles and
to the calculus matched with observation. Previous Iailure to achieve this was
due not to the principles set up by the philosophers but to the imperIections oI
observation based on sensual data.
139
Ibn Bjja made this clear in the terms oI logical methodology in his short
treatise f Ilm al-haya, starting out Irom Aristotle`s Posterior Analytics: The
astronomers have to Iall back on data arrived at indirectly, through observation
and calculation. In trying to build a universal prooI oI their haya, they set up
a syllogism in which the inIerences oI such fndings will Iorm the middle term
oI a syllogism; hence, their results must be at variance with the principles oI
physical science.
140
Ibn Rushd was not enough oI an astronomer to evaluate the theories
oI his predecessors and contemporaries mathematically. He used to say so
himselI.
141
His approach is dogmatical: Aristotle`s physical and metaphysical
doctrine, seen and interpreted as a closed system, advancing Irom the evidence
and induction oI existence in physics to the causes and principles oI essence in
metaphysics, required the astronomersas in Plato`s assignmentto provide
calculable models which would link the data oI observation with the essential,
unchanging principles oI the eternal circular movement: principles 'pointing
in their turn to the cosmic essences oI the spheres, immaterial soul-intellects
moved by the desire to emulate the Eirst Mover.
Aristotle`s theory oI homocentric spheres,
142
and Averroes` deIense oI
the Aristotelian model against the Ptolemaic system, proceed Irom two basic
assumptions: The eternal movement oI the celestial bodies must be absolutely
regular;
143
and the theories explaining their apparent movements should not
regard the spheres and planets as mere mathematical entities, but as animate
substances, 'enjoying liIe and action.
144
Ibn Rushd`s attempt to reconstruct
Aristotle`s true systeman attempt hampered by the overwhelming success oI
the Ptolemaic system in mathematical astronomy, and compounded by errors
oI translation in the Arabic version oI the Metaphysics
145
, is accompanied
by constant attacks against Ptolemy`s use oI epicycles and excentric cycles in
his interpretation oI the planetary movements, starting Irom his Epitome oI
Aristotle`s De Caelo:
'The apparent advancing and receding motion oI the (sphere oI the fxed) stars
cannot exist in their actual motion. . . . This precession and recession was not
observed by the ancient Greeks except in the case oI the planets, nor were many
oI the multiple motions established by Ptolemy observed by the Babylonians,
such as the movement oI the epicycles.
146
The movement oI the spheres must be homocentric, because the center
oI the earth must coincide with the centre oI the universe.
147
Ptolemy deviated
Irom this principle, because in his model, the center oI the deIerent axis (car-
rying the epicycles) must be excentric against the centre oI the earth, and the
centre oI the epicycle moves on the eccentric with varying velocity, in such
a way that only when seen Irom the punctum aequansthat is, the point on
the line oI apsides whose distance Irom the earth is the double oI the linear
eccentricitythe motion in the eccentric appears to be uniIorm.
148
Averroes
objects:
'That the earth is in the center and at rest is attested by the demonstrations which
the mathematicians are accustomed to apply to these matters. II it were not in
the centre, as Ptolemy claimed, three possibilities would obtain, all oI which
would result in absurdities and are in evident contradiction with the evidence oI
observation.
149
The problem accompanied Averroes all his liIe and is prominent in the
Iong Commentaries on both the Metaphysics and the De Caelo, the magiste-
rial works oI his last years, containing numerous reIerences to the problems oI
celestial mechanics where the Almagest was at variance with Aristotle.
In his early work, the project oI Ibn Rushd was limited to a 'state oI the art sum-
mary oI the encyclopaedia oI the rational sciences, limited to 'what is necessary
Ior the human perIection as a rational being, among the basic methods oI juris-
prudence, oI logic, and oI natural philosophy.
150
The same is explicitely indicated
Ior his 'Summary oI the Almagest,
151
pointing to Ibn al-Haytham as an authority
on this science, while denying the astronomers the philosophers` competence in
the methods oI demonstrative science.
152
In his early Compendium (Jawmi) oI
the Metaphysics (written in the early fIties oI the 6th century A.H.), Iollowing
closely the cosmology oI al-Erb (and revised in a second version), Averroes
seems to fnd epicycles acceptable. But attacks on Ptolemy start about this time
already, in the physical Compendia as well as in the Summary oI the Almagest:
The method Iollowed by the mathematicians oIIers neither 'signs (dalil) nor
demonstrations; most oI what they attribute to the celestial bodies is impossible,
notably the epicycle.
153
The most detailed discussion oI the celestial movements,
based on Metaph. XII.8, is Iound in his Great Commentary (1afsr) oI the
Metaph., completed a Iew years beIore his death c. 590/1194. The commenta-
tor Iounds his attack on the principle oI simple, circular movement Iound in the
celestial bodies, a principle in confict with the system oI eccentric circles and
epicycles, to be replaced by a system oI homocentric circles oI each planet where
the poles oI one circle rotate in the plane oI the adjacent one.
154
Ieaving aside the intricacies oI these discussions with regard to the tech-
nicalities oI astronomy and mathematics
155
what is striking, and relevant Ior a
fnal perspective oI Averroes` scientifc approach, is the apparent subordination
oI applied science to physical theory: Aristotle`s true philosophy was Iounded
upon true science; iI this science oI the Ancients could be restored, all pieces in
the cosmic puzzle would Iall into place. Basically, the issue under discussion is
sound philosophic method.
ReIerring to Aristotle`s De Caelo, chapter II.3, Ibn Rushd elaborates
Aristotle`s note that 'we have to pursue our inquiries at a distancea distance
created not so much by our spatial position as by the Iact that our senses enable
us to perceive very Iew oI the attributes oI the heavenly bodies (268a47):
With regard to these things, only certain premises are available to human
induction,
'and the things Irom which are acquired the premises, by which man scrutinizes
many oI the things concerning the heavenly body and through which he aspires
to know their causes, are derived Irom the things which are closest to those in
resemblance, viz. the animate bodies, and especially man, since it has been made
clear that this body is animate. However, it is evident that this (kind oI statement)
is ambiguous (yuql bil-tashkk) |being ambiguous| about what is prior and pos-
terior, and thereIore this kind oI concept and judgment is weak.
156
The systematic reasoning behind Averroes` cross-reIerences to the other
parts oI philosophy is based on rules oI the burhn: The 'demonstration oI
existence will be provided in the various natural sciences, as, Ior example, the
parts oI the soul in the psychological part oI natural philosophy; the 'demon-
stration oI causes is the privilege oI a discipline higher in rank with regard to
the lower levels, and Ior the very highest can be given only by induction Irom
dalil perceived in the posterior ones.
157
This systematic relationship between natural philosophy and metaphys-
ics is underlined in many statements Irom the very earliest phase oI Averroes`
activity, as in the Compendia oI the Tabiyyt and oI the Metaphysica:
'Demonstration in an absolute way (demonstratio simpliciter, al-burhn al-
mutlaq)
158
in this science is Iounded on the propositions accepted Irom natural
science and theological science: It has been explained in the Physics that the
mover oI the celestial bodies is not in matter, and in the book De Anima that what
is oI this kind is intellect, and in the frst treatise |oI De Anima| that the intel-
ligible Iorm |i.e., soul| is moved only through desire coming Irom its intellect;
hence this must have its object in imaginationit is a celestial body exercising
desire.
159
The Iundamentals oI epistemology, where the demonstration oI exis-
tence, essence and cause constitutes a hierarchy and interdependence oI the
sciences, oI physics and metaphysics, not only justiIy, but require systematic
cross-reIerences between the disciplines; these principles were established by
Aristotle in Metaph. VI.1, and are constantly called upon by Averroes as a
guideline Ior philosophic method.
160
Another case in point is a remark on the order oI the planets: Contrary to
the doctrine oI Ptolemaeic astronomy, the sphere oI the Sun must be assumed
between the Moon and the remaining planets in order to conIorm with the
principles established by Aristotele. In Iact, however, Aristotle had not dis-
cussed the relative positions, distances, and velocities oI the stars and planets
in detail; these topics empirical astronomy was leIt to deal with adequately (De
Caelo II. 10, 291a2932). The astronomer, Averroes explains (Comm mag. De
Caelo, II c. 57, Iol. 64ra3356 ad locum), demonstrates the existence oI data
apparent Irom or indicated by sense perception, regarded as mathematical enti-
ties abstracted Irom matter, while the natural philosopher gives the causes oI
the same subjects regarded as natural substances. But the philosopher reIers to
astronomy, as Aristotle does in De Caelo II.10, because he considers the causes
oI those things whose existence has been established in astronomy:
'Now both the natural philosopher and the astronomer engage in the study oI
these questions; however, the astronomer mainly gives the existence |quia;
Arabic, anna l-shay| while the natural philosopher gives the cause |propter quid;
Arabic, li-ma l-shay|. What the astronomer mainly gives is based only on those
things that appear to the senses . . . , the natural philosopher, however, endeavours
to give the cause why this is so |propter quam hoc est supra ipsam; Arabic, li-ma
huwa al hdh|.
161
The relation between the angular velocity oI the planets and their dis-
tances with regard to the frst heaven underlies a general principle stated by
Aristotle: The absolute speed oI the planet nearest to the frst revolution (the
circle oI the fxed stars)that is, Saturnis greatest, while the others are
slower, the decrease in velocity being in proportion with their distance (De
Caelo II 10, 291b610).
162
While recognizing this general principle, Ptolemy
tried to establish the precise relative order oI the planets with respect to the
Sun.
163
Based on observation and computation oI the relative distance, and
on the apparent eccentricity oI the spheres oI Mercury and the Moon, he con-
cluded that the spheres oI Moon, Mercury, and Venus lie below the Sun while
Mars, Jupiter and Saturn lie above.
164
Unabashed by the reIerence oI the Arabic version to the ashb al-
MafistAristotle`s mathematikoi (291b9)Ibn Rushd declares that the
conditions underlying Aristotle`s exposition are reconcilable only with the
'opinion oI those who say that the Sun is above Mercury and Venus, and not
belowhere the astronomers are at variance, and the truth oI the matter has
not yet been established.
165
However, Aristotle`s statement on the connection
between the planet`s velocity and its distance Irom the frst heaven does not
imply a mathematical, proportional ratio; even though the Sun`s motion may
be quicker than the motions oI Mercury and Venus, it may nevertheless move
in a sphere above, 'because its potency surpasses theirs . . . because there, local
proximity is similar to mutual proximity oI the essences, and this is proximity
in knowledge and in rational cognition: the stronger the cognition oI the frst
movement, the more perIect the desire toward it will be, and the stronger the
desire, the quicker its motion will be.
166
In linking the cosmic order with Aristotle`s metaphysicsthe cosmic
motion originates Irom the conscious desire oI rational souls toward the
Unmoved MoverAverroes puts Iorward the reasoning oI philosophy
against the celestial mechanics oI the astronomers: isti enim motus quos ponit
1holomeus fundantur super . . . fundamenta quae non conveniunt scientiae nat-
urali.
167
But does he betray his own principles, making the natural phaenomena
and the Iacts established by natural science subordinate to a-priori postulates oI
metaphysics? The doubt and caution expressed, again and again, in view oI the
diIfcult and controversial feld oI astronomy may convince us that he is not:
The very passion oI his aporia between astronomy and metaphysics betrays his
awareness that only a true understanding oI the astronomical cosmos will yield
true answers to the ultimate questions oI metaphysics. Indeed, it is the task oI
metaphysics to 'save the phenomena oI observation.
However, the Commentator deplores the astronomers` (and his own)
inability to reconstruct the true Aristotelian cosmos, that is, to provide not just
a mathematical emulation which does not contradict his physics, but a true
physical astronomy. Regarding the order oI the planets, as also the models con-
tradicting Aristotle`s doctrine oI homocentric spheres, 'the necessary move-
ments in these things have not yet been demonstrated in this science: Ior the
movements assumed by Ptolemy are based on premisses not reconcilable with
natural science, sc. eccentrics and epicycles, which are both Ialse.
168
He Ielt
the truth to be near at hand:
'Maybe, iI God will grant me liIe, I shall investigate the |science oI| the sphere
oI Aristotle`s age (al-hayatu llat knat f ahdi Arist); and it will turn out that
it did not contain any |such| absurdities with respect to physical science. This
is |the system Iounded upon| what Aristotle called spiral motions` (al-harakt
al-lawlabiyya). It is, as I think, a movement where the poles oI one sphere turn
about the poles oI another sphere, so that its own movement will proceed along a
spiral line, as Ior example the motion oI the sun with regard to the diurnal motion.
On the basis oI |the assumption oI| this motion we can give |an explanation oI|
what happens to the star, such as the diIIerences in movement, backward move-
ment and movement in a straight line. We shall make a study oI these motions,
Ior it is impossible that there should be irregularity in the celestial bodies unless
|explained| in this way.
169
It is a tragic irony that this 'spiral motion (haraka lawlabiyya), the term
behind which Averroes suspected the fnal solution, goes back to a mistake
oI the translator.
170
But in his old age he despaired oI this hope. A Iew years
beIore his death, he returned to the problem oI the planetary movements in his
Commentarium magnum oI the Metaphysics (in the context oI Aristotle`s dis-
cussion oI the number oI the eternal moving principles oI the heavens, Metaph.
A. 8.1073a14II.).
171
Averroes is ready to concede to Aristotle that 'the specialist oI this sci-
ence must accept Irom astronomy the inIormation it gives about the number
oI the motions, but he insists that he need not accept 'the other matters it
comprises.
172
Eor the discipline oI astronomy which inquires into the celestial
motions 'cannot establish, on the basis oI the spherical motions apparent to us,
the course oI the causes unless it does not contradict the principles of physics
(ill m laysa yalhaqu min wadih muhlun f l-ilmi l-tab |i.e., physical
philosophy|).
173
The apparent irregularities in the planetary motions 'do not
conIorm to the nature oI the motions oI the celestial bodies; that is, it has
become evident in the physical science that all their motions are regular.
174
The astronomer must thereIore postulate a model Irom which the phenomena
(al-ahwl al-zhira) would result without violating the principles oI natural
philosophy (min ghayri an yal:ama minh muhlun f l-ilmi l-tab).
175
The
complexities involvedcomposite movements, excentricities, epicycles
result in disagreeement among the 'modern mathematicians (al-hadathu
mina l-talmiyyn) as to the number oI these motions: an undisguised allu-
sion to Ibn al-Haythams`s Shukk; and here, Ibn Rushd is eager to note that
the calculus based on such mathematical models is, in certain cases at least,
in blatant disagreement with observed phenomena.
176
The mathematicians oI
Aristotle`s time, who provided him with a tentative solution, are commended
Ior being in closer agreement with physical philosophy than Ptolemy and the
'moderns. The theory oI epicycles, positing centres oI motion beside the
centre oI the world, and the model oI excentric spheres, involving 'superfuous
bodies in heaven with no purpose but as flling (hashwan) are both 'contrary
to nature.
177
The only solution Averroes can propose is a closer examination oI the
mathematical astronomy Aristotle himselI relied upon, that is, the theories
oI Eudoxus and Callippus; perhaps the model oI 'spiral motion he ascribes
to Aristotle 'would allow us to do without these two things |i.e., epicycles
and eccentric spheres|. . . . Ptolemy Iailed to notice what had compelled the
Ancients to accept spiral motions, namely the impossibility oI the epicycle
and the eccentric sphere. Butand here we return to the starting point oI our
inquiry'in our time, astronomy is no longer something real; the model exist-
ing in our time is a model conIorming to calculation, not to reality.
178
In this, Ibn Rushd not only observed the letter oI Aristotle`s doctrine, but
also the spirit oI his science, where metaphysics investigated a cosmic reality,
not just 'units with a serial order, but 'enjoying liIe and actionin the fnal
analysis, Plato`s heritage in the Aristotelian encyclopaedia.
3 Ix +nr l:cr or +nr Aixin+.: 1nroio. :xi Scirxcr
The models oI concentric planetary spheres, postulated by Ibn Bjja, under-
stood but vaguely by Ibn Rushd, and constructed, however imperIectly, by
al-Bitrj, had little infuence beyond al-Andalus and the oncoming decline oI
Western Muslim civilization. Neither here, nor indeed in the coming bloom oI
the mathematical sciences in the East, the philosophical Weltbild played a rle
in the improvement oI the old approaches, let alone a renewal apt to launch a
new paradigm. Science thrived at the hands oI theologians who leIt philosophy
to the falsifa who in vain hat whetted their steel to tackle the apories oI the
haya, and discarded with principles which had proved useless Ior the practical
tasks oI the art.
*
The apories admitted, however unwillingly, by Ibn Rushd, were stated
bluntly and uncompromisingly, in a remarkable passage oI the 'Guide oI the
Perplexed, Dallat al-hirn, oI the Jewish jurist, physician and philosopher
Maimonides, on the authority oI Ibn Bjja:
179
'As Iar as the action oI ordering the motions and making the course oI the stars
conIorm to what is seen is concerned, everything depends on two principles:
either that oI the epicycles or that oI the eccentric spheres or on both oI them. . . .
Both those principles are entirely outside the bounds oI qiys |logos, the method
oI demonstrative science| and opposed to all that has been made clear in natural
science |al-ilm al-tab, i.e., the part oI philosophy demonstrating rationally the
processes oI the physical world|.
While explaining the diIfculties involved, controverting the principles
established by Aristotle, and aIter going into the problems oI planetary motion
in detail, Maimonides concludes:
'Consider how great these diIfculties are. II what Aristotle has stated with
regard to natural science is true, there are no epicycles or eccentric circles and
everything revolves around the center oI the earth. But in that case how can the
various motions oI the stars come about? Is it in any way possible that motion
should be on the one hand circular, uniIorm, and perIect, and on the other hand
the things that are observable should be observed in consequence oI it, unless
this be accounted Ior by making use oI one oI the two principles |sc. epicycles
or eccentric circles| or oI both oI them? This consideration is all the stronger
because what is calculated on the hypotheses oI the two principles is not at Iault
even by a minute. . . . This is the true perplexity. However, . . . this does not affect
the astronomer. Eor his purpose is not to tell us in which way the spheres truly
are |laysa maqsduh an yukhbiran bi-srati wufdi l-aki kayfa hiya|, but to
posit an astronomical system in which it would be possible Ior the motions to be
circular and uniIorm and to correspond to what is apprehended through sight |an
yafrida hayatan yumkinu bih an takna l-haraktu dawriyyatan wa-mustawi-
yatan wa-tutbiqa m yudraku iynan|, regardless oI whether or not things are
thus in Iact |kna l-amru ka-dhlika aw lam yakun|.
Ibn Bjja already had expressed doubts whether Aristotle was aware oI
some oI the intricacies oI planetary motion, sc. the eccentricity oI the sun:
indeed, he was not, and would have been baIfed iI he had beenresorting, as
in other cases, to 'guessing and conjecturing.
'However, regarding all that is in heaven, man grasps nothing but a small
measure of what is mathematical |ill bi-hdh l-qadri l-talmiyyi l-yasr, i.e.,
the little that can be established mathematically|. I shall accordingly say in the
manner oI poetical preciousness: The heavens are the heavens oI the Iord, but
the earth hath He given to the sons oI man` |Ps. 115:16|. I mean thereby that the
deity alone knows the true reality |haqqat al-sam|, the nature, the substance,
the Iorm, the motions, and the causes oI the heavens.
180
The decision taken by the great Jewish thinker is in tune as well with the
overall development oI Muslim intellectual attitudes in the later Middle Ages:
while theology becomes scientifc, science becomes theological. The agencies
oI this transition are maniIold, and comport a Iar-reaching shiIt in the social
and intellectual milieu oI science in Islam. Erom the perspective oI philosophy,
the main Iactors are the Hellenization oI Kalm in the school oI al-Ghazl,
notably in the attendance oI Eakr-al-Dn al-Rz in Sunn, and oI Nasr-al-Dn
al-Ts in Sh Islam, the adoption oI Avicenna`s encyclopaedia oI intellectual
knowledge in the religious community, and at the same the elimination oI the
stumbling stones in Aristotelian physicsthe eternity oI the world, and the
laws oI physical causality submitting God to a necessity imposed by reason. In
consequence, the deIense oI the rational sciences was undertaken by members
oI the same religious community who regarded the methods oI demonstration
as an indispensable basis oI sound argument in the service oI Islam. As a Iur-
ther consequence oI this development in the social and intellectual communi-
ties oI rational science, the teaching traditions oI kalm, falsafa, and riydiyyt
grow Irom diverse branches to become parts oI an integrated curriculum oI
learning in the Iranian, and later on in the Mughal and Ottoman madrasas.
181
Eollowing al-Ghazl, the Asharite interpretation oI scientifc knowl-
edge
182
deIended logical method as a 'just scale (al-qists al-mustaqm)
and 'vessel oI knowledge (miyr al-ilm), and in the same vein, the teach-
ers oI the sciencesmany oI whom were powerIul authorities in theology as
wellreIerred to the epistemology oI the Analytica Posteriora, as restated by
al-Erb and realized in his system oI the sciences. As al-Ts states in his
epochal 1adhkira on theoretic astronomy, 'every science has a subject which
is investigated in that discipline, and principles, either selI-evident, or else
obscurein which case they are proved in another science and are taken Ior
granted in this science, and in the case oI astronomy, 'its principles that need
prooI are demonstrated in three sciences: metaphysics, geometry, and natural
philosophy (al-tabiyyt).
183
This is closely reminiscent oI al-Erb`s Kitb
al-Burhn.
184
In the same vein, Adud-al-Dn al-j (d. 1355), in his monumental
Summa oI Asharite Kalm, presents the principles oI astronomy as neutral
with respect to religious law, 'prohibition does not extend to them, being
neither an object oI belieI nor subject to aIfrmation or negation. But this
compromise is reached at a price: 'These |sc. the hypotheses oI mathemati-
cal astronomy| are imaginary things that have no external existence, . . . mere
imaginings which are more tenuous than a spider`s web. Iike in the apories
stated by Averroes and admitted by Maimonides, we come back to the basic
confict between philosophy and astronomy: the confict about the reality oI
their object oI study.
185
There was, however, an alternative open to the theolo-
gian-astronomer, assuming Iull responsibility Ior his science while discarding
with the principles taken Irom falsafa. In a commentary on al-Ts`s dogmatic,
1afrd al-aqid, al-Qshjan astronomer Irom the circle oI Ulugh-Bek,
active in Istanbul until his death in 879/1474seeks to establish his science
in a creation not governed by the physical laws oI causality, but by the will
oI God.
186
'What is stated in the science oI astronomy does not depend upon
physical (tabiyya) and metaphysical (ilhiyya) premises. . . . Eor oI what is
stated in this science, some things are geometrical premises, which are not
open to doubt; others are premises arrived at through intuition (muqaddamt
hadsiyya),
187
as we have mentioned; others are premises determined by reason
in accordance with the apprehension oI what is most suitable and appropriate.
On this basis, 'it is suIfcient Ior the scientist to conceive, Irom among the
possible models, the one by which the circumstances oI the planets with their
maniIold irregularities may be put in order in such a way as to Iacilitate their
determination oI the positions and conjunctions oI these planets Ior any time
they might wish and so as to conIorm with perception and sight, this in a way
that the intellect and the mind fnd wondrous.
188
Bent to describe adequately
the opihcium mundi, the theologian-scientist Ielt Iree to discard the causes pro-
claimed by the philosophers as universal principles, and going Iurther, Ielt Iree
to explore possibilities contradicting the principles oI Aristotelian physics.
189
It was a proud and competent astronomer who took Ptolemy himselI
to task Ior an metabasis eis allo genostrespassing into the Ioreign domain
oI cosmology: al-Brn who in his Qnn al-Masd upbraided Ptolemy Ior
having, in his Planetary Hypotheses, 'deviated Irom the path he had Iollowed
in the Almagest, |having taken up| that which related to opinions outside oI this
science, that is in the belieI oI people that the celestial bodies have liIe, percep-
tion, sensation, and the choice oI the noblest motions.
190
The alternative, chosen by many generations oI creative mathematicians
and astronomers, was to Iollow the principles oI demonstrative science as they
had been established by Aristotle, reIormulated by al-Erb as supreme scien-
tifc method and restated by Ibn Sn, and at the same time, leaving cosmology
aside: Conceding that every science has its own principles; that one science
will build upon the other; and that metaphysics comes not frst, but next (iI
at all). The winds oI change rose not Irom the side oI philosophers, but Irom
mathematicians who observed, described and calculated the phenomena, and
contemplated the wonder oI creation in the eternal splendor oI the cosmic
order.
No+rs
1. Ibn Rushd, Sharh k. al-Sam wa-l-lam, I, c. 90; ms. Tunis, Ahmadiyya, no. 5538,
Iol. 70a ed. in Iacsimile by G. Endress, Commentary on Aristotle's Book On the
Heaven and the Universe, by Ibn Rushd (ErankIurt am Main: Institute Ior the History
oI Arabic-Islamic Science, 1994), 47 (passage missing in the Iatin version oI Michael
Scot, and perhaps added by the author at a later date); cI. another passage in the same
spirit, ibid., II c. 35, Iol. 38b2224 Iacs. ed., p. 210 (in ansaa |MS. in sha| Llhu
j l-umr |see below, p. 267|; v. G. Endress, 'Averroes De Caelo, Arabic Sciences and
Philosophy, 5 (Cambridge, 1995): 949, esp. p. 45, and also Jaml-al-Dn al-Alaw,
al-Matn al-rushd (Casablanca, 1987), 107.
2. Ibn Rushd, Tajsr M bad al-taba, XII c. 45, ed. M. Bouyges, Bibliotheca Arabica
Scholasticorum, VVII (Beyrouth, 193842), p. 1663.1112, 1664.27; cI. English
trans. by Ch. Genequand, /bn Rushd's Metaphysics (Ieiden: Brill, 1984), 179.
3. 'Nobody who is ignorant oI geometry shall enteraccording to Hellenistic tradi-
tion, the inscription oI the Platonic Academy at Athens; see, e.g., Elias, Comm. in Arist.
Cat., ed. A. Busse (Berlin, 1900), 118.19.
4. Eacs. oI the frst edition (/oannis Keppleri Harmonices mundi libri \, Iincii Austriae:
Plancus, 1619) in Johannes Kepler, Gesammelte Berke, Bd VI: Harmonice mundi, ed.
Max Caspar (Mnchen: Beck, 1940), 13 (Iacsimile oI the original title).
5. Proclus, /n / Euclidis Elem., ed. G. Eriedlein (Iipsiae: Teubner, 1873): prologus I, p.
22, 1720: apo= or tpv uoixpv 0rmpiov to ryioto ouortoi, tpv tmv oymv
rutoiov ovooivouoo, xo0' pv oropioupyptoi to aov, xoi ovooyiov tpv aovto to
rv tm xoom ouvopoooov. English translation by Glenn R. Morrow, Proclus, A Com-
mentary on the First Book oj Euclid's Elements (Princeton, N.J.: Princeton University
Press, 1970), 19.
6. See Albert Dihle, Philosophie als Lebenskunst, Rheinisch-WestIlische Akademie
der WissenschaIten, GeisteswissenschaIten, Vortrge, G 304 (Opladen: Westdeutscher
Verlag,1990), 12, on the superior claim oI the Platonists Ior the study oI physical phi-
losophybuilding upon mathematics, but transcending Irom the mathematicals, veri-
fed empirically, to the purely intelligible principles (Proclus, /n Tim., 23.9 II.); Pierre
Hadot, Qu'est-ce que la philosophie antique? (Paris: Gallimard, 1995), 100 I.
7. See Ilsetraut Hadot, 'Ies aspects sociaux et institutionels des sciences et de la mede-
cine dans lÀntiquite tardive, Antiquite tardive, 6 (Turnhout, 1998): 233250; ead.,
Arts liberaux et philosophie dans la pensee antique (Paris: Etudes Augustiniennes,
1984).
8. 'Encyclopaedia, that is, a general, 'all-round education encompassing the felds oI
knowledge preparing the way to higher learning.
9. I. Hadot, ibid., 242244.
10. Hunayn b. Ishq, Risla il Al b. ahy, ed. G. Bergstrsser: Hunain b. /shq uber
die syrischen und arabischen Galenuberset:ungen (Ieipzig, 1925), Arabic text, 18;
German trans., 15. The earliest translations oI Aristotelian logic were commissioned to
scholars oI the Christian churches, who had kept up the teaching oI logic in Syriac, and
oI the isagogic tradition oI the Alexandrian school; most prominentbut quite inde-
pendent Irom those oI the astrologers and physiciansare the activities oI the patriarch
Timothy, working by commission Ior the caliph al-Mahd; v. Henri Hugonnard-Roche,
'Ies traductions du grec au syriaque et du syriaque a làrabe (a propos de lÒrganon
dÀristote), in Rencontre de cultures dans la philosophie medievale. traductions et tra-
ducteurs de lAntiquite tardive au XIJ
e
siecle (Iouvain-la-Neuve; Cassino, 1990), 133
147; Sebastian Brock, 'Two letters oI the Patriarch Timothy Irom the late eighth century
on translations Irom Greek, Arabic Sciences and Philosophy, 9 (1999): 233246.
11. CI. Albrecht Dihle, 'PhilosophieEachwissenschaItAllgemeinbildung, in
Aspects de la philosophie hellenistique, Entretiens sur lÀntiquite classique, 32
(Vandouvres-Geneve: Eondation Hardt, 1986), 188232.
12. to o0poto xoi oi opi0oi , Pythag. Ir. 14 A 17 Diels/Kranz, IromApollon, Mira-
bilia 6 Keller.
13. Philolaos, Ir. 44 B 4.
14. opi0ov rivoi tpv ouoiov aovtmv, Arist. Metaph. A 5, 987a1319. CI. the sources
presented by C. J. de Vogel, Greek philosophy. a collection of texts, vol. 1 (Ieiden,
1963), 10II., esp. p. 16I. (nos. 3742).
15. Metaph. A.6, 987b10; e 8, 1073a18: opi0ou= yop ryouoi to= ioro= oi ryovtr=
ioro=.
16. CI. the introduction oI H. G. Zekl to his translation oI the Analytica posteri-
ora, Aristoteles, Erste Analytik, Zweite Analytik, in Organon/Aristoteles; Bd 3/4,
griechisch-deutsch, Philosophische Bibliothek, Bd. 494/495 (Hamburg: Meiner, 1998),
lviiilxxxvi.
17. Resp. VII, 526a6: mv oiovop0pvoi ovov rympri .
18. Resp. VII, 526a13: apooovoyxoov outp tp vopori ppo0oi tpv upv ra'
outpv tpv op0riov.
19. Resp. VII, 527B7, 910: tou ori ovto=.
20. On mathematics in the history oI Platonism, see Heinrich Drrie and Matthias
Baltes, Der Platonismus im 2. und 3. Jahrhundert nach Christus, Der Platonismus
in der Antike, Bd 3 (Stuttgart-Bad Cannstatt: Erommann-Holzboog, 1993), 6871,
266279.
21. Thomas I. Heath, A history of Greek mathematics, I, 98; Drrie and Baltes, op.
cit., 267I.
22. On the Hellenistic reading and transmission oI Nicomachus, v. Ieendert Gerrit
Westerink, 'Deux commentaires sur Nicomaque: Asclepius et Jean Philopon, Revue
des etudes grecques, 77 (1964): 526535; Etienne Evrard, Jean Philopon, son Com-
mentaire sur Nicomaque et ses rapports avec Ammonius, ibid. 78 (1965): 592598;
Ieonardo Tarn |ed.|, Asclepius oI Tralles, Commentary to Nicomachus Introduction
to Arithmetic (Philadelphia: American Philosophical Society, 1969). A close contempo-
rary was Theon oI Smyrna, who wrote an 'Exposition oI mathematical matters useIul
Ior the study oI Plato (ed. E. Hiller, Ieipzig 1878).
23. Dominic J. O`Meara, Pythagoras Revisited. mathematics and philosophy in late
Antiquity (OxIord: Clarendon Press, 1989), 212.
24. Ibid., 166169; see also Ian Mueller, 'Mathematics and philosophy in Proclus`
commentary on book I oI Euclid`s Elements, in Proclus. lecteur et interprete des
anciens, Actes du colloque international du CNRS (24 October 1985) publies par Jean
Pepin et H. D. SaIIrey (Paris: CNRS, 1987), 305318.
25. I. G. Westerink, 'Philosophy and medicine in Iate Antiquity, Janus, 51 (1964):
169177; id., 'Ein astrologisches Kolleg aus dem Jahre 564, By:antinische Zeitschrift,
64 (1971): 621.
26. CI. G. Endress, 'Die wissenschaItliche Iiteratur, in Grundriss der Arabischen
Philologie, 2 (Wiesbaden: Reichert, 1987), 407409.
27. The textual evidence does not support the thesis proposed by Michel Tardieu that
Simplicius, on his return Irom the court oI Khosrow Anshrwn, retired to Harrn,
and that the Sbiat al-nniyyn mentioned by al-Masd should have kept up a
Platonic Academy until the 10th century; v. Tardieu, 'Sbiens coraniques et Sabiens
de Harran, Journal asiatique, 274 (1986): 144; idem, 'Simplicius et les calendriers
de Harrn dàpres les sources arabes et le commentaire de Simplicius a la Physique
dÀristote, in Simplicius. sa vie, son oeuvre, sa survie (Berlin: de Gruyter, 1987), 40
51; cI. Dimitri Gutas, 'Plato`s Symposion in the Arabic tradition, Oriens, 31 (1988):
3660 (p. 44); Concetta Iuna, review oI: Rainer Thiel, Simplikios und das Ende der
neuplatonischen Schule in Athen (Stuttgart: Steiner, 1999), in: Mnemosyne, ser. 4, vol.
54 (Ieiden, 2002): 482504.
28. Through his pupil al-Sarakhs; v. Eranz Rosenthal, Ahmad b. at-Tayyib as-Saras
(New Haven: American Oriental Society, 1943), 4151.
29. See G. Endress, 'Al-Kind ber die Wiedererinnerung der Seele: arabischer Pla-
tonismus und die Iegitimation der WissenschaIten im Islam, Oriens, 34 (Ieiden,
1994): 174221.
30. Ed., with a Turkish translation, by Aydin Sayili, 'Sbit ibn Kurra`nin Pitagor teor-
emini tamimi, Belleten, 22 (Ankara: Trk Tarih Kurumu, 1958): 526549.
31. Ibid., p. 541.
32. Qutotations are Iound in the Arabic scholia oI al-Nayrz, who in his turn relied
not on the Iull commentary oI Procluswhom he does not mention by namebut
rather on the scholia Iound in his manuscript, containing explicit reIerences to Heron
and transmitting the doctrine oI ProclusSimplicius; v. Rdiger Arnzen, Abul-
Abbs an-Nayr:s Ex:erpte aus (Ps.-?) Simplicius Kommentar :u den Dehnitionen,
Postulaten und Axiomen in Euclids Elementa I, eingeleitet, ediert und mit arabischen
und lateinischen Glossaren versehen (Kln, Essen, 2002), xxvixxxvii and Index nomi-
num. These concern specifc problems, but convey scarcely any traces oI the philo-
sophic prolegomena. Proclus` Elementatio physica was also known; a selection oI the
propositions, in Arabic translation, is quoted in a treatise oI Yahy ibn Ad on the reIu-
tation oI atomism; v. G. Endress, 'Yahy ibn Ad`s Critique oI Atomism: three treatises
on the indivisible part, ed. with introduction and notes, Zeitschrift fur Geschichte der
Arabisch-Islamischen Wissenschaften, 1 (ErankIurt a. M., 1984): 155179.
33. Translated Irom the Syriac by the Nestorian Habb (Abdyash) ibn Bihrz; v.
M. Steinschneider, Die hebraischen berset:ungen des Mittelalters und die Juden
als Dolmetscher (Berlin, 1893; repr. Graz: Akad. Druck- und Verlagsanstalt, 1956),
320, p. 516519; this version, which is no longer extant, was replaced by that oI
Thbit ibn Qurra (d. 288/901), ed. W. Kutsch: Jbit b. Qurras arabische berset-
:ung der 'Aot0aj:txj Etoo,o,j des Nikomachus von Gerasa (Beyrouth: Imprimerie
Catholique, 1959). A Hebrew commentary-paraphrase, based on al-Kind`s redaction, is
extant in manuscript (v. Steinschneider, loc. cit.), and is being prepared Ior publication
by Gad Ereudenthal.
34. Kitb al-Msq al-kabr, v. Ibn al-Nadm, al-Fihrist, ed. Elgel, 269.23 (wa-li-
hdh l-kitbi mukhtasart).
35. al-Kind, K. il l-Mutasim bi-Llh j l-jalsaja al-l, ed. M. Abd-al-Hd Ab Rda
(Cairo, 195053), 1:97162, p. 103.1; Risla j kammiyyat kutub Aristtls, ed. M.
Guidi, R. Walzer, Studi su al-Kind /: Uno scritto introduttivo allo studio di Aristotele
(Roma: Accademia Nazionale dei Iincei, 1940), 6, p. 393 ed. Ab Rda, Rasil,
372I.
36. Ab Mashar, K. al-Madkhal al-kabr il ilm ahkm al-nujm: Liber introductorius
maior ad scientiam iudiciorum astrorum, ed. Richard Iemay (Napoli: Istituto Universi-
tario Orientale, 1995), 1:2332 (introd.).
37. On the position oI the mathematicals, esp. on the doctrine oI the three kinds oI
being, see Heinrich Drrie and Matthias Baltes, Die philosophische Lehre des Pla-
tonismus. einige grundlegende Axiome, platonische Physik (im antiken Jerstandnis), I,
Der Platonismus in der Antike, Bd 4 (StuttgartBad Cannstatt: Erommann-Holzboog,
1996), 4866 (texts), 26690 (commentary); Philip Merlan, From Platonism to Neopla-
tonism (Den Haag 1953; 1960), 22025; G. Endress, 'al-Kind ber die Wiedererin-
nerung |supra, n. 29|, 182.
38. al-Kind, Risla j kammiyyat kutub Aristtls wa-m yuhtj ilayhi j tahsl al-
falsafa, edd. M. Guidi and R. Walzer: Uno scritto introduttivo allo studio di Aristotele,
Studi su al-Kind I (Roma: Accademia Nazionale dei Iincei, 1940); v. Christel Hein,
Dehnition und Einteilung der Philosophie. von der spatantiken Einleitungsliteratur :ur
arabischen En:yklopadie, Europische HochschulschriIten, 177 (ErankIurt a.M. |etc.|:
Iang, 1985), 174177.
39. CI. Ilsetraut Hadot, 'Ia division neoplatonicienne des ecrits dÀristote, in Simpli-
cius, Commentaire aux Categories dAristote, trad. commentee, (Ieiden: Brill, 1990),
Iasc. 1:6393, p. 91.
40. Op. cit., p. 391, Iollowed by a detailed explanation oI the quadrivium, p. 394. See
the reIerences to the Greek commentators given by R. Walzer, ibid., p. 377.
41. Ibn al-Nadm, al-Fihrist, ed. Elgel, 255.28.
42. Ed. by Zakariyy YsuI, Muallajt al-Kind al-msqiyya (Baghdd: al-Majma al-
Ilm al-Irq, 1962), 6792; cI. Amnon Shiloah, 1he theory of music in Arabic writings
(c. 9001900) (Mnchen: Henle, 1979), 254I.
43. al-Kind, al-Musawwitt al-watariyya, l. c., p. 70; on the classifcation oI mathe-
matics, see C. Hein, op. cit. |above, n. 38|, p. 182II.
44. al-Musawwitt, 71.1020.
45. Ibid. 72.
46. Ibid. 77.510.
47. Ibid. 78.1079.20.
48. Third maqla, p. 85II.: j mushkalat al-awtr.
49. Ed., with introduction and Erench translation, by Ieon Gauthier, Antecedents greco-
arabes de la psychophysique (Beyrouth: Imprimerie catholique, 1938); see also M. Ull-
mann, Die Medi:in im Islam (Ieiden: Brill, 1970), 302, with Iurther reIerences.
50. Tzvi Iangermann, 'Another Andalusian revolt? Ibn Rushd`s critique oI al-Kind`s
pharmacological computus, infra, chapter 12.
51. Ibn Rushd, al-Kulliyyt j l-tibb, edd. Sad Shaybn, Ammr al-Tlib (Cairo: al-
Majlis al-Al li-l-ThaqIa, 1989), 309313, quotation Irom p. 312.8; corresponding
to ed. J. M. Eorneas Besteiro, C. Alvarez de Morales (Madrid: Consejo Superior de
Investigaciones Cientifcas; Granada: Escuela des Estudios Arabes, 1987), 389392
(391.6).
52. A detailed exposition was given by Paul Kraus, 1bir ibn Hayyn: contribution
a lhistoire des idees scientihques dans lIslam, t. 2. Jbir et la science grecque (Ie
Caire, 1942; repr. Paris: Ies Belles Iettres, 1986), ch. V, 'Ia theorie de la balance,
187303.
53. Eor these 'orders (tori=) and 'degrees (oaootoori=) cI. Galen, De comp. med.
per genera, II.1 (Opera, ed. Kuehn, vol. 13: 464I.); III.2 (p. 572I.); De simpl. med., V.
27 (vol. 11: 786788); v. quotations in Kraus, Jbir, 189, nn. 24.
54. See the excerpt Irom K. al-Bahth, edited by Paul Kraus, 1bir ibn Hayyn: textes
choisis (Paris, Ie Caire: IEAO, 1935), 510513, and the relevant passage in Kraus,
1bir, II:194I.
55. CI. Kraus, op. cit., 20II., adding abundant reIerences to Greek philosophical and
mathematical literature, and also to the use oI arithmetical symbolism in Christian
Greek authors.
56. Or Kitb /khwn al-Saj: ed. Eriedrich Dieterici, Die Abhandlungen der Ichwan es-
Safa (Ieipzig, 1883); ed. Khayr-al-Dn al-Zirikl (Cairo, 1928); repr. with an introd. by
Butrus al-Bustn (Bayrt: Dr Sdir, 1957), and other editions; cI. Amnon Shiloah, 1he
theory of music in Arabic writings (c. 9001900) (Mnchen: Henle, 1979), 230233.
57. Explicitly: mithlam kna yajaluh l-hukam al-Fthghriyyn, I:48.11 (Risla
I, 1).
58. Nicomachus Gerasenus, Introductio arithmetica, ed. Hoche (Iipsiae: Teubner,
1866), 12, reIerring to Pythagoras; cI. Martin Iuther DÒoge (trans.), Nicomachus oI
Gerasa, Introduction to Arithmetic (NewYork: Macmillan, 1926), 181.
59. Rasil, PreIace, I:21.1022.2. CI. Nicomachus, Introd. Arithm., p. 1 (trans. DÒoge,
p. 182): 'This wisdom he |sc. Pythagoras| defned as the knowledge, or science, oI the
truth in real things, conceiving science to be a steadIast and frm apprehension oI the
underlying substance, and real things to be those which continue uniIormly and the
same in the universe |etc.|; oI the Iour mathematical sciences, one should start with
the one 'which naturally exists beIore them all, is superior and takes the place oI origin
and root; 'this is arithmetic, not solely bcause we said that it existed beIore all the
others in the mind oI the creating God like some universal and exemplary plan, relying
upon which as a design and archetypal example the creator oI the universe sets in order
his material creations and makes them attain to their proper ends, but also because it is
naturally prior in birth, inasmuch as it abolishes other sciences with itselI (p. 9; trans.
DÒoge, p. 187).
60. The Epistle on Music oj the /khwn al-Saj (Baghdad, 10th century), transl. Amnon
Shiloah, Documentation and Studies: Publications oI the Department oI Musicology
and the Chaim Rosenberg School oI Jewish Studies, Tel Aviv University (Tel-Aviv: Tel
Aviv University, Eaculty oI Eine Arts, School oI Jewish Studies, 1978). ReIerences to
the Arabic text are to the edition printed at Beirut, Dr Bayrt & Dr Sdir, 1957.
61. The Hellenistic antecedents oI this attitude are Ioundapart Irom Plato himselI and
older Pythagoreanismin the ps.-Platonic Epinomis, in the Neoplatonic commentaries
on Euclid`s Elementa and on NicomachusIamblichus, Proclus, Philoponusand in
Porphyry`s In Ptolem. Harm. (quoting, like Nicomachus, the treatise On Harmony by
Archytas oI Tarentum); cI. the documentation by E. E. Robbins and I. Ch. Karpinski in
M. I. DÒoge |trans.|, Nicomachus oI Gerasa, Introduction to Arithmetic |supra, n. 58|.
62. Shiloah, 1he Epistle on Music, introd., 7.
63. Rasil, I:237.26; tr. Shiloah, p. 68 (aphorism 18); cI. also pp. 23439 Shiloah,
669, aphorisms no. 7, 9, 11, 20, 21.
64. A detailed bibliography has been presented by Everett K. Rowson in his analysis
oI the teaching tradition leading Irom al-Kind to the later heirs oI Balkh`s school:
A Muslim philosopher on the soul and its fate. al-mir's Kitb al-Amad al l-abad
(New Haven, Conn.: AOS, 1988), 119; see also the same, 'The philosopher as litte-
rateur, Zeitschrift fur Geschichte der Arabisch-Islamischen Wissenschaften 6 (1990):
4092, esp. pp. 17 on Ab Zayd al-Balkh.
65. Abu `l-Hasan M. b. YsuI al-mir, K. al-/lm bi-manqib al-/slm, ed. A.
Abdalhamd Ghurb (Cairo: Dr al-Kitb al-Arab, 1967).
66. Manqib, 88.5.
67. Manqib, 90.3, 90.810.
68. The epithet al-muallim al-awwal does not seem to occur beIore al-Erb, but
al-Kind already calls him mubarri: al-nniyyn j l-jalsaja (al-Falsaja al-l, ed.
Ab Rda, 102I.) Eor the stages oI Arabic Aristotelianism, cI. G. Endress, 'IÀristote
arabe: reception, autorite et transIormation du Premier Matre, Medioevo, 24 (Padova,
1998): 142.
69. raiotpp raiotpmv, ilm al-ulm; v. Hein, Dehnition, p. 39.
70. An. post., I.22, 83a3334; cI. J. Barnes |trans.|, Aristotles Posterior Analytics
(OxIord: Clarendon Press, 1975), 34, 169.
71. Ibid., I.1, 71a14.
72. CI. Kurt von Eritz, 'Die opoi in der griechischen Mathematik, Archiv fur Begriffs-
geschichte 1 (Bonn, 1955): 13103, repr. in von Eritz, Grundprobleme der Geschichte
der antiken Wissenschaft (Berlin, 1971), Arpad Szabo, 'The transIormation oI math-
ematics into deductive science and the beginnings oI its Ioundation on defnitions and
axioms, Scripta Mathematica, 27 (1964), 1:2748, 2:11339 (quotation Irom p. 137).
73. Walter Ieszl, 'Mathematics, axiomatization, and the hypotheses, in Aristotle on
science. the Posterior Analytics, ed. by Enrico Berti, Studia Aristotelica, 9 (Padova:
Antenore, 1983), 270328 (v. conclusions pp. 304I., 313I., 326I.); WolIgang Kullmann,
'Die Eunktion der mathematischen Beispiele in Aristoteles` Analytica posteriora,
ibid., 24570 (conclusions p. 267I.).
74. Edward Booth, Aristotelian aporetic ontology in Islamic and Christian thinkers
(Cambridge |etc.|: Cambridge University Press, 1983), 2.
75. J. Barnes, Aristotles Posterior Analytics, xi.
76. Trans. Thomas Heath, Mathematics in Aristotle (OxIord: OxIord University Press,
1949), 8I.
77. Aristotle, Phys. II.2, 193b22194b15 (194a8: to uoixmtrpo tmv o0potmv).
78. Ibid. 193b34: mpioto yop tp vopori xivporm=.
79. al-Erb, K. al-1am bayna rayay al-hakmayn, ed. Dieterici/Ndir (Beirut 1960),
97.1998.8.
80. CI. G. Endress, 'The deIense oI reason: the plea Ior philosophy in the religious
community, Zeitschrift fur Geschichte der Arabisch-Islamischen Wissenschaften, 6
(ErankIurt a. M., 1990): 149 (esp. 1623).
81. al-Erb, K. al-Burhn, ed. Mjid Eakhr, al-Mantiq ind al-Frb (Bayrt: Dr
al-Mashriq, 1987), 65.1822.
82. Ibid. 70.1214.
83. al-Erb, K. al-Burhn, 68.ult.70.11.
84. al-Erb, /hs al-ulm, ed. Uthmn Amn (Cairo, 1949), 74.
85. But v. Sarah Stroumsa, 'Al-Erb and Maimonides on medicine as a science,
Arabic Sciences and Philosophy, 3 (1993): 235249.
86. /hs, 7778.
87. Ibid., 7983, esp. 80.113.
88. This is analyzed on the basis oI al-Erb`s /hs, and interpreted in the light oI
related texts Irom the philosophical and the scientifc traditions, in the enlightening
contribution to the present volume by Elaheh Kheirandish, 'The many aspects oI
Appearances: Arabic optics to 950 A.D., supra, pp. 5381. I owe to her article the reIer-
ences given below, notes 9091.
89. v. al-Erb, K. al-Burhn, 25I. CI. Aristotle, An. Post., I.2.71b18: oaooriiv or
rym ouoyioov raiotpovixov.
90. An. Post. I.13.78a30II.; Probl. Phys. XV.6.911b1934.The Arabic version oI
Probl. Phys. XV.6, reIerred to in the original Greek by E. Kheirandish (infra, p. 56, text
P), is now available in the edition oI I. S. Eilius, 1he Problemata Physica attributed to
Aristotle. the Arabic version of Hunain ibn /shq and the Hebrew version of Moses ibn
1ibbon, Aristoteles Semitico-Iatinus, 11 (Ieiden: Brill, 1999), 658/659.
91. Quoted by E. Kheirandish, infra, 59 (text I), ed. Roshdi Rashed, Cuvres phi-
losophiques et scientihques dal-Kind, vol. 1 (Ieiden: Brill, 1997), appendice II:
'Ia catoptrique de Qust ibn Iq, 572/573. The addressee is the Abbasid regent
al-MuwaIIaq (regn. 84391).
92. Ibn Sn, Tis rasil j l-hikma wa-l-tabiyyt li-l-shaykh al-ras, ed. Hasan s
(Beirt: Dr Qbis, 1986) |a reprinting oI one oI several early Cairo editions oI the
'nine epistles|, no. 5, 8394, on the mathematical science: pp. 84 and 89.
93. William E. Gohlmann, 1he life of Ibn Sina, Studies in Islamic Philosophy and
Science (Albany: The University oI New York Press, 1974), 2027; Dimitri Gutas,
Avicenna and the Aristotelian tradition, Islamic Philosophy and Theology, 4 (Ieiden:
Brill, 1988), 26.
94. Ibn Sn, al-Shif, al-Ilhiyyt, ed. |George Shehta| Qanawt |Anawati|, Sad
Zyid (Cairo: al-Haya al-mma li-shun al-matbi al-amriyya, 1960), 10I.; Erench
translation by Georges C. Anawati: Avicenne, La Metaphysique du Shif, Etudes
musulmanes, 21 (Paris: Vrin, 1978), 91II.
95. al-/lhiyyt, 10.1014.
96. Ibid., 11.1212.2.
97. Ibid., 2124; cI. Anawati, 100102.
98. Ibid., 24.12.
99. ed. Anawati and Zyid, 303324.
100. Quoted Irom the translation oI A. I. Sabra in The Optics oj /bn al-Haytham, books
IIII (Iondon: Warburg Institute, 1989), II: Commentary, 3; cI. also Shlomo Pines, 'Ibn
al-Haytham`s critique oI Ptolemy, in Proceedings oj the Xth /nternational Congress oj
the History oj Science (Paris, 1964), 54750 Studies in Arabic versions oj Greek texts
and in mediaeval science, The Collected works oI Shlomo Pines, vol. 2 (Jerusalem:
Magnes Press; Ieiden: Brill, 1986), 43639.
101. Eom the introduction to Ibn Sn`s Mantiq al-mashriqiyyn, translated by Dimitri
Gutas, Avicenna and the Aristotelian tradition (Ieiden |etc.|: Brill, 1988), 44I.
102. al-Hasan Ibn al-Haytham, al-Shukk al Batlamys, edd. Abd-al-Hamd Sabra,
Nabl al-Shahbn (Cairo: Dr al-Kutub, 1971), 19.12, 37.6.
103. Ibn al-Haytham as quoted Irom his autobiography by Ibn Ab Usaybia, Uyn
al-anb j tabaqt al-atibb, ed. A. Mller (Cairo, Knigsberg, 1884), 2:93.4; cI.
Matthias Schramm, Ibn al-Haythams Weg :ur Physik, Boethius, 1 (Wiesbaden: Steiner,
1963), 141, cI. 285.
104. Ibn al-Haytham, On the Conhguration of the World, edition, translation, and com-
mentary by Y. Tzvi Iangermann (NewYork: Garland; 1990).
105. CI. M. Schramm, Ibn al-Haytham, 136I.
106. M. Schramm, Ibn al-Haytham, 136138.
107. Ibn al-Haytham, Kitb j hall shukk kitb Uqldis j l-usl, Iacsimile ed. by E.
Sezgin, Ibn al-Haytham, On the Resolution of Doubts in Euclids Elements and Inter-
pretation of Its Special Meanings (ErankIurt a.M., 1985), 4.47. CI. Proclus, In I Eucl.,
31.23, 31.2232.2, trans. Morrow, 26: 'Every Iorm oI knowledge which apprehends
the logos, or cause, oI the thing it knows is a science. . . . Eor genuine science is the
one, the science by which we are able to know all things, the science Irom which come
the principles oI all other sciences, some immediately and some at Iurther remove.
And cI. Aristotle, Phys. II.3.195b21; Metaph. H.4.1044b12: the physicist looks Ior the
proximate, most pertinent cause.
108. Ibn al-Haytham, K. al-Manzir, al-maqlt 13, ed. Abdalhamd Sabra (Kuwait,
1983), |6|, p. 62.79: wa-nabtadiu j l-bahthi bi-stiqri l-mawjdti wa-tasajjuhi
ahwli l-mubsarti, wa-numayyi:u khawssa l-ju:iyyti wa-naltaqitu bi-stiqri m
yakhussu l-basara j hli l-ibsr, etc.; cI. Sabra, 1he Optics of Ibn al-Haytham, I
(Translation): quoted by Sabra in this volume, p. 85.
109. CI. A. I. Sabra, loc. cit., II, 12I.: 'inspection or examination oI particulars, reIer-
ring to al-Erb`s al-Aljz al-mustamala j l-mantiq (ed. M. Mahd, 93) Ior the use oI
istiqr in conjunction with tasajjuh 'to review, to survey.
110. Joep Iameer, Al-Frb and Aristotelian syllogistics. Greek theory and Islamic
practice, Islamic Philosophy, Theology and Science, 20 (Ieiden |etc.|: Brill, 1994),
144confrming A. I. Sabra`s analysis oI the terminology. It is true that this under-
standing may be claimed Ior the Aristotelian epagge as well.
112. Ibid., 285289.
113. Ibn al-Haytham, al-Manzir [supra, n.108|, book II, section 3, |200232|
pp. 307 316; v. A. I. Sabra, 1he Optics of Ibn al-Haytham |supra, n. 100|, I (trans.),
pp. 200206; II (commentary), pp. 97201.
114. Emriv to oivorvo, Simplicius, In De Caelo comm., ed. Heiberg, 488.1824,
492.25II.; Proclus, Hypotyposis astronomicarum positionum, ed. C. Manitius (Ieipzig,
1909), 5.10; cI. Harold Cherniss, 'The philosophical economy oI the theory oI ideas,
American Journal of Philology, 57 (1936): 44556 Cherniss, Selected papers (Ieiden,
1977), 121132; Shmuel Sambursky, 1he physical world of the Greeks (Iondon, 1956),
59; idem, 1he physical world of late Antiquity (Iondon, 1962); Jrgen Mittelstrass, Die
Rettung der Phanomene. Ursprung und Geschichte eines antiken Forschungsprin:ips
(Berlin, 1962).
115. Willy Hartner, 'Medival views on cosmic dimensions and Ptolemy`s Kitb
al-Manshrt, in Melanges Alexandre Koyre, 1 (Paris: Hermann, 1964), 254282
W. Hartner OriensOccidens (Hildesheim: Olms, 1968), 319348; Bernard R. Gold-
stein, 1he Arabic version of Ptolemys Planetary Hypotheses, Transactions oI the
American Philosophical Society, new series, vol. 57.4 (Philadelphia, 1967).
116. Willy Hartner, 'Ealak, in Encyclopdia of Islam, new ed., 2 (1963): 761763.
117. The position oI Sosigenes, the teacher oI Alexander oI Aphrodisias, was analyzed,
as an introduction to his study oI Ibn al-Haytham`s theory, by M. Schramm, /bn al-
Haytham, 1563.
118. Alain Philippe Segonds, 'Proclus: astronomie et philosophie, in Proclus, lecteur
et interprete des anciens |as quoted supra, note 24|, 317334; quotation p. 332 Irom
Proclus, Hypotyposis astronomicarum positionum, ed. C. Manitius (Ieipzig: Teubner,
1909), 236. See also W. Hartner, 'Medival views, quoted supra, n. 115; Iucas Sior-
vanes, Proclus. neo-Platonic philosophy and science (Edinburgh: Edinburgh Univer-
sity Press, 1996), 266.
119. Ibn al-Haytham, Shukk |supra, n. 102|, 16.2
120. Ibid., 33.24.
121. Ibid., 34.38.
122. Ibid. 37.1415, cI. also 33.5, 63.14II.; the reIerence to the Almagest, book IX.2,
is Iound in Claudii Ptolemaei opera quae exstant omnia, vol. 1: Syntaxis mathematica,
ed. J. I. Heiberg (Iipsiae: Teubner, 18981903), 2:211.2224 (xotopporo0oi tivi
aopo tov oyov); Ptolemys Almagest, trans. and annotated by G. J. Toomer (Iondon:
Duckworth, 1984), 422; Regis Morelon, Ia version arabe du Iivre des Hypotheses de
Ptolemee,` Institut Dominicain dEtudes Orientales du Caire. Melanges (MIDEO), 21
(1993): 785 |Bk I|.
123. Ibid. 38.2II.
124. Ibid. 41.ult.42.8.
125. See the detailed and penetrating discussion oI these passages by A. I. Sabra, 'Con-
fguring the universe: aporetic, problem solving, and kinematic modeling as themes
oI Arabic astronomy, Perspectives on Science, 6:1998 (Cambridge, MA: MIT, 1999):
288330, esp. pp. 298305: 'Ibn al-Haytham and the aporetic argument: Ptolemy`s
dilemma, preceded by a review oI Ibn Haytham`s Maqla j Hayat al-lam, ed. and
trans. by Y. Tzvi Iangermann, Ibn al-Haythams On the Conhguration of the World,
Harvard Dissertations in the History oI Science (NewYork: Garland, 1990).
126. A. I. Sabra, 'Confguring the universe |see preceding note|, 304.
128. Ibn al-Haytham, Shukk, 38.5.
129. Ibn Bjja, Kalm baatha bih li-Ab 1ajar suj ibn Hasdy, in Rasil falsahy-
ya li-/bn Bjja, ed. Jaml-al-Dn al-Alaw (Beirut, Casablanca, 1983), 78. On the
5th/11th century Andalusian astronomer (Ibn) al-Zarqlluh, v. J. Vernet in Dictionary of
Scientihc Biography, XIV:592 s.n.
130. Comm. med. Meteor., b. III latine, ex hebraico Kalonymi> (Venetiis: apud Iunc-
tas, 1562), t. IV: Iol. 448v I. The Iatin diIIers Irom the Arabic, Ibn Rushd, Talkhs al-
thr al-ulwiyya, ed. Jaml-al-Dn al-Alaw (Bayrt: Dr al-Gharb al-Islm, 1994), p.
141.1012, but this is corrupted, being in evident contradiction to p. 145.110 (cI. the
comments oI the editor, ibid., p. 144, note 4, and p. 145, note 4).
131. This was pointed out by A. I. Sabra, 1he Optics of Ibn al-Haytham, books I-III,
2: Commentary (Iondon: Warburg Institute, 1989), p. 6, quoting the medival Iatin
translation oI Averroes` commentary. The reIerence in Ibn al-Haytham, as quoted by
Sabra, ibid., p. 4, is his treatise On the Rainbow and the Halo (completed in Rajab,
419/1028), as reIerred to by Kaml-al-Dn al-Eris, Tanqh al-Manzir (Haydarbd:
Dirat al-MariI al-Uthmniyya, 134748/192830), II, 258279. The text identifed
by the editor, Jaml-al-Dn al-Alaw, as being a possible source oI Ibn Rushd`s reIer-
ence (Maqlat al-Hasan b. al-Hasan /bn al-Haytham j l-Aar al-zhir j wajh al-qamar,
ed. Abd-al-Hamd Sabra, Journal of the History of Arabic Science, 1 |Aleppo, 1977|:
519) treats a diIIerent topic, and does not contain specifc statements on method.
132. Translated by A. I. Sabra, op. cit., 4, Irom the quotation Iound in Kaml-al-Dn
al-Eris, Tanqh al-Manzir, ed. MustaI Hijz, II, p. 259, 78.
133. Ibn Rushd, Talkhs al-thr al-ulwiyya, ed. Alaw, p. 145.110 (l. 45: wa-laysa
hlu ilmi l-manziri maa |sic leg. pro al-munzirn| hdh l-ilmi f iti hdihi l-
asbbi ka-hli ilmi l-manziri maa ilmi l-handasa). Eor Ibn al-Haytham`s conception
oI the causes determined in mathematical demonstrations, cI. the text quoted above,
note 107).
134. Talkhs al-thr al-ulwiyya, p. 144.
135. Talkhs al-thr al-ulwiyya, p. 145.ult.146.6.
136. J. supra, n. 78.
137. Epitome of Platos Republic and its Scientihc Arguments |written, perhaps, in the
same period as the Middle Commentary on the Nicomachean Ethics, fnished in 572
H.|, Hebrew text ed. Erwin I. J. Rosenthal, Averroes Commentary on Platos Republic,
Cambridge Oriental Publications, 1 (Cambridge: Cambridge University Press, 1956;
repr. 1966, 1969), b. II, ch. XV: 75.1635; trans. Ralph Ierner, Averroes on Platos
Republic (Ithaca: Cornell, 1974), 96I.
138. CI. Aristotle, Phys. II.2, quoted supra, note 77. On the diIIerent 'causes deter-
mined in physics and astronomy, see the passage Irom Averroes` commentary on De
Caelo, quoted infra, n. 161.
139. al-Bitrj, De motibus celorum. critical edition oI the Iatin translation by Michael
Scot, ed. by Erancis J. Carmody (Berkeley and Ios Angeles, 1952); id.: On the Prin-
ciples of Astronomy, an ed. oI the Arabic and Hebrew versions with translation, analysis,
and an Arabic-Hebrew-English glossary by Bernard R. Goldstein (New Haven, 1971).
A superb analysis oI the Andalusian 'revolt against Ptolemyor should we say 'resto-
ration oI Aristotle?in planetary theory has been given by Abdalhamd I. Sabra in his
article 'The Andalusian revolt against Ptolemaic astronomy: Averroes and al-Bitrj,
in 1ransformation and 1radition in the Sciences. essays in honor of I. Bernard Cohen,
ed. by E. Mendelsohn (Cambridge |etc.|: Cambridge University Press, 1984), 133153.
The Iollowing synopsis oI Ibn Rushd`s positions is expounded in more detail in my
article 'Averroes De Caelo: Ibn Rushd`s cosmology in his commentaries on Aristotle`s
On the Heavens, Arabic Sciences and Philosophy, 5 (1995): 141.
140. Kalm li-Ab Bakr ibn ahy j l-haya, MS. Berlin: Staatsbibliothek, Wetzstein,
87 (Ahlwardt, Jer:eichni, no. 5060; now at Cracow, Biblioteka Jagiello nska), Iol.
203a204b.
141. Ibn Rushd, K. M bad al-taba, IV 13, ed. Quiros, 133 ed. Amn, 130; see also
below, 21.
142. Eully developed in Metaph., A 8.
143. De Caelo II.6.
144. De Caelo II.12, 292a21.
145. See Ch. Genequand, Ibn Rushds Metaphysics. a translation, with introduction, of
Ibn Rushds commentary on Aristotles Metaphysics, book Lm (Ieiden, 1984), 54I.
146. Ibn Rushd, 1awmi al-Sam wa-l-lam (Haydarbd, 1365/1946), p. 47.
147. See also Henri Hugonnard-Roche: 'Remarques sur lèvolution doctrinale
dÀverroes dans les commentaires au De caelo: le probleme du mouvement de la terre,
Melanges de la Casa de Jela:que: 13 (Madrid, 1977): pp. 103117 (pp. 105108: I.
Que la terre est au centre; pp. 108111: II. Que le centre de la terre est le centre du
monde; pp. 111113: III. Cause du mouvement et du repos de la terre; pp. 113115: IV.
Sur la theorie du equilibre par indiIIerence).
148. Willy Hartner, OriensOccidens (Hildesheim, 1968), 461.
149. Ibn Rushd, Talkhs al-Sam wa-l-lam, ed. Jaml-al-Dn al-Alaw (Ees, 1984),
272.515 on De Caelo II 14.
150. On the development oI Averroes, and on the indications given with regard to this
primary scope oI his early work, v. Jaml-al-Dn al-Alaw, al-Matn al-Rushd (Casa-
blanca: Tbql, 1986), 49II., 140143, 159I.; G. Endress, 'Ie projet dÀverroes, in
Averroes and the Aristotelian Tradition, ed. by G. Endress and J. A. Aertsen, in the
series 'Islamic Philosophy, Theology and Science, vol. 31 (Ieiden: Brill, 1999), 331;
id., 'Law ansaa Llhu j l-umr: the project oI Ibn Rushd, in La Obra de Averroes. el
pensamiento hlosohco y cientihco de Averroes en su tiempo, Proceedings, VIII Cente-
nario de Averroes, Cordoba, 1998 (Iorthcoming).
151. Juliane Iay, 'LAbrege de lAlmageste: un inedit dÀverroes en version hebraque,
Arabic Sciences and Philosophy, 6 (1996): 2361, p. 26. Only the Hebrew version oI
Ibn Rushd`s Mukhtasar al-Majist (Qissr al-Magist) is extant.
152. Ibid., 42I., 53.
153. Ibid., p. 54I.; p. 54 n. 87 on the terminology: demonstratio per signum (demonstra-
tio quia) and demonstratio propter quid; see also p. 154.
154. E. g., Comm mag. De Caelo II c. 49, Iol. 62ra2326: et si ponantur centra diversa,
accidit impossibile quod diximus, et hoc quod faciunt mathematici, quod ponunt eccen-
tricos, numquam dixit hoc Aristoteles, sed causa diversitatis apud ipsum sunt motus
leulab. On lawlab, see p. 156 and note 178.
155. They have been dealt with in some detail by E. E. Carmody |ed.|, Al-Bitrj, De
motibus celorum |see above, n. 139|; idem, Innovations in Averroes De Caelo (Berke-
ley, Cal., 1982). See also the article oI A. I. Sabra, reIerred to above, n. 139.
156. Ibn Rushd, 1awmi al-Sam wa-l-lam (Haydarbd 1365/1946), p. 42.410.
157. CI. also 1awmi al-Sam wa-l-lam, p. 10: sa-nubayyin hdh j l-jalsaja al-
l |. . .|, hhun innam huwa al jihat al-musdara al m tabayyana j kitb al-
Burhn. Eor a systematic presentation oI the epistemological principles underlying this
procedurewhich go back ultimately to Aristotle`s Analytica posteriora and al-Erb`s
Kitb al-Burhnsee Ibn Rushd, Sharh al-Burhn, ed. Badaw (al-Kuwayt, 1984),
348II., on Analytica Posteriora, I, 13.
158. Eor the Arabic terminology see Ibn Rushd: Tajsr M bad al-taba, ed. Bouyges
|infra, n. 171|, 703.11; on barhn mutlaqa cI. Ibn Rushd, Sharh al-Burhn, l.c. |see
preceding note|.
159. Eor the Iatin text, see Averroes, Commentarium Magnum in Aristotelis De Caelo,
in Aristotelis Opera cum Averrois Cordubensis Commentariis (Venice: apud Junctas,
1550), vol. 5: b. II, c. 61, Iol. 65v3442. Eor the reIerence to De Anima, cI. Commen-
tarium Magnum in Aristotelis De Anima Libros, ed. E. S. CrawIord (Cambridge 1953),
I, c. 89, p. 119, and II, c. 20, p. 159.
160. CI. Jaml-al-Dn al-Alaw: 'Ishkl al-alqa bayn al-ilm al-tab wa-M bad al-
taba ind Ibn Rushd, in: Majallat Kulliyyat al-db wa-l-ulm al-insniyya; adad
khss 3 (Ees: Jmiat Sd Muhammad b. Abdallh, 1988), 751.
161. Commentarium magnum in De Caelo |Iatin version|, c. 57 ad De Caelo II.10,
291a29-b11 (text according to the unpublished ed. oI E. E. Carmody, corresponding to
Aristotelis opera cumAverrois commentariis |Venetiis: apud Junctas, 1562, repr. Erank-
Iurt am Main.: Minerva, 1962|, vol. V, I. 136rEE: naturalis vero et astrologus com-
municant in consideratione istarum quaestionum, tamen astrologus in maiori parte dat
quia, naturalis vero dat propter quid, et quod dat astrologus in maiori parte non est nisi
eis quae sensui apparent . . ., naturalis autem laborat in dando causam propter quam hoc
est supra ipsam . . . . Astrologus enim considerat de causis abstractis sermone idest rati-
one a materia scilicet causis doctrinalibus |Arabic, talmiyya 'mathematical causes|,
Naturalis vero de causis que sunt cum materia; verbi gratia in utraque scientia queritur
quare celum est sphericum: Naturalis dicit quia est corpus neque grave neque leve,
Astrologus dicit quia linee exeuntes a centro ad circumIerentiam sunt equales.Eor the
principles oI method observed by the philosopher, cI. al-Erb, Kitb al-Burhn, ed.
M. Eakhr (Beirut, 1987), p. 66.1317. Consider also Aristotle, Analytica priora, I 30,
46a1722: 'It Ialls to experience to provide the principles oI any subject. In astronomy,
Ior instance, it was astronomical experience that provided the principles oI the science,
Ior it was only when the phainomena were adequately grasped that the prooIs in astron-
omy were discovered, and see G. E. I. Owen, 'Tithenai ta Phainomena, Aristote et
les problemes de methode, ed. par S. Mansion (Iouvain, 1961), 83103.
162. See Elders, Aristotles cosmology, p. 227I.; Moraux |ed.|, Aristote, Du ciel, introd.,
p. civ I.
163. Ptolemy, Planetary Hypotheses, ed. B. Goldstein, p. 31 Iol. 90a36.
164. Ibid. 2729 Iol. 88b2490a6; Ior the Arabic tradition, see Ibn Sn, al-Shij,
|III, 4|: /lm al-haya, edd. Muhammad Mudawwar, Imm Ibrhm Ahmad (Cairo,
1980), p. 463 (Iollowing Ptolemy); also E. J. Carmody |ed.| Al-Bitrj, De motibus celo-
rum, introd. p. 62I., Iatin text p. 127II. (Venus above the sun, Mercury below).
165. Comm. mag. in De Caelo, II, c. 58, Iol. 64r4548, cI. 64va30.
166. Ibid., Iol. 64va4859. We may compare this with Alexander, apud Simplicium: In
De Caelo, p. 472.10II.: The planetary sphere is not moved unwillingly, but in accordance
with its purpose and desire; it may be necessity, but is also recognized as good.
167. Ibid., II c. 62, Iol. 66ra57.
168. Ibid., Iol. 66ra559 ad De Caelo II 12.
169. Tajsr al-Sam wa-l-lam, II c. 35, Iol. 38b2239a4.
170. See Charles Genequand, op. cit., supra, n. 145, pp. 54I.
171. Averroes, Tajsr M bad al-tabat, ed. Maurice Bouyges, Bibliotheca Arabica
Scholasticorum, serie arabe, VVII (Beyrouth: Imprimerie catholique, 193852), vol.
3:1639II. (comm. 42II.); trans. Ch. Genequand, Ibn Rushds Metaphysics (v. n. 145),
168II.On the problem, v. W. Ross, Aristotles Metaphysics (OxIord, 1924), II: 382I.)
172. Tajsr M bad al-taba, 1654.36 ed. Bouyges.
173. Ibid., 1655.1113.
174. Ibid., 1655.151656.1.
175. Ibid., 1656.34.
176. Ibid., 1656.813.
177. Ibid., 1661.81662.7.
178. Tajsr M bad al-taba XII c. 45 (on Metaph. XII.8), ed. Bouyges p. 16621664
(cI. above, note 2); trans. Ch. Genequand, Ibn Rushds Metaphysics, p. 178I. (correct
178.6 'Ptolemy was Iree Irom : 'P. Iailed to notice |1663.3 dhahaba alayhi|).The
word in question is lawlab, meaning circular, and more especially heliocoidal or spiral
motion. The underlying Greek word, however, is ovrittouoo 'backward-rolling`. In
Metaph. A 8 (1074a2), putting the purely geometrical account oI Eudoxus and Callip-
pus into a mechanical model, Aristotle postulated Ior each planet except the moon the
existence oI certain spheres which 'roll back the outer sphere oI the planet, preventing
the deIerent sphere oI one planet Irom aIIecting the next. Now the Arabic translation
Averroes had beIore him translated ooipo= ovrittouoo= by ukarun tadru bi-
dawrin lawlab (Tajsr, p. 1669.7, 1670.45). This gave him a quite diIIerent idea oI the
model Aristotle had in mind (cI. the passage Irom his commentary on De Caelo, quoted
above, p. 245)and deluded him with Ialse hopes. CI. Ch. Genequand, Ibn Rushds
Metaphysics, 55; W. D. Ross, Aristotles Metaphysics, II: 391I.; Th. Heath, Mathematics
in Aristotle, 218.
179. The passage has been discussed, and compared with relevant statements in Mai-
monides` Mishneh 1orah, by Y. Tzvi Iangermann, 'The True Perplexity: the Guide of
the Perplexed, part II, chapter 24, in Perspectives on Maimonides. philosophical and
historical studies, ed. by Joel I. Kraemer, The Iittman Iibrary oI Jewish Civilization
(OxIord: OxIord University Press, Ior the Iittman Iibrary, 1991), 159174.
180. Maimonides, Dallat al-hirn, book II, ch. 24, ed. S. Munk, 51b54b ed.
S. Munk, Y. Yoel (Jerusalem: Junovitch, 1931), 225229; ed. Hseyin Atay (Ankara:
Ankara niversitesi, 1974), 349353; trans. S. Pines, 1he Guide of the Perplexed (Chi-
cago: The University oI Chicago Press, 1963), 322327.
181. CI. G. Endress, 'Philosophische Ein-Band-Bibliotheken aus IsIahan, Oriens, 36
(2001): 1058.
182. This is treated, in the context oI our present subject, by A. I. Sabra in his ground-
breaking article, 'Science and philosophy in medieval Islamic theology: the evidence
oI the Iourteenth century, Zeitschrift fur Geschichte der Arabisch-Islamischen Wissen-
schaften, 9 (1994): 142.
183. Nasr al-Dn al-Ts's Memoir on Astronomy (al-Tadhkira j ilm al-haya), ed.,
trans. E. J. Ragep, in Sources in the History oI Mathematics and Physical Sciences, 12
(NewYork |etc.|: Springer, 1993), 1:90/91.
184. J. supra, p. 167 (notes 81 and 82).
185. Al-j, al-Mawqij j ilm al-kalm, quoted IromA. I. Sabra, 'Science and philoso-
phy in medieval Islamic theology, supra, n. 182, 37.
186. Eirst presented, and put into context, by E. Jamil Ragep, 'Ereeing astronomy Irom
philosophy: an aspect oI Islamic infuence on science, Osiris 2nd series, 16 (2001):
4971. I am grateIul to Jamil Ragep Ior letting me read his article beIore publication.
187. On hads, v. D. Gutas, Avicenna |as in n. 93|, 159177.
188. Quoted Irom the excerpts edited and translated by Ragep, l.c., 68, 70.
189. Including the Earth`s rotation; v. Ragep, ibid.
Grrn:ri Lxirrss 176
190. al-Brn, al-Qnn al-Masd (Hyderabad 19541956), I, p. 27, translated by E.
J. Ragep in Nasr al-Dn al-Ts's Memoir on Astronomy, 1:40 (see above n. 183), see
p. 260 and p. 302 note 144 in this work Ior the reIerence to Aristotle`s De Caelo, II.12.
I Ix+roiuc+iox
Ab Sahl Wjan ibn Rustam al-Kh
1
was a mathematician Irom Tabaristan
2
who fourished in the second halI oI the tenth century C.E. and whose work
was well-known among the mathematicians oI his age working in the Byid
domains. He had as patrons at least three kings oI the Byid Dynasty: Adud
al-Dawlah, Samsm al-Dawlah, and SharaI al-Dawlah, whose combined reigns
cover the period 962989. In the times immediately Iollowing his, Ibn al-Hay-
tham and al-Brn knew oI several oI his works, and Umar al-Khayym cites
him as one oI the 'distinguished mathematicians oI Iraq.
3
Iater, some oI his
works were studied by Muhammad ibn Sirtq in the frst halI oI the Iourteenth
century, and the eighteenth century Egyptian scholar MustaI Sidq not only
copied several oI his works but was suIfciently interested in his treatise on the
volume oI the paraboloid to make a shorter version oI it.
Al-Kh had, however, not only undoubted ability but the good Iortune to
be a geometer in a century that was one oI the most active periods oI geometri-
cal research in medieval Islam. The tenth century included most oI the working
lives oI geometers such as Ibrhm ibn Sinn (d. 946) Ahmad al-Saghn, Ab
Sad al-Al ibn Sahl, Ahmad al-Sijz (f. 970), Ab Nasr ibn Irq and Abu`l-
WaI al-Bzjn, geometers whose works variously cite,
4
complement, and
contrast with those oI Ab Sahl. (We know that al-Kh was in personal contact
with at least three oI these, who worked with him on observations oI the sun,
which he supervised, during the reign oI SharaI al-Dawlah in 988. And he was
in correspondence with at least one more.) In view oI this rich intellectual
working environment we are Iortunate in having available not only some thirty-
two works by the great Byid geometer, as well as excerpts and quotations
Irom not-yet-discovered works oI his, but letters that he wrote and preIaces to
his works, in which he makes a number oI comments on the mathematics oI
his time. Taken together these provide a view oI the work oI the better part oI a
century through the eyes oI one oI its major fgures, and it is some elements oI
this view that I want to sketch out in this chapter.
6
1rx+n-Crx+ur. M:+nrx:+ics +nroun +nr L.rs or
Aiu S:ni :i-Kuni
. Iennart Berggren
. Irxx:r+ Brrrrx 178 1rx+n-Crx+ur. M:+nrx:+ics +nroun +nr L.rs or Aiu S:ni :i-Kuni 179
We have already spoken oI Ab Sahl as a geometer, rather than a math-
ematician generally, and something that stands out as one reads his treatises
is that all oI them are devoted to geometry. Indeed, al-Kh was known as
'Master oI his age in the art oI geometry by his two younger contemporaries
Abu`l-Jd and al-Shann.
5
In the preIace to his treatise on the regular heptagon
that is preserved in MS. Cairo MR 40 Ab Sahl writes oI the beneft oI geom-
etry to the mathematical sciences and describes it as 'the example which ought
to be imitated in the pursuit oI truth (in the theoretical sciences) and the leader
which is to be Iollowed when it comes to honesty.
Perhaps the reason Ior his devotion to geometry lay in what he saw as
the surety oI the knowledge gained Irom it. In the preIace to his treatise on the
regular heptagon just cited, he writes, 'Its (geometry`s) Ioundation is frm and
its rules are consistent and unchanging. It can be aIIected neither by reIuta-
tion nor can it be aIficted by infrmity. In contrast to this certainty, he writes
in his treatise on the distance to the shooting stars oI Galen`s account oI the
variety oI explanations oIIered Ior shooting stars, and the controversies these
opinions occasioned. He then says, 'But the other group, whom neither Galen
nor anyone else could criticize . . . because they depended on prooIs . . . were
the mathematicians, and he strikes a similar note in his correspondence with
al-Sb where he contrasts 'opinion in physics with 'demonstrative science.
It was, however, geometry in a wide sense that interested him, not only
the works oI Euclid, Apollonius, and Archimedes. Eor example in his treatise
on the possibility oI infnite motion in a fnite amount oI time he studies the
movement oI the shadow oI a sundial`s gnomon, and in the aIorementioned On
Shooting Stars he is interested in fnding the distance to and size oI the shoot-
ing stars. Moreover, he tells us in his treatise on the computation oI rising times
that he has investigated astronomy as well as centers oI gravity and optics,
and his treatise on the complete compass was, as he himselI notes, intended
to describe a mathematical instrument useIul Ior drawing conic sections on
sundials and astrolabes.
Ai-Kuni :xi Proiirxs rrox Hriirxis+ic Groxr+r.
However, these latter treatises are details against an overwhelming background
oI the rich geometrical heritage Irom the Hellenistic world, and the intellectual
presuppositions oI that heritage, as Iound in and implied by the works oI Euclid,
Archimedes and Apollonius. Indeed, no Iewer than fIteen oI his works deal
directly with problems discussed by these three mathematicians and many oI the
others are either very much in the spirit oI their works generally or are specifc
extensions oI classical problems. Two examples oI such extensions are:
1. Al-Kh`s problem oI constructing circles 'tangent to two given objects (any
one oI which may be a point, a straight line, or circle)
6
and having its center on a
line given in position.
7
This is just Apollonius`s problem on constructing a circle
'tangent to three given objects with the three objects replaced by two but the
requirement added that the center oI the circle to be constructed must lie on a given
(possibly curved) line (as in fgure 6.1 above).
2. Al-Kh`s extension oI two problems Archimedes solves in On the Sphere and
Cylinder II, namely:
a) To construct a segment oI a sphere equal in volume to a segment oI a given
sphere and similar to another segment oI that sphere, and
b) Do the same as in (a) with the word 'surIace substituted Ior 'volume,
Al-Kh then solved the Iollowing problem:
c) To construct a sphere having a segment equal in surIace area to a given
segment oI one sphere and in volume to a given segment oI another sphere.
Other examples are his extensions (not all successIul, as indicated
below) oI Archimedes` results on centers oI gravity oI fgures having an axis
oI symmetry.
OI course it is possible to fnd other tenth-century geometrical works
that might be described as extending Greek problems. Eor example, the work
oI Abu`l-WaI on inscribing and circumscribing fgures with some symmetry
properties in other such fgures comes to mind, a problem also treated by Ab
Kmil and al-Kh,
8
but al-Kh seemed to be unique in fnding and (generally)
solving problems oI some real depth.
Eigure 6.1
However, other Islamic geometers Irequently made creative contribu-
tions in extending the Greek tradition, as Hogendijk (1984) points out in his
survey oI Islamic work on the regular heptagon. Although this work peaked in
the latter halI oI the tenth century, when it involved a halI-dozen well-known
geometers,
9
it also stimulated fve solutions oI the problem by Ibn al-Haytham,
whose most important work Iell in the eleventh century, and even work by
Kamal al-Dn ibn Ynus at least 200 years later.
The problem oI constructing the side oI a regular heptagon in a given
circle seems to have arisen because Abu`l-Jd, whose fattering opinion oI al-
Kh we quoted above, criticized the solution Iound in the Arabic treatise titled
Book of the Construction of the Circle Divided into Seven Equal Parts. In its
incipit the treatise is said to have been translated Irom an original oI Archime-
des by Thbit ibn Qurra. However there are a number oI perplexing Ieatures oI
the treatise, not the least oI which are the Iacts that oI its eighteen propositions
only two concern the subject announced in its title, the regular heptagon, and
the Iact that in these two propositions the author used a particularly vexing
verging construction to divide a line segment into three segments satisIying
certain proportions. These proportions, in turn, guaranteed that the resulting
segments would, as sides oI a triangle, produce a regular heptagon in the cir-
cumscribed circle.
10
Almost as damning as Abu`l-Jd, who described the solution given in
the Construction of the Circle Divided into Seven Equal Parts as something
that appeared more diIfcult than the original problem, al-Kh wrote in the
introduction to his frst treatise on the heptagon (Iound in the Cairo ms. MR40)
that 'it appeared Irom the book Archimedes authored on the subjecta subtle
bookthat he did not Iulfll his aim and achieve his desire in his solution in
any way. Some, al-Kh continued, believed that Archimedes` treatment was
incomplete because it was impossible Ior him to do more, others because he
flled-in the gap in a treatise that had not come down to the Arabic writers.
(Abu`l-Jd went Iurther and suggested that Archimedes had made a mistake,
a suggestion tantamount to heresy among tenth-century geometers, and on Ior
which al-Sijz roundly criticized him.)
The question was, however, what to do next. Some heldaccording to
al-Khthat 'there was no way in which and no method by which one could
fnd
11
what Archimedes had Iailed to fnd. Others believed it could be done,
and it appears that both al-Saghn and al-Kh solved the problem at close to
the same time. (Hogendijk (1984) argues that al-Saghn`s solution was earlier,
and this may very well be the case since his solution bears clear traces oI the
verging Iound in the Construction of the Circle Divided into Seven Equal Parts,
whereas al-Kh`s solution removed all trace oI the verging and divided the line
segment directly by means oI conic sections.)
Two other interesting Ieatures oI tenth-century geometry that emerge
Irom the various works on the heptagon are:
1. The importance oI the patronage oI the Byid kings in supporting geometrical
research, and, something that A. Anbouba has pointed out,
2. The role correspondence played in the development oI mathematics at that time.
12
As Ior the importance oI patronage, the initial satisIactory solutions oI
the heptagon problem, that oI Ab Sahl as well as that oI al-Saghn,
13
were
both dedicated to the king Adud al-Daula, and Ab Sahl wrote another one
dedicated to his son, Prince Abu`l-Eawris, who later took the title SharaI al-
Dawla. Moreover, his 1reatise on the Ratio Between 1hree Lines is dedicated
to SharaI al-Dawla, and the introduction to his treatise on rising times makes it
clear it was written in the context oI activity at the Royal Palace.
As Ior the matter oI correspondence, one is struck by how much oI what
has survived just on the topic oI the heptagon is the subject oI letters, Ior exam-
ple: Ietters by Abu`l-Jd ibn al-Iayth to one al-Hsib (in Bukhara) in which
the Iormer describes the methods oI al-Saghn and al-Kh as well as his own
method, letters to Abu`l-Hasan al-Ghd, and to al-Brn, as well as correspon
dence between al-Sijz and Ab Sad al-Al ibn Sahl. The topic is, in addition,
one oI those treated in the letter oI al-Sijz to 'the people oI Khorasan.
Al-Sijz was particularly active as correspondent. Twenty-two oI the
thirty-fve treatises listed under his name in Sezgin (1974) are letters (or
'answers), and in several the correspondent is named.
Al-Kh himselI participated in the scientifc correspondence oI his
epoch, as is witnessed by the partially extant correspondence he had with al-
Sb. This seems to have had the beneft Ior al-Kh oI providing him a Iorum
in which to try out ideas, and evidently saved him Irom the embarrassment oI
publishing his alleged discovery on the center oI gravity oI a semicircle and its
consequence that the value oI a was 3
1
9
.
This topic, centers oI gravity, is another example oI one which interested
the Greeks, and in particular Archimedes, and also elicited interest Irom some
oI the tenth-century geometers, including al-Kh, although the level oI inter-
est hardly rivaled that surrounding the regular heptagon. Thus al-Kh gives us
the precious inIormation that he possessed a book attributed to Archimedes`
'On Centers oI Gravity, but one that did not contain any prooI oI the law
oI the lever, since he reIers to it as a 'premiss Ior Archimedes` treatment oI
barycentric questions. We also learn that Ab Sad al-Al ibn Sahl wrote on
centers oI gravity and also inquired on the status oI the law oI the lever. Einally,
it appears Irom al-Kh`s treatise on the volume oI a segment oI a paraboloid
that his work on centers oI gravity motivated his work on that question in men-
surational geometry.
Two versions oI his treatise on the volume oI a segment oI a paraboloid
(a third topic in which the tenth-century geometers knew they were Iollowing
in Archimedes` Iootsteps) have come down to us, and it is interesting to note
al-Kh`s almost apologetic tone in his introduction to the frst oI these. He
recognizes that Thbit wrote on the subject, but he said that he is working on
it because he needs the result Ior work he is doing on centers oI gravity and
because he Ieels Thbit`s arguments are not easily understood. One wonders iI
the somewhat deIensive tone might be because Ibrhm ibn Sinn had not taken
well to the suggestion that the work oI his grandIather, Thbit, on the parabola
was less than clear.
Iike his frst treatise on the heptagon, al-Kh`s treatise on the measure-
ment oI the paraboloid also attracted serious study by the Iollowing generation
oI mathematicians. Thus, Ibn al-Haytham says in his Discourse on the Mea-
surement of the Paraboloid that he knew al-Kh`s work on the subject.
14
Einally, to close this discussion on centers oI gravity, we remark that
it appears Irom al-Khzin`s 12th-century treatise, 1he Balance of Wisdom,
that Ibn al-Haytham also wrote on centers oI gravity, Ior al-Khzin summa-
rizes both his and al-Kh`s treatises on this matter in Chapter 1 oI the Eirst
Discourse oI that work.
15
However, al-Kh`s treatise has not survived and is
known only because al-Khzin summarized it together with that oI Ibn al-
Haytham, which has also not survived. This suggests that the topic was not
regarded as one oI major interest at the time, despite its impeccable credentials
oI being among the topics that interested Archimedes, and it did not create a
lasting tradition. It seems that this was one aspect oI the Archimedean tradition
that did not fourish in the tenth century.
Ai-Kuni :xi Mr+nois or Hriirxis+ic Groxr+r.
It was, however, not just the problems oI Hellenistic mathematics that attracted
the geometers oI the tenth century but its methods as well. Thus, the method oI
analysis and synthesis was standard in the tenth century and one that al-Kh
and his contemporaries used Ireely. As a result oI this activity at least one trea-
tise on the method was written during that century, namely that oI Ibrhm ibn
Sinn, and one in the Iollowing century, namely that oI Ibn al-Haytham.
16
One subject that appears to have occasioned some dispute during the
tenth century was that oI the relation oI analysis to synthesis. According to
the standard, simple account oI the relationship between the two, as given,
Ior example, by al-Sijz in his 1reatise on Geometrical Problem Solving
(Hogendijk 1996, p. 4), the synthesis was just the analysis done backward, so
the steps oI the two should be in virtual one-to-one correspondence. However
that is not what one fnds, and Ibrhm ibn Sinn recorded the dismay oI some
geometers on seeing geometrical objects appear in the synthesis that did not
occur in the analysis.
Ibn Snn reIerred to this phenomenon in the Iollowing passage
17
I have Iound that modern geometers have neglected the method oI Apollonius in
analysis and syntheses, as they have in most oI the things I have brought Iorward,
and that they have limited themselves to analysis alone in so restrictive a manner
that they have led people to believe that this analysis did not correspond to the
synthesis eIIected.
Ibrhm`s answer was that the lack oI correspondence was only superf-
cial, and that the reason it seemed so was that the analysis stated the steps in an
abbreviated version but one which was perIectly clear to the experts.
As an example oI this consider al-Kh`s analysis oI a problem Irom
Drawing two lines from a known point.
18
The problem is the Iollowing (fgure
6.2 illustrates the analysis and subsequent synthesis Ior this problem):
A point A and a line BG are given. Draw two line segments Irom A to the line,
containing a given angle, so that the two segments AB and AG have to each other
a given ratio.
Analysis: Assume the line segments AB and AG are drawn. (i) Since
AB:AG is known and ZBAG is known, A(ABG) is known in Iorm. (ii) There-
Iore ZABG is known. (iii) But point A is known, hence line AB is known in
position. (iv) And since ZBAG is known, line AG is also known in position. (v)
Since line BG is known in position, then, points B and G are both known.
In Iact no synthesis is given Ior this analysis, but it is not hard to supply
one, as Iollows:
Eigure 6.2
A
D B G L N M
P
O
K
Synthesis: (i) Iet the given ratio be the ratio oI the lines P and Q. Draw P and Q
(as IK, KM) so that the angle they contain (ZIKM) is the given angle. Draw
IM. (ii) Drop a perpendicular Irom K to IM, defning N. Also drop a perpen-
dicular Irom A to the given line, defning D. (iii) Draw DB so that AD:DB
KN:IN. Draw AB. (iv) Draw AG so that ZBAG is the given angle. Claim:
AB:AGP:Q. (v) A(ABD) - A(KIN), because both have a right angle and the
same ratio oI sides containing that angle. (vi) Hence ZABG ZKIM; but
ZBAG ZIKM the given angle, so A(ABG) is similar to A(KIM). (vii)
ThereIore AB:AG KI:KM P:Q.
Such a synthesis, using what we have elsewhere
19
called a scale model oI
the solution, al-Kh has also given Ior another problem, and it is exactly the
kind oI synthesis Ibrhm was talking about when he wrote:
As Ior the Iact that they draw lines which were not drawn in analysis, it does not
establish in any way any diIIerence between analysis and synthesis; Ior example,
when their analysis ends up with the Iact that some triangle is known in Iorm,
because its angles are known, though in the analysis it is not drawn on a line
known in magnitude, they do nevertheless deduce Irom the ratios oI its sides,
one to another, a thing by which the problem gets solved. Can they avoid then,
in their synthesis, setting up a triangle whose angles are equal to these known
angles, in order to come to know the ratios oI its sides, and make Irom it what
gets the problem solved?
Thus auxiliary diagrams such as KLM above are entirely appropriate in
syntheses, and do not impair the similarity between analysis and synthesis, the
Iormer being, in Ibrhm`s word, an abridgement oI the latter.
Al-Kh`s treatise that we reIerred to above is unusual in that it considers
about a dozen problems (variations on a single theme, admittedly) entirely by
the method oI analysis, where the analyses Iorm a series oI successive reduc-
tions, in the sense that there are analyses Ior the initial problems in the two
cases when the given line BG is a straight line or a circular arc, and, aIter that,
the analyses are reductions oI the problem under discussion to the problem just
analyzed. By the time one has about halI a dozen analyses, each resting on the
previous one, it becomes something oI a puzzle to imagine what the synthesis
oI the last problem would look like.
On the other hand, it is possible to see in another treatise by al-Kh, his
Ratio from a Single Line Falling between1hree Lines, a more direct response to
the criticismthat Ibrhmascribes to unknown parties, namely that the structure
oI and diagrams Ior the synthesis should be as nearly identical as possible with
those oI the analysis. Indeed, in that treatise al-Kh ensured that the analyses
and syntheses contain virtually the same steps in every proposition. Perhaps he
had read the critique Ibrhm reIers to aIter he had written his On Drawing 1wo
Lines from a Single Point, and took the criticism seriously. (Perhaps, in Iact,
that treatise was one oI the writings that provoked the criticism.)
It is apparent Irom al-Kh`s analysis that a Iundamental concept in an
argument by analysis is the idea oI a mathematical entity (such as position,
magnitude or shape) being 'known. The word is used almost exclusively in
Arabic treatises where the Greeks would have used the word 'given, and the
basic text was Euclid`s work Givens, best known by its Iatin title Data. Al-
Kh studied the work extensively and had not only thoroughly mastered its use
but had written a treatise adding several propositions to it. (Since al-Kh`s use
Ior those propositions is not known the treatise is one oI his least interesting,
but that may not have been how he viewed the matter.)
However, whatever sense (or senses) the word given` may have had to
the Hellenistic Greeks, the corresponding Arabic term, 'known, had acquired
a variety oI diIIerent senses to the tenth-century Arabic geometers. A particu-
larly vexing concept was the concept oI a known ratio. According to Euclid`s
Data a ratio was given when it was equal to the ratio oI two given magnitudes
(a magnitude being given when it was possible to fnd its equal). (A classical
analysis problem that depends on this latter notion is the Iamous problem in
the Meno that asks Ior a triangle equal to a given triangle to be inscribed in a
given circle.)
By al-Kh`s time the idea oI a known ratio had been divided into two
separate notions: a ratio known Irom the point oI view oI quantity (nisbat
al-kamm), where the numerical measure oI the antecedent relative to the con-
sequent is known, and a ratio simply 'known, in the sense oI Euclid`s Data.
Al-Kh called this latter an 'existent ratio, and he had a long discussion with
al-Sb about these two diIIerent senses, the frst oI which he ascribes to the
algebraists and astronomers, and with which he wants nothing to do!
Al-Kh writes as Iollows to al-Sb:
Eor we do not mean by this |geometric| aspect oI 'known the amount oI a thing,
nor by 'known ratio the measure oI one oI them as compared with the other
. . . . As Ior the existent ratio, according to the sense in which we use it, how could
it not be known between the circle and the square when each oI them is known.
II two magnitudes are known then indeed the ratio oI one oI them to the other
is, in our opinion, knownas Euclid proved in the frst theorem oI the Data. I
am astonished at people who claim that the area oI a circle is not permitted to be
equal to the area oI a square, and that there is no ratio between the two oI them,
and who say (so) since the circumIerence oI the circle is curved and not oI the
same kind as the circumIerence oI the square. . . .
Iikewise, Ibn al-Haytham, reIerring to the same problem and, arguing
Ior the existence oI a square with the same area as the circle, states:
The essence oI the knowable notions does not require that one perceive them or
that they be actually produced. Rather, iI the prooI oI the possibility oI the notion
has been provided, the notion is sound, whether one has actually produced it or
not. (Quoted Irom Hogendijk 1985a, 96.)
This debate touches directly on the Iamous problems oI trisecting the
angle and squaring the circle, Ior consider the Iollowing argument Ior squar-
ing the circle:
20
A circle is given in position and magnitude, and, according to
analysis, a square with the same magnitude is postulated. The circle and square
have the known ratio oI 1:1 and the circle is known in magnitude, hence by
Data, 2 the square is known in magnitude. Then, by Data, 55 the side oI the
square is known in magnitude. A similar argument, also relying on Data, 2, just
as easily trisects the angle.
The apparent faw in the above analyses occurs in the application oI
Data, 2, which states that 'II a known magnitude has a known ratio to another
magnitude, the latter is also known in magnitude. The prooI oI this proposi-
tion involves a rearrangement oI the ratios, which in the case oI the quadrature
oI the circle would imply Iorming a ratio oI a circle to a square. Pursuing a
resolution oI these anomalies, then, Iorces us to restate Data, 2 (and some other
propositions) to Iorbid ratios between angles, and between circles and squares!
This clearly contradicts how Data, 2 and other Data propositions were used,
and puts us in the diIfcult position oI having to rewrite the statement oI a basic,
heavily utilized proposition with no apparent evidence oI ancient recognition
oI the anomaly in mathematical texts.
We believe that this conundrum provides a context Ior the remarks both
oI al-Kh and the later remarks oI Ibn al-Haytham.
All oI these issues, oI course, relate to the question oI when a problem
has been solved or, to put it diIIerently, what constitutes a valid solution to a
problem. We have seen some oI this discussion in tenth-century reaction to
Archimedes` construction oI the heptagon, and Hogendijk pointed out the
discussion that occurred about the validity oI 'moving geometry, with its
verging constructions, in comparison with that oI the fxed geometry, that is,
constructions in the Euclidean manner.
21
It appears that the adherents oI fxed
geometry won the day, but it was not a geometry limited to Euclidean methods.
Eor example, Ab Sahl writes that the chord oI 1 is known according to him
who trisects the angle,
22
and al-Kh`s trisection oI the angle uses an elegant
intersection oI a hyperbola and a circle.
23
This having been said, however, it must be added that the Hellenistic
tradition expressed by the condemnation by anonymous 'geometers as 'no
small error,
24
oI using conic sections when circles and straight lines will solve
the problem, still has a lingering eIIect on al-Kh. This is apparent Irom his
discussion in his aIorementioned treatise Ratio from a Single Line, where he
considers the Iollowing problem (fgure 6.3):
To draw a straight line Irom a known point G to cut three lines ED, EH, and BZ
known in position so that the ratio DH:HK oI one oI the two segments which are
Iormed between two oI the three lines to another segment so Iormed is known.
At the end oI the treatise, aIter he has dealt with the case shown above
when the three given lines are straight lines in general position, al-Kh extends
the problem by considering the case when two oI the given lines are straight
(and non-parallel) and the third is replaced by a curve (which al-Kh reIers
to as a non-straight line`). In this case he again, as in the trisection problem,
eIIects the solution by means oI a hyperbola, this time one passing through
the given point and whose asymptotes are two lines parallel to the two given
straight lines. The hyperbola is implicity assumed to intersect the given curve
but, since that curve has no description, any consideration oI when this might
in Iact happen is impossible.
There are two other treatises that we know oI in which al-Kh speaks oI
curves other than straight lines or conic sections, one oI them being his treatise
on the astrolabe (Berggren 1994). But he does not do anything with them in
that treatise, other than to reIer to them, since, as he says, 'this is not our aim
in this treatise.
25
A true use oI an arbitrary 'non-straight line is in his On
Centers of 1angent Circles, in which (as in his Ratio from a Single Line Falling
between 1hree Lines) al-Kh uses conics to solve, Ior the case in which one oI
the givens is an arbitrary curve, a problem he has already solved without conics
when that given is a straight line. The above evidence would seem to indicate
that he still regarded a problem solved by conics as a problem solved, but a
solution without conics was to be preIerred to one with conics, the very view
we Iound expressed by Pappus`s anonymous 'geometers.
Eigure 6.3
We noted above the amount oI geometric work in the tenth century that
involved analysis, and the question arises oI why one so oIten fnds analyses in
the Arabic texts oI the tenth century. On the one hand it could be argued that
this was done because oI the style set by Apollonius (and perhaps Archimedes
in the frst Iew propositions oI SC II), and al-Kh certainly Iollowed their lead
in several oI his treatises, Ior example, in his brilliant solution to an extension
oI a problem oI Archimedes Irom SC II (discussed on p. 179 above) and in his
construction oI an equilateral pentagon in a given square. And the example oI
Apollonius is cited by Ibrhm ibn Sinn, when he wrote, 'I have Iound that
modern geometers have neglected the method oI Apollonius in analysis and
syntheses, but in Iact Ibrhm is complaining about the modern practice oI not
giving the syntheses. And, indeed, al-Kh writes, at the end oI his treatise on
two lines Irom a known point:
And iI we were to go into partition (into cases), diorismos, synthesis and the posi-
tions oI the points by Apollonius` method in some oI his books a very big book
would be produced, but we hope to have the leisure to do it later, God willing.
And in the second book, chapter six, oI his treatise on the astrolabe he
omits all syntheses. Thus, it seems to us that at least with some authors analy-
sis alone was an accepted method oI giving prooIs and one is perhaps seeing
creative mathematicians, Iamiliar with how to develop a synthesis Irom an
analysis, saying that the synthesis can wait until they have some time. In the
meantime, the excitement oI new discoveries and solutions to new problems
must receive their attention.
Eor the tenth century was a creative time in geometry, and al-Kh em-
phasizes in a treatise on the heptagon
26
that he has solved in many ways that
which Archimedes was unable to complete in even one way. According to al-
Samawal 'the construction oI the regular heptagon in the circle (was done)
by Wayjan ibn Rustam al-Kh,
27
and al-Samawal wrote this to argue Ior his
belieI in the continued progress oI the sciences, a belieI that was shared not
only by Umar al-Khayymi and Abu`l-Qsim al-Asturlb, his predecessors in
the twelth century, but by Ab Sahl who wrote in his preIace to the treatise on
the regular heptagon, 'They (earlier geometers who Iailed to fnd the side oI
the regular heptagon) were sure, nevertheless, according to what their convic-
tion has promised them, that the science oI geometry will endure and it will
continue to grow in contrast to man`s liIe-span, which comes to an end.
O+nrr Works or :i-Kuni
We leave the domain where the great Hellenistic geometers wrote when we
come to Ab Sahl`s treatise Constructing the Astrolabe with Proofs. Here Ab
Sahl is working in what was virtually an Islamic tradition, and his work was the
subject oI a commentary by his contemporary, Ab Sad, and was also known
to al-Brn.
It is instructive to compare this work with other tenth-century works
on the astrolabe, such as that oI al-Sijz, whose exhaustive treatise on astro-
labes was a sort oI encyclopedia oI astrolabes, and those oI Habash on the
melon-Iorm astrolabe and al-Saghn on planispheric astrolabes in which the
projection oI the sphere onto the plane through its equator is not Irom a pole
oI the sphere but Irom some other point on the axis. Ab Sahl`s treatise is
distinguished Irom much oI the surviving Islamic literature on the astrolabe
written up to his time in all oI the Iollowing three Ieatures: a general treatment
oI stereographic projection, a discussion oI a considerable number oI variants
oI stereographic and non-stereographic projection, and being principally con-
cerned with an extensive body oI problems oI slight interest to the astrolabe
maker but oI considerable interest to a geometer. As regards the treatment oI
stereographic projection the reader fnds in this treatise a systematic exposition
oI the general theory oI stereographic projection developed in the context oI
Book I oI Apollonius`s Conics. Al-Kh proves the circle-preserving property
oI the projection in its most general Iorm,
28
and, in using the theory to produce
the curves on the plate oI the astrolabe, he shows that only one technique is
necessary to project any circle symmetric to a meridian provided its inclination
to the equator is known. (See Berggren 1991 and 1994.)
Thus, it was all branches oI geometry that attracted al-Kh, not only
those represented in the Elements, Conics and Sphere and Cylinder but also
problems related to the construction oI such astronomical instruments as the
astrolabe and the sundial. Thus, in addition to the Iact that his treatise on the
astrolabe is the longest oI his known works, he closes his Complete Compass
as Iollows. AIter he has shown how to use this instrument that he designed Ior
drawing lines, circles and conic sections satisIying certain given conditions
he says, 'It is now evident how to draw by means oI this compass, the inter-
sections oI conic surIaces with any one oI a variety oI surIaces, according to
a given position.
29
Consequently Irom this it will be easy Ior us to construct
astrolabes on plane or axial surIaces, and to construct sundials on any surIaces,
and in the same way all the instruments on which are Iound lines oI intersection
oI conic surIaces with an arbitrary surIace. But God knows best.
As regards al-Kh`s treatment oI projections other than that oI the
planispheric astrolabe it suIfces to say here that he recognized that either
orthographic projection or any projection Irom a point on the axis oI the sphere
onto a surIace symmetric about the axis oI the sphere
30
produces a possible
astrolabe. He also states that the projection, which al-Brn inIorms us was
invented by al-Saghn,
31
oI circles on the sphere Irom a point on the axis other
than a pole produces on the plane oI projection conic sections (other than cir-
cles), and he cites the relevant proposition Irom the Conics. Einally, he points
out the cases in which these mappings are, as we would say, not one-to-one.
It is these Ieatures oI the treatise that provoked the most extensive comment
by Ab Sad.
32
Ab Sahl`s treatise resembles those oI Habash and al-Saghn
in being about mathematical curiosities, but instead oI developing new math-
ematical projections Ab Sahl regards the ordinary planispheric astrolabe as
providing an opportunity Ior using Euclid`s Data to analyze, and his Elements
to solve, problems in geometrical construction.
Einally, the Iollowing three works deal with problems not Iound in the
Greek literature, but with techniques that are continuations oI Hellenistic tech-
niques.
33
I know oI no other treatises like them in the tenth-century geometrical
literature. They are:
1. On the distance Irom the center oI the earth to the shooting stars.
2. On the area oI the earth`s surIace visible Irom a given height.
3. On the possibility oI an infnite motion in a fnite time.
34
Al-Kh seems to have been utterly uninterested in the number-theoretic
Iacets oI the Greek tradition, exemplifed by Elements VIIIX, and the writings
oI Nicomachus and Diophantus, and in this he separates himselI Irom Abu`l-
WaI, Ab JaIar al-Khzin and the later Ibn al-Haytham. Neither did he con-
tribute to the development oI algebra, as did al-Karaj, nor to any oI the lively
traditions oI arithmetic,
35
which was a popular topic with many tenth-century
writers, and his works show no interest in the rapidly developing trigonometry
or in the numerical methods oI which there was a strong tradition going back to
antiquity. Indeed, he specifcally denies any expertise in numerical methods in
a passage in a work without any title.
36
The work begins 'Some oI our Iriends
. . . and the passage comes just aIter he has shown how by applying the Trans-
versal Theorem one may calculate successively, Ior a given degree oI the eclip-
tic, its declination, where it rises on a given horizon, its rising time both at the
given locality and at the equator, and the equation oI daylight on the day when
the sun is in that degree. He says:
Notwithstanding . . . I am not sure iI there is not among the solutions oI this
problem an easier, shorter or more expeditious |route| than the one we Iollowed,
used by the skilled mathematicians and the expert zj-makers, especially iI they
give priority to multiplication over division, or delay the one until aIter the other,
or substitute the one Ior the other. . . .
In Iact, in this treatise he spells out his interests Iairly explicitly. He states
that although he has a suIfcient knowledge oI Ptolemaic astronomy he has not
spent the time in detailed work and calculations that, alone, can lead to real
expertise. In Iact he says,
Despite the sublimity oI these astronomical investigations we do not confne
ourselves to their investigation and ignore the other sciences which scholars usu-
ally investigate, Ior example, the science oI the centers oI gravity, the science oI
optics, and the science oI the characteristics oI the Iorms oI the conic sections,
which is the most astonishing oI them all. . . . Add to that the theories oI Archi-
medes and the derivation oI the geometrical theories . . . a single science, which
stands by itselI. The obsession oI scholars with the derivation oI such things (i.e.,
the sciences mentioned above) is greater than their obsession with the derivation
oI the other sciences beside these. We, likewise, [follow in the same path] without
any boasting or insinuation.
Thus al-Kh obviously thinks oI himselI as being in the mainstream
oI mathematical investigations oI his time, and one notes, again, no mention
oI the algebraic or numerical side oI mathematics. It is possible, I think, to
explain why he does not think oI the numerical side oI mathematics as being
in the mainstream, a reason suggested by the characterization he gives oI his
mathematics as 'demonstrative, and his denouncing the mathematics Iound in
Archimedes` Measurement of a Circle as approximative and oI a type that does
Archimedes no creditso much so that he says that he believes the treatise
is spurious. Although he mentions trigonometric tables specifcally when he
compares the value oI a implicit in Ptolemy`s Almagest with the much cruder
value in Measurement of a Circle, and although he says Ptolemy uses a fner
division, it is clear that he Ieels is is only comparing various degrees oI bad
mathematicsnot the bad with the good.
Apart Irom the reIerence to optics (an interesting allusion in view oI
other known connections between Ab Sad al-Al ibn Sahl and al-Kh as
well as the latter`s infuence on Ibn al-Haytham), Ab Sahl`s account oI his
scientifc interests accords very well with what we would have deduced Irom
his existing treatises. This suggests that our list, iI not complete, is at least
representative.
Coxciuiix Prx:rks
I would also like to add something to this discussion about Ab Sahl`s char-
acter. Although aIficted by no Ialse humilityhe knew when he had done a
good piece oI work and was not reluctant to say sohe appears to have been
a man who was, Irom some points oI view, dedicated to the search Ior truth
in the sciences he investigated and who was quick to point out when he was
speaking without really expert knowledge. Eor example in his discussion oI
the complete compass he admits that he might be using terminology that is not
completely standard because there might be a treatise on the subject which he
has not seen. In his work on calculating the rising times oI arcs oI the ecliptic
he is quick to remind the reader oI his own lack oI expertise in the 'tricks oI the
trade oI the expert table-makers, and says at the end oI the treatise, 'Now iI
there is anything in this discourse which is redundant to the argument, and not
in its right place, please accept my apology. This was mentioned only because
one statement led to another.
He was also acquainted with Iailure. Eor example, Umar al-Khayym
mentions him as a one oI the mathematicians around Adud al-Dawlah who
were unable to solve the problem oI dividing 10 into two parts a and b (a > b)
so that, as we would state it, a
2
b
2
a/b 72. And there was also the incident
in which he misled himselI on the value oI a because he convinced himselI, on
the basis oI a numerical pattern, oI an incorrect result Ior the location oI the
center oI gravity oI a semicircle. (This latter incident was somewhat discredit-
able to him because oI what can only be described as his stubborn reIusal to
acknowledge that he had made a mistake that contradicted the estimate Ior a in
Archimedes` Measurement of a Circle.)
To close this discussion, I would like to raise a question, one which
is phrased specifcally in terms oI al-Kh`s work but which equally could
be asked oI that oI several oI his contemporaries in the tenth century, and
that is: What, iI anything, about al-Kh`s work refects its origins in Islamic
civilization? Or, to put it diIIerently, iI al-Kh`s work had been translated
anonymously into Iatin and stripped, as such works sometimes were, oI the
dedications and the occasional 'By Allah, what is there that might cause us
to suspect that we were dealing with a translation Irom Arabic and not Irom
Greek? Such a question could not be asked oI any zj, with its reIerences to
the coordinates oI Mecca or interest in azimuths. It could not be asked oI the
geometrical solutions to the problem oI fnding the direction oI Mecca given
by Ibn al-Haytham. Neither would one ask it oI Ibrhm ibn Sinn`s work on
sundials, nor oI Abu`l-WaI`s work on the geometry oI craItsmen.
There are, oI course, some approaches to problems that defnitely are not
Iound in the classical literature, and his use oI conic sections Ior constructing
a regular heptagon, his consideration oI arbitrary curves, and his approach to
analysis seem to be products oI his own age. One even fnds a hint oI a specif-
cally Islamic context in his interest in azimuths in his treatise on astrolabes.
But it would appear, on the whole, Irom the problems he addressed and the
methods used to address them that his works could have been written by any
suIfciently able Greek geometer aIter the times oI Archimedes and Apollo-
nius. And I conIess that I have a Ieeling on reading his works that I am reading
the works oI one who might have been considered at his time as one oI the 'old
guard. Certainly, one can hardly avoid such a Ieeling in looking at Kh`s work
on rising times, where he explicitly states that he wrote it to show that one can
do very nicely with Menelaus`s theorem what was beginning to be done in his
time by means oI the newly developing trigonometric theory. Perhaps, to adapt
Keynes`s description oI Newton, al-Kh was the last mathematician to look on
mathematics with the eyes oI the great Hellenistic geometers.
No+rs
1. This is simply one possible version oI his name, based on a number oI variants in
the manuscripts. The most common reading is, in Iact, Ab Sahl Wjan ibn Wustam al-
Qh. Some standard bio/bibliographical sources on al-Kh are Sezgin 1974, 314321;
Dold-Samplonius Dictionary of Scientihc Biography, vol. XI (1975), pp 239241.
2. According to Bayard Dodge, trans., 1he Fihrist of al-Nadim, NewYork 1970, 669.
3. Quoted Irom Sesiano 1979, 281.
4. According to Sezgin 1974, 339, Ab Nasr ibn Irq cites Kh`s Points on Lines in the
Ratio of Areas in his Al-Masil al-Handasiyya.
5. Quoted in Hogendijk 1985b, 101, n. 5.
6. The requirement that a circle be 'tangent to a point means that it passes through the
point.
7. This is No. 9 in Sezgin 1974, 319. It has been published by Abgrall 1995.
8. Hogendijk 1985b.
9. Abu`l-Jd Muhammad ibn al-Iayth, al-Sijz, al-Saghn, al-Shann, al-Al ibn Sahl,
and Ab Sahl al-Kh.
10. Eor details on the medieval and modern literature on this treatise and its history see
Hogendijk 1984.
11. Hogendijk`s translation in Hogendijk 1984, 214.
12. Anbouba pointed this out in Anbouba 1977.
13. The Iact that both al-Kh and al-Saghn`s treatises were dedicated to the same
patron supports, in view oI the other evidence, the supposition that al-Kh developed
al-Saghn`s solution, rather than discovering the whole approach by conic sections
independently.
14. Rashed 1981, 258.
15. KhanikoII 1857, pp. 2633.
16. Recently published by Rashed in MIDEO 20 (1991), 21231 along with Ibn al-
Haytham`s treatise on knowns in MIDEO 21 (1993), 87275.
17. Quoted Irom Rashed`s article on Ibrhm ibn Sinn in the DSB VII, 23.
18. The author, jointly with Glen Van Brummelen, has published this in Suhayl 2, 2001
pp. 161198.
19. Berggren 1983.
20. I thank Glen Van Brummelen Ior this argument.
21. See Hogendijk 1984, esp. p. 200 note 3.
22. The passage Irom al-Kh`s correspondence is in Berggren 1983, 54.
23. See the accounts in Sayili 1962 and in Knorr 1989, 301309.
24. In Book iv oI Pappus`s Math. Coll. (Vol. I, pp. 270272). I am not, oI course, sug-
gesting that Pappus was a source Ior al-Kh.
25. Berggren 1994, 150.
26. Preserved in BN 4821 and dedicated to Abu`l-Eawris, the son oI Adud al-Dawlah.
27. This ascription oI the construction oI the regular heptagon to Ab Sahl al-Kh sup-
ports (and perhaps refects) al-Kh`s own words in his treatise on the regular heptagon,
'Now the easiest among those pursuits is the science oI the side oI the regular heptagon
in a circle to which the intellects oI the renowned among the Iamous geometricians
have been applied . . . . Nevertheless, no one among them was able to attain an iota oI
it. Nevertheless, when the servant oI our lord . . . Adud al-Dawlah . . . looked into it he
was able to fnd it.
28. That is, in the case that any circle on the sphere not containing the pole is projected
Irom either pole onto any plane perpendicular to the axis.
29. Woepcke 1874, 11 notes in a Iootnote here: 'To give to the preceding constructions
this extension it will suIfce to determine the position and the two openings oI the com-
pass relative to the plane tangent to the given surIace at the point where the axis oI the
compass rests on the surIace.
30. Ab Sahl`s choice oI words ('surIaces having an axis which coincides with that oI
the sphere) shows he intended to be completely general, but the only specifc examples
he mentions are cones, cylinders or planes.
31. See Berggren 1982.
32. We shall publish this commentary separately.
33. One would normally add here his treatise On the a:imuth of the qibla, but, despite
Sezgin`s attribution oI it to al-Kh, it seems unlikely that it is by him. The treatise is not
attributed to him in the manuscript and the Iact that the author oI the treatise gives the
working-out oI an example in numerical terms seems utterly unlike al-Kh.
34. Discussed in Sayili 1956, and recently in Rashed 1999.
35. Hindu-Arabic, sexagesimal, and the so-called 'fnger arithemetic.
36. Number 13 in Sezgin`s list oI his works. (Sezgin 1974, 319.)
Biiiior:rn.
Anbouba, Adel. 1977. 'Tasb al-dira (Construction oI the Regular Heptagon), Jour-
nal for the History of Arabic Science 1, no. 1, 73105 (Arabic pagination).
Abgrall, Philippe. 1995. 'Ies cercles tangents dàl-Quhi. Arabic Sciences and Philoso-
phy 5, no. 2 (September), 263295.
Bellosta, Helene. 1991. 'Ibrhm ibn Sinn: On Analysis and Synthesis. Arabic Sci-
ences and Philosophy 1, 211232.
Berggren, J. I. 1982. 'Al-Brn on Plane Maps oI the Sphere. Journal for the History
of Arabic Science 6, 47112.
Berggren, J. I. 1983. 'The Correspondence oI Ab Sahl al-Kh and Ab Ishq al-
Sb: A translation with commentaries. Journal for the History of Arabic Science 7,
39124.
Berggren, J. I. 1991. 'Medieval Islamic Methods Ior Drawing Azimuth Circles on the
Astrolabe. Centaurus 34, 309344.
Berggren, J. I. 1994. 'Ab Sahl al-Kh`s Treatise on the Construction oI the Astrolabe
with ProoI: Text, Translation and Commentary. Physis 31, 141252.
Hogendijk, J. P. 1984. 'Greek and Arabic Constructions oI the Regular Heptagon.
Archive for History of Exact Sciences 30, no. 3/4, 197330.
Hogendijk, J. P. 1985a. Ibn al-Haythams Completion of the Conics. New York:
Springer-Verlag.
Hogendijk, J. P. 1985b. 'Al-Kh`s Construction oI an Equilateral Pentagon in a Given
Square. Zeitschrift fur Geschichte der Arabisch-Islamischen Wissenschaften 1 (1985),
100144.
Hogendijk, J. P. 1996. Al-Sif:s Geometrical 1reatise on Problem Solving. Teheran:
Eatemi Publishing Co.
Jones, A. 1986. Pappus of Alexandria Book 7 of the Collection. Springer-Verlag: New
York.
. Irxx:r+ Brrrrx 196
KhanikoII, N. 1857. 'Analysis and extracts oI Book oI the Balance oI Wisdom . . . by
al-Khzin in the TwelIth Century. Journal of the American Oriental Society 6, 1128.
Knorr, W. 1989. 1extual Studies in Ancient and Medieval Geometry. Boston: Birk-
haser.
Pappi Alexandrini Collectionis quae supersunt. (3 vols.; E. Hultsch, ed.) Berlin, 1876
1878.
Rashed, R. 1979. 'Ibn al-Haytham wa amal al-musabbai I al-dirati and 'Ia Con-
struction de l`heptagone regulier par Ibn al-Haytham. Journal for the History of Arabic
Science 3, no. 2, 309357 (Erench) and 358376 (Arabic).
Rashed, R. 1981. 'Ibn al-Haytham et la mesure du parabolode. Journal for History of
Arabic Science 5, 91262
Rashed, R. 1999. 'Al-Qh vs. Aristotle on Motion, Arabic Sciences and Philosophy
9, 724.
Sayili, A. 1956. 'A Short Article oI Ab Sahl Waijan ibn Rustam al-Qh on the Pos-
sibility oI Infnite Motion in a Einite Time. Actes JIII Congres International dHistoire
des Sciences. Elorence-Milan, 248249
Sayili, A. 1962. 'The trisection oI the angle by Ab Sahl Wayjan ibn Rustam al Kh.
Proceedings of the 10th International Congress on the History of Science. Ithaca,
546547.
Sesiano, J. 1979. 'Note sur trois theoremes de Mecanique dàl-Qh et leur con-
sequence. Centaurus 22, no. 4, 281297.
Sezgin, E. 1974. Geschichte des Arabischen Schrifttums, Vol. V: Mathematik. Ieiden:
E. J. Brill.
Woepcke, E. 1851. Lalgebre dOmar al-Khayyami. Paris.
Woepcke, E. 1874. Trois traites Arabes sur le compas parIait, Notices et Extraits des
Manuscrits de la Bibliotheque Imperiale et autres bibliotheques 22, 1175.
IV
Numbers, Geometry, and Architecture
1 Prriixix:rirs
1.1 Introduction
One oI the most impressive achievements in Islamic mathematics is the devel-
opment oI general methods Ior constructing magic squares. A magic square oI
order n is a square divided into n
2
cells in which diIIerent natural numbers must
be arranged in such a way that the same sum appears in each oI the rows, col-
umns, and two main diagonals (fgures 7.1 and 7.2); mostly, the n
2
frst natural
numbers are placed, and then the constant sum amounts to
1
2
n(n
2
1), the n-th
part oI the sum oI the natural numbers Irom 1 to n
2
. II, in addition to this basic
property oI simple magic squares, the square remains magic when the borders
are successively removed, it is called a bordered square (fgure 7.3). II the sum
in any pair oI complementary diagonals (i.e., pairs oI parallel diagonals lying
on each side oI a main diagonal and having together n cells) shows the constant
sum, the square will be called pandiagonal (fgure 7.4).
Squares are usually divided into three categories: squares oI odd
orderalso called odd squares(n 2k 1, k natural); evenly-even squares
(n 4k); oddly-even squares (n 4k 2). There are general methods that make
it possible to construct squares oI any order Irom one oI these three categories.
Except Ior the smallest possible order, n 3, there are numerous possibilities
oI Iorming magic squares oI any given order. There may be, however, some
limitations concerning additional magical properties; Ior instance, bordered
squares cannot be constructed iI n 4, and there are no pandiagonal squares
oI oddly-even order.
InIormation about the beginning oI Islamic research on magic squares is
lacking; it may have been connected with the introduction oI chess into Persia.
Initially, the problem was a purely mathematical one; thus, the Arabic ancient
designation Ior magic squares is wafq al-add, that is, 'harmonious disposi-
tion oI the numbers. We know that treatises on magic squares were written
in the ninth century, but the two earliest extant texts date back to the tenth
7
Qu:ir:+us Mir:iiiis
acques Sesiano
:cqurs Srsi:xo 200 Qu:ir:+us Mir:iiiis 201
century: the 1reatise on the magic disposition of numbers in squares by Abu`l-
WaI al-Bzjn (940997 or 998) and a chapter in Book III oI Al b. Ahmad
al-Antk`s (d. 987) Commentary on Nicomachoss Arithmetic.
1
It appears that,
by that time, the science oI magic squares was already established; bordered
squares oI any order n > 5 could be constructed, while simple magic squares
could be obtained Ior small orders.
2
Various general methods Ior the construc-
tion oI simple magic squares oI odd and evenly-even orders, and also Ior pan-
diagonal squares oI evenly-even orders, were devised in the early eleventh cen-
tury. The science oI magic squares can be said to be at its apogee around 1100;
by that time, the remaining problem oI constructing simple magic squares oI
oddly-even order had been solved. Erom the thirteenth century onwards, magic
squares become increasingly associated with magic purposes. Consequently,
some texts merely picture squares and mention their attributes. Some others
do keep the general theory alive, though oIten only to enable the reader to
construct amulets Ior himselI.
Eigure 7.2
Eigure 7.1
Eigure 7.3
In Europe, interest in magic squares was frst aroused toward the end oI
the Middle Ages, when two sets oI squares associated with the seven planets
were learned oI through Arabic magic texts (whence the name), but without
any indication as to their construction. Methods oI construction spread east-
ward around the twelIth century toward India and China; and, to a lesser extent,
toward Byzantium (c. 1300).
It is interesting to note that, Irom the very beginning oI the Islamic stud-
ies on magic squares, attempts were made to consider squares with additional
conditions, constructed by adapting or modiIying the general methods. The
present study will show how elementary procedures led, already in the tenth
century, to a highly intricate Iorm oI construction.
A magic square oI order three flled with the frst nine natural numbers
can only take one Iorm, where the even numbers occupy the corner cells and
thus surround the area containing the odd numbers (fgure 7.5). Since the
square oI order three was the smallest possible and the frst to be obtained,
this property might have given rise to the idea oI extending such a separation
to squares oI higher odd order. In this case the odd numbers are to occupy a
central rhombus with corners meeting the middle oI the square`s sides and the
even numbers will be in the Iour remaining triangles.
It is easy to obtain a simple magic square in which the even and odd
numbers are separated thus. Principles oI construction were devised at about
the same time as those Ior common simple magic squares, probably in the
early eleventh century, and simplifed rules appeared in the centuries that Iol-
lowed.
3
Concern Ior distinction by parity is also seen in some eleventh-century
constructions oI squares oI even orders, and in one case the odd numbers Iorm
a kind oI hexagon in each quadrant.
4
The idea oI constructing magic squares
with separation by parity was thereIore common in the eleventh century. The
Eigure 7.4
Eigure 7.5
same idea already occurs in the tenth century, at a time when general methods
were only known Ior bordered magic squares. Now the construction oI a bor-
dered square oI odd order showing in addition the above-mentioned separation
between odd numbers and even numbers is much more diIfcult. A frst attempt
seems to survive in Abu`l-WaI`s treatise. A general method oI construction
is explained by al-Antk, but the unequal value oI his treatise suggests that this
very elaborate construction did not originate with him. Whoever the author,
it is one oI the gems oI the Islamic science oI magic squares (or, indeed, oI
Islamic science in general).
There is no specifc Arabic denomination Ior these squares or Ior their
construction. I have chosen to call a square thus constructed quadratus mira-
bilis owing to a wonder comparable to that oI my countryman Jakob Ber-
noulli three hundred years ago when he discovered the properties oI his spira
mirabilis.
AIter some preliminaries (1, 24), we shall analyze this construction (2),
and this will be Iollowed by the translation oI al-Antk`s relevant section (3).
1.2 Construction ot Bordered Squares ot Odd Orders in the 1enth
Century
Abu`l-WaI and al-Antk start by showing how to construct bordered squares
oI any odd order n (n > 5) as Iollows (fgure 7.6, with n 11).
5
Beginning next
to a corner cell in the outer border, write the frst odd and even numbers alter-
nately along the column and row meeting at this corner, and continue as Iar as
Eigure 7.6
their middle cells. Write the next (odd) number in the middle cell last reached,
then put the Iollowing number in the corner cell at the other end oI the column
where 1 was placed, the next one ( n) in the middle cell on the other side, the
Iollowing number in the other upper corner cell; and fnally arrange the subse-
quent numbers along the remaining column and row, alternately as beIore, but
this time choosing to start next to the middle cell oI the column. At this point,
halI oI the border cells are occupied. Eill each oI the remaining blank cells
with n
2
1 minus the number in the opposite cell.
6
Repeat this procedure until
the central square oI order 3 is reached, and fll it with the remaining numbers
arranged as in fgure 7.5. This completes the construction oI the square. Note
that, with this arrangement, in each row oI the successive squares all numbers
are even except the number in the middle and in each column they are all odd
except the two numbers at the corners.
1.3 Geometrical Structure ot a Bordered Square vith Separation
Iet the main square have the order n 2k 1. The rhombus which is to contain
the odd numbers consists then oI 2k 1 'rows and 'columns in which the
number oI cells is alternately k 1 and k starting with the outer rows and col-
umns. Hence, within the rhombus there is a total oI (k 1)
2
k
2
2k
2
2k 1
cells, which is the quantity oI odd numbers among the frst n
2
natural numbers;
the remaining 2k
2
2k even numbers will then occupy the Iour corner trian-
gles, each oI which contains
1
2
k(k 1) cells. The rhombus itselI includes inner
squares, the largest oI which has the order k iI k is odd and k 1 iI k is even.
Consequently, the largest inner square Ior both n 4t 1 and n 4t 3 has the
order 2t 1, the only diIIerence being that its corner cells are not Iully included
in the frst case but are included in the second case (fgure 7.7a and 7.7b, Ior
t 2). Consider fnally the borders surrounding this largest inner square. The
quantity oI cells they contain which do not lie within the rhombus depends
upon the Iorm oI n. Eor the p-th border, counted Irom the one surrounding the
inner square, it will be
Ior order n 4t 1: 16p 4 altogether, and 4p in each row and column (with a
common corner cell), where p 1, ., t;
Ior order n 4t 3: 16p 12 altogether, and 4p 2 in each row and column,
where p 1, ., t 1.
The inner square is easy to fll in, Iollowing the above-mentioned method
Ior bordered squares (fgure 7.8): aIter putting
n
2
1
2
(odd) in the central cell,
write next to a corner cell the number 4t
2
1 when n(> 9) has the Iorm 4t 1,
and continue as explained above but using odd numbers only. With the comple-
ments placed in the remaining blank cells, the inner square becomes a bordered
magic square, occupied by a continuous sequence oI odd numbers. Thus Iar
our two tenth-century authors proceed in the same way. Erom now on, however,
they take a very diIIerent approach.
1.4 Abu`l-\ata`s Construction
AIter completing the inner square as above, Abu`l-WaI suggests using Ior the
remaining part the arrangement obtained by the method Ior bordered squares
but omitting those numbers which do not satisIy the requirements oI parity
and those which have already been used Ior the inner square (fgure 7.9; cI.
Eigure 7.7a
Eigure 7.7b
fgure 7.6). The blank cells are then flled with the remaining numbers, with
due consideration to parity and the amount needed to complete the magic sum.
(One possibility is seen in fgure 7.10.)
But this procedure does not resolve the main problem, which resides in
the increasing complexity oI the computations as the order increases: in the
columns outside the rhombus and in the rows inside it most oI the cells will
remain empty because oI the unequal distribution by parity mentioned at the
end oI 1.2 (and seen in fgure 7.6). To sum up, this (apparently) early approach
Eigure 7.8
Eigure 7.9
does not meet the Iundamental requirement oI methods Ior constructing magic
squares, which is to leave no room Ior uncertainty in placing the numbers once
the order is known.
2 Dirrc+ Coxs+ruc+iox
2.1 Placing the Remaining Odd Numbers in the Rhombus
The method reported by al-Antk is reminiscent oI, and no doubt inspired
by, the method Ior common bordered squares described in 1.2, the main diI-
Ierence being the points oI departure in the border oI the rhombus. Al-Antk`s
method is valid Ior any rhombus, irrespective oI the parity oI k in the order
n 2k 1 oI the main square. It should also be noted that Ior any two consecu-
tive orders n 4t 1 the arrangement oI the remaining smaller odd numbers
(1 to 4t
2
1) in the corner triangles oI the rhombus will be the same.
To write these numbers in the rhombus, put 1 next to (say) the lower leIt
corner cell oI the inner square, 3 next to its upper leIt corner cell and then alter-
nate the next odd numbers along the border row and column oI the rhombus as
Iar as the corner cells. Ieave these blank and fll in the other corner cells, frst
the lower cell and then the right cell. The subsequent numbers are placed alter-
nately along the other border row and column, taking the corner cell just flled
as point oI departure and stopping at the sides oI the inner square.
7
The process,
which is repeated in the subsequent borders, ends when only two cells are leIt
(one in each oI the main diagonals oI the rhombus); the lower cell should be
flled in frst, then the right cell. AIter the complements have been written in,
Eigure 7.10
the rhombus is complete, and no odd number is leIt (fgures 7.11 and 7.12, with
orders oI the Iorm 4t 1).
2.2 Placing the Even Numbers. Preliminaries
A 2EQUIRED SUM IN EACH "ORDER
The magic sum being
1
2
n(n
2
1), we may compare the sum in m cells already
flled in with the sum they should contain on the average, namely
1
2
m(n
2
1).
(We shall reIer to this hereaIter as the 'sum due Ior m cells.) II the odd num-
bers are arranged in the rhombus as above, each upper border row oI the main
square shows an excess over the sum due, and so does each oI its leIt col-
umns, while the corresponding opposite rows and columns, being flled with
complements, show a defcit oI the same amount. Eor the p-th upper border
row, counted Irom the border surrounding the inner square, the excess I
p
is as
Iollows:
Eor order n 4t 1:
I
p
8pt 4, thus
I
1
8t 4
Eigure 7.11
I
2
16t 4
I
3
24t 4
I
4
32t 4
.
Eor the leIt columns, the excesses are each 2 less than the corresponding
excesses Ior the rows; they are thereIore
I
'
p
8pt 2, thus
I
'
1
8t 2
I
'
2
16t 2
.
Eor order n 4t 3, the excess A
p
is:
A
p
8(p 1)(t 1) 4 Ior the p-th upper row, thus
A
1
4
A
2
8(t 1) 4
A
3
16(t 1) 4
A
4
24(t 1) 4
Eigure 7.12
.
and Ior the p-th leIt column, the excess is:
A
'
p
8(p 1)(t 1) 2, thus
A
'
1
2
A
'
2
8(t 1) 2
.
We have already determined the quantity oI even numbers needed Ior
each border row and column (1.3) and have now Iound the excess or defcit in
each incomplete border row and column aIter placing the odd numbers. Thus
we may calculate the sum oI the even numbers to be placed in each border row
or column: this is the sum due (
n
2
1
2
multiplied by the number oI empty cells
leIt in each incomplete border row or column) reduced or increased by the
excess or defcit I or A.
b) Placing the even numbers. basic rules
Iet us consider the set oI even numbers to be placed. Eor any order n 2k 1,
there is an even quantity 2k
2
2k oI even numbers: k(k 1) smaller ones (less
than
n
2
1
2
) and their k(k 1) complements. Since k(k 1) is even, all these
numbers can not only be aligned vertically in pairs oI complements but also
grouped horizontally by pairs, as Iollows:
2 4 ; . ;
n
2
8f3
2
n
2
8f7
2
; . ;
n
2
5
2
n
2
1
2
;
n
2
1 n
2
3 ; . ;
n
2
8f1
2
n
2
8f5
2
; . ;
n
2
7
2
n
2
3
2
.
We shall henceIorth call 'dyad any such pair oI even numbers and char-
acterize it by the value oI f (f 1, 2, .,
1
2
k(k1) Ior those oI the upper line,
Irom right to leIt).
The method Ior the initial elimination oI the excesses and defcits and the
subsequent flling oI the remaining empty cells rests on the appropriate choice
and arrangement oI these dyads.
The initial elimination oI the excesses and defcits (reIerred to hereaIter as
equali:ation) relies Ior n 4t 1 and n 4t 3 on Iour rules:
I. Writing the number o in one side (row or column) and o 2s in the
other and then flling the opposite cells with the complements will pro-
duce in the side oI o the sum
o |n
2
1 (o 2s)| n
2
1 2s
and in the other side
=
o
=
o
Eigure 7.13
=
o
=
o
Eigure 7.14
2
+1 - (o + 6) o + 6
o + 2
2
+1 - (o + 2)
2
+1 - o o + 4
2
+1 - (o + 4) o
2
+1 - 4
2
+1 - 4
2
+1 + 4
2
+1 + 4
Eigure 7.15
=
2
+1 - (8 - 4)
=
2
+1 + (8 - 4)
2
- 8 + 7
2
2
- 8 + 3
2
2
+ 8 - 5
2
2
+ 8 - 1
2
Eigure 7.16
|n
2
1 o| o 2s n
2
1 2s (fgure 7.13).
This rule enables us to eliminate any diIIerence 2s by using two even num-
bers which diIIer by 2s. It should be noted, however, that taking s = 1 'breaks
two dyads, since this uses only one element oI each, which is a situation to
be avoided iI possible, as we shall see below. The Iollowing particular cases,
which only involve dyads that have been leIt complete, are mainly used by al-
Antk Ior the equalization:
II. Taking any dyad oI smaller numbers and writing its elements in
opposite sides, then flling the opposite cells with the complements, will
produce the smallest diIIerence possible, 2, in the Iorm oI a defcit in the
side containing the smallest element (fgure 7.14).
III. Placing two consecutive dyads oI smaller numbers around the border,
then flling the opposite cells with the complements, will produce a diI-
Ierence oI 4, in the Iorm oI a defcit in the row and column containing
the frst dyad (fgure 7.15).
IV. Writing the f-th dyad oI smaller numbers in one side and their com-
plements in the opposite side will produce a defcit oI 8f 4 in the side
oI the f-th dyad and a corresponding excess in the other (fgure 7.16). II
this dyad is written in the corner cells, the same will hold, but in addi-
tion there will be a defcit oI 2 in the perpendicular side containing the
smallest element, and a corresponding excess in the opposite side (fgure
7.17).
The application oI these Iour rules allows us to eliminate any diIIerence
oI the Iorm 8u v with v 0, 2, 4, 6, thus all the possible diIIerences, and
without breaking dyads unless we apply the frst rule.
Erom rule I we Iurther deduce a 'neutral arrangement, which produces the
2
+1 - (8 - 4)
2
+1 + (8 - 4)
2
+1 - 2
2
+1 + 2
2
- 8 + 3
2
2
+ 8 - 5
2
2
- 8 + 7
2
2
+ 8 - 1
2
Eigure 7.17
sum due, without any excess or defcit:
V. Consider two pairs oI numbers having an equal diIIerence: o, o 2s;
, 2s. Writing the two extreme terms in one side and the two middle
terms in the other, then flling the opposite cells with the complements,
will produce in the two sides the sum 2(n
2
1), which is the sum due Ior
Iour cells (fgure 7.18). Here again, we should choose s 1, so that the
two pairs oI numbers are two (separate or consecutive) dyads.
1hus, if by applying rules IIJ we are able to hll a certain (even) number
of cells in each border row or column, including the cells in the corners, so as
to arrive at the sumdue for the number of cells already hlled, and if the number
of remaining (empty) cells in this row or column is divisible by 4 while the set
of still available even numbers consists of dyads, these cells can be easily hlled
using ruleJ. If the method is to be generally applicable, the hrst step (the equal-
i:ation process) must be uniform and involve the smallest possible number of
cells so as to apply to all orders, from the smallest one possible.
This is the method Iollowed by al-Antk. In order to show that this
method is indeed generally applicable, it will be described below in modern
symbolism.
3.3 Placing the Even Numbers tor the Order N = 4T - 1 (T > 2)
A 2OWS
We shall consider only the upper rows, since the corresponding lower rows will
be occupied by the complements oI the numbers in the upper rows.
As we have seen (1.3), in each border row and column there is a number
oI cells remaining empty which is divisible by 4. Since the diIIerences Irom the
sum due must be eliminated by flling the smallest possible number oI cells,
including those at the corners, and there must remain a number oI empty cells
divisible by 4, we shall equalize each border row and column by means oI Iour
or at most eight numbers. Al-Antk`s frst step is thus to fll Iour cells in each
o
o
o
o
Eigure 7.18
row, including the two cells at the corners, so as to leave some excess Ior p > 2
but none in the row closest to the inner square since it has only Iour empty
cells; the next step will be to eliminate the remaining excess by means oI Iour
other numbers, thus leaving Ior p > 3 a number oI empty cells divisible by 4
and no excess or defcit. The Iour numbers to be initially placed are chosen as
Iollows (see fgure 7.19 n 17, thus t 4 , numbers in Roman type).
Eirst (lowest) upper row (p 1): Put
n
2
5
2
in the leIt corner
n
2
1
2
in the right corner.
These numbers Iorm the largest dyad oI smaller numbers (corresponding
to f 1). Since their sum is n
2
3 instead oI the sum due n
2
1, while the initial
excess was I
1
8t 4, the remaining excess is 8t. By rule I we can eliminate
this excess by means oI any two even numbers with a diIIerence oI 8t. We
may, Ior instance, put the smallest even number 2 in the upper row and 2 8t
in the opposite side. The frst upper row will then contain as even numbers, in
addition to the ones in the corner cells, 2 and n
2
1 (2 8t), and will thus no
longer show an excess.
Eor the other rows (p > 2), the initial excess is I
p
8pt 4. This may be, at
least partly, compensated using an appropriate dyad. Rule IV will enable us to
determine which f will suit. Iet us thus put
8pt 4 8f 4.
Since we have to eliminate the excess by means oI either Iour or eight
numbers (but in any case not two), we take f pt (and not f pt 1) and put
the pt-th dyad oI smaller numbers
n
2
8pt3
2
,
n
2
8pt7
2
in the p-th upper row. Then the remaining excess in each upper row is 8.
Above, we used the frst dyad (f 1) Ior the frst border. Iet us accord-
ingly take the p-th dyad and put its smaller element in the upper leIt corner oI
the p-th border and the larger element in the lower corner in the same column.
Since the sum oI this smaller element, namely
n
2
8p3
2
,
and the complement oI the larger element, namely
n
2
8p5
2
,
is (n
2
1) 2, the excess leIt in each upper p-th row (p > 2) will be equal to 6,
and 4(p 1) cells in this row remain empty.
B #OLUMNS
We shall consider only the leIt columns, since the right columns are to be flled
with the complements oI the numbers in the leIt column. Now the corner cells
are already occupied; we shall thereIore eliminate the diIIerences in the p-th
column by means oI six numbers Ior p > 2 but only two Ior the case p 1,
where just two empty cells are leIt. This is done as Iollows (see fgure 7.19,
numbers in italics).
Eigure 7.19
Eirst column (p 1): Since the corner cells have been occupied by
n
2
5
2
and
n
2
3
2
,
with sumn
2
1, the initial excess I
'
1
8t 2 has been reduced to 8t. Proceeding
as Ior the frst row, we put 4 in the leIt column and 4 8t in the right column.
Thus the frst leIt column, now completed by 4 and n
2
1 (4 8t), has no
excess, and with the choice oI 4 there are no more broken dyads: the dyad
consisting oI the smallest numbers 2 and 4 (corresponding to f 2t
2
t) and
the associated dyad Iormed by 2 8t and 4 8t (corresponding to f 2t
2
t)
have both been used.
Other columns (p > 2): As we know, the initial excess is I
'
p
8pt 2. Now the
two numbers in the corners are
n
2
8p3
2
and
n
2
8p7
2
,
with sum n
2
8p 5, and so the excess remaining in each leIt column has
become 8pt 2 8p 4 8p(t 1) 6. By rule IV, we can reduce each oI these
excesses to a constant: putting
8p(t 1) 6 8f 4,
we choose f p(t 1) 1 and write the dyad
n
2
8(ptp1)3
2
,
n
2
8(ptp1)7
2
in the p-th leIt column (p > 2). Thus the excesses are now reduced to the
constant amount oI 2.
Remark. AIter placing the odd numbers in a square with order n 4t 1,
t > 2, we were leIt with t unfnished borders, numbered p 1, ., t. Now we
have seen that all the dyads oI smaller even numbers initially at our disposal
can be written as
n
2
8f3
2
,
n
2
8f7
2
with f 1, ., 2t
2
t.
Among these dyads, the Iollowing ones have been used:
(i) f 1 (in the corner cells oI the frst upper row, thus Ior p 1)
(ii) f 2t
2
t (elsewhere in the frst border, thus Ior p 1)
(iii) f p, that is, f 2, 3, ., t (in the leIt corner cells Ior p > 2)
(iv) f pt, that is, f 2t, 3t, ., t
2
(in the upper rows Ior p > 2)
(v) f p(t 1) 1, that is, f 2t 1, 3t 2, ., t
2
t 1 (in the leIt
columns Ior p > 2).
Each oI the dyads used corresponds to exactly one f among these 3t
values, none oI which is identical to another. Thus none oI the dyads used can
occur twice.
C #OMPLETING THE SQUARE
There is now leIt in each row and column (Ior p > 2) a number oI empty cells
divisible by 4. Eurthermore, the excess is 6 Ior each incomplete upper row and
2 Ior each incomplete leIt column (fgure 7.20). The set oI remaining even
numbers consists solely oI dyads or also tetrads oI consecutive terms; Ior the
numbers placed previously always Iormed dyads, which were either grouped
or separated by one or two (or more) unplaced dyads. The remainder oI each
border may now be flled in as Iollows.
(1) Take, according to rule III, one tetrad oI consecutive even numbers (o,
o 2, o 4, o 6) and place them (anti-clockwise) around the border, starting
with the top row, and then write the respective complements. This will increase
the sums in the border rows and columns as in fgure 7.15. Thus, each oI the
Iormer excesses and defcits is replaced by 2 (fgure 7.21).
(2) Take any two (consecutive or separate) dyads (, 2, ,, , 2) and, Iol-
lowing rule II, write each dyad in a pair oI opposite sides, with the smaller
elements on the side oI the excesses, and write the complements. This will
eliminate the previous diIIerences, since the sums in the border rows and col-
umns are increased as seen in fgure 7.14. Thus, the rows and columns oI the
borders now show the sum due Ior the quantity oI cells already flled in: that is,
as many times
1
2
(n
2
1) as there are occupied cells; and, Ior p > 3, in each row
and column the number oI remaining empty cells is divisible by 4.
(3) Take any two (consecutive or separate) dyads (o, o 2, t, t 2) and, Iol-
lowing rule V, place the extreme terms in one row or column and the middle
terms in the opposite row or column. AIter the complements have been written
in, this will produce the sum due Ior Iour cells. This enables us to fll the empty
cells with the numbers which have not yet been used.
The above steps are seen in fgure 7.22 (where the borders, starting
Irom the outer one, have been equalized successively, using the frst two steps
Eigure 7.20
Eigure 7.21
described above, by means oI the available numbers ranging discontinuously
Irom 6 to 56; and the third step was applied frst to the rows and then to the
columns). The completed squares oI the smaller orders 9 and 13 are also shown
in fgures 7.23 and 7.24. Al-Antk restricts himselI to constructing the square
Ior order 9.
Remark. These three fnal steps appeared in the tenth century Ior the con-
struction oI bordered squares oI even order, which implies that they were not
specifcally devised Ior the quadratus mirabilis.
2.4 Placing the Even Numbers tor the Order N = 4T - 3 (T > 1)
A 2OWS
We shall again consider only the upper rows, since the lower rows are to be
flled with the complements oI the numbers in the upper rows.
As we have seen (1.3), the number oI empty cells is oI the Iorm4p 2. In
order to eliminate the diIIerence Irom the sum due, we wish to fll the smallest
Eigure 7.22
possible number oI cells, including those at the corners, and leave a number oI
empty cells divisible by 4. Thus we may ask whether a single dyad placed in the
upper corner cells can directly produce the sum due. Since the initial excess is
A
p
8(p 1)(t 1) 4, we apply rule IV and put 8(p 1)(t 1) 4 8f 4.
Thus, A
p
will be eliminated at once iI we take f (p 1)(t 1) 1 pt p t.
We thereIore write in the corner cells oI the p-th upper row (p > 1) the dyad
n
2
8(ptpt)3
2
,
n
2
8(ptpt)7
2
,
Eigure 7.23
Eigure 7.24
with the smaller term in the column with the excess (see fgure 7.25, numbers
in Roman type). Thus we shall choose
Ior p 1:
n
2
5
2
,
n
2
1
2
;
Ior p 2:
n
2
2n7
2
,
n
2
2n3
2
;
Ior p 3:
n
2
4n9
2
,
n
2
4n5
2
;
and so on.
B #OLUMNS
We shall again consider only the leIt columns, Ior the right columns are to be
flled with the complements oI the numbers in the leIt columns.
Since the corner cells are already occupied, the number oI remaining
empty cells is a multiple oI 4 (Ior p > 2). We shall thereIore eliminate the excess
leIt in the p-th column by means oI Iour numbers (see fgure 7.25, numbers
in italics). Since the initial excess beIore the corner cells were flled was A
'
p

8(p 1)(t 1) 2 and they are now occupied by
Eigure 7.25
n
2
8(ptpt)3
2
,
n
2
8(ptpt)5
2
,
with sum (n
2
1) 2, the remaining excess is 8(p 1)(t 1) 8(pt p t 1).
As we must place Iour numbers in each column, it might be possible to
start by choosing a dyad which leaves an excess oI the Iorm 8u 4 and then
by rule IV another dyad to eliminate the remaining diIIerence. Now we
have already used the frst dyad Ior the frst border; iI we use accordingly the
p-th dyad Ior the p-th border (p > 2) and write it in the right column, this
column will contain the smaller numbers
n
2
8p3
2
,
n
2
8p7
2
and the leIt p-th column the corresponding p-th dyad oI larger numbers
n
2
8p1
2
,
n
2
8p5
2
.
As a result, the previous excess in the p-th leIt column (p > 2) has now
increased to 8(pt p t 1) 8p 4 8(pt 2p t 1) 4. Eollowing rule
IV, we put 8(pt 2p t 1) 4 8f 4 and take accordingly f pt 2p t 1;
we thereIore write in the p-th leIt column the dyad
n
2
8(pt2pt1)3
2
,
n
2
8(pt2pt1)7
2
.
Thus, we have reached a point where Ior each column the cells flled in
show the sum due and the number oI empty cells leIt (Ior p > 3) is divisible
by 4.
Remark. AIter placing the odd numbers in a square with order n 4t 3, we
are leIt with t 1 incomplete borders, numbered p 1, ., t 1. OI all the
dyads oI smaller even numbers initially at our disposal, that is,
n
2
8f3
2
,
n
2
8f7
2
with f 1, ., 2t
2
3t 1,
the Iollowing ones have been used:
(i) f pt p t, that is, f 1, t 2, 2t 3, ., t
2
t 1, t
2
, t
2
t 1 (in the
corner cells oI the upper rows, thus Ior p > 1)
(ii) f p, that is, f 2, 3, ., t 1 (in the right columns Ior p > 2)
(iii) f pt 2p t 1, that is, f t 3, 2t 5, 3t 7, ., t
2
t 1, t
2
2t 1
(in the leIt columns Ior p > 2).
Each oI the dyads used corresponds to exactly one f among these 3t 1
values, none oI which is identical to another. Thus none oI the dyads used can
occur twice.
C #OMPLETING THE SQUARE
Since all the rows and columns contain the sum due Ior the quantity oI occu-
pied cells and the number oI remaining empty cells is divisible by 4, we shall
conclude by applying rule V with the remaining dyads. The result is shown in
fgure 7.26 (where the tetrads oI available even numbers have been used frst
Ior the rows, starting with the top row, then Ior the columns). Eor the smaller
orders n 7 and n 11 see fgures 7.27 and 7.28. Both squares are constructed
by al-Antk.
2.3 Placing the Even Numbers tor the Order N = 3
As al-Antk mentioned, the above construction Ior the case n 4t 1 is
invalid Ior t 1. (The numbers to be used in the corner cells appear elsewhere
in the border.) Thus the square he presents (fgure 7.29) is not constructed
according to these rules. Abu`l-WaI`s simplifcation is oI no help either. The
square oI order fve he reIers to is presumably the same as al-Antk`s (the only
extant manuscript omits the fgures altogether). Abu`l-WaI does, however,
point out that iI we put 2 in one oI the upper corner cells, the other one can be
Eigure 7.26
occupied by 4, 6, 8 or 12. There are in Iact twenty-one possibilities, which are
listed below (see fgure 7.30).
Take o 2 and
4 with 14, 20 in the upper row and 8, 10 in the leIt column, or with
16, 18 in the upper row and 6, 12 in the leIt column (as in fgure 7.29);
6 with 10, 22 and 8, 12 or with 14, 18 and 4, 16;
20 with 4, 14 and 16, 18 or with 8, 10 and 12, 22.
8
Take o 4 and
6 with 12, 18 (upper row) and 2, 16 (leIt column).
Eigure 7.27
Eigure 7.28
Take o 6 and
10 with 2, 22 and 8, 12;
22 with 2, 10 and 14, 18.
Take o 8 and
2 with 14, 16 and 4, 6;
10 with 2, 20 and 4, 14;
16 with 2, 14 and 4, 20.
Take o 12 and
4 with 8, 16 and 2, 6.
Take o 14 and
8 with 2, 16 and 4, 6;
18 with 2, 6 and 4, 16.
Take o 16 and
8 with 4, 12 and 2, 6;
18 with 2, 4 and 6, 12.
3 Drscrir+iox or +nr Coxs+ruc+iox i. :i-Ax+
.
:ki
Al-Antk`s description oI the construction leading to the quadratus mirabilis
is very concise, but suIfcient to confrm our mathematical reconstruction. We
think it in any case appropriate to add a translation oI the relevant text. This
is the second part oI the section in which al-Antk deals with squares oI odd
orders. (In the frst part, he explains the construction oI bordered squares sum-
marized in 1.2.)
Consider the odd numbers Irom 1 to the last oI those which will be in the
square. Arrange them inside the square so as to Iorm the shape oI a rhombus
within the large square, thus leaving (empty) cells Iorming triangles with a
same number oI cells on each side. Write there the even numbers Irom 2 to
the last to be Iound in the square in such a way that the sums be equal every-
where. Then the odd numbers will be inside the larger square in a rhomboid
fgure and the even numbers will surround them on the Iour sides, as is
shown in this fgure (fgure 7.31).
Eigure 7.29 Eigure 7.30
The way to place the odd numbers is the Iollowing.
9
Take 1 and the
last term belonging to this square, namely its largest number, then 3 and the
number preceding the largest one, and so Iorth until you reach its middle
term. You put this middle term in the central cell oI the square. You put its
adjacent odd terms in the cells where you placed the two oI the nine terms
which were adjacent to the middle term in the square oI 3 by 3 (fgure 7.32).
You do the same Ior the remaining numbers until the square oI 3 by 3 is
completed, iI this is the object oI your treatment, or you do the same Ior
the squares oI 5 by 5, 7 by 7, 9 by 9, iI this is the object oI your treatment.
You will always do this until the whole square contained by the rhombus is
completed.
This being done,
10
take the two odd terms reached, and put the smaller
one in the middle cell oI the frst leIt-hand line and its complement Iacing
it in the frst right-hand line, the next small number in the middle cell oI
the lower line, and Iacing it in the upper line its complement. Take then the
two terms reached and put the smaller one below, next to the middle cell
on its leIt, and, Iacing it above, its complementary term. Put then the next
small term on the leIt, just above the middle cell, and Iacing it on the right
its complement. Then put the next small number in the middle cell oI the
second leIt-hand line, and Iacing it in the second right-hand line its comple-
ment. Then put: the small number Iollowing this term in the middle cell oI
the second lower line, and Iacing it its complement; then the Iollowing small
term next to the middle cell in the frst upper line, and Iacing it in the frst
lower line its complement; the Iollowing small number in the second right-
hand line, below the middle cell, and Iacing it on the leIt its complement.
When this is done Ior the two squares (oI order 7 and 9), 1 and the last term
are reached and the placing has been perIormed in the desired way. In the
case oI the square oI 9, the treatment is then completed.
In the case oI the squares oI 11 and 13,
11
do the same until the above
arrangement is attained. Then you write the small number Iollowing the
small number placed lastly in the third cell Irom the middle cell,
12
on the
leIt oI it, in the frst lower line, and Iacing it its complement. Then you put
the Iollowing small number in the third cell above the middle cell in the frst
leIt-hand line, and its complement Iacing it on the right. Then you put the
subsequent small term in the second lower line, next to the middle cell, and
its complement Iacing it above. Then you put the subsequent small term next
Eigure 7.31 Eigure 7.32

to the middle cell in the second leIt-hand line, and its complement Iacing it
on the right. Then you put the small term reached in the middle cell oI the
third leIt-hand line, and Iacing it its complement. Then you put the subse-
quent small term in the middle cell oI the third lower line, and Iacing it above
its complement. Then you put the subsequent small term next to the middle
cell oI the second upper line, and Iacing it below its complement. Then you
put the subsequent small number in the second right-hand line, next to the
middle cell, and Iacing it on the leIt its complement. Then you put the sub-
sequent small number in the frst upper line, in the third cell Irom the middle
cell, and Iacing it below its complement. Then you put the subsequent small
number in the third cell Irom the middle cell in the frst right-hand line, and
Iacing it on the leIt its complement. When this is done, the odd numbers are
placed in these two squares, starting Irom the center.
In the case oI the squares oI 15 and 17, you do the same as Ior the
square oI 13 until you reach the third border. Then you put the subsequent
small number in the Iourth cell Irom the middle one in the frst lower line,
and Iacing it its complement; then the subsequent number in the Iourth cell
Irom the middle cell in the frst leIt-hand line, and Iacing it its complement;
then the subsequent number in the third cell Irom the middle cell in the
second lower line, and Iacing it its complement; then the subsequent number
in the third cell Irom the middle cell in the second leIt-hand line, and Iacing
it its complement; then you put the subsequent number next to the middle
cell in the third lower line, and Iacing it its complement; then the subsequent
number next to the middle cell in the third leIt-hand line, and Iacing it its
complement; then the subsequent number in the middle cell oI the Iourth
leIt-hand line, and Iacing it its complement; then the subsequent number in
the middle cell in the Iourth lower line, and Iacing it its complement; then
the subsequent number next to the middle cell in the third upper line, and
Iacing it its complement; then the subsequent number next to the middle cell
in the third right-hand line, and Iacing it its complement; then the subsequent
number in the third cell Irom the middle cell in the second upper line, and
Iacing it its complement; then the subsequent number in the third cell Irom
the middle cell in the second right-hand line, and Iacing it its complement;
then the subsequent number in the Iourth cell Irom the middle cell in the
frst upper line, and Iacing it its complement; then the subsequent number in
the Iourth cell Irom the middle cell in the frst right-hand line, and Iacing it
its complement. When this is done, you have fnished with the odd numbers
Ior these two squares in the desired way. II you wish to proceed with larger
squares, continue placing step by step in the same way.
At this point we fnd that the squares are divided into classes requir-
ing each its own treatment Ior the arrangement oI the even numbers in the
corners. There are those oI (order) 5, 9, 13, 17, and so Iorth by steps oI 4; and
those oI (order) 7, 11, 15, 19, and so Iorth by steps oI 4.
Eor constructing that oI 5 and those oI the same kind (observe this).
13
The remaining empty cells in the frst border, which surrounds the inner
square, are the Iour at the corners and eight adjacent to them, two on each
side; in the second border, the empty cells are the Iour at the corners and
twenty-Iour cells adjacent to them. It will always be like that: each border has
16 angular cells more than the border below.
Eor the square oI 5 only, the treatment Iails: it requires a diIIerent
method. Eor the squares belonging to the same class, the situation is the Iol-
lowing.
Square oI 9.
14
AIter placing the odd numbers as indicated, the frst
upper line is in excess over the sum due by 20, the lower line is in defcit by
20; the frst right-hand line is in excess over the sum due by 18, and the leIt-
hand line is in defcit by 18. The second upper line is in excess by 36, and the
lower line is in defcit by the same amount; the second right-hand line is in
excess by 34, and the leIt-hand line is in defcit by the same amount.
Square oI 13. The frst upper line is in excess by 28, and the lower line
is in defcit by the same amount; the frst right-hand line is in excess by 26,
and the leIt-hand line is in defcit by the same amount. The second upper line
is in excess by 52, and the lower line is in defcit by the same amount; the
second right-hand line is in excess by 50, and the leIt-hand line is in defcit
by the same amount. The third upper line is in excess by 76, and the lower
line is in defcit by the same amount; the third right-hand line is in excess
by 74, and the leIt-hand line is in defcit by the same amount. Similarly Ior
the others.
In the square oI 9, a second line has 16 more than the line beIore. In
the square oI 13, a frst line has 8 more than the frst oI the square oI 9, then
each line has 24 more than the line beIore. In the square oI 17, a frst line
has 8 more than the frst oI the square oI 13, then each line 32 more than the
line beIore. Each line will always have more than the line beIore in the same
manner.
'Excess over the sum due means the Iollowing. The required amount
Ior the central cell equals the number in the middle. Thus each cell oI the
square oI 5 has a sum due oI 13. ThereIore you will add the odd numbers in
each line and divide the result by the number oI cells flled; iI the quotient
is less than the middle term, the sum in the cells will be less than the sum
due by the product oI this defcit and the number oI cells; analogously iI the
quotient is in excess. So the subsequent placing oI the even numbers in the
empty cells oI each line must be such that it compensates the defcit, iI any,
or Ialls short by the amount oI the excess, iI any. We shall show this in the
appropriate place, when explaining how to deal with the even numbers.
Eor constructing the square oI 7 and the others oI this class (observe
this).
15
The remaining empty cells are as Iollows. In the frst border, which
surrounds the inner square, there are the Iour cells at the corners; in the
second border, there are the cells at the corners and 16 cells adjacent to them,
4 on each side; in the third border, there are the cells at the corners and 32
cells adjacent to them. And so Iorth always: each border has 16 angular cells
more than the border below.
In this class oI squares, the frst upper line has an excess oI 4 over the
sum due, and the lower line has a defcit oI the same amount;
16
the frst right-
hand line has an excess oI 2 over the sum due, and the leIt-hand line has a
defcit oI the same amount.
The second upper line oI the square oI 7 has an excess oI 20 over the
sum due, and the second right-hand line an excess oI 18.
In the square oI 11, the second upper line has an excess oI 28 over the
sum due, the second right-hand line an excess oI 26, the third upper line an
excess oI 52, the third right-hand line an excess oI 50, and all the opposite
lines have a defcit equal to the excess.
Square oI 15. The second upper line exceeds the frst upper line by 32,
the third the second by 32.
And so Iorth Ior the others: the excess increases each time by 8.
17
Thus, writing the even numbers in the lines must be done so as to
equalize them, and we must then search Ior numbers the sum oI which will
produce the required excess or the required defcit.
To determine this, you associate 2 and the last even term, then 4 and
the corresponding opposite term, and so Iorth until you reach the two middle
terms oI these even numbers.
18
This being done, you see that iI the last small
number and the preceding one are placed on one side, and Iacing them their
complements, the side containing the two small numbers will be less than
the sum due by 4, and the other more by 4; placing the next two numbers,
the diIIerences will be 12 and 12, with the next two numbers 20 and 20,
then 28, 36, 44, 52, 60 and so Iorth to the numbers 4 and 2.
19
Now you see
that the borders which require the placing oI even numbers have indeed, on
two oI their sides, this succession oI excesses and defcits, namely 4, |12,|
20, 28, 36 and so Iorth. Thus, bringing the excess oI these even numbers to
where the defcit is will equalize the borders on two sides, and the other two
sides will need to be equalized with the remaining even numbers (by means
oI what Iollows).
Taking Iour (consecutive) small numbers, the sum oI the frst and the
second has a certain defcit, and the sum oI the third and the Iourth a defcit
smaller by 8. Then adding the frst and the Iourth you fnd that they have a
defcit equal to halI the sum oI the two defcits; adding similarly the second
and the third gives the same result. Eor instance, the sum oI the last two small
numbers
20
is less than their sum due by 4, and the sum oI the two preceding
numbers less by 12; thus, adding the last and the frst, and the second and the
number beIore, will produce a defcit oI 8, that is, halI the sum oI 12 and 4.
The knowledge oI this is necessary, Ior you will use it constantly.
21
When the frst small number, or any arbitrary small number, is written
on one side and the subsequent number is written on the other side, and the
two large numbers which are their complements are written on the opposite,
the side where the frst small number is written will be less than the sum due
by 2, whereas the other side will be in excess by 2.
22
II some small number
is written on one side and on the other side the third small number counted
Irom this one is written, the side containing the frst small number will be
less than the sum due by 4. And so Iorth: whenever the distance between
them is increased by one number the defcit is increased by 2.
23
Eor instance,
putting 4 on one side and 6 on the other side, and Iacing them their comple-
ments, the side containing 4 will be less than the sum due by 2 and the side
containing 6 will be more by 2. Writing 8 instead oI 6 and doing the same,
the side containing 4 will be less by 4 and the side oI 8 more by 4, and the
amount will increase together with the interval between the two numbers.
Placing Iour numbers in the corners, two in consecutive corners and
their complements in the corners diagonally opposite, iI the two consecutive
small numbers are on the upper side this side will be less than the sum due by
4 iI they are the last two, by 12 iI they are the two previous numbers, and so
Iorth with a regular increment oI 8 until 2 and 4 are reached. The right-hand
sides will have a uniIorm diIIerence, excess or defcit, oI 2 Irom the sum due,
without any increment or diminution.
24
All this must be understood: it will be necessary Ior the writing oI the
even numbers in this class oI squares.
Examples of treatments for all that we have explained.
Treatment Ior the square oI 5 by 5 (fgure 7.33).
25
You put the odd numbers in
the inner square oI 3 as explained. Those remaining are 1, 25, 3, 23. You put
1 in the lower middle cell and 25 Iacing it above, 3 in the middle leIt-hand
cell and 23 on the opposite side, on the right. Next, you put 2 in the upper
right-hand corner and Iacing it diagonally, in the lower leIt-hand corner, its
complement, namely 24. You put 4 in the upper leIt-hand corner and Iacing it
diagonally, in the lower right-hand corner, its complement, namely the even
number 22. You put 6 on the right side and its complement, 20, Iacing it on
the leIt. You put 10 and 8 below, 12 on the right, and you put Iacing each its
complement.
Passing Irom 5 to 9, 13, 17 and those oI this kind, put the last small
even term in the upper leIt-hand corner oI the frst border, that is, the border
Iollowing the inner square flled with odd numbers, and Iacing it diagonally,
in the lower right-hand corner oI the frst border, its complement.
26
Put the
preceding small term in the upper right-hand corner oI the frst border, and
Iacing it diagonally, in the lower leIt-hand corner, the term which is its
complement. When this is done, you fnd that the excess oI the upper line is
16,
27
the excess oI the right-hand line 16, thus the excesses oI the upper line
and the right-hand line are the same. Put then 2 in the upper line; consider the
excess oI the upper line, take its halI, and count aIter 2 as many small even
numbers as this halI, and put the number reached in the lower line. Put Iacing
each oI these two numbers its complementary term. Put 4 on the right side;
28
then count aIter 4 as many small numbers as halI oI the excess, and put the
number reached in the leIt-hand line. Put Iacing each oI these two numbers
its complementary term. AIter doing this all the sides oI the frst border will
be equalized Ior this kind oI square.
Put then the two small terms the sum oI which is less than their sum
due by 12 in the right-hand corners oI the second border, with the lesser one
above;
29
put in the diagonally opposite corners, on the leIt, their comple-
ments. Then look Ior the pair oI small numbers such that their sum is less
than their sum due by an amount equal to the (initial) excess oI the second
right-hand line less 14; put them in the second right-hand line and, Iacing
them on the leIt, their complements. Then look Ior the pair oI small numbers
such that their sum is less than their sum due by an amount equal to the
(initial) excess oI the upper line less 8; put them in this line and, Iacing them
below, their complements.
Put then the pair oI small numbers the sum oI which is less than their
sum due by 20 in the right-hand corners oI the third border, with the lesser
one in the upper corner; put Iacing them diagonally, in the corners oI the
third leIt-hand line, their complements. Then look Ior the pair oI small num-
bers which have a sum less than their sum due by an amount equal to the (ini-
tial) excess oI the third right-hand line less 22; put them in this line, and their
complements Iacing them, in the third leIt-hand line. Then look Ior the pair
oI small numbers such that their sum is less than their sum due by an amount
equal to the (initial) excess oI the third upper line less 8; you put them in this
line and, Iacing them in the third lower line, their complements.
Put then the pair oI small numbers such that their sum is less than their
sum due by 28 in the right-hand corners oI the Iourth border, with the lesser
one above, and Iacing them diagonally on the leIt their complements. Then
look Ior the pair oI small numbers such that their sum is less than their sum
due by an amount equal to the (initial) excess oI the Iourth right-hand line
less 30; put them in this line and, Iacing them in the Iourth leIt-hand line,
Eigure 7.33
their complements. Then look Ior the pair oI small numbers such that their
sum is less than their sum due by an amount equal to the (initial) excess oI
the Iourth upper line less 8; put them in this line and, Iacing them below,
their complements.
Proceed always likewise, with increments oI 8 Ior the defcits in a
right-hand line and uniIorm defcits oI 8 in an upper line. When this has been
done, the excess oI each upper line over its sum due will be 6, the excess oI
each right-hand line over its sum due will be 2, and the number oI remaining
empty cells will be Iour cells in each second line, eight in each third line,
twelve in each Iourth line, and so Iorth by successive additions oI Iour.
Iet us now turn our attention to the remaining numbers.
30
You take a
tetrad oI small numbers and put the frst number in an upper line, the third
in the opposite line below, the second on the right and the Iourth on the
leIt, and you put Iacing each oI these Iour numbers its complement. This
placing is perIormed Ior each border. When this is done, you take a pair oI
small numbers; put the frst above, the second below, and Iacing them their
complements. Then take another pair; put the frst on the leIt side oI the same
border, the second on the right, and Iacing them their complements. Do the
same Ior all borders. When this is done, each line and its opposite will be
equalized and none will be in excess.
II there are remaining empty cells, it can be only Iour Iacing Iour,
eight Iacing eight, twelve Iacing twelve (and so Iorth); they will be equal-
ized by groups oI Iour with available sequences oI Iour numbers, in the way
explained at the beginning oI the section Ior Iour numbers oI which each
pair is in progression: you put the frst oI the frst pair and the second oI the
second pair on one side, the second oI the frst pair and the frst oI the second
pair on the Iacing side, and opposite to each its complement (fgure 7.34).
Eigure 7.34
1reatment for 7, 11, 15, 19 and the like.
You put the last oI the small even terms in the still empty upper leIt-hand
corner oI the frst border, which surrounds the inner square flled with odd
numbers, and Iacing it diagonally, in the lower right-hand corner oI the same
border, its complement. Put the preceding small term in the upper right-hand
corner oI the same border, and, Iacing it diagonally, in the lower leIt-hand
corner, its complement. When this is done the frst border is equalized Ior
all squares oI this kind.
31
AIter that, consider the excess oI each upper line
and look Ior the pair oI small numbers such that their sum is less than their
sum due by the same amount. Put them in the corners oI this line, the lesser
number on the right, and Iacing them diagonally below their complements.
Complete in this way all the remaining corners. When this is done all the
upper and lower lines will be equalized, and their remaining empty cells will
be Iour Iacing Iour, and so Iorth by additions oI 4. Each group oI Iour is then
equalized with Iour numbers in the way we have explained previously.
We are leIt with the right-hand lines, which exceed the sum due by
16, 24, 32, 40, 48, and so Iorth by increments oI 8, and the remaining empty
cells are Iour Iacing Iour, eight Iacing eight, and so Iorth by additions oI 4.
32
You then look Ior a pair oI large numbers such that their sum exceeds their
sum due by an amount which, when added to the excess oI the right-hand
line, equals the defcit oI the sum oI two small numbers; put then the two
small numbers on the same side (as the large numbers), and put on the leIt
the complements oI the Iour numbers, each pair Iacing its complements.
When this is done, each pair oI corresponding sides will be equalized, and
the remaining empty cells will be Iour Iacing Iour, and so Iorth by successive
additions oI 4. Each group oI Iour is then equalized with Iour numbers as we
have explained previously.
Example of the treatment of the right-hand line for the square of 7.
You fnd that the excess oI the upper line is 20 and the excess oI the right-
hand line, 18. Putting the two numbers having a sum less than their sum due
by 20 in the two upper corners, with the lesser one on the right, leaves 16 as
the excess in the right-hand line. Consider then the two large numbers with
an excess oI 12; Ior adding 12 to 16, which is the right-hand excess, gives
an excess oI 28, equal to the defcit oI the sum oI two small numbers. You
put then the two large numbers on the right, as also the two small numbers
having a sum less than their sum due by 28, and you put on the leIt, Iacing
each one, its complement (fgure 7.35). The treatment is the same Ior the
other squares (oI this kind) (fgure 7.36).
No+rs
1. Eor the frst text, see J. Sesiano, 'Ie traite dÀbu`l-WaI sur les carres magiques,
Zeitschrift fur Geschichte der arabisch-islamischen Wissenschaften, 12 (1998), pp.121
244; an edition oI the second treatise, in MS (Ankara) Saib 5311, Iol. 1
r
36
r
will be
published in the same journal.
2. Although methods Ior simple magic squares are easier to apply than methods Ior
bordered ones, the latter are easier to discover; this explains why they appeared frst.
3. See J. Sesiano, Un traite medieval sur les carres magiques, Iausanne 1996, pp.
3540.
4. Ibid., p. 68.
5. Here, and whenever the original (Arabic) order is Irom right to leIt, we have chosen
to reverse it; the rules Ior construction have been changed accordingly.
6. Throughout this study, we call 'opposite cells those at either end oI any border row,
column or main diagonal oI a bordered magic square. Such pairs oI cells will always
contain a 'smaller number o (o
n
2
1
2
) and its 'complement n
2
1 o.
7. The need to reserve opposite cells in the main square Ior complements makes it
impossible to proceed here in exactly the same manner as Ior a bordered square.
8. Thus, Abu`l-WaI has indeed covered all possibilities Ior two consecutive corner
cells when one oI these numbers is taken to be 2, as can be seen by considering the
complements oI the last three values oI .
9. This explains the arrangement in the inner square. See 1.3, in hne.
Eigure 7.35
Eigure 7.36
10. Here begins the arrangement oI the remaining odd numbers in the rhombus. The
explanation can be Iollowed on fgures 7.117.12 (provided the terms 'leIt and 'right
are exchanged), starting with 63.
11. Squares are taken by pairs since the arrangement Ior orders 4t 1 is the same. See
2.1.
12. In the 'third cell Irom the middle, the middle cell is also counted.
13. See 1.3 (n 4t 1).
14. See 2.2.a (I
p
, I'
p
).
15. See 1.3 (n 4t 3).
16. See 2.2.a (A
p
, A'
p
).
17. That is: the excess oI one row or column over the row or column underneath is
constant Ior the same order, but increases by 8 Ior each order.
18. See 2.2.b (pairs oI complements and dyads).
19. See 2.2.b, Rule IV, frst part.
20. That is,
n
2
5
2
and
n
2
1
2
.
21. See Rule V.
22. See Rule II.
23. See Rule I.
24. See Rule IV, second part.
25. See 2.5.
26. See 2.3.a (p 1).
27. Case n 9 only; the excess increases by 8 Ior the other orders.
28. See 2.3.b (p 1).
29. See 2.3.a-b (p > 2).
30. See 2.3.c (and Rule III).
31. See 2.4.a & c.
32. See 2.4.a & c. The values oI the excesses are 16, 24, 32, 40, . . . iI p 2; 48, 64,
80, . . . iI p 3; and so Iorth.
Ix+roiuc+iox
As long as man has been constructing his dwellings, he has wanted to knowhow
many bricks are needed or how much earth has to be removed Ior the Iounda-
tions oI his house. Problems oI this kind are already Iound in the oldest texts on
arithmetic, such as the Chinese Nine Chapters on Arithmetical 1echniques. In
chapter V,
1
entitled Shang kung (= Evaluation oI Work), there are calculations
oI volumes oI regular solids, and oIten also oI the number oI people required
Ior excavating and transport. The same kind oI problems are treated in Arabic
manuals, such as al-Karaj`s Sufhcient Arithmetic |Kj jl-hisb|. This work
explains how to determine the amount oI sun-dried or fred bricks needed Ior
a building or how to level the ground. In this paper I do not want to talk about
these general construction problems but rather about calculations in connection
with other specifc elements oI Islamic architecture.
Isi:xic Arcni+rc+urr
What does 'Islamic architecture mean? As we can defne 'Christian architec-
ture mainly by churches, chapels, monasteries, etc., so we can defne 'Islamic
architecture by the principal Iorms oI Islamic buildings, which are not only,
but mainly religious. The oldest oI these buildings are mosques which date
Irom the frst beginnings oI Islam. According to Creswell`s account,
2
Arabia,
at the rise oI Islam, did not possess anything worthy oI the name oI architec-
ture. Only a small proportion oI the population was settled, and these lived in
dwellings which were scarcely more than hovels. When Muhammad migrated
to Medina he built a house Ior himselI and his Iamily in AD 623. Iike many
mud-brick houses in the Middle East, Muhammad`s house consisted oI a square
courtyard with two rooms (later increased to nine to accommodate his wives)
on the south-east side. The frst communal prayers were held in this courtyard.
Eor the comIort oI the worshippers a portico made oI palm-trunks and branches
was built on the north side oI the courtyard, together with a smaller one which
8
C:icui:+ix Surr:cr Arr:s :xiVoiuxrs
ix Isi:xic Arcni+rc+urr
Yvonne Dold-Samplonius
Yvoxxr Doii-S:xrioxius 236 C:icui:+ix Surr:cr Arr:s :xiVoiuxrs ix Isi:xic Arcni+rc+urr 237
gave shelter to visitors who sometimes spent the night there. The portico also
served as a place Ior deliberations on community aIIairs; hence to this day
the mosque has retained its multivalent role as a place oI prayer, social activi-
ties and political debate. Practical needs thus contributed to the house oI the
Prophet becoming the frst mosque oI Islam.
The men who Iormed the Arab armies oI conquest were mainly Bedouin,
but even those who came Irom permanent settlements, such as Mecca and
Medina, knew nothing oI art or architecture. In these early days, the Muslims,
when they conquered a town in Syria, usually took one oI the churches and used
it as a mosque, or merely shared one oI the churches iI the town had surrendered
without resistance. At Jerusalem they made use oI the remains oI the basilical
hall oI Herod, which ran along the south side oI the Temple enclosure, but had
been destroyed by the army oI Titus. In Persia, at Persepolis and Qazwn, they
appear to have taken hypostyle audience-halls oI the Persian kings, which had
fat rooIs resting on columns with double bull-headed capitals. But the situation
was diIIerent in Iraq, Ior there the Arabs Iounded new towns, so pre-existing
buildings could not be employed, and they had to construct some sort oI place
Ior themselves. What kind oI buildings were the frst mosques oI the earliest
towns in Islam?
At Basra, Iounded about 635, the frst mosque was simply an area marked
out on the ground and the people prayed there without any building. At KuIa,
Iounded in 638, the frst mosque was equally primitive. The frst mosques to be
worthy oI the name oI architecture were the second Great Mosques at Basra
(665) and KuIa (670). According to Creswell, it is apparent that the roofng
system resembled that oI a hall oI columns oI the Achaemenian kings, exactly
as was the case in the frst Great Mosque at Baghdad. The oldest extant monu-
ment oI Muslim architecture is the Dome oI the Rock at Jerusalem, completed
in 691, and restored and embellished during the Iollowing centuries. The har-
mony oI its proportions and the richness oI its decorations make the Dome oI
the Rock one oI the most beautiIul buildings in the world.
Style and methods oI construction changed Irom generation to genera-
tion, especially in respect to the material, the gates, the Iaades and minarets,
the profle oI the arches in the interior, and the ornamentation. But the ground-
plan oI the mosque remained largely the same. In fgure 8.1 two basic catego-
ries oI mosque design in distinctive regional style are shown:
3
on the leIt the
hypostyle hall and open courtyard, as Iound in the Arabian heartland (Syria,
Iraq, Saudi Arabia, Egypt, and Yemen), Spain and North AIrica, and on the
right, the bi-axial Iour-iwan type, as Iound in Iran and Central Asia, with Iour
iwans. Specifc elements are the minarets, cupolas, and arches.
Both Ior the role they play in respect oI the mosque and in their own right,
there are three felds in which Islam has made a unique contribution to archi-
tecture and the architectural arts: calligraphy, garden design, and geometry. The
use oI geometry in decoration is ubiquitous and serves to cover fat, curved
and convoluted surIaces in two- or three-dimensional Iorms. Thus it can enrich
and beautiIy an interior through the use oI uniquely Islamic devices such as
the muqarnas (stalactite vault). The muqarnas was an invention oI the Islamic
world. Although it can be oI structural value, associated with the transitional
zone oI a dome, more oIten it is a purely decorative Ieature connected to, or
suspended Irom, a structural member.
Thus the specifc elements characterizing Islamic architecture are the
muqarnas, and the qubba (cupola), as well as arches and vaults. These we want
to calculate. Another specifc aspect oI Islamic architecture, ornament, as Iound
in mosaics, tilings, geometrical writing, will not be treated in this paper. In
earlier papers I explained the calculation oI muqarnas and qubba, hence in the
present chapter the emphasis is on the calculation oI arches and vaults. These
calculations were not the basis Ior constructions but an appraisal oI the neces-
sary manpower and building materials.
1nr C:icui:+ors
Much is still unknown about the tradition oI practical mathematics in Islamic
civilization. In 1992 Rebstock published an important survey on calculation
in the Islamic Orient, based on a study oI more than 100 Arabic and Persian
Eigure 8.1
Two categories oI mosques, as Iound in the Arabian heartland (leIt) and in Iran and
Central Asia (right) (Erishman).
C:icui:+ix Surr:cr Arr:s :xiVoiuxrs ix Isi:xic Arcni+rc+urr 239
arithmetic texts, most oI which remain to be edited. This study has consider-
ably improved our insight into the practical use oI mathematics in the medieval
Islamic tradition.
Under the reign oI Caliph al-Mamn (813833) the mathematician and
astronomer Ab JaIar Muhammad b. Ms al-Khwrizm
4
wrote in Baghdad
his well-known Algebra, a work oI elementary practical mathematics. Its pur-
pose is, according to the author, to provide 'what is easiest and most useIul in
arithmetic, such as men constantly require in cases oI inheritance, legacies, par-
tition, lawsuits, and trade, and in all their dealings with one another, or where
the measuring oI lands, the digging oI canals, geometrical computations, and
other objects oI various sorts and kinds are concerned.
5
Only the small sec-
ond part oI al-Khwrizm`s treatise concerns practical mensuration.
6
He gives
rules Ior fnding the areas oI various plane fgures, including the circle, and Ior
fnding the volume oI a number oI solids, including the cone, the pillar with
a circular base, the pyramid, and the truncated pyramid. This section is really
concerned with the practical application oI mensuration, as the frst Iew lines
already demonstrate: 'Know that the meaning oI the expression one by one is
mensuration: one cubit (in length) by one cubit (in width) understood.
The earliest manual on practical arithmetic is a textbook by Abu`l-WaI
al-Bzjn,
7
written between 961 and 976. This Book on Settling What Is Nec-
essary from the Science of Arithmetic for Secretaries and Businessmen |K. al-
man:il f-m yahtfu ilayhi al-kuttb wa al-umml min ilmal-hisb|
8
consists
oI seven parts. The frst three are purely mathematical: on ratio, multiplication
and division, and mensuration. The other Iour contain the solutions oI practi-
cal problems: on taxes, exchange and shares related to the harvest, problems
concerning payment Ior work, and construction estimates.
9
The other manual
by Abu`l-WaI, 1he Book on What the Artisan Needs to Know of Geometric
Constructions |K. f-m yahtfu ilayhi al-sni min aml al-handasa| seems
to have been compiled Irom class notes by one oI his pupils. This practical
geometry outlines basic mechanical methods Ior constructing, proportionally
subdividing, and symmetrically extending geometric fgures, Iurther simplifed
by the use oI only a single opening oI the compass.
10
Interesting problems on mensuration are Iound in Ab Bakr Muhammad
al-Karaj`s
11
Sufhcient Arithmetic |Kf fl-hisb|, written between 1010 and
1016. The composition oI this treatise was part oI his duties as a mathematician
holding an oIfcial position: to write a simple textbook in a way accessible to
civil servants. In the introduction al-Karaj states that 'his work presents what
people oI diIIerent classes need Ior their various activities. Al-Karaj`s aim was
to write a practical manual, like al-Khwrizm and Abu`l-WaI beIore him, but
he does not deal with the important practical problem oI taxes and heritage. The
treatise contains the elements oI arithmetic with integers and Iractions (common
Yvoxxr Doii-S:xrioxius 238
and sexagesimal), the extraction oI square roots, the determination oI areas and
volumes, and elementary algebra. Chapters 44 to 49 deal with measuring plane
fgures. Chapter 50 is on the mensuration oI solids, in chapter 52 the amount oI
sun-dried or baked bricks needed Ior a construction is determined, and chapter
53 describes three instruments Ior levelling the ground.
12
Al-Karaj explains the
same three instruments in more detail in his treatise Locating Hidden Waters
|K. inbt al-miyh al-khahyya|.
13
Two recently studied texts are the Book of the Levels in the Explanation of
the Measurements |K. al-tabaqt f sharh al-misht| by Qd Ab Bakr,
14
writ-
ten beIore the end oI the twelIth century, and the Calculators Riches |Ghunyat
al-hussb| by Ahmad b. Thabt who died 1234.
15
Their calculations oI domes
concern hemispheric domes and are not Iaultless; in the case oI vaults their
calculations are simple and correct.
Ghiyth al-Dn Jamshd Masd al-Ksh ranks among the greatest
mathematicians and astronomers in the Islamic world. He was a master cal-
culator/ mathematician oI extraordinary ability, he applied iterative Iunctions
widely, he laid out his calculations in such a careIul way that errors could not
creep in undetected, and he did a running check at every stage. In short, his
talent to optimize a problem let him appear as the frst modern mathematician.
When Ulgh Beg decided to construct the Samarkand observatory, he invited
al-Ksh to his courtsome time aIter 1416. Al-Ksh died in June 1429 out-
side the observatory, probably murdered on the command oI Ulgh Beg.
16
Two
years earlier he had fnished the Key of Arithmetic |Mifth al-hisb|, one oI his
major works. The work is intended Ior everyday use, as is clear Irom al-Ksh`s
remark: 'I redacted this book and collected in it all that is needed Ior the one
who calculates careIully, avoiding tedious length and annoying brevity. By Iar
the most extensive part is Book IV, On Measurements.
The main aim oI the last chapter, Measuring Structures and Buildings,
is practical: 'The specialists merely spoke about this way oI measuring Ior the
arch and the vault, and did not think that anything else was necessary. But I
present this application among the necessities together with the rest, because it
is more oIten required in measuring buildings than in the rest. Al-Ksh uses
geometry as a tool Ior his calculations, not Ior constructions. Besides arches/
vaults and qubbas, al-Ksh calculates here the surIace area oI a muqarnas,
that is to say, he establishes approximate values Ior such a surIace. He is able
to do so because a muqarnas, although it is a complex architectural structure, is
based on relatively simple geometrical elements. Only elementary geometrical
rules are used in the calculation.
To summarize: In the extant Arabic literature we fnd several calcula-
tions oI arches and vaults as well as treatises on the measurement oI the qubba.
The most accomplished explanations and calculations, however, are those oI
al-Ksh, who also considers more elaborate Iorms oI arches and domes than
the others. His skill in fnding practical approximations is equally shown in
measuring the muqarnas. No other treatise on the calculation oI a muqarnas
surIace has yet been Iound.
C:icui:+ix Arcnrs :xiV:ui+s
Erom about 200 BC to AD 1100 the main structural components oI all large
buildings in the Roman and Byzantine Empires were the semicircular arch
(the barrel vault is an elongated arch and the dome is a rotated arch), columns
and walls.
17
Erom the Byzantine Empire the Umayyads inherited the system
oI round arcading which rarely showed a tendency towards becoming slightly
pointed.
18
The innovation oI the pointed, or ogival, arch came Irom the East.
The ogival arch is frst Iound in Buddhist India in the second century AD
and had reached Syria, possibly by way oI Sassanid Iran, by AD 561. A number
oI such arches appear in Syrian buildings oI the eighth century and it became
common in Egypt in the ninth century. A beautiIul early example is the mosque
oI Ibn Tln in Cairo, built in 876879. Under Umayyad rule the round arch
persisted, but developed into the two-centered Iorm showing an increasing
tendency towards pointedness. A round arch is struck Irom a single center; a
pointed arch has more than one center and can be thought oI in its simplest
Iorm as being struck Irom two centers with overlapping arcs which produce an
increasingly pointed arch as they are moved apart horizontally (fgure 8.2). In
the great mosque at Damascus, built in 715, the arches are very slightly pointed
with the two centers being only one-eleventh oI the span apart.
In the succeeding two centuries the same trend is still apparent, but
is complicated by the three- and Iour-centered arch. Eigure 8.4 shows a three-
centered arch, with point Eas a (double) center and the other two centers situated
in the two lower points Z and H. These two points are obtained by intersecting
the extensions oI the radii through E with the perpendiculars dropped Irom the
extremities oI the span oI the arch. When the two lower centers move, the arch
will change its acuteness. The Iour-centered arch is similar to the three-centered
one except that the (double) center splits into two points, which are displaced
Irom the center oI the frst circle toward the extremes. The smaller the displace-
ment oI the two centers, the closer the profle will be to the profle oI the three
centered arch. The greater the displacement, the shallower the profle.
Since ogival arches are the exception in the frst centuries oI Islamic
architecture, only the calculation oI semicircular arches is Iound in the early
arithmetic books. This practice continued well into al-Ksh`s time, probably
because most manuals were based on earlier manuals. Al-Ksh remarks that,
'Our predecessors determined those (i.e., arch and vault) as halI a circular hol-
low cylinder, but we did not see anything like it, either in old or in new build-
ings. We have mostly seen ones that are pointed in the middle, and in a Iew
cases they are smaller than halI a hollow cylinder.
Among these predecessors we fnd Ahmad b. Thabt (d. 1234), who cal-
culates arches as Iollows
19
in fgure 8.3; all given and calculated quantities are
written out in words |MS AyasoIya 2728, Iol. 117b|:
The authors oI these computations use 'arch and 'vault interchange-
ably. The diIIerence between an arch and a vault is that the depth oI an arch is
not greater than its span, but in the case oI the vault it is greater. The distance
between the two Iaades is called the depth oI the arch. What we call in the arch
its depth, is called in the vault its length.
Ahmad b. Thabt discusses the Iollowing calculation (fgure 8.3): When
a vault is said to have an exterior curve oI |u
1
| 20d |where d stands Ior
one dhir, i. e., one l|, and an interior curve oI |u
2
| 12d, and the distance
between the two curves amounts to |b | 2d, and its length to |l | 50d, then
how much is its surIace area and its volume?
Solution: Ahmad b. Thabt considers the surIace area oI the vault as the
sum oI 'plane surIaces. These consist oI the visible surIaces inside and outside
Eigure 8.2
Diagram showing pointed arches Iormed with constant radii on centers with successive
separation oI one tenth, one seventh, one fIth, and one third oI a span (Warren).
the vault together with the surIaces at both ends, plus the invisible ones, namely,
the two surIaces which support the vault. Thus we have:
Outside surIace area: l u
1
50d 20d 1000d
2
,
Inside surIace area: l u
2
50d 12d 600d
2
,
Two ends: 2
u
1
u
2
2
b (20d 12d) 2d 64d
2
,
Two supports: 2 l b 2 50d 2d 200d
2
, thereIore the total surIace
area equals 1864d
2
.
20
The volume oI the vault is obtained by multiplying the surIace area oI one
end oI the vault by the length oI the vault:
u
1
u
2
2
b l 16d 2d 50d 1600d
3
.
The same method is applied by the later mathematician Ibn al-Hanbal (d.
1564) in his Book on Measurements |Kitb al-misha|, sometimes considered
as a comprehensive commentary to Ahmad b. Thabt`s work.
21
The calculation oI the volume oI a wreath or discus is already Iormulated
by al-Karaj as Iollows |chapter 50, section 4|: 'Multiply halI the sum oI the
outer and inner circumIerence by the width oI the solid to be measured and
multiply this product by the length.
This Iormula is identical to the one applied above. In a barrel vault with
radii r
1
> r
2
, the circumIerences oI the outer halI circle and the inner halI circle
are ar
1
and ar
2
, and the surIace areas oI the outer and inner circle are ar
1
2
and
ar
2
2
.
Hence the surIace area oI the halI ring is
Eigure 8.3
Calculation oI a barrel vault.

1
]
1
1
+

( )

+

Multiplying this product by the length (width) yields the Iormula
u
1
u
2
2
b l.
The vault or arch are not mentioned by al-Karaj in this context.
How does al-Ksh deal with these matters? At frst he explains at length
the diIIerent elements oI an arch and how these are connected and which part
could disappear into a wall. This is Iollowed by fve methods Ior drawing the
Iaade oI an arch.
22
The frst two are three-centered arches, the third is a Iour-
centered arch, the Iourth and fIth are two-centered. The Iourth method is the
one shown in fgure 8.2 on the lower right, its two centers being separated by
one third oI the span. According to al-Ksh, the second Iaade was the most
common in his time; he thereIore uses it to illustrate his calculation method.
This kind oI Iaade is convenient, as al-Ksh remarks, when you need a span
oI fve to ten, or up to fIteen cubits.
Construction oI the second Iaade (fgure 8.4 is taken Irom the oldest
extant ms. with Roman letters added):
draw a semicircle on diameter AD , the span oI the arch
extend AD at both sides by the thickness oI the arch to the points I and M. E is the
center oI the semicircle;
divide this semicircle into Iour equal parts through the points A, B, C, G, D
extend BE and GE by EZ and HE, equal to AC, and by BK and GL, equal to DM,
the thickness oI the arch
on center E draw the arcs ML and KI, on center H the arc G1, and on center Z
the arc B1
connect H1 and Z1 and extend them by the thickness oI the arch to the points O
and S
draw arc LO on center H and arc KS on center Z
erect the perpendiculars SN and ON on the lines 1S and 1O
The sections AK, K1, 1N, 1L, and LD Iorm together the Iaade oI the arch
with 1N as the keystone.
When we construct area AFQD with parallel sides and right angles, we
obtain the spandrels oI the arch, the sections tNQ and JNF. Section tMD as well
as section JIA could disappear into a wall.
AIter al-Ksh has explained and carried out all fve methods Ior con-
structing the Iaade oI an arch and has completed the characterization oI the
arch and the vault, he continues by surveying them. He explains that he has
calculated Iactors relating some measurements oI an arch to its span and to its
thickness. He has put these Iactors down in a table together with an explanation
oI the method. These Iactors have also been transIormed into Indian numerals
and been added to the table. He then gives an example oI how to calculate with
these Iactors. This is Iollowed by a detailed account oI how these Iactors were
accomplished.
23
Example, How to use the table:
Al-Ksh assumes (fgure 8.4) the span AD oI the second Iaade to be equal to
20 and the width DM oI the arch to be equal to 5. I call the exterior curve u
1
the line oI convexity and the interior curve u
2
the line oI concavity, and b is the
width oI the arch at its base. Al-Ksh does not render the calculations but gives
only the rounded results. I now list the approximations given by al-Ksh:
column 1: With this Iactor al-Ksh calculates the interior curve :
AB1GD 1.651 AD 1.651 20 33.02 = 33.
column 2: Multiplying this Iactor with the width al-Ksh obtains halI the
diIIerence between the exterior and the interior curve. Adding this amount to
the interior curve, Iound by means oI column 1, he obtains halI the sum oI the
Eigure 8.4
Construction oI the second Iaade in a ms. oI al-Kashi`s treatise.
exterior and interior curve. Multiplying this quantity with the width oI the arch
he obtains the surIace area (A) oI the Iaade:
1.599 b
u
1
u
2
2
;
S
u
1
u
2
2
u
2
D
b
u
1
u
2
2
b A
In our practical example: u
1
IJNtM; u
2
AB1GD;
u
1
u
2
2
1.599 5 = 8;
u
1
u
2
2
u
2
8 33 41; A 41 5 .
column 3: Inner height oI the arch:
E1 AD 0.598 20 0.598 11.96 = 12.
column 4: Upper width oI the arch:
1N AI 1.099 5 1.099 5.495 = 5.5.
Table to calculate arches and their parts.
column 5: To obtain the surIace area oI the concavity, AB1GDE, al-Ksh
multiplies the square oI the span by 5 and divides by 12. This calculation is
equivalent to multiplying with the value Iound in the table, 0.419, or 0;25,9,13
= 5/12.
With these values we can now calculate many diIIerent parts oI the arch:
To calculate the volume oI the arch we proceed in the same way as Ahmad
b. Thabt above: AIter the surIace area oI the arch has been Iound, by means oI
the second column oI the table, we multipy this number with the depth oI the
arch and obtain its volume.
Sometimes the arch disappears partly inside a wall and we want to know
how much is visible and how extensive the segments inside the wall are, namely
section tDM and section JAI: These segments are calculated by taking the diI-
Ierence oI the circle segment MtE and the triangle tDE, which can be Iound
Irom MD and ED.
When we subtract this amount Irom the total surIace area oI the arch we
obtain the surIace area oI the visible part of the arch.
It might be necessary to calculate the spandrels, section NQt and section
NFJ: In this case we calculate the area AFQD and subtract Irom this amount the
area oI the visible part oI the arch, calculated above, and the area oI the opening
oI the arch, area AB1GDE, Iound by means oI the fIth column. The diIIerence
gives the surIace area oI the spandrels. When we multiply this amount with the
depth oI the spandrels, we obtain the volume oI the two spandrels.
As explained above, al-Ksh`s book is Ior practical use. Hence he rightly
shows how to make liIe easier by working with rounded values. When arches
other than the fve models given by al-Ksh are involved, approximation again
is used, and one takes the model closest to the required arch. Golombek and
Wilber
24
have considered existing examples oI Timurid arches in the order
outlined by al-Ksh. Examples have been recorded Ior all but the fIth model,
which was, however, common in small windows. In comparing the models
described by al-Ksh with actual examples oI Timurid arches we have to bear
in mind that al-Ksh`s purpose was to calculate volumes and surIaces, not to
construct them. This means that an elegant approximation, which leads us to an
easy calculation, is the ultimate goal.
Bulatow
25
has analyzed arches Irom the twelIth to fIteenth centuries in
Central Asia and he suggests that some pointed arches were constructed as
intersections oI ellipses. Eor he notes that Ior spans exceeding 10 m. these were
easier to construct than Iour-centered arches. The architects were, in addition,
Iamiliar with the stability oI the ellipse, because its construction was known
Irom Sasanian examples. According to his analysis, this kind oI arch is Iound in
some oI the most important Timurid buildings oI the period, such as the Gur-i
Amir in Samarkand, the mausoleum oI Tamerlane and Ulgh Beg, namely in
its dome, interior niches, arches in zones oI transition, and entrance portal. The
same arches have elsewhere been identifed as three- and Iour-centered arches
and can be considered as such Ior all practical purposes.
Al-Ksh does not mention an elliptical profle either Ior arches or Ior
domes. There are a number oI domes Ior which the profle may be interpreted
as the intersection oI symmetrical elliptical arcs. Bulatow has demonstrated
that the dome oI the Gur-i Amir was probably designed by using a pair oI Ioci
and a string. However, an analysis oI the Gur-i Amir (fgure 8.5) suggests that
this dome could also have been originated by the Iourth method (fgure 8.2,
Iar right): With line AD as the span and the points B and Z dividing the span
into three equal parts we obtain the circle segments just inside the curve drawn
by Bulatow. The diIIerence between the two curves lies within the error range
accepted by modern architects. The ellipse may be easier to construct but circu-
lar segments are easier to calculate. Eor calculating an elliptical dome, like the
dome oI the Gur-i Amir, al-Ksh`s Iactors are excellent.
The section on calculating arches ends in al-Ksh`s Key of Arithmetic
with the Iollowing remark (my translation) 'I have talked a lot about the subject
oI this section, as this section is very important, and my predecessors did not
treat it as they should have done.
Eigure 8.5
Dome oI the Gur-i Amir (Bulatow`s analysis with additions).
1nr Quii:
Besides mosques and madrasasthe second important religious building
monasteries, drinking Iountains and elementary schools, the mausoleum, or
qubba, is an important Islamic monument. Erom the earliest times most qubbas
were oI the same style: a cubic room on a square base with a vaulted rooI. The
problem oI erecting a cupola on a square base fnds in Muslim architecture a
large variety oI solutions. The space oI transition Irom the square, or polygon,
to the circle is occupied either by planks covered with stucco, or by corner-
trompes made at frst oI bricks and later oI stone, or by stalactite pendentives,
called muqarnas. The Arabic word Ior dome or cupola is qubba pl. qibb or
qubab. By extension qubba also means: cupolaed structure, dome-shaped edi-
fce; domed shrine, memorial shrine (esp. oI a saint). At the end oI this section
we shall see that the term parabolic qubba |al-qubba al-mukhya| is used in a
purely mathematical sense to denote a paraboloid.
Calculation oI the qubba Iollows the same pattern as that oI vaults: In
the common arithmetic texts only hemispheric domes are calculated. Al-Ksh
however, gives an elegant and precise method to calculate a dome originated
by turning the Iaade oI an arch around its axis. His exposition is based on his
Iourth model, that is, on a two-centered arch with the two centers one-third oI
the span apart (fgure 8.2, below right). Applying al-Ksh`s Iactors, excellent
results can be obtained with an error range oI less than 3 percent. 'To simpliIy
the procedure as al-Ksh says, he does not explain how he arrives at these
Iactors. Eor a practical application the rules are enough on their own. The Iol-
lowing question arises: Why is al-Ksh so concise in the case oI domes, when
he uses such elaborate calculations Ior surveying arches? Could these Iactors
be used Ior several kinds oI domes, which deviated to varying extents Irom the
model discussed by al-Ksh?
In all extant Arabic treatises the qubba is assumed to consist oI a solid
shell between two concentric surIaces. In practice, however, the inner and outer
surIaces oI the shell are never really parallel, because pressure occurs in the
lower part (up to 61) and pull in the upper part. Al-Ksh does not carry out the
calculation oI the hemispheric qubba but reIers to his calculation oI the sphere.
There he uses, as expected, the right Iormulas Ior area and volume, expressing
a as the ratio between the circumIerence and the diameter oI a circle. Abu`l-
WaI, who is the frst to mention the measurement oI the qubba, also gives the
right defnitions and calculation method, mentioning Archimedes.
26
Erom then
on the calculation oI volume and surIace areas oI hollow hemispheric qubbas
Iorm a regular part oI practical arithmetical textbooks, although a wrong 'Ior-
mula Ior the calculation oI the volume Irequently occurs.
When the inner and outer diameter oI a hemispheric qubba are known, its
volume and the inner and outer surIace areas can be calculated Irom the volume
and the surIace area oI a sphere.
The correct Iormulas Ior the area (A) and the volume (J) oI a sphere with
diameter 2r are:
A (2r)
2
a; J r
A
3

4
3
ar
3
.
In Islamic mathematics a is usually expressed as: discard one seventh and
halI oI one seventh and multiply with Iour:
a
S
1
1
7

1
2

1
7
D
4
11
14
4 3
1
7
.
Mathematically, it is more diIfcult to prove a Iormula Ior calculating the
surIace area oI a sphere than Ior the volume. However, in the common arithme-
tic books the computation oI the inner and outer surIace areas oI the qubba is
based on the correct Iormula:
A (2r)
2
S
1
1
7

1
2

1
7
D
4
The volume oI the qubba, on the contrary, is usually based on the Iollow-
ing wrong Iormula:
J (2r)
3
S
1
1
7

1
2

1
7
D

S
1
1
7

1
2

1
7
D
Comparing the two results calculated by means oI the correct and the
erroneous Iormula, the diIIerence (D) is:
D (2r)
3
,
HS
11
14
D
2
a
6
J
, and with a 3
1
7
,
D (2r)
3
S
121
196

22
42
D
0.0935(2r)
3
.
This means that the calculated volume is 17.86 percent(!) more than the
correct volume, independent oI the size oI the diameter. This is especially sig-
nifcant when the architect is paid according to the measurements oI the build-
ing (see conclusion).
Even al-Karaj in Sufhcient Arithmetic gives this wrong Iormula, but he
adds a second method, in which the volume oI a sphere is calculated by mak-
ing a waxen model:
27
'You take a rectangular solid, made out oI wax and with
its three dimensions oI equal length, and weigh it. Iet its weight be 30 dirham.
Now you make Irom this (material) a sphere, as perIect as possible, with the
diameter equal to the dimension oI the solid. The weight oI this sphere is Iound
to be a little less than 18
2
3
.
Hence, the diameter is raised to the third power and then
1
3
and
1
9

2
5
is subtracted.
The diIIerence between the two results is very small.
This implies the Iormula:
J (2r)
3
S
1
1
3

1
9

2
5
D
Although this sounds like practical mathematics, it gives a result which is
even worse than that oI the frst methodbeing oII by almost 20 percent! Thus
one wonders whether the experiment was made at all.
Muhammad Bah al-Dn al-mul (15471622) says in his Essence of
Arithmetic (chapter 6, section 3, Mensuration of Solids, ca. 1600),
28
'To mea-
sure the sphere, multiply halI its diameter with one third oI its surIace area;
or, subtract three Iourteenths Irom the cube oI its diameter, and again Irom the
remainder, and again Irom the remainder.
Hence two Iormulas are given:
J r
A
3
, the correct mathematical Iormula, or,
J (2r)
3
S
1
3
14
D

S
1
3
14
D F
S
1
3
14
DG
,
the Iormula Ior practical use.
Here we fnd the two Iormulas side by side, each in its own right, the
mathematically correct Iormula and the common practical Iormula. The brack-
ets have been added, because I have assumed, with only Nesselmann`s edition
at my disposal, that the second 'and again Irom the remainder is due to the
scribe.
29
Whether Bah al-Dn or the scribe made this change, this Iormula
gives a result that is out oI the question, the calculated Iormula being 7.4 per-
cent less than the correct volume. But who wants this? Throughout the centu-
ries constructors, architects, and artisans, probably profted by using the wrong
Iormula.
In her commentary to the Topkapi scroll Necipoglu reasons:
30
It is unIortunate that Ibn al-Haytham`s Book of Buildings and Constructions
|K. al-Abniya wal-uqd| and al-Karaj`s Book of Architectural Construc-
tions |K. Uqd al-abniya| have not survived. The latter, which dealt with the
construction oI buildings, bridges, and engineering, was described by Ibn al-
AkIn as a work exclusively based on practical geometry. Ibn al-Haytham`s
lost work, on the other hand, was reIerred to by the physician and biographer
Ibn Ab Usaybia (12031270) as a 'treatise on the construction oI ditches
and buildings, with all the fgures oI geometry joined together, a work con-
cluding with the three conic sections, namely, the parabola, hyperbola, and
ellipse. These lost treatises relying on practical geometry, thereIore, appear
to have been illustrated with geometric constructions that included conic sec-
tions, constructions probably adapted to simplifed mechanical procedures.
Ibn Khaldn testifed to the use oI conic sections in architecture and
those craIts dealing with bodies, 'Conic sections are a branch oI geometry.
This discipline is concerned with the study oI fgures and sections occurring
in connection with cones. It proves the property oI cones by means oI geo-
metrical prooIs based upon elementary geometry. Its useIulness is appar-
ent in practical craIts that have to do with bodies, such as carpentry and
architecture. It is also useIul Ior making remarkable statues |monuments|
and large objects |haykil, 'eIfgies or 'edifces| and Ior moving loads
and transporting large objects |haykil| with the help oI mechanical contriv-
ances, engineering |techniques|, pulleys and similar things.. . .
Treatises on the mensuration oI parabolas and paraboloids were writ-
ten by such mathematicians as Thbit b. Qurra (836901), Ibrhm b. Sinn
(909946), al-Sijz (2nd halI 10th c.), Ab Sahl al-Qh (f. 980100), and
Ibn al-Haytham (died c. 1040), some oI them including sections on arches,
vaults and domes. Al-Sijz, Ior example, wrote a work exclusively dealing
with the mensuration oI domes, entitled Epistle about the Characteristics of
Hyperbolic and Parabolic Domes |R. f khawss al-qubba al-:ida wal-
mukhya|.
The many Arabic treatises on conic sections, written aIter Apollonius
oI Perga`s Conica was translated, oIten deal with practical application. Their
contribution to architectural practice (particularly designing pointed arches,
vaults and domes) and to the decorative arts awaits assessment by historians
oI science so that the ways in which theory and praxis interacted can be
understood more clearly.
So Iar, the inspection oI the above-mentioned treatises has not confrmed
that they were Ior practical use. Jan Hogendijk has looked at al-Sijz`s treatise
on the hyperbolic and parabolic domes. According to him, the work is on pure
mathematics only, with no connection to architecture. The other treatises have
been studied by Suter
31
and recently edited by Rashed.
Thbit b. Qurra`s
32
long treatise On Measuring Paraboloids |F mishat
al-mufassamt al-mukhya|
33
deals with solids created by rotating diIIerent
sections oI a parabola. In proposition 36, the last proposition oI the treatise, he
proves the Iollowing property oI the 'parabolic qubba (fgure 8.6, top), that is:
The volume oI every parabolic qubba is equal to halI the volume oI the cyl-
inder, with as base the circular base oI the qubba, iI the qubba has a regular
vertex, or oI the lower basic circle, iI the qubba does not have a regular ver-
tex, and with the height equal to the axis oI the qubba.
Ab Sahl al-Qh
34
wrote a much shorter treatise on the subject Deter-
mining the Jolume of the Paraboloid |F istikhrf mishat al-mufassam al-
mukf|.
35
Here he criticizes the 'Iamous and well-known treatise by Thbit
b. Qurra, 'the only one on the subject, saying that it is 'too voluminous and
long, there are about Iorty propositions, numerical, geometrical and others,
there are all these lemmata Ior just one proposition: how to know the volume
oI a paraboloid.
36
Al-Qh expounds that examining Thbit`s treatise, he Iound
it very diIfcult to understand, whereas Archimedes` treatise On the Sphere and
the Cylinder seemed much easier to him. Thinking that many people might
have gained the same impression, he Ieels compelled to determine the volume
oI the paraboloid aIresh. He is able to do this by using a method that is easily
understandable and does not need any lemmata. Al-Qh proves in only three
propositions that: 'Every paraboloid is equal to halI its cylinder.
Al-Qh`s treatise was criticized by Ibn al-Haytham
37
in his 1reatise on
Measuring the Paraboloids |Maqla f mishat al-mufassamt al-mukhya|.
38
In an almost philosophical introduction he writes: 'The person who pronounces
or composes any discourse or essay has a motive which leads him to say what
Eigure 8.6
Parabolic domes aIter Ibn al-Haytham (Suter).
he says or write what he writes. We examined with great care Thbit b. Qurra`s
treatise On Measuring Paraboloids. We Iound that he Iollowed a road devoid oI
any plan, and that his way oI explaining was long and involved painIul diIfcul-
ties. We then got hold oI a treatise on Determining the Jolume of the Parabo-
loid by Ab Sahl al-Qh. We Iound it bare and concise and we learned that,
according to the author, the reason he did his research and wrote the treatise,
was the diIfculties encountered in Thbit b. Qurra`s treatise. However, we have
established that Ab Sahl`s treatise, although easier and simpler, only contains
the prooI oI the measurement oI one oI the two kinds oI paraboloids. (In Iact,
Thbit b. Qurra`s treatise too discusses only the frst kind oI paraboloid.)
Ibn al-Haytham continues by saying that there are two kinds oI parabo-
loids, one is comprehensible and easy, the other diIfcult and painIul. His prooI
Ior the frst kind oI paraboloid needs several lemmata and is more detailed than
al-Qh`s demonstration. Then he turns to the second kind oI paraboloid gener-
ated by rotating the parabola around its ordinate (fgure 8.6), and proves in an
elegant way that: 'The paraboloid generated by rotating the parabola around its
ordinate is equal to one third plus one fIth oI the circumscribed cylinder.
Both Iormulas are easy and exact, so they could well be used Ior practical
calculations but they do not provide the surIace areas. Also, we have not yet
Iound traces oI these easy Iormulas in arithmetic manuals. Until we fnd these,
we should consider the above results on paraboloids as highschool mathemat-
ics. Another serious question is, were parabolic qubbas ever built? The parabo-
loid, shown in fgure 8.6, above, would certainly make a beautiIul qubba, but
where do we fnd an example oI a parabolic qubba? Elliptical arches and qub-
bas existed in pre-Islamic times. Exact Iormulas Ior ellipsoids are more diIfcult
to develop, but, as we have seen in the section on arches, their surIace areas and
volumes can be approximated. However, how these structures were calculated
beIore al-Ksh established his Iactors, remains to be looked at.
These treatises on conic sections are oI great interest Irom the point oI
view oI the history oI mathematics but were probably never used by artisans.
They do not give the impression oI being meant Ior practical use.
One important question remains: why are the above-mentioned treatises
by Ibn al-Haytham and al-Karaj lost? Were they never studied, and thereIore
not copied? Or, were they used so intensively, that the soiled and worn out cop-
ies were thrown away? During the Dibner conIerence in November 1998, David
King remarked that the extant astrolabes might well be those that were never
used. Could this apply similarly to treatises on construction or containing orna-
ments, like the Topkapi scroll?
Arrroxix:+ix +nr Muq:rx:s
The muqarnas (fgure 8.7) developed around the eleventh century in Iran and
North AIrica. Whether or not the two developments are related, two points
about these new Iorms must be made. One is that, Irom the late eleventh cen-
tury onward, all Muslim lands adopted and developed the muqarnas, which
became almost as common a Ieature oI an elevation as the Corinthian capital
was in Antiquity. The second and Iar more important point is that, Irom the
moment oI its frst appearance, the muqarnas acquired Iour characteristic
attributes whose evolution and characteristics Iorm its history: it was three-
dimensional and thereIore provided volume wherever it was used, the nature
and depth oI the volume being leIt to the discretion oI the maker; it could be
used both as an architectonic Iorm, because oI its relationship to vaults, and as
an applied ornament, because its depth could be controlled; it had no intrinsic
limits, since not one oI its elements is a fnite unit oI composition and there is
no logical or mathematical limitation to the scale oI any one composition; and
it was a volume that could be a solid or a void, a projecting mass oI complex
shapes or a complex outlinea three-dimensional unit which could be resolved
into a two-dimensional outline.
39
The muqarnas was used in domes, in niches, on arches, and as an almost
fat decorative Irieze. In each instance the module as well as the depth oI the
composition is diIIerent and adapts to the size oI the area involved or to the
required purpose. In ceilings it serves a clear architectonic aim, or at the very
least provides the structural illusion oI ascending movement culminating in a
small cupola. The muqarnas is at the same time a linear system and an organi-
zation oI masses. Despite occasional textual reIerences to plans, there are no
known Islamic architectural working drawings Irom the pre-Mongol era. The
earliest known example oI a construction plan is a 50 cm. stucco plate show-
ing the projection oI a quarter muqarnas vault which was Iound at the Takht-i
Suleiman in Iran. This plate Irom the 1270`s was the basis Ior Ulrich Harb`s
reconstruction oI the collapsed vault. Eourteenth-century sources Irequently
mention architectural drawings produced either on clay tablets or on paper.
In the fIteenth-century Timurid world drawings seem to have been more
widely used than beIore. Their extensive use had become essential because
oI the increasing intricacy oI the geometric design. Up until Necipoglu`s dis-
covery oI the Topkapi scroll the earliest known examples oI such architectural
drawings were a collection oI Iragmentary post-Timurid design scrolls made
oI sixteenth-century Samarqand paper, preserved at the Uzbek Academy oI
Sciences in Tashkent. These scrolls almost certainly refect the sophisticated
Timurid draIting methods oI the fIteenth century. In 1876 the English architect
C. Purdon Clarke brought back Irom Teheran some scrolls and working draw-
ings Irom the eighteenth and the nineteenth centuries that he had collected Iol-
lowing the death oI the oIfcial state architect, Mirza Akbar; these scrolls are
now preserved in the Victoria and Albert Museum. In 1981 similar material,
still in the hands oI the master-artisan, was examined by W. K. Chorbachi in
two Arab towns. The collection oI these scrolls was not only the basic reIerence
manual but also served as a design book. A Iew years ago I visited a workshop
at Eez/Morocco, where the artisans used a construction-plan Ior a muqarnas
on a scale oI 1:1. The pieces cut out Ior constructing the muqarnas could
actually be put on the draIt such that the cross section oI the element, i.e., the
cross section oI the beam, matched the fgure in the draIt exactly. The Timurid
scrolls show a decisive switch to the Iar more complex radial muqarnas, with
an increasing variety oI polygons and star polygons. Also the Akbar scrolls are
more elaborate than the twentieth-century Eez drawing. Despite their simplic-
ity, however, the more recent scrolls testiIy to a relatively unbroken tradition
oI architectural practice in the Islamic world Irom at least the Timurid period
Eigure 8.7
Curved Muqarnas with its Plane Projection.
onward. A continuous tradition Irom the thirteenth-century Takht-i Suleiman
plate to the twentieth-century Eez drawing is evident. As in the Ilkhanid period,
700 years earlier, the plane projection oI the elements in the Moroccan plan
consists oI simple geometrical fgures: squares, halI-squares, rhombuses, halI-
rhombuses, rectangles, almonds, bipeds (fgure 8.7).
The contents oI the Topkapi and Tashkent scrolls support the commonly
expressed view that the key to TimuridTurkmen architecture lies not only in its
Iascination with complicated vaulting systems but also in its extensive surIace
decoration. The numerous two-dimensional geometric patterns and epigraphic
compositions in these scrolls condense complex compositions into shorthand
Iormulas meant to act as guidelines Ior the simpler working methods employed
on the construction site. They thus provide a valuable glimpse into the pro-
cesses oI design and execution. The contents oI the Topkapi scroll, which
resemble those oI a pattern book, can be seen as an index oI the unprecedented
Timurid-Turkmen emphasis on surIace decoration, an emphasis that turned the
fat Iaades oI buildings into stage props Ior the display oI virtuoso ornamental
panels and Iragmented vaults into multiIaceted compartments with no struc-
tural role. The Topkapi scroll refects a 'painterly aesthetic oI architecture
inIormed by the cultural prestige oI drawings on paper.
Mohammad al-Asad describes
40
the muqarnas as:
. . . a vaulting system based on the replication oI units arranged in tiers, each
oI which supports another one corbeled on top oI it. The fnal result is a stair-
like arrangement that is sometimes reIerred to as honeycomb or stalactite
vaulting. The units are made oI wood, brick, plaster, or stone and can be
painted, or, as in the case oI the brick or plaster ones, covered with glazed
tiles. Muqarnas compositions can be located in diIIerent parts oI a building,
articulating a column capital, supporting a minaret`s balcony, or vaulting
over an entry portal, niche, or hall. Muqarnas vaults are usually part oI a
doubleshell arrangement and are thereIore visible only Irom the inside oI
a building. In some cases, as in the mausoleums oI Nr al-Dn in Damascus
(1172) and Imam Dur in Samarra (circa 1085), the muqarnas is also refected
on the outside.
Al-Ksh defnes the muqarnas in his practical way as Iollows
41
(fgure
8.7):
The muqarnas is a ceiling like a staircase with Iacets and a fat rooI.
Every Iacet intersects the adjacent one at either a right angle, or halI a
right angle, or their sum, or another combination oI these two. The two
Iacets can be thought oI as standing on a plane parallel to the horizon.
Above them is built either a fat surIace, not parallel to the horizon, or
two surIaces, either fat or curved, that constitute their rooI. Both Iacets
together with their rooI are called one cell. Adjacent cells, which have
their bases on one and the same surIace parallel to the horizon, are called
one tier.
Hence the elements oI a muqarnas consist oI cells and intermediate ele-
ments, connecting the rooIs oI two adjacent cells.
The plane projection oI an element consists oI simple geometrical Iorms,
called by al-Ksh squares, halI-squares (cut along the diagonal), rhombuses,
halI-rhombuses (isosceles triangles with the shorter diagonal oI the rhombus
as their base), almonds (deltoids), and large and small bipeds. Rectangles also
occur. To make these elements ft, they have to be constructed according to the
same unit oI measure. Al-Ksh uses in his computation the so-called module
oI the muqarnas, defned as the base oI the largest Iacet, that being the side oI
the square.
On this Ioundation he calculates the surIace area oI the muqarnas. In the
Iollowing I shall only sketch his methodthe Iull explanation can be Iound in
two Iormer papers.
42
Drrixi+ioxs
A Iacet oI a cell is a vertical side.
A rooI oI a cell is a surIace, not parallel to the horizon, or two joined surIaces,
either fat or curved.
A cell consists oI two Iacets plus their rooI.
An intermediate element is a surIace, or two joint surIaces, connecting the rooIs
oI two connecting cells.
An element is a cell or an intermediate element.
A tier is a row oI cells, with their bases on the same surIace parallel to the hori-
zon.
The module is defned as the base oI the largest Iacet, that being the side oI the
square. It is the measure-unit oI the muqarnas and is equated with 'one.
A rhombus is a parallelogram with all sides equal to the module and with the
acute angles equal to 45.
A halI-rhombus is an isosceles triangle with an angle oI 45 at its vertex and with
a base equal to the shorter diagonal oI the whole rhombus.
A rhomboid is a parallelogram with two opposite sides equal to the module and
with the acute angles equal to 45.
An almond is a quadrilateral with two opposite right angles, an acute angle oI 45,
and the two sides adjoining at the acute angle equal to the module.
A small biped is the complement oI an almond to a rhombus.
A jug is a quadrilateral with two opposite angles oI 6730, a right angle, and the
two sides adjoining at the right angle equal to the module. It consists oI two halI-
rhombuses connected along the sides equal to the module.
A large biped is the complement oI a jug to a square.
A barley-kernel is a quadrilateral with two opposite equal obtuse angles, a right
angle, and the two sides adjoining at the right angle equal to the module.
Al-Ksh distinguishes Iour types oI muqarnas: The simple muqarnas,
the clay-plastered |mutayyan| muqarnas, both have plane Iacets and rooIs. In
the curved muqarnas and the Shrz muqarnas the rooIs oI the cells and the
intermediate elements are curved.
The simple muqarnas is the one in which the underlying elements in the
plan are simple. The surIaces oI the cells` Iacets in the lowest tier are based on
a square, a rhombus, or a rhomboid; in the other tiers they have an additional
almond or halI-rhombus, and their rooIs are shaped like squares, halI-squares,
rhombuses, halI-rhombuses, almonds, bipeds, or barley-kernels. The barley-
kernels do not occur except on the upper tier. All these elements had been
treated earlier in chapter two oI the same book IV, which is entitled, Measuring
Quadrilaterals and What Is Connected with It.
To measure the area oI the simple muqarnas, called minbar-like by the
masons, al-Ksh tells us to proceed tier by tier. Al-Ksh computes the area oI
the Iacets oI a tier by multiplying the sum oI the bases oI the Iacets oI the tier by
the height, which Ior the simple muqarnas is in most cases equal to the module.
The next step is to count all the surIaces oI the rooIs oI the cells. Their sum plus
the area oI the Iacets is the area oI one tier. Then he adds all the tiers to obtain
the surIace area oI the muqarnas. Al-Ksh continues: 'II we measure the sur-
Iace on which the muqarnas is constructed, we would obtain the surIace area
oI the entire rooI oI the muqarnas. This means that Ior al-Ksh the muqarnas
is an ornament, a decoration. We already have a rooI, or a vault, on which the
muqarnas is constructed. The surIace area oI the entire structure consists oI the
muqarnas plus the rooI or vault on which the muqarnas is constructed.
Note: Al-Ksh gives in his calculations the sexagesimal as well as the
decimal values in order to avoid errors in later copies oI his work. He must
have calculated in the sexagesimal system, as these are the only numbers Iound
in the addendum, Method of the Masons, although the result, that is, the coeI-
fcient, is given in the sexagesimal as well as in the decimal Iorm.
N.B. The sexagesimal results are more accurate than the decimal ones.
Al-Ksh saw the clay-plastered or mutayyan muqarnas in ancient build-
ings in IsIahan. It is similar to the simple muqarnas, except that the heights
oI its tiers might diIIer and a Iew tiers might have only a rooI and no Iacets.
Calculations on the mutayyan muqarnas are analogous to the calculations on
the simple muqarnas.
The curved muqarnas is like a simple muqarnas in which the rooIs oI the
cells are curved. Curved surIaces are located between two rooIs. These interme-
diate elements are shaped as a triangle or as two triangles, which together Iorm
a biped. Such triangles are also Iound on its rooI as well as curved almonds
and barley-kernels. The Iacets oI the cells are only squares or rectangles. Their
bases are either equal to the module oI that muqarnas (i.e., the cell is standing
on a square or a rhombus), or halI the diameter oI its square (i.e., standing on a
halI-square), or the diIIerence between the diagonal and its side (which equals
the shorter side oI an almond or biped), or the side oI an octagon with radius
equal to the module. There are no other bases besides these Iour. One eighth
oI an octagon with radius equal to the module is an isosceles triangle with an
angle oI 45 at its vertex, that is, a halI-rhombus, with a base equal to the shorter
diagonal oI the whole rhombus. That al-Ksh fnds it necessary to mention the
octagon probably means that octagons occurred rather Irequently, or that he had
a muqarnas in mind where octagons occurred. Two eighths oI an octagon with
radius equal to the module give a fug, the complement to a large biped.
To measure the area oI the curved muqarnas we add the bases oI all the
cells. Now we multiply the sum by the coeIfcient and obtain the area oI all the
cells, i.e., their Iacets plus their curved rooIs. To this amount we add the areas
oI all the intermediate elements and thus obtain the area oI the muqarnas. See
my above-mentioned papers on how to obtain the coeIfcient and calculate the
intermediate elements.
The Shr: muqarnas is like a curved muqarnas but has a greater variety
oI elements. In the previous kind, only Iour possible measures Ior the bases oI
the Iacets occur, as was explained, but in the Shrz muqarnas the possibilities
are innumerable. Also many diIIerent elements are Iound on the rooI: besides
the curved rooIs oI the cells with intermediate triangles and bipeds one fnds tri-
angles, squares, pentagons, hexagons, star polygons, etc., fat as well as curved.
Sometimes a Iacet without a rooI is Iound with a mihrb (niche) drawn on it.
In Timur`s time, when building activity exploded, local constructors
could manage the simpler buildings. But Ior the special and more artistic
monuments architects and artisans were imported Irom the conquered lands,
frst Khwrizm, then Tabrz and Shrz, and fnally India and Syria. It is known
that Timur brought in architects Irom Shrz in 1388 and 1393, and that many
migrated oI their own Iree will.
43
The names oI several Shrz architects have
been transmitted, the most Iamous being Qawm al-Dn ibn Zayn al-Dn al-
Shrz, the only active builder whose surviving structures display a distinctive
architectural style. This might well be the reason why the type oI muqarnas
constructed with many variations, 'innumerable possibilities as al-Ksh
explains, was called Shrz.
To measure the area oI the Shrz muqarnas we proceed in principle
in the same way as in the case oI the curved muqarnas. We frst make a ruler
corresponding to the module oI that muqarnas and divide it into sixty sections,
iI we calculate in the sexagesimal system, or into ten sections, iI we calculate
'with Indian numbers, that is, in the decimal system. With this ruler we mea-
sure all the various elements and then compute as beIore.
Although the appearance oI a muqarnas is complex, they usually consist
oI only a Iew basic elements. These might have been preIabricated, as in the
case oI the collapsed vault on the Takht-i Suleiman, or as in the decorations
still made in Eez. Al-Ksh`s Iour types are more or less oI the same kind, and
constructed with similar elements. The elements oI a muqarnas are standard-
ized. Apart Irom the decoration, the diIIerence in the appearance results mainly
Irom the diIIerent ways in which these standardized elements are put together.
The practice oI using standardized elements, making construction Iaster and
cheaper, is widely Iound in ancient China. In Islamic architectural practice
many monuments, especially palaces, were built rapidly, either because inse-
curity oI power made lengthy building programs unlikely to be completed or
because they tended to be personal rather than dynastic and were not meant to
or expected to survive their original patron. Hence standardized elements were
a necessity. When we want to study practical mathematics, we have to take into
consideration the practice.
Coxciusioxs
In seventeenth-century SaIavid Iran architects were paid a percentage on each
building based on the cubit measure oI the height and thickness oI the walls:
44
The Persians determine the price Ior masons on the basis oI the height and
thickness oI walls, which they measure by the cubit, like cloth. The king
imposes no tax on the sale oI buildings, but the Master Architect, that is ChieI
oI Masons, takes two percent oI inheritance allotments and sales. This oIf-
cer also has a right to fve percent on all edifces commissioned by the king.
These are appraised when they are completed and the Master Architect, who
has directed the construction, receives as his right and salary as much as fve
percent oI the construction cost oI each edifce.
Iikewise in medieval Italy it was common practice to pay the artisans
according to the surIace area they had completed. The same custom seems to
have existed in the Arab world. It is also useIul to know in advance, more or
less, how much material is needed like gold Ior gilding, bricks Ior construc-
tion or paint. Payment per cubit was common in Ottoman architectural prac-
tice where a team oI architects and surveyors had to make cost estimates Ior
projected buildings and supply preliminary drawings Ior various options. In
addition to Iacilitating estimates oI wages and building materials beIore the
construction began, al-Ksh`s Iormulas may also have been used in appraising
the price oI a building aIter its completion. His sophisticated Iormulas were,
like the simple Iormulas Iound in the manuals oI arithmetics, useIul Ior every-
day liIe. This was al-Ksh`s objective: to present 'all that is needed Ior the one
who calculates careIully.
Al-Ksh`s Key of Arithmetic is the only known treatise in which an
attempt is made to calculate the surIace area oI a muqarnas. Iooking at a
muqarnas structure (fgure 8.8) we can see why. Although based on a relatively
simple two-dimensional plan, the three-dimensional muqarnas is a complex,
intricate vaulting. To calculate the curved surIaces oI the diIIerent cells was
a near impossibility Ior all but a master mathematician such as al-Ksh. By
suitably chosen approximations he worked out Iactors appropriate Ior calcula-
tion in everyday liIe, optimizing the problem like a modern mathematician. I
have some doubts, however, whether these muqarnas Iactors were oIten applied
in practice, as they are given in decimal Iractions to the sixth position, or in
sexagesimal Iractions to the Iourth position. Eor the calculation oI arches, the
Iactors contain decimal Iractions to the third position only; this is suIfcient
Ior all practical purposes. Another possibility why al-Ksh worked out Iactors
Ior calculating a muqarnas could be that he knew how to do it. As in modern
mathematical research, Iormulas are discovered Ior their own sake, only later,
may practical applications be Iound and developed.
No+rs
1. Vogel, pp. 4353.
2. Creswell, pp. 608610. Compare also G. R. D. King`s criticismon Creswell`s appraisal
oI the early level architecture in the Arabian peninsula.
3. Erishman, p. 13.
4. See Toomer, DSBVII, pp. 358365.
5. Rosen |1831| trans., p. 3.
6. Rosen l.c. trans., pp. 7086.
7. See Youschkevitch, DSB I, pp. 3943.
8. Edited by Saidan |1971|, based on the Ieiden ms. Or. 103, which contains only the
frst three parts oI the ms., and the Cairo ms. Riyda 42 M.
9. Saidan |1974| pp. 369375.
10. Necipoglu |1995| pp. 133138 and p. 176, note 13.
11. See Sesiano, Enc. Hist. pp. 475476.
12. Hochheim, Repr. pp. 204222.
13. See Vernet and Bruin.
14. Hogendijk |1990|.
15. Rebstock |1993|.
16. According to Amn Ahmad Rz (d. 1010 A.H.) in 1adhkira-ye haft iqlm |The
account oI the seven climates|, Ms. Sepahslr (Tehran) No. 2733, Iol. 774. (Courtesy
oI M. Bagheri).
17. Hill |1996| pp. 98101.
18. This is essentially Creswell`s theory, see Warren, p. 59I.
19. Rebstock |1993| pp. 133134.
20. This example implies a 4!
21. Rebstock |1992| pp. 202218.
22. All fve constructions are perIormed in the video 'Qubba Ior al-Ksh.
23. This detailed account will be published in a Iorthcoming paper together with a dis-
cussion on the accuracy oI the Iactors and the calculations.
24. Golombek |1988| pp. 153157.
25. Bulatow |1978|.
26. Saidan |1971| l.c., p. 268.
27. Hochheim |187780| part 2, p. 28.
28. Nesselmann |1843| p. 33 and note 19. Muhammad Bah al-Dn al-mul is men-
tioned neither in DSB nor in Enc. Hist. mul is a town in Syria.
29. CI. Rebstock |1992| p. 216: In the edition Riydyt, Aleppo 1976, p.88, the term in
brackets does not occur. However, it is written in the Ieningrad Ms.
30. Necipoglu |1995| pp. 140141,with some minor changes.
31. Reprinted in 1986.
32. See RosenIeld/Grigorian, DSB XIII, pp. 288295.
33. Rashed, Vol. I, pp. 319457; Suter, Repr. Vol. II, pp. 435476.
34. See Dold-Samplonis, DSB IX, pp. 239241.
35. Rashed, Vol. I, pp. 850871; Suter, Repr. Vol. II, pp. 435476.
36. Rashed, Vol. I, pp. 850852; Suter, Repr. Vol. II, p. 463.
37. See Sabra, DSBVI, pp. 189210.
38. Rashed, Vol. II, pp. 207293; Suter, Repr. Vol. II, pp. 369412.
39. Grabar |1992| p. 147.
40. Necipoglu |1995| p. 349.
41. Dold-Samplonius |1992/3| p. 202.
42. Dold-Samplonius 1992/93 and 1996, see also zdural.
43. Golombek/Wilber |1988| pp. 187194.
44. Necipoglu |1995| pp. 44, 159.
Biiiior:rn.
Bruin, Erans. 1970. Surveying and Surveying Instruments, being chapters 2630 oI the
book On Finding Hidden Waters by Ab Bakr Muhammad al-Karaj (AD 1029), Brn
Newsletter No. 39. Beirut.
Bulatow, M. S. 1978. Geometric Harmony in the Architecture of Central Asia, 9th15th
century (Russian). Moscow.
Creswell, K. A. C. 1958. A Short Account of Early Muslim Architecture. Pelican Paper-
back.
Creswell, K. A. C. 1960. 'Architecture, Encyclopaedia of Islam. 2nd ed. Vol. I, pp.
608624.
DSB. 19701980. Dictionary of Scientihc Biography, 16 vols. NewYork.
Diez, E. 1938/1986. 'Kubba, Encyclopaedia of Islam, Supplement, pp. 139146, repr.
in the 2nd ed. Vol. V, pp. 289296.
Dold-Samplonius, Yvonne. 1992. 'The XVth Century Timurid Mathematician Ghiyth
al-Dn Jamshd al-Ksh and his Computation oI the Qubba, Amphora, Festschrift for
Hans Wussing on the Occasion of his 65th Birthday. Ed. S. S. Demidov, M. Eolkerts, D.
E. Rowe, and Ch. J. Scriba. Basel et al., pp. 171181.
Dold-Samplonius, Yvonne. 1992/3. 'Practical Arabic Mathematics: Measuring the
Muqarnas by al-Ksh, Centaurus 35, pp. 193242.
Dold-Samplonius, Yvonne. 1993. 'The Volume oI Domes in Arabic Mathematics Jes-
tigia Mathematica, Studies in medieval and early modern mathematics in honour of
H.L.L. Busard. Ed. M. Eolkerts and J. P. Hogendijk. Amsterdam &Atlanta, pp. 93106.
Dold-Samplonius, Yvonne. 1996. 'How al-Ksh Measures the Muqarnas: A Second
Iook, Mathematische Probleme im MittelalterDer lateinische und arabische Sprach-
bereich, Ed. M. Eolkerts (Harrassowitz Verlag in Kommission) Wiesbaden, pp. 5790.
Enc. Hist. 1997. Encyclopaedia of the History of Science, 1echnology, and Medicine in
Non-Western Cultures. Ed. H. Selin. Dordrecht.
Erishman, Martin, and Khan, Hasan-Uddin (Eds.). 1994. 1he Mosque. History, Architec-
tural Development & Regional Diversity. Iondon.
Golombek, Iisa, and Wilber, Donald. 1988. 1he 1imurid Architecture of Iran and 1uran.
2 Vols. Princeton, N.J.
Grabar, Oleg. 1978/1992. 1he Alhambra. Iondon/Sebastopol, CaliI. |2nd ed.|
Hill, Donald,1984/1996. A History of Engineering in Classical and Medieval 1imes.
Iondon/NewYork. |Paperback ed.|
Hochheim, AdolI, 18771880/1998. Kf fl Hisb (Gengendes ber Arithmetik).
(German) Halle. Repr. Islamic Mathematics and Astronomy, Vol. 38. ErankIurt/ Main.
Hogendijk, Jan P. 1990. 'AMedieval Arabic Treatise on Mensuration by Qd Ab Bakr,
Zeitschrift fur Geschichte der arabisch-islamischen Wissenschaften 6, pp. 130150.
Al-Ksh, Ghiyth al-Dn, 1427. Mifth al-Hisb (Key oI Arithmetic). Ms. Malek
Iibrary 3180/1, Tehran, dated 830 AH(!), copied by Mon al-Dn al-Ksh, who went
with al-Ksh Irom Kashan to Samarqand. Ms. Or. 185, Ieiden, dated AD 1558.
King, G. R. D. 1991. 'Creswell`s Appreciation oI Arabian Architecture, Muqarnas, Vol.
8, pp. 94102.
Iuckey, Paul. 1951/1998. Die Rechenkunst bei Jamshd b. Masd al-Ksh mit Ruck-
blicken auf die altere Geschichte des Rechnens, (German). Wiesbaden. Repr. Islamic
Mathematics and Astronomy, Vol. 56, pp. 75226. ErankIurt/Main.
Nader, Nabulsi. 1977. Al-Ksh, Ghiyth al-Dn. Mifth al-Hisb (Key oI Arithmetic).
Arabic edition, with Erench notes and introduction. Damascus.
Necipoglu, Glru. 1995. 1he 1opkapi ScrollGeometry and Ornament in Islamic Archi-
tecture. Santa Monica, CaliI.
Nesselmann, Georg H. E., 1843/1998. Essen: der Rechenkunst von Mohammed Beha-
eddin al-mul. (Arabic edition, with German translation and notes.) Berlin. Repr.
Islamic Mathematics and Astronomy, Vol. 59, pp. 29166. ErankIurt/ Main.
zdural, Alpay. 1991. 'Analysis oI the Geometry oI Stalactites: Buruciye Medresse in
Sivas. ME1U Journal of Faculty of Architecture 11 (Middle East Technical University,
Ankara), pp. 5771.
Rashed, Roshdi. 1993/1996. Les mathematiques inhnitesimales du IX
e
au XI
e
siecle. Vol.
I: Eondateurs et Commentateurs. Vol. II: Ibn al-Haytham |Travaux en mathematiques
infnitesimales| (Erench and Arabic). Iondon.
Rebstock, Ulrich. 1992. Rechnen im islamischen Orient: Die literarischen Spuren der
praktischen Rechenkunst (German). Darmstadt.
Rebstock, Ulrich. 1993. Die Reichtumer der Rechner von Ahmad b. 1habt (German).
Beitrge zur Sprach- und Kulturgeschichte des Orients, Vol. 32. WalldorI-Hessen.
Rosen, Erederic (ed. and trans.), 1831/1997. 1he Algebra of Mohammed Ben Musa. Ion-
don. Repr. Islamic Mathematics and Astronomy, Vol. 1. ErankIurt/Main.
RosenIeld, Boris A., andYouschkevitch, AdolI P. 1954. Al-Ksh, Ghiyth al-Dn. Mifth
al-Hisb (Key oI Arithmetic). Russian Translation and Commentary. Matematitcheskije
Traktaty, Istoriko-matematitcheskife Issledovaniya 7, pp. 13326.
Saidan, Ahmad S. 1971. 1he Arithmetic of Abl-Waf al-B:fn, (Arabic) Edition,
Introduction, Commentaries and ReIerence to 1he Arithmetic of al-Karaf. Amman.
Saidan, Ahmad S. 1974. 'The Arithmetic oI Ab`l-WaI`, Isis 65, pp. 367375.
Suter, Heinrich. 1986. Beitrage :ur Geschichte der Mathematik und Astronomie im
Islam, Nachdruck seiner SchriIten aus den Jahren 18921922, 2 Vols. (German). Ed.
Euat Sezgin. ErankIurt am Main.
Vernet, Juan (in collaboration with A. Catala). 1970/1979. 'Un ingeniero arabe del siglo
XI: al-Karaj, Al-Andalus 35, pp. 6991. Repr. in Estudios sobre historia de la ciencia
medieval, pp.147169 (Spanish). Barcelona.
Vogel, Kurt (ed.). 1968. Chiu Chang Suan Shu (Neun Bcher arithmetischer Technik),
with commentary (German). Braunschweig.
Warren, John. 1991. 'Creswell`s Use oI the Theory oI Dating by the Acuteness oI the
Pointed Arches in Early MuslimArchitecture, Muqarnas, Vol. 8, pp. 5965.
Jideo: 'Qubba Ior al-Ksh (16 min.), Yvonne Dold-Samplonius, Technics: Christoph
Kindel and Kurt Saetzler, IWR, Heidelberg 1995/6. Distributed by AMS 1997, see
www.ams.org.
V
Seventeenth-Century 1ransmission ot Astronomy
One oI the most neglected areas in the history oI Islamic astronomy is the
development oI that science in South Asia, the infuence oI Sanskrit astronomy
on it, and its impact on the older Indian tradition oI the siddhntas.
1
Vast quanti-
ties oI relevant manuscripts in Persian, Arabic, and Sanskrit survive in public
and private collections in South Asia and, Ior the Persian and Arabic material,
in Iran and Central Asia as well, but little attention has until very recently been
paid to this important material. This chapter will deal with one aspect oI this
general question, namely, with the presentation oI Muslim planetary models by
the Brhmana scholar, Nitynanda, who wrote at Delhi during the reign oI Shh
Jahn, his sources, and his infuence.
2
Eard al-Dn Masd ibn Ibrhm al-Dihlaw apparently leIt the service
oI the dil Shh oI Bjpr on the death oI his patron, Ibrhm dil Shh II,
in 1627, and seems to have joined the entourage oI Shh Jahn at Junnar in
Mahrstra beIore the Iuture Emperor marched toward Delhi on 2 December
oI that year.
3
He cast the horoscope, presumably beIorehand, Ior Shh Jahn`s
enthronement, which took place on 14 Eebruary 1628.
4
Shortly aIter that event,
he was asked by Shh Jahn`s wazr, saI Khn, to prepare a new zj, the Zf-i-
Shh-Jahn, that would be based on Ulugh Beg`s Zf-i-fadd, but would employ
a new calendar, the tarkh-i-Ilh Shhishn.
5
Eard al-Dn was able to pres-
ent the enormous Zf-i-Shh-Jahn to the Emperor in October 1629, the very
month in which he is said to have died.
6
saI Khn was impressed by the new zj, and decided to have it translated
into Sanskrit. The task oI making the translation was assigned to Nitynanda,
a Brhmana residing in Delhi; he completed the translation, which he entitled
Siddhntasindhu, in the early 1630s. UniIorm copies oI gigantic size (45 33
cm. and approximately 440 Iolia) were prepared,
7
and at least eleven were dis-
tributed to worthy individuals (mostly Muslim nobles) in Northern India.
8
Eour
copies oI the Siddhntasindhu are kept at the Palace Iibrary in Jaipur. Three
are perhaps Irom among the original production; one bears the seal oI Shh
Jahn.
9
The Iourth was copied, in identical style, by GaQgrma oI Ksmra
Ior Mahrja Jayasim
.
ha in 1727.
10
The Iour copies at Jayapura
11
are the only
9
1nr P1IDDH^lP] or Ni+.:x:xi:
David Pingree
D:vii Pixrrr 270 1nr SP1IDDH^lP] or Ni+.:x:xi: 271
complete copies extant; Iour incomplete manuscripts are preserved in other
libraries in Rjasthn (Alwar and Bikaner) and Madhya Pradesa (Ujjain).
12
The Iailure oI the Siddhntasindhu to fnd Iavor among the Hindu adher-
ents oI the traditional siddhntas inspired Nitynanda to write an elaborate
apology Ior using Muslim astronomy, the Sarvasiddhntarfa that he com-
pleted in 1639.
13
The second and third chapters oI this work, on the computa-
tion oI the mean and true longitudes oI the planets, is the part oI this gigantic
and unpublished work on which we will Iocus, using manuscripts Irom Benares
and Iondon.
14
Whereas the Siddhntasindhu, Iollowing Ulugh Beg, presents the mean
motions oI the Sun, the Moon, and the planets as motions in multiples oI 30
Arab years, single Arab years, Arab months, days, and hours, and also presents
motions in solar years and solar months given Persian names, Nitynanda in the
Sarvasiddhntarfa has converted these mean motions approximately into inte-
ger numbers oI revolutions in a Kalpa oI 4,320,000,000 years in order to make
them comparable and comprehensible to his expected audience oI siddhntins.
In the course oI his presentation oI the parameters and models oI the Siddhn-
tasindhu, which he calls the Romakasiddhnta or Roman Zf, he compares them
with those oI the Brhmasphutasiddhnta composed by Brahmagupta in 628
15
and those oI the Sryasiddhnta composed by an unknown author in about
800.
16
Though he does not provide the numbers oI rotations oI these planets in
a Kalpa according to these two representatives oI the Indian tradition, I include
them Ior comparison in table 9.1.
The shorter year-length in the Romaka should diminish the number oI
rotations each planet makes in a Kalpa. Eor instance, the diIIerence in days
1able 9.1
Romaka Brhmasphuta Srya
Sun 4,320,000,000 4,320,000,000 4,320,000,000
Moon 57,750,968,965 57,753,300,000 57,753,336,000
Mars 2,296,968,639 2,296,828,522 2,296,832,000
Mercury`s sghra 17,936,534,114 17,936,998,984 17,937,060,000
Jupiter 364,356,698 364,226,455 364,220,000
Venus` sghra 7,022,180,538 7,022,389,492 7,022,376,000
Saturn 146,835,981 146,567,298 146,568,000
Days 1,577,847,748,101 1,577,916,450,000 1,577,917,828,000
Year in days 6,5;14,33,7,24,31,... 6,5;15,30,22,30 6,5;15,31,31,24
between the Romaka and the Brhmasphuta amounts to about 188,095 years, in
which there are about 2,508,000 rotations oI the Moon; the Romaka`s rotations
oI the Moon are 2,331,035 less than those oI the Brhmasphuta. However, the
rotations oI Mars, Jupiter, and Saturn according to the Romaka are all greater
in number than they are according to the Brhmasphuta.
How Nitynanda derived his rotations in a Kalpa Irom the mean motions
oI the Siddhntasindhuthat is, oI the Zf-i-Shh-Jahnhas yet to be deter-
mined. But the yearly mean motions they imply are easily computed Irom
360 R
4,320,000,000
(the Sun obviously travels 360 in a year).
Moon 13 rotations 2, 12; 34, 50, 39, . . .
Mars 3, 11; 24, 50, 34, . . .
Mercury`s sghra 4 rotations 54; 42, 40, 12, . . .
Jupiter 30; 21, 46, 58, . . .
Venus` sghra 1 rotation 3, 45; 10, 54, 9, . . .
Saturn 12; 14, 10, 47, . . .
Iater on, in order to 'simpliIy the calculations, Nitynanda, in the
Iashion oI an Indian karana, imagines a period oI 10,000,000 days, which he
equates with 27,379 solar years so that each year equals 6,5;14,36,20, . . . days
instead oI the earlier 6,5;14,33,7, . . . days. Using the ratio oI the solar years in
10,000,000 days to those in a Kalpa, one can easily compute the 'integer rota-
tions oI each planet in the shorter period (Table 9.2).
The epoch that Nitynanda chooses is the beginning oI the Indian month
Caitra in the year in which Shh Jahn was enthronedthat is, 25 Eebruary
1628at noon at Iank. He gives the epoch mean longitudes oI the planets as
numbers oI lapsed days in a period oI 10,000,000 days. Since these numbers all
1able 9.2
Planet Text Computation
Sun 27,379 (27,379)
Moon 366,011 366,010.1...
Mars 14,557 14,557.5...
Mercury`s sghra 113,677 113,676.9...
Jupiter 2,309 2,309.1...
Venus` sghra 44,504 44,504.6...
Saturn 930 930.6...
have seven digits, on division by 10,000,000 they become decimal Iractions oI
a rotation. In the Iollowing table the true longitudes oI the planets as given in
the Tuckerman tables are provided Ior comparison (Table 9.3).
Mars is so Iar oII because its elongation Irom the Sun is 119, which
produces an equation oI about 39.
In the midst oI this computation Nitynanda reports on the theory oI
precession according to Maya the Asura.
17
This Maya was the person to whom
the Sun revealed the Sryasiddhnta; he is mentioned here because Nitynanda
claims that the Romaka was revealed to a Yavana by the Sun.
18
This theory
certainly comes Irom a Muslim, not an Indian source. Nitynanda states that,
according to Maya, precession amounted to 16;6,32 in 1628; this puts the date
oI coincidence oI the tropical and sidereal zero-points in AD 500. He Iurther
notes that the zero-point was once in Aries 8, at another time in Aries 10;
these are the Babylonian norms Ior Systems B and A oI the Moon.
19
It is odd to
fnd them surIacing in a seventeenth century Sanskrit poem.
Since the computation oI the rotations oI the planets in 10,000,000 days
was somewhat crude, Nitynanda oIIers a daily bja or correction Ior each
(Table 9.4).
The longitudes oI the apogees oI the Sun and the fve star-planets and the
nodes oI the latter in 1628 are presented in the Iollowing table as they are given
by Nitynanda (Table 9.5).
The lunar apogee and node rotate respectively 488,327,103 and
232,088,311 times in a Kalpa or 3095 and 1471 times respectively in
10,000,000 days; their epoch positions are given as 7,134,658 or 256;50, . . .
and 6,765,363 or 243;33, . . . 116;26, . . ..
In an attempt to persuade his readers that the diIIerences between his
three systems are small, Nitynanda gives the corrections to derive the lon-
gitudes oI the apogees and nodes oI the Sryasiddhnta Irom those oI the
Romakasiddhnta, and states those oI the Brhmasphutasiddhnta. He does
not inIorm his readers that he has added precession (16;6,32), as he should, to
the longitudes in the Sryasiddhnta (Table 9.6).
The corrections to the Romaka`s longitudes oI the nodes are, in general,
Iar too large to convince the reader that there is little diIIerence between that
text and the Sryasiddhnta.
Nitynanda also gives the bjas or corrections to go Irom the Romaka`s
epoch mean longitudes to those oI the Sryasiddhnta corrected by precession
(he does not speciIy this latter correction), and Irom the mean longitudes oI the
Sryasiddhnta to those oI the Brhmasphutasiddhnta (Table 9.7).
Chapter 3 oI the Sarvasiddhntarfa deals with the computation oI the
true longitudes oI the planets Irom their mean longitudes. It begins with a bald
1able 9.3
Planet Iapsed days Epoch mean True DiIIerences
longitudes longitudes
Sun 9,565,549 344;21,35 346 2
Moon 9,589,692 345;13,44 355 10
Mars 2,897,712 104;19,3 73 31
Mercury`s sghra 9,381,010 337;36,58 (335)
Jupiter 7,363,135 265;4,22 270 5
Venus` sghra 1,087,113 39;8,9 (8)
Saturn 5,054,485 181;57,41 189 7
1able 9.4
Sun
243
10,000,000
0;0,0,5,14,55,40,48
Moon
121
10,000,000
0;0,0,2,36,48,57,36
Mars
2166
10,000,000
0;0,0,46,47,8,9,36
Mercury`s sghra
1246
10,000,000
0;0,0,26,54,48,57,36
Jupiter
716
10,000,000
0;0,0,15,27,56,9,36
Venus` sghra
2200
10,000,000
0;0,0,47,31,12
Saturn
2186
10,000,000
0;0,0,47,13,3,21,36
1able 9.3
Planet Apogees Nodes
Sun Cancer 5;9,33,15
Mars Ieo 24;40,23 Taurus 20;40,23
Mercury Scorpio 7;12,3 Aquarius 7;12,3
Jupiter Iibra 2;15,6 Cancer 10;15,6
Venus Gemini 25;9,0 Pisces 25;9,0
Saturn Sagittarius 19;39,6 Cancer 19;39,6
statement oI the radii oI the epicycles in the three siddhntas (the eccentrici-
ties in place oI the manda-epicycles Ior the Romaka); however, the pulsating
epicycles oI the Sun and the Moon in the Brhmasphutasiddhnta are ignored;
only their values Ior meridian-crossings are presented (Table 9.8).
The author makes no comment on the Iact that the Romaka`s lunar model
has a crank-mechanism. This, however, like the other elements in Muslim
astronomy that are not Iound in Indian science, is called by a Sanskrit name,
pksika (relating to the halI-months between the syzygies) in order to camou-
fage its Ioreignness. We shall comment more on this strategy oI Nitynanda
later.
AIter a long discursus on the geometrical method oI computing a Sine-
table in which he is clearly indebted to a Muslim source, he turns to the method
oI computing and applying the equations according to the Romaka.
In the case oI the Sun, using Ptolemy and a table oI Sines (R 60 1,0)
instead oI Chords, one should Iorm (see fgure 9.1) a triangle whose sides are
Sinx
e
R
and R Cosx
e
R
1able 9.6
Apogees
Planet Romaka Correction Srya with Brhmasphuta
precession
Sun 95;93,15 1;46,5,15 93;23,28 93;23,32 77;56,56
Mars 144;40,23 1;28,34 146;8,57 146;8,32 128;25,8
Mercury 217;12,3 19;22,23 236;34,26 236;33,40 224;54,53
Jupiter 182;15,6 5;12,43 187;27,49 187;25,58 172;35,49
Venus 85;9,0 10;49,11 95;58,11 95;57,5 81;17,35
Saturn 259;39,6 6;55,3 252;44,3 252;43,58 238;16,9
Nodes
Planet Romaka Correction Srya with Brhmasphuta
precession
Mars 50;40,23 5;29,30 56;9,53 56;10,19 21;53,27
Mercury 307;12,3 89;35,45 36;47,48 36;48,48 21;8,34
Jupiter 100;15,6 355;31,33 95;46,39 95;47,11 82;1,9
Venus 355;9,0 80;37,36 75;46,36 75;48,50 59;43,55
Saturn 109;39,6 6;48,58 116;46,36 116;29,17 103;3,9,42
1
a
b
l
e
9
.
7
P
l
a
n
e
t
R
o
m
a
k
a
C
o
r
r
e
c
t
i
o
n
P
r
e
c
e
s
s
e
d
P
r
e
c
e
s
s
i
o
n
C
o
r
r
e
c
t
e
d
C
o
r
r
e
c
t
i
o
n
C
o
r
r
e
c
t
e
d
S
r
y
a
S
r
y
a
B
r
h
m
a
s
p
h
u
t
a
S
u
n
3
4
4
;
2
1
,
3
5
1
;
5
5
,
1
1
3
4
2
;
2
6
,
2
4
1
6
;
6
,
3
2
3
2
6
;
1
9
,
5
2
0
;
3
,
2
9
3
2
6
;
2
3
,
2
1
M
o
o
n
3
4
5
;
1
3
,
4
4
1
;
4
6
,
3
2
3
4
3
;
2
7
,
1
2
1
6
;
6
,
3
2
3
2
7
;
2
0
,
4
0
0
;
2
5
,
4
3
3
2
7
;
4
6
,
2
3
M
a
r
s
1
0
4
;
1
9
,
3
2
;
5
3
,
5
7
1
0
1
;
2
5
,
6
1
6
;
6
,
3
2
8
5
;
1
8
,
3
4
1
;
1
5
,
1
3
8
4
;
3
,
2
1
M
e
r
c
u
r
y
`
s
s
g
h
r
a
3
3
7
;
3
6
,
5
8
5
;
0
,
1
9
3
3
2
;
3
6
,
3
9
1
6
;
6
,
3
2
3
1
6
;
3
0
,
7
3
1
5
;
3
0
,
7
J
u
p
i
t
e
r
2
6
5
;
4
,
2
2
1
;
1
3
,
4
7
2
6
5
;
1
8
,
9
1
6
;
6
,
3
2
2
5
0
;
1
1
,
3
7
0
;
8
,
2
4
2
5
0
;
2
0
,
1
V
e
n
u
s
`
s
g
h
r
a
3
9
;
8
,
9
5
;
3
5
,
1
0
4
4
;
4
3
,
1
9
1
6
;
6
,
3
2
2
8
;
3
6
,
4
7
1
;
5
3
,
1
8
2
6
;
4
3
,
2
9
S
a
t
u
r
n
1
8
1
;
5
7
,
4
1
6
;
5
9
,
2
4
1
7
4
;
5
8
,
1
7
1
6
;
6
,
3
2
1
5
8
;
5
1
,
4
5
0
;
7
,
5
1
5
8
;
5
8
,
5
0
as the argument, x, lies in the semicircle beginning at 270 Irom the apogee
or in that beginning at 90 Irom the apogee. Nitynanda, since R 1,0, simply
Iorms Sinx e and Cosx e without inIorming his reader that the sexagesimal
point must be shiIted one place to the leIt. The square-root oI the sum oI the
squares oI these two sides is the distance oI the Sun Irom the earth, SO. Then
Sino Sinx
e
R

R
SO
Sinx
e
SO
.
Without a hint oI irony, Nitynanda states that some extremely clever
people advocate the ancient Indian 'Method oI Sines, in which
Sino Sinx
e
R
.
Ptolemy`s solution to the problem oI fnding geometrically the angle
which corrects the anomaly in the lunar epicycle is quite straightIorward. In
fgure 9.2 P
1
O P
2
O 10;19 and P
1
C 49;41. Since angle 2j is given, in the
1able 9.8
Eccentricities/Manda epicycles` radii
Planets Romaka Sryasiddhnta Brhmasphutasiddhnta
Sun 2;1,20 2;202;16,40 2;16,40
Moon 5;12,24 5;205;16,40 5;16
Crank mechanism 10;24,48
Mars 6;4 or 6;14 12;3012;40 11;40
Mercury 3 5;04;40 6;20
Jupiter 2;47 5;305;20 5;30
Venus 0;52 2;01;50 1;501;30
Saturn 3;29 8;108;0 5
Sghra epicycles` radii
Planets Romaka Sryasiddhnta Brhmasphutasiddhnta
Mars 39;43 39;1038;40 40;36,4039;30
Mercury 22;30 22;1022 22
Jupiter 11;47 11;4012 11;20
Venus 43;10 43;4043;20 4343;50
Saturn 6;51 6;306;40 5;50
right triangle P
2
OL
2
the sides P
2
L
2
and L
2
O can be computed. Similarly, in the
right triangle P
1
OL
1
the sides P
1
L
1
and L
1
O are respectively equal to P
2
L
2
and
L
2
O. Then CL
2
1
CP
2
1
P
1
L
2
1
; and CO CL
1
L
1
O. Moreover, CL
2
CO OL
2
;
and CP
2
2
CL
2
2
L
2
P
2
2
. Now, in right triangle CL
2
P
2
angle L
2
CP
2
, which equals
angle ACA', can be computed.
Nitynanda`s solution, based on the use oI the Sine and Cosine Iunctions,
is more complex, and its interpretation is made diIfcult by the terminology
that he has invented to name the parts oI the fgure. P
1
O OP
2
he calls the
pksa epicycle-radius and measures at 10;24,48; P
1
C, which is 6010;24,48
49;35,12, is called the radius oI the pksa eccentric. The elongation oI the mean
Moon Irom the mean Sun is the pksa argument; it is to be doubled. The Sine
oI the double elongation is multiplied by the epicycle-radius, P
1
O, and divided
by the radius oI the pksa (here one must understand the pksa eccentric, P
1
C).
This produces P
1
L
1
, the Sine oI angle P
1
CL
1
measured in the units oI a circle
whose radius is 60.
Sin
60
2j
P
1
O
60
Sin
P
1
O
2j P
1
L
1
. Also P
1
L
1
Sin
P
1
C
ZP
1
CL
1
.
ThereIore,
Sin
60
ZP
1
CL
1
Sin
60
2j
P
1
O
60

60
P
1
C
Sin2j
P
1
O
P
1
C
.
Cos
2

Sin
2

Sin
1

Cos
1

Eigure 9.1
Eor my understanding oI the rest oI Nitynanda`s rules I am indebted
to my colleague, Kim PloIker. The frst rule is to take the Sine Irom the argu-
ment (that is, 2j) increased or decreased by the arc oI the previous result (that
is, angle P
1
CL
1
) as the argument lies between 0 and 180 or 180 and 360
respectively. In fgure 9.2 angle CP
1
O 180 (2j ZP
1
CL
1
). ThereIore,
2j ZP
1
CL
1
180 ZCP
1
O Z CP
1
Q, and
QC Sin(2j ZP
1
CL
1
)
P
1
C
R
.
Right triangles OQC and OL
1
P
1
are similar to each other. ThereIore,
OC
QC

P
1
O
P
1
L
1
, or OC
QC P
1
O
P
1
L
1
.
Since P
1
L
1
Sin2j
P
1
O
R
,
OC Sin(2j ZP
1
CL
1
)
P
1
C
R
P
1
O
R
Sin2j P
1
O
,
which reduces to:
OC Sin(2j ZP
1
CL
1
)
P
1
C
Sin2j
.
Indeed, Nitynanda`s instructions are to multiply Sin(2j ZP
1
CL
1
) by
the radius oI the eccentric, P
1
C, and to divide the product by the Sine arising
Eigure 9.2
Irom the original argumentthat is, by Sin2j. The result he correctly states to
be the pksa hypotenuse, OC.
Now one is instructed to multiply the original Sin2j and Cos2j by the
epicycle-radius, P
1
O, and to divide the results by 60; this converts the two Sines
Irom base 60 to base 10;24,48, to the lengths P
1
L
1
P
2
L
2
and L
1
O L
2
O in
fgure 9.2. Then the sum or diIIerence oI OC and LO is CL; and CL
2
PL
2
P
2
C
2
, the hypotenuse Irom the opposite point (abhimukhacihna). Einally,
L
2
P
2

60
P
2
C
SinZL
2
CP
2
;
its arc is the correction to the apogee oI the epicycle, ZACA.
The rest is straightIorward, but expressed as Iar as possible in terms
Iamiliar Irom traditional Sanskrit siddhntas. The anomaly oI the Moon on
its epicycle is counted Irom the corrected epicycle-apogee, and the corrected
anomaly, x, Irom the uncorrected apogee. The Sine and the Cosine oI x are
multiplied by the radius oI the epicycle, r, and divided by the Radius, R 60.
The mean pksika hypotenuse, CO, is increased or decreased by Cosx
r
R
and
S
CO Cosx
r
R
D
2
S
Sinx
r
R
D
2
H
2
,
where H is the distance oI the Moon Irom the earth. Then the lunar equation is
the arc oI
Sinx
r
R

R
H
Sinx
r
H
.
Eollowing Muslim practice, Nitynanda instructs his readers to reduce the lunar
longitude on its orbit to its corresponding longitude on the ecliptic.
The computation oI the equation oI the center according to Nitynanda is
illustrated in fgure 9.3, where A is the apogee, E the equant, D the center oI the
deIerent, and O the center oI the earth. In order to convert the anomaly at E, x,
to that at D, x, we have, analogously to the computation oI the solar equation,
Sino
1
Sinx
e
R
and x x o
1
. Then CH is Sinx and HD is Cosx. In the
right triangle CHO, HO HD e, and CO
2
CH
2
HO
2
. Then,
OG Sino
2

CO
R
Sinx
2e
R
Irom which Sino
2
Sinx
2e
CO
.
The equation oI the anomaly is computed in a manner similar to that used
in traditional Indian siddhntas. The diIIerence is that the distance oI the center
oI the epicycle Irom the earth is now CO rather than R; in this way the infu-
ence oI the frst, manda, equation on the second, sghra, equation is accounted
Ior. In fgure 9.4 CA CP r is the radius oI the epicycle and ZACP ,,
the anomaly. Then
BP Sin,
r
R
and BC Cos,
r
R
. Then
BO CO BC; PO
2
BO
2
BP
2
; and Sino BP
R
PO
.
Nitynanda`s algorithm Ior computing the equation oI the center oI Mer-
cury is, oI course, more complex. It can be understood with the help oI fgure
9.5.
Here OEEDDMDNe, NCR, and ZMDNZDECZOEHx.
Nitynanda begins by subtracting x Irom 180 (or 180 Irom x); this yields
ZNDE. Since in triangle NDE the perpendicular, DF, Irom D to NE bisects
ZNDE,
ZNDF ZFDE
180 x
2
.
Then 2
S
Sin
180 x
2

e
R
D
NE.
Eurther, since in the right triangle EDF.
ZEDF
180 x
2
, ZDEF 90
180 x
2

x
2
;
A
H
E
F
G
D
C
O
e
e
Eigure 9.3
thereIore ZCEF
3x
2
.
Then NG Sin
3x
2

NE
R
.
In triangle ECN ZCNE 180
S
3x
2
ZNCE
D
;
and SinZNCE NG Sin
3x
2

NE
R
.
Eigure 9.4
C
Q
O
e
H
M
E
N
F
G
D
e
e
Eigure 9.5
Thus ZCNQ 180 ZCNE 180
S
180
S
3x
2
arc
S
Sin
3x
2

NE
R
DDD
;
and SinZCNQ CQ Sin
S
3x
2
arc
S
Sin
3x
2

NE
R
DD
.
But right triangle CQE is similar to right triangle NGE; thereIore, according to
Nitynanda,
CE CQ
NE
NG
NE
R
Sin
S
3x
2
arc
S
Sin
3x
2

NE
R
DD

NE
NG

R
NE
Sin
S
3x
2
arc
S
Sin
3x
2

NE
R
DD

R
NG
.
It is clear that HO Sinx
e
R
and EH Cosx
e
R
;
and CH CE EH. Then CO
2
CH
2
HO
2
; and Sino HO
R
CO
.
Having provided this much, Nitynanda omits the rules Ior computing
Mercury`s equation oI the anomaly since they do not diIIer Irom those Ior the
other planets.
In all oI these algorithms Ior computing planetary equations Nitynanda
has provided only the fnal bare rulespresumably derived Irom a Persian
commentary on Ulugh Beg`s zjbut no hint oI the geometrical rationales that
lie behind them. Iater on he describes the Muslim system oI internesting plan-
etary spheres, and, in the section where he provides the algorithms Ior these,
in most manuscripts there are spaces leIt blank Ior diagrams; however, in only
one oI the halI-dozen or so manuscripts that I have been able to inspect, Alwar
2005, which was copied in 1846, are any diagrams actually drawn, and they
completely misrepresent the geometry that we can reconstruct Irom the rules.
ThereIore, though the algorithms are mostly correct and though Nityn-
anda, as we demonstrated at the beginning oI this chapter, is doing his best to
persuade the traditionalists that the so-called Romakasiddhnta is not very diI-
Ierent Irom the Sryasiddhnta, he, or the scribes oI the Sarvasiddhntarfa iI
they simply Iailed to copy his diagrams, has or have Iailed to provide suIfcient
inIormation Ior an expert in the Sryasiddhnta to understand how the Romaka
arrives at its results. This problem is made worse by Nitynanda`s introduction
oI a new vocabulary, based on Sanskrit, Ior reIerring to the elements oI the
Ptolemaic models without defning these new terms; it had been the practice oI
other scholars attempting to describe Muslim models in Sanskrit to transliterate
the Arabic/Persian technical terms, and to defne them. Thus, Nitynanda reIers
to the various elements oI the lunar crank-mechanism by applying the adjective
pksa (pertaining to a halI-month) to the normal Sanskrit terms: pksakarna
(the distance oI the center oI the Moon`s epicycle Irom the earth), pksakendra
(the double elongation oI the Moon Irom the Sun), pksapara (the radius oI the
concentric circle bearing the center oI the Moon`s deIerent), and pksapratiman
dalavysrddha (the radius oI the Moon`s deIerent). The term 'opposite point
(nuqtat al-muhdh) is translated as 'abhimukhacihna. Curiously, the equant
(markaz muaddal al-masr) is called by Nitynanda the 'abhicrabindu or
'point oI magic.
The most noteworthy later use oI chapter 3 oI Nitynanda`s Sarvasid-
dhntarfa is in a manuscript I Iound in the Paundarka Collection in the Palace
Iibrary at Jaipur.
20
This contains the earliest version, datable to 1726 or 1727,
oI Jaganntha`s Siddhntakaustubha in which, on II. 20v25v, Jayasim
.
ha`s
guru
21
paraphrases or copies out all the verses oI this chapter containing the
algorithms and the parameters. He only substitutes the equivalent 'sammukha
and 'unmukha Ior 'abhimukha; even 'abhicra is preserved as the name oI
the equant. Some oI the terminology applied by Nitynanda to the elements
oI the lunar crank-mechanism combining the adjective 'pksa or 'pksika
with common Sanskrit astronomical terms such as 'karna (hypotenuse) and
'kendra (anomaly) appear in about 1732 in Kevalarma`s Drkpaksasrin,
22
the poetic version, and in about 1734 in the Sanskrit prose version oI de Ia
Hire`s 1abulae astronomicae
23
and in the Yavanacandracchedyakopayogin;
24
all three were products also oI Jayasim
.
ha`s court. The Sarvasiddhntarfa, then,
provided Jaganntha with the Sanskrit words to describe Ulugh Beg`s planetary
models, and Jaganntha`s successors with some oI the terminology with which
they wrote oI European astronomy.
No+rs
1. See D. Pingree, 'Islamic Astronomy in Sanskrit, JHAS 2, 1978, 315330, and
'Indian Reception oI Muslim Versions oI Ptolemaic Astronomy, 1radition, 1ransmis-
sion, 1ransformation, ed. E. J. Ragep et. al, Ieiden 1996, pp. 471485.
2. D. Pingree, Census of the Exact Sciences in Sanskrit (henceIorth CESS), Series A,
vols. 15, Philadelphia 19701994, A3, 173b174a; A4, 141a141b; and A5, 184a;
'The Indian Reception, pp. 476480; and 'Amrtalahar oI Nitynanda, Sciamus 1,
2000, pp. 209217.
3. 1he Shah Jahan Nama of Inayat Khan, translated by A. R. Euller and edited and
completed by W. E. Begley and Z. A. Desai, Delhi 1990, p. 13.
4. Abd al-Hamd Ihr, Pdshh Nma, ed. K. Ahmad and Abd al-Rahm, BI 56, 2
vols., Calcutta 18671868, vol. 1, p. 97.
5. Ibid., vol. 1, pp. 286287.
6. 1he Shah Jahan Nama, p. 35.
D:vii Pixrrr 284
7. Manuscripts 266 and 268 in D. Pingree et al., A Descriptive Catalogue of the Sanskrit
Astronomical Manuscripts Preserved at the Maharafa Man Singh II Museum in Jaipur,
India, Iorthcoming.
8. Named in manuscripts 266 and 267 in Ngar transliterations are: zam Khn oI
Bangla, Abdallh Khn oI Patana, Da Shib oI Vnras, Itiqad Khn oI Dill, Khn
Dauran oI Ujjayin, Mahabat Khn Khn-i-khnn oI Burhnapura, Ujra Khn (?) oI
Ihora, JaIar Khn oI Ksmra, an unnamed person oI Multna, and Nitynanda him-
selI.
9. Manuscript 266 (Khasmohor 4960).
10. Manuscript 267 (Museum 23).
11. The Iourth copy, manuscript 269 (Khasmohor 4962), belonged to Jagannthaper-
haps Jaganntha Samrt.
12. Alwar 2627 and Anup 5332 and 5333; and SOI 9410.
13. See the verses cited in CESS A3, 174a.
14. Benares 34466 and Wellcome V.36.
15. CESS A4, 254b255b, and A5, 239b.
16. CESS A6 (in preparation).
17. CESS A4, 358b.
18. See 'The Indian Reception, pp. 477478.
19. H. Hunger and D. Pingree, 1he Astral Sciences of Mesopotamia, Ieiden 1999, pp.
229 and 237.
20. Manuscript 45 (Pundarka (jyotisa) 5).
21. CESS A3, 56a58a; A4, 95a; and A5, 113b114a.
22. Edition in preparation by S. Ikeyama and D. Pingree.
23. Edition in preparation by D. Pingree.
24. Edition in preparation by D. Pingree.
1 Ix+roiuc+or. Prx:rks
Nothing is known about the characteristics oI the earliest known Maghrib :f,
Ibn Ab l-Rijl al-Qayrawn`s Hall al-aqd wa-bayn al-rasd (beginning oI the
eleventh century), which seems to have been lost. Two centuries later, Abu`l-
Abbs Ibn Ishq al-Tamm al-Tunis (f. Tunis and Marrakesh ca. 11931222),
leIt an unfnished set oI tables which survive in a unique manuscript oI Hyder-
abad, discovered in 1978 by D. A. King.
1
The predominant infuence in Ibn
Ishq`s :f was that oI the Andalusian school represented by Ibn al-Zarqlluh
(d. 1100), Ibn al-Kammd (f. Cordova 111617) and Ibn al-Him (f. 1205).
This infuence continued along the thirteenth and Iourteenth centuries through
several 'editions oI the :f oI Ibn Ishq such as the one prepared by the
anonymous compiler oI the aIorementioned collection extant in the Hyder-
abad manuscript (ca. 665 H./1266680 H./1281), the Minhf oI Ibn al-Bann
oI Marrakesh (12561321),
2
and the two :fes composed by the Tunisian-
Andalusian astronomer Muhammad ibn al-Raqqm (d. 1315).
3
Andalusian
infuence is also present in two other Iourteenth century :fes, written by two
astronomers oI Constantine, in the Central Maghrib, who were active in Eez:
the one compiled by Abu`l-Hasan Al ibn Ab Al al-Qusantn, the canons
oI which were written in verse,
4
and the Zf al-Muwhq oI Abu`l-Qsim ibn
Azzz al-Qusantn (d. 1354),
5
partially based on observations made in Eez
with an armillary sphere ca. 1344.
All the aIorementioned :fes share a certain number oI characteristics
among which we should mention that their mean motion tables are sidereal, and
that they contain tables based on the theory oI trepidation that enable the user to
calculate the amount oI precession Ior a given date and, thus, obtain the tropical
longitudes oI heavenly bodies. Trepidation implies a variation in the obliquity
oI the ecliptic and these :fes also include tables, based on the model designed
by Ibn al-Zarqlluh, which implies the existence oI cycles that regulate the
diminution and the expected Iuture increase oI the obliquity oI the ecliptic. The
analysis oI a limited number oI new sources dated between the fIteenth and the
10
Ox +nr Iux:r 1:iirs ix S:xJ:q D:r`s 2I] LHPII
ulio Sams
uiio S:xs 286 Ox +nr Iux:r 1:iirs ix S:xJ:q D:r`s 2I] LHPII 287
early seventeenth centuries (three commentaries on the poem on timekeeping
written in 1391 by al-Jadhr,
6
and the Kitb al-adwr f tasyr al-anwr, writ-
ten by Ab Abd Allh al-Baqqr in 821/1418)
7
give some evidence that obser-
vations were made in the Maghrib in the 13th and 14th centuries. The result oI
these observations was that astronomers realized that the observed precession
oI the equinoxes was much larger than the values that could be computed with
trepidation tables and that the obliquity oI the ecliptic stubbornly continued to
diminish below the limits fxed in obliquity tables. The situation being thus, we
fnd a new contact with the East whichat least in the case oI al-Andalushad
been interrupted in the eleventh century. This brings to the Maghrib oI the late
Iourteenth century the 1f al-a:yf oI Muhy al-Dn al-Maghrib (d. 1283)
and the Zf al-Jadd oI Ibn al-Shtir (d.1375), and there is evidence oI adapta-
tions oI these :fes Ior their use in specifc cities oI the Maghrib. These :fes
compute directly tropical longitudes, are based on constant precession, reject
the existence oI cycles regulating the obliquity oI the ecliptic and they oIIer
results that agree with observations much better than those computed with the
tables oI the Andalusian school. In spite oI this the Andalusian tradition stayed
alive until the nineteenth century and coexisted with the 'new Eastern :fes.
The reason is clearly explained by the astrologer al-Baqqr: horoscopes were
cast using sidereal longitudes and my impression is that astrologers used Ibn
al-Bann`s Minhf and other similar sources, while astronomers and conscien-
cious muwaqqits preIered oriental :fes.
8
The last Eastern :f which was introduced in the Maghrib was Ulugh
Begh`s Zf-i Sultn (13931449),
9
usually called al-Zf al-Jadd in Maghrib
sources. I have no evidence that this :f was known in the Maghrib beIore the
end oI the seventeenth century, but it is obvious that it became very popular
during the eighteenth and nineteenth centuries. The Arabic translation oI the
introduction to this :f, made by Hasan b. Muhammad, known as Qd Hasan
al-Makk (f. 17th c.) reached Morocco, Ior the Hasaniyya Iibrary preserves a
manuscript oI this work dated in 1291/1874.
10
Interestingly, however, there were,
at least, two Tunisian recensions oI this :f prepared by Muhammad al-SharI,
called Sanjaq Dr al-Tnis and by Abd Allh Husayn Qusa b. Muhammad b.
Husayn al-HanaI al-Tnis.
11
Both authors are undated but MS 16650 oI the
Tunis National Iibrary (Khaldniyya collection) contains a copy oI Qusa`s
recensionunder the title Ghunyat al-tlib f taqwm al-kawkibwhere it
is stated that the work (not the copy) was fnished on Thursday 22nd Shawwl
1091 H./15 November 1680 (which was a Eriday), and that the tables have
been adapted to the longitude oI Tunis (41;45 Irom the western meridian).
12
As Ior Sanjaq Dr`s Zf al-Sharfalso called al-Zf al-Mukhtasar f ilm
al-tadl wa l-taqwmI have been able to study two manuscripts oI this work,
extant in MSS 1816 oI the National Iibrary in Tunis
13
(MS Ta hereaIter) and
18104 oI the library oI the late Tunisian scholar Hasan Husn Abd al-Wahhb,
now also in the Tunis National Iibrary
14
(MS Tb hereaIter). MS Tb is dated in
1107/16951696 (Iol. 62 r).
15
In this :f the mean positions oI the Sun, Moon
and planets are given, in intervals oI thirty years, between 1080/ 16691670 and
1290/18731874; on the other hand, the positions (?) oI the lunar mansions are
given, according to the observations oI Ulugh Begh, Ior the end oI year 1090/
16791680 (MS Ta, Iol. 36r; MS Tb, Iol 90 r) and the day, hour and minute oI
the spring equinox is given between years 1090/1679 and 1205/1790. All this
evidence points to a work made in the late seventeenth century,
16
and this agrees
with the presumed date Ior Qusa`s Ghunya: commentaries on the introduction
oI this :f were written towards the end oI the seventeenth or the beginning oI
the eighteenth century. One oI them was written by Ahmad b. Muhammad B
Daydah al-Qdr al-Qayrawn (f. beIore 1150/17371738), who praised, in
his introduction, the importance oI Sanjaq Dr`s :f but added that this work
was diIfcult to understand and to use. Eor that reason commentaries had been
written by his masters al-hff Ab Abd Allh Muhammad al-Qal and, beIore
him, by Abu`l-Hasan Al b. Mm al-HanaI, known as Karbso (or Karbsa)
(beIore 1163/17491750).
17
Al-Ql`s commentary was lost in the time oI al-
Qadr, while the work oI Karbso contained many mistakes and al-Qadr
wrote, Ior that reason, his Hshiya al sharh al-Sharf al-1unis Sanfaq Dr
li-:f al-Sultn Ulugh Beg,
18
which, apparently, is not a very interesting work.
Sanjaq Dr`s :f, like Qusa`s Ghunya, is computed Ior 41;45 oI longi-
tude and it contains tables oI oblique ascensions Ior a latitude oI 36;50 (MS
Ta, Iol. 28 r; MS Tb, Iol. 82 r), id. Ior the division oI the houses Ior the same
latitude (MS Ta, Iols. 28 v31 r; MS Tb, Iols. 82 v85 r), id. oI halI oI daylight
Ior the same latitude (MS Ta, Iol. 33 v; MS Tb, Iol. 87 v), lunar longitude at
sunset Ior a latitude 36;40 (MS Ta, Iol. 35 r; MS Tb, Iol. 89 r), lunar parallax
in longitude and latitude also Ior 36;40 (MS Ta, Iol. 36 v; MS Tb, Iol. 90 v)
and ascendants oI anniversaries Ior the same latitude (MSTA, Iol. 45 v; MSTb,
Iol. 98 r)): 36;50 is the modern value Ior the latitude oI Tunis and it does not
appear attested in any other historical source, while 36;40 is one oI the three
values Ior the latitude oI Tunis given by al-Khwrizm
19
and it is systematically
used by the compiler oI the Hyderabad recension oI the :f oI Ibn Ishq. Apart
Irom this, it is interesting to note that the prologue oI the Zf al-Sharf states that
the author is going to Iollow the habtaq method, which means that planetary
equations are computed in a simplifed way using double argument tables. This
type oI tables is documented in the East Irom the time oI IbnYnus (d. 1009),
20
but I do not know oI any instance in which they appear in the Maghrib beIore
the Zf al-Sharf Ior the computation oI lunar or planetary equations. In spite
oI this, Ibn al-Kammd (f. Cordova 1116) used double argument tables Ior
the computation oI time Irom mean to true syzygy
21
and Ior the calculation oI
solar eclipses,
22
and the Hyderabad recension oI Ibn Ishq`s :f contains a set
oI double argument planetary latitude tables, based on Ptolemaic parameters,
attributed to an otherwise unknown Ibn al-Baytr.
23
In the equation tables oI the
Zf al-Sharf one enters the table in the vertical sense with the mean anomaly oI
the Moon or the planet (tabulated Irom 0 to 360 with 6 intervals) and in the
horizontal sense with the :odiacal sign oI the marka: (the double elongation in
the case oI the Moon, and the mean longitude oI the planet computed Irom the
apogee): the total equation (center anomaly) will be read directly in the inter-
section oI the two and will be added to the mean longitude oI the planet.
The present chapter intends to begin the study oI this set oI double argu-
ment equation tables with the case oI the Moon, which seems particularly
accessible. The Zf al-Sharf contains two diIIerent sets oI tables Ior the compu-
tation oI the solar and lunar longitudes: on the one hand, mean motions oI the
Sun, Moon, solar apogee, lunar nodes, and double elongation calculated to the
precision oI minutes (MS Ta, Iols. 10 r and v; MS Tb, Iols. 63 v64 r); a table
oI the solar equation (MS Ta, Iol. 10 r; MS Tb, Iol. 63 v), also calculated to the
precision oI minutes and with an interval oI 6 oI the argument (tabular maxi-
mum 3;52 Ior arguments 264270, the table being obviously displaced verti-
cally 1;56); the aIorementioned double argument table Ior the lunar equation
to the precision oI minutes (MS Ta, Iols. 11 r and v; MS Tb, Iols. 65 v66 r).
The second set oI solar and lunar tables is Iar more precise. The solar
mean motion tables (MS Ta, Iol. 23 r; MS Tb, Iol. 78 r) give the mean motion
oI the Sun and oI the solar apogee (to the precision oI thirds), Ior hours, days,
lunar months, lunar years mabsta, and the positions oI the marka: (solar lon-
gitude Irom the apogee) Ior the end oI years 10801290 H. in 30-year intervals.
Here, as in the rest oI the :f, the system oI intercalation is the same as that used
by Ulugh Beg, in which years 2, 5, 7, 10, 13, 15, 18, 21, 24, 26, and 29 oI the
30-year cycle are leap years (kabsa). The corresponding solar equation table
(MS Ta, Iol. 24 r; MS Tb, Iol. 78 r) is calculated Ior every degree oI the argu-
ment Irom 0 to 359, to the precision oI thirds, it has a vertical displacement
oI 1;55,53,12,
24
and it reaches a tabular maximum oI 3;51,46,24 Ior an argu-
ment oI 268. As Ior the Moon, the mean motion tables (MS Ta, Iols. 24 v25
r; MS Tb, Iols. 78 v79 r) give the corresponding mean motions in longitude
(wasat), anomaly (khssa), double elongation (marka:) and nodes (faw:ahar)
Ior the same periods, computed to the precision oI thirds (longitude) or seconds
(the rest). Three other tables (MS Ta, Iols. 25 v26 v; MS Tb, Iols. 79 v80 v)
tabulate more or less standard lunar equations which will be described in detail
below. It is easy to see thatwith the exception oI the double argument equa-
tion tables oI the frst set and the lunar equations oI the secondthe frst set is
just the result oI a rounding oI the second one to the precision oI minutes.
2 Mr:x Soi:r :xi Iux:r Mo+ioxs
25
All the mean motion values derive Irom Ulugh Beg`s Zf-i Sultn.
26
Precession/motion of the apogees: the analysis oI the table oI single years and that
oI positions Ior 10801290H. gives two diIIerent results
0;0,0,8,27,14,19 (years)
and a value between
0;0,0,8,27,14,24,31,41,45,32
and
0;0,0,8,27,14,28,52,55,30,50
Ior the table oI periods. This corresponds,
precisely, to a precession oI 1 in 70 Persian years.
Sun: two slightly diIIerent parameters appear to have been used.
0;59,8,11,10,28,20 (days, months, positions 10801290H.)
0;59,8,11,10,28,37 (single years)
Lunar longitude: the diIIerent sets oI tables are mutually compatible and the mean
motion parameter used lies between
13;10,35,1,47,53,43,49,40,23,58
and
13;10,35,1,47,53,45,38,5,22,28
Lunar anomaly: the underlying parameter is
13;3,53,55,54,24
Double elongation: the underlying parameter is
24;22,53,23,46,8,59,46,47,36
It is easy to check that this daily parameter does not correspond to the
double value oI the diIIerence between the mean motion in longitude oI the
Moon and that oI the Sun. Eor the computation we begin subtracting the daily
motion oI the apogee Irom the rounded value oI the lunar mean motion in lon-
gitude:
13;10,35,1,47,53,45 (0,1) 0;0,0,8,27,14,27 (0,2)
13;10,34,53,20,39,18 (0,3).
27
and then,
13;10,34,53,20,39,18 ( 0,3) 0;59,8,11,10,28,20
12;11,26,42,10,10,58 ( 0,3)
The corresponding mean motion in double elongation will be
24;22,53,24,20,21,56 ( 0,6)
This inconsistency is already present in Ulugh Beg`s tables and it may
be the reason (?) oI the incoherence we fnd in the tables oI positions Ior years
10801290. Table 10.1 collects all the inIormation on positions oI the solar
center (col. 1), solar apogee (col. 2), lunar longitude (col. 4), and double elon-
gation (col. 7): the values transcribed are those oI the two manuscripts oI the Zf
al-Sharf. A control oI errors has been made in all the cases: amounts between
parentheses in col. 1 correspond to the diIIerence (expressed in thirds) between
the tabular and the recomputed value. Columns 3, 5, 6, 8 have been calculated
Irom the corresponding tabular values and column 9 contains line-by-line diI-
Ierences in column 8.
Table 10.2 corresponds to positions also oI the solar center (marka:, col.
1), solar apogee (col. 2), lunar longitude (col. 4), and double elongation (col. 7)
Ior years 841 H.871 H., at one year intervals, as they appear in Ulugh Beg`s
Zf-i Sultn. An asterisk (*) marks the leap-years (kabsa). Here, once more,
we fnd a disagreement between the values oI the double elongation computed
Irom the solar and lunar positions and those appearing in the table. It is easy
to check, however, that the positions in the Zf al-Sharf derive Irom those in
the Zf-i Sultn. To establish this, I have compared two sets oI positions: those
corresponding to 841 H. nqisa (which means midday oI the last day oI 840)
and those corresponding to the last day oI 1080: the interval between these two
dates should be 85048 days but the actual interval used Ior the recomputation is
one day less (?). On the other hand a correction has been introduced to account
Ior the diIIerence in geographical longitude between Samarqand (longitude
99;16 in Ulugh Beg`s Zf) and Tunis (41;45 in the Zf al-Sharf): the diIIer-
ence in longitude (57;31) corresponds to 3
h
50
m
4
s
or 0;9,35,10
d
. The positions
used in the two :fes are:
Solar marka: Solar apogee Iunar longitude Double elongation
841 18;26,0,13 90;30,4,48 115;51,56,28 25;12,49
1080 321;32,2,24 93;49,47,58 51;1,29,2 2;26,56
Eor n number oI revolutions in each case, we can see that:
Solar markaz:
18;26,0,13 0;59,8,11,10,28,20 * 85047;9,35,10
d
360 * n 321;32,2,27,9
Solar apogee:
90;30,4,48 0;0,0,8,27,14,26 * 85047;9,35,10
d
93;49,47,57,28
Lunar longitude:
115;51,56,28 13;10,35,1,47,53,45 * 85047;9,35,10
d
360 * n 51;1,29,2,25
Double elongation:
25;12,49 24;22,53,23,46,9 * 85047;9,35,10
d
360 * n 2;26,55,2
T
a
b
l
e
1
0
.
1
P
o
s
i
t
i
o
n
s
i
n
t
h
e
Z
f
a
l
-
S
h
a
r
f
(
1
)
(
2
)
(
3
)
(
4
)
(
5
)
(
6
)
(
7
)
(
8
)
(
9
)
(
1
)
(
2
)
(
4
)
(
3
)
(
5
)
*
2
(
7
)
(
6
)
S
u
n
(
c
e
n
t
e
r
)
A
p
o
g
e
e
|
S
o
l
a
r
I
o
n
g
.
I
u
n
a
r
I
o
n
g
.
|
E
l
o
n
g
a
t
i
o
n
|
D
o
u
b
l
e
e
l
.
D
o
u
b
l
e
e
l
.
(
t
a
b
.
)
D
i
I
I
.
1
0
8
0
3
2
1
;
3
2
,
2
,
2
4
9
3
;
4
9
,
4
7
,
5
8
5
5
;
2
1
,
5
0
,
2
2
5
1
;
1
,
2
9
,
2
3
5
5
;
3
9
,
3
8
,
4
0
3
5
1
;
1
9
,
1
7
,
2
0
2
;
2
6
,
5
6
1
1
;
7
,
3
9
1
1
1
0
3
5
9
;
3
1
,
3
0
,
2
1
9
4
;
1
4
,
4
5
,
5
3
9
3
;
4
6
,
1
6
,
1
4
8
9
;
1
8
,
1
2
,
3
9
3
5
5
;
3
1
,
5
6
,
2
5
3
5
1
;
3
,
5
2
,
5
0
2
;
9
,
5
1
1
1
;
5
,
5
8
0
;
1
,
4
1
1
1
4
0
3
7
;
3
0
,
5
8
,
1
8
(
1
)
9
4
;
3
9
,
4
3
,
4
8
1
3
2
;
1
0
,
4
2
,
6
1
2
7
;
3
4
,
5
6
,
1
7
3
5
5
;
2
4
,
1
4
,
1
1
3
5
0
;
4
8
,
2
8
,
2
2
1
;
5
2
,
4
4
1
1
1
;
4
,
1
6
0
;
1
,
4
2
1
1
7
0
7
5
;
3
0
,
2
6
,
1
3
9
5
;
4
,
4
1
,
4
2
1
7
0
;
3
5
,
7
,
5
6
1
6
5
;
5
1
,
3
9
,
5
4
3
5
5
;
1
6
,
3
1
,
5
8
3
5
0
;
3
3
,
3
,
5
6
1
;
3
5
,
4
0
1
1
;
2
,
3
6
0
;
1
,
4
0
1
2
0
0
1
1
3
;
2
9
,
5
4
,
9
(
1
)
9
5
;
2
9
,
3
9
,
3
6
2
0
8
;
5
9
,
3
3
,
4
5
2
0
4
;
8
,
2
3
,
3
2
3
5
5
;
8
,
4
9
,
4
7
3
5
0
;
1
7
,
3
9
,
3
4
1
;
1
8
,
3
4
1
1
;
0
,
5
4
0
;
1
,
4
2
1
2
3
0
1
5
1
;
2
9
,
2
2
,
5
(
1
)
9
5
;
5
4
,
3
7
,
3
1
2
4
7
;
2
3
,
5
9
,
3
6
2
4
2
;
2
5
,
7
,
9
3
5
5
;
1
,
7
,
3
3
3
5
0
;
2
,
1
5
,
6
1
;
1
,
2
9
1
0
;
5
9
,
1
4
0
;
1
,
4
0
1
2
6
0
1
8
9
;
2
8
,
5
0
,
2
(
1
)
9
6
;
1
9
,
3
5
,
2
6
2
8
5
;
4
8
,
2
5
,
2
8
2
8
0
;
4
1
,
5
0
2
,
4
6
3
5
4
;
5
3
,
2
5
,
1
8
3
4
9
;
4
6
,
5
0
,
3
6
0
:
4
4
,
2
2
3
1
0
;
5
7
,
3
1
0
;
1
,
4
3
1
2
9
0
2
2
7
;
2
8
,
1
7
,
5
8
(
1
)
9
6
;
4
4
,
3
3
,
2
0
3
2
4
;
1
2
,
5
1
,
1
8
|
3
1
8
;
5
8
,
3
4
,
2
4
3
5
4
;
4
5
,
4
3
,
6
|
3
4
9
;
3
1
,
2
6
,
1
2
|
0
:
2
7
,
1
8
1
0
;
5
5
,
5
2
0
;
1
,
3
9
1
.
M
S
T
b
4
5
"
2
.
M
s
.
T
a
5
6
"
3
.
M
S
T
b
2
4
"
T
a
b
l
e
1
0
.
2
P
o
s
i
t
i
o
n
s
i
n
U
l
u
g
h
B
e
g
`
s
Z
j
-
i
S
u
l
t
a
n
(
1
)
(
2
)
(
3
)
(
1
)
(
2
)
(
4
)
(
5
)
(
4
)
(
3
)
(
6
)
(
5
)
*
2
(
7
)
(
8
)
(
7
)
(
6
)
(
9
S
u
n
(
C
e
n
t
e
r
)
A
p
o
g
e
e
|
S
o
l
a
r
l
o
n
g
.
I
u
n
a
r
l
o
n
g
.
|
E
l
o
n
g
a
t
i
o
n
|
D
o
u
b
l
e
e
l
.
D
o
u
b
l
e
e
l
.
(
t
a
b
.
)
D
i
I
I
.
8
4
1
1
8
;
2
6
,
0
,
1
3
9
0
;
3
0
,
4
,
4
8
1
0
8
;
5
6
,
5
,
1
1
1
5
;
5
1
,
5
6
,
2
8
1
6
;
5
5
,
5
1
,
2
7
1
3
;
5
1
,
4
2
,
5
4
2
5
;
1
2
,
4
9
1
1
;
2
1
,
6
8
4
2
7
;
2
0
,
1
8
,
1
9
9
0
;
3
0
,
5
4
,
4
1
9
7
;
5
1
,
1
3
,
0
1
0
0
;
1
8
,
3
7
,
5
2
2
;
2
7
,
2
4
,
5
4
;
5
4
,
4
8
,
1
0
1
6
;
1
5
,
5
1
1
1
;
2
1
,
3
0
;
0
,
3
8
4
3
*
3
5
7
;
1
3
,
4
4
,
1
6
9
0
;
3
1
,
4
4
,
4
2
8
7
;
4
5
,
2
8
,
5
8
9
7
;
5
5
,
5
2
,
4
3
3
1
0
;
1
0
,
2
3
,
4
5
2
0
;
2
0
,
4
7
,
3
0
3
1
;
4
1
,
4
7
1
1
;
2
1
,
0
0
;
0
,
3
8
4
4
3
4
6
;
8
,
2
,
1
2
9
0
;
3
2
,
3
4
,
3
5
7
6
;
4
0
,
3
6
,
4
7
8
2
;
2
2
,
3
3
,
2
0
5
;
4
1
,
5
6
,
3
3
1
1
;
2
3
,
5
3
,
6
2
2
;
4
4
,
4
9
1
1
;
2
0
,
5
3
0
;
0
,
7
8
4
5
3
3
5
;
2
,
2
0
,
7
9
0
;
3
3
,
2
4
,
2
7
6
5
;
3
5
,
4
4
,
3
4
6
6
;
4
9
,
1
3
,
5
7
1
;
1
3
,
2
9
,
2
3
2
;
2
6
,
5
8
,
4
6
1
3
;
4
7
,
5
1
1
1
;
2
0
,
5
2
0
;
0
,
1
8
4
6
*
3
2
4
;
5
5
,
4
6
,
1
4
9
0
;
3
4
,
1
4
,
2
9
5
5
;
3
0
;
0
,
4
3
6
4
;
2
6
,
2
9
,
3
5
8
;
5
6
,
2
8
,
5
2
1
7
;
5
2
,
5
7
,
4
4
2
9
;
1
3
,
4
7
1
1
;
2
0
,
4
9
0
;
0
,
3
8
4
7
3
1
3
;
5
0
,
4
,
1
0
9
0
;
3
5
,
4
,
2
1
4
4
;
2
5
,
8
,
3
1
4
8
;
5
3
,
1
0
,
1
1
4
;
2
8
,
1
,
4
0
8
;
5
6
,
3
,
2
0
2
0
;
1
6
,
4
9
1
1
;
2
0
,
4
6
0
;
0
,
3
8
4
8
*
3
0
3
;
4
3
,
3
0
,
1
7
9
0
;
3
5
,
5
4
,
2
2
3
4
;
1
9
,
2
4
,
3
9
4
6
;
3
0
,
2
5
,
5
0
1
2
;
1
1
,
1
,
1
1
2
4
;
2
2
,
2
,
2
2
3
5
;
4
2
,
4
5
1
1
;
2
0
,
4
3
0
;
0
,
3
8
4
9
2
9
2
;
3
7
,
4
8
,
1
3
9
0
;
3
6
,
4
4
,
1
5
2
3
;
1
4
,
3
2
,
2
8
3
0
;
5
7
,
6
,
2
6
7
;
4
2
,
3
3
,
5
8
1
5
;
2
5
,
7
,
5
6
2
6
;
4
5
,
4
7
1
1
;
2
0
,
3
9
0
;
0
,
4
8
5
0
2
8
1
;
3
2
,
6
,
9
9
0
;
3
7
,
3
4
,
9
1
2
;
9
,
4
0
,
1
8
1
5
;
2
3
,
4
7
,
3
3
;
1
4
,
6
,
4
5
6
;
2
8
,
1
3
,
3
0
1
7
;
4
8
,
4
9
1
1
;
2
0
,
3
6
0
;
0
,
3
8
5
1
*
2
7
1
;
2
5
,
3
4
,
1
6
9
0
;
3
8
,
2
4
,
9
2
;
3
,
5
8
,
2
5
1
3
;
1
,
2
,
4
1
1
0
;
5
7
,
4
,
1
6
2
1
;
5
4
,
8
,
3
2
3
3
;
|
1
|
4
,
4
5
4
1
1
;
2
0
,
3
6
0
;
0
,
0
8
5
2
2
6
0
;
1
9
,
5
0
,
5
1
1
9
0
;
3
9
,
1
4
,
2
3
5
0
;
5
9
,
4
,
1
3
3
5
7
;
2
7
,
4
3
,
1
8
6
;
2
8
,
3
9
,
5
1
2
;
5
7
,
1
8
,
1
0
2
4
;
1
7
,
4
6
1
1
;
2
0
,
2
8
0
;
0
,
8
8
5
3
2
4
9
;
|
1
4
|
6
,
8
,
7
9
0
;
4
0
,
3
,
5
4
3
3
9
;
5
4
,
1
2
,
1
3
4
1
;
5
4
,
2
3
,
5
4
2
;
0
,
1
1
,
5
3
4
;
0
,
2
3
,
4
6
1
5
;
2
0
,
4
9
1
1
;
2
0
,
2
5
0
;
0
,
3
8
5
4
*
2
3
9
;
7
,
3
6
,
1
4
9
0
;
4
0
,
5
3
,
5
6
3
2
9
;
4
8
,
3
0
,
1
0
3
3
9
;
3
1
,
3
9
,
3
3
9
;
4
3
,
9
,
2
3
1
9
;
2
6
,
1
8
,
4
6
3
0
;
4
6
,
4
5
1
1
;
2
0
,
2
6
0
;
0
,
1
8
5
5
2
2
8
;
1
,
5
|
4
|
7
,
1
0
9
0
;
4
1
,
4
3
,
4
8
3
1
8
;
4
3
,
3
7
,
5
8
3
2
3
;
5
8
,
2
0
,
9
5
;
1
4
,
4
2
,
1
1
1
0
;
2
9
,
2
4
,
2
2
2
1
;
4
9
,
4
7
1
1
;
2
0
,
2
3
0
;
0
,
3
8
5
6
*
2
1
7
;
5
5
,
|
2
2
|
8
,
7
9
0
;
4
2
,
3
3
,
5
0
3
0
8
;
3
7
,
5
5
,
5
7
3
2
1
;
3
5
,
3
5
,
|
4
|
8
9
1
2
;
5
7
,
3
9
,
5
1
2
5
;
5
5
,
1
9
,
4
2
3
7
;
1
5
,
4
3
1
1
;
2
0
,
2
3
0
;
0
,
0
8
5
7
2
0
6
;
4
9
,
3
6
,
1
0
1
3
9
0
;
4
3
,
2
3
,
4
2
2
9
7
;
3
2
,
5
9
,
5
5
3
0
6
;
2
,
1
6
,
2
5
8
;
2
9
,
1
6
,
3
0
1
6
;
5
8
,
3
3
,
0
2
8
;
1
8
,
4
5
1
1
;
2
0
,
1
2
0
;
0
,
9
8
5
8
1
9
5
;
4
3
,
5
4
,
8
9
0
;
4
4
,
1
3
,
3
4
2
8
6
;
2
8
,
7
,
4
2
2
9
0
;
2
8
,
5
7
,
2
4
;
0
,
4
9
,
2
0
8
;
1
,
3
8
,
4
0
1
9
;
2
1
,
4
7
1
1
;
2
0
,
8
0
;
0
,
4
8
5
9
*
1
8
5
;
3
7
,
2
0
,
1
5
9
0
;
4
5
,
3
,
3
6
2
7
6
;
2
2
,
2
3
,
5
1
2
8
8
;
6
,
1
2
,
3
9
1
1
;
4
3
,
4
8
,
4
8
2
3
;
2
7
,
3
7
,
3
6
3
4
;
4
7
,
4
3
1
1
;
2
0
,
5
0
;
0
,
3
8
6
0
1
7
4
;
3
1
,
3
8
,
1
1
9
0
;
4
5
,
5
3
,
2
9
2
6
5
;
1
7
,
3
1
,
4
0
2
7
2
;
3
2
,
5
3
,
1
6
7
;
1
5
,
2
1
,
3
6
1
4
;
3
0
,
4
3
,
1
2
2
5
;
5
0
,
4
5
1
1
;
2
0
,
2
0
;
0
,
3
8
6
1
1
6
3
;
2
5
,
5
6
,
7
9
0
;
4
6
;
4
3
,
2
2
2
5
4
;
1
2
,
3
9
,
2
9
2
5
6
;
5
9
,
3
3
,
5
2
2
;
4
6
,
5
4
,
2
3
5
;
3
3
,
4
8
,
4
6
1
6
;
5
3
,
4
8
1
1
;
1
9
,
5
9
0
;
0
,
3
8
6
2
*
1
5
3
;
1
9
,
2
2
,
1
4
9
0
;
4
7
,
3
3
,
2
3
2
4
4
;
6
,
5
5
,
3
7
2
5
4
;
3
6
,
4
9
,
3
1
1
0
;
2
9
,
5
3
,
5
4
2
0
;
5
9
,
4
7
,
4
8
3
2
;
1
9
,
4
3
1
1
;
1
9
,
5
5
0
;
0
,
4
8
6
3
1
4
2
;
1
3
,
4
0
,
1
0
9
0
;
4
8
,
2
3
,
1
5
2
3
3
;
2
,
3
,
2
5
2
3
9
;
3
,
3
0
,
7
6
;
1
,
2
6
,
4
2
1
2
;
2
,
5
3
,
2
4
2
3
;
2
2
,
4
5
1
1
;
1
9
,
5
2
0
;
0
,
3
8
6
4
1
3
1
;
7
,
5
8
,
6
9
0
;
4
9
,
1
3
,
8
2
2
1
;
5
7
,
1
1
,
1
4
2
2
3
;
3
0
,
1
0
,
4
4
1
;
3
2
,
5
9
,
3
0
3
;
5
,
5
9
,
0
1
4
;
2
5
,
4
8
1
1
;
1
9
,
4
9
0
;
0
,
3
8
6
5
*
1
2
1
;
1
,
2
4
,
1
2
9
0
;
5
0
,
3
,
9
2
1
1
;
5
1
,
2
7
,
2
1
2
2
1
;
7
,
2
6
,
2
2
9
;
1
5
,
5
9
,
1
1
8
;
3
1
,
5
8
,
2
2
9
;
5
1
,
4
3
1
1
;
1
9
,
4
5
0
;
0
,
4
8
6
6
1
0
9
;
5
5
,
4
2
,
8
9
0
;
5
0
,
5
3
,
2
2
0
0
;
4
6
,
3
5
,
1
0
2
0
5
;
3
4
,
6
,
5
9
4
;
4
7
,
3
1
,
4
9
9
;
3
5
,
3
,
3
8
2
0
;
5
4
,
4
6
1
1
;
1
9
,
4
2
0
;
0
,
3
8
6
7
*
9
9
;
4
9
,
8
,
1
5
9
0
;
5
1
,
4
3
,
3
1
9
0
;
4
0
,
5
1
,
1
8
2
0
3
;
1
1
,
2
2
,
3
7
1
2
;
3
0
,
3
1
,
1
9
2
5
;
1
,
2
,
3
8
3
6
;
2
0
,
4
1
1
1
;
1
9
,
3
8
0
;
0
,
4
8
6
8
8
8
;
4
3
,
2
6
,
1
1
9
0
;
5
2
,
3
2
,
5
6
1
7
9
;
3
5
,
5
9
,
7
1
8
7
;
3
8
,
3
,
1
4
8
;
2
,
4
,
7
1
6
;
4
,
8
,
1
4
2
7
:
2
3
,
4
3
1
1
;
1
9
,
3
5
0
;
0
,
3
8
6
9
7
7
;
3
7
,
4
4
,
7
9
0
;
5
3
,
2
2
,
4
9
1
6
8
;
3
1
,
6
,
5
6
1
7
2
;
4
,
4
3
,
5
0
3
;
3
3
,
3
6
,
5
4
7
;
7
,
1
3
,
4
6
1
8
;
2
6
,
4
6
1
1
;
1
9
,
3
2
0
;
0
,
3
8
7
0
*
6
7
;
3
1
,
1
0
,
1
4
9
0
;
5
4
,
1
2
,
5
0
1
5
8
;
2
5
,
2
3
,
4
1
6
9
;
4
1
,
5
9
,
2
9
1
1
;
1
6
,
3
6
,
2
5
2
2
;
3
3
,
1
2
,
5
0
3
2
;
5
2
,
4
1
1
1
;
1
9
,
2
8
0
;
0
,
4
8
7
1
5
6
;
2
5
,
2
8
,
1
0
9
0
;
5
5
,
2
,
4
3
1
4
7
;
2
0
,
3
0
,
5
3
|
1
5
4
;
8
,
4
0
,
5
6
;
4
8
,
9
,
1
2
1
3
;
3
6
,
1
8
,
2
4
|
2
4
;
5
5
,
4
4
1
1
;
1
9
,
2
6
0
;
0
,
2
1
.
4
s
2
5
;
5
1
,
5
6
,
1
1
i
n
t
h
e
A
l
e
p
p
o
m
s
.
2
.
4
s
i
n
t
h
e
A
l
e
p
p
o
m
s
.
3
.
4
s
i
n
t
h
e
A
l
e
p
p
o
m
s
.
4
.
4
'
i
n
t
h
e
A
l
e
p
p
o
m
s
.
5
.
5
0
i
n
t
h
e
A
l
e
p
o
m
s
.
T
h
e
c
o
r
r
e
c
t
v
a
l
u
e
s
h
o
u
l
d
b
e
5
2
"
.
6
.
|
1
4
|
i
s
t
h
e
c
o
r
r
e
c
t
v
a
l
u
e
:
t
h
e
A
l
e
p
p
o
m
s
.
h
a
s
4
0
.
7
.
5
2
i
n
t
h
e
A
l
e
p
p
o
m
s
.
8
.
4
8
i
n
t
h
e
A
l
e
p
p
o
m
s
.
9
.
1
8
t
h
i
r
d
s
i
n
t
h
e
A
l
e
p
p
o
m
s
.
1
0
.
3
6
i
n
t
h
e
A
l
e
p
p
o
m
s
.
T
h
e
c
o
r
r
e
c
t
v
a
l
u
e
s
h
o
u
l
d
b
e
4
0
"
.
3 Soi:r :xi Iux:r Lqu:+ioxs
As I have previously stated, the Zf al-Sharf contains two displaced tables oI
the solar equation, the frst being the result oI a rounding, to the approxima-
tion oI minutes, oI the second one, which was copied Irom Ulugh Beg`s Zf-i
Sultn. Table 10.3 contains excerpts oI the aIorementioned second solar equa-
tion table oI the Zf al-Sharf: the amounts between parentheses correspond to
the diIIerences, in thirds, between tabular and recomputed values, an eccentric-
ity oI 2;1,20
p
having been used Ior the recomputation.
28
The set oI lunar equation tables also derive IromUlugh Beg`s Zf-i Sultn.
The lunar equation oI the center, called tadl awwal ('frst equation) in the Zf
al-Sharf (excerpts in table 10.4), although displaced vertically 13;15,34, is
a standard table calculated Ior an eccentricity oI 12;33,22
p
, the radius oI the
deIerent being 60
p
. The Iour tables (see excerpts below in tables 10.5, 10.6,
Table 10.3
Solar Equation
0 1;55,53,12 180 1;55,53,12
10 1;36,24,49 190 2;16,41,50 (1)
20 1;17,28,37 (1) 200 2;36,48,32
30 0;59,36,8 210 2;55,33,19 (2)
40 0;43,17,44 220 3;12,19,28
50 0;29,1,47 230 3;26,35,19
60 0;17,14,5 240 3;37,55,7 (1)
70 0;8,17,7 250 3;45,59,46 (1)
80 0;2,29,14 260 3;50,37,13
268 3;51,46,24
269 3;51,|45|,29
1
90 0:0,3,57 270 3;51,|4|2,27
2
271 3;51,37,19
92 0;0,0,0
100 0;1,9,11 280 3;49,17,10
110 0;5,46,38 (1) 290 3;43,29,17
120 0;13,51,17 (1) 300 3;34,32,19
130 0;25,11,5 310 3;22,44,37
140 0;39,26,57 320 3;8,28,40
150 0;56,13,17 330 2;52,10,16
160 1;14,57,52 340 2;34,17,47 (1)
170 1;35,4,34 (1) 350 2;15,21,35
1. 45' is legible in MS Tb Iol. 78 r.
2. 42' is also legible in MS Tb, Iol. 78 r.
10.7, 10.8) designed Ior the computation oI the equation oI anomaly are not so
standard:
29
the frst, called tadl thn (second equation), is a table displaced
vertically 7;37,28, but, iI we subtract this constant Irom the tabular values, we
will discover that the two halves oI the table are not symmetrical. The solution
to the diIfculties posed by this table can be Iound in the instructions, given in
the canons, Ior the computation oI the lunar equation oI anomaly (MS Ta, Iol.
3v; MS Tb, Iols. 56 v57 r):
30
With the mean motion tables obtain the mean longitude (!
m
) and the mean anom-
aly (o
m
) oI the Moon as well as the marka: (double elongation, 2f ).
Enter the table oI the lunar equation oI the center (tadl awwal) with 2f and obtain
the equation oI the center (j).
Add j o
m
o
v
(true anomaly, al-khssa al-muaddala).
Enter with o
v
the table oI the 'second equation (tadl thn) and obtain ,.
Table 10.4
Iunar Equation oI the Center
0 13;15,34 180 13;15,34
10 14;44,3 190 9;43,4 (1)
20 16;|1|2;2|8|
1
200 6;30,10 (1)
30 17;40,8 (1) 210 3;51,33 (1)
40 19;6,57 (1)
2
220 1;54,|4|3
3
50 20;31,57 230 0;40,26
60 21;53,58 240 0;5,1 (1)
246 0;0,0
70 23;11,11 250 0;2,34
80 24;21,6 260 0;26,32
90 25;20,20 270 1;10,|4|8
4
100 26;4,36 280 2;10,2
110 26;28,34 290 3;19,57
114 26;31,8
120 26;26,7 (1) 300 4;37,10
130 25;50,42 310 5;59,11
140 24;36,25 320 7;24,13 (1)
150 22;39,35 (1) 330 8;51,0 (1)
160 20;0,58 (1) 340 10;18,48
170 16;48,4 (1) 350 11;47,5
1. 16;12,20 in MS Tb, Iol. 79 v.
2. 19;6,55 in MS Tb, Iol. 79 v.
3. 43" are legible in MS Tb Iol. 79 v.
4. 48" are clearly legible in MS Tb, Iol. 79 v.
Table 10.5
Tadl Thn ('Second Equation)
Eirst Table Ior the Computation oI the Iunar Equation oI Anomaly
0 7;37,28 180 7;37,28
10 6;27,24 (1) 190 8;34,7 (1)
20 5;18,48 (1) 200 9;28,32 (1)
30 4;13,7 210 10;18,38 (1)
40 3;11,49 220 11;2,38
50 2;16,23 (1) 230 11;39,5
1
60 1;28,22 (1) 240 12;6,59
70 0;49,|1|7
2
(2) 250 12;25,43 (2)
80 0;2|4|,40
3
260 12;34,59 (1)
265 12;36,10 (1)
90 0;3,54 270 12;3|5|,3
4
(1)
98 0;0,0 (6)
100 0;0,1|8|
5
280 12;26,13 (1)
110 0;10,53 (2) 290 12;|9|,10 (1)
6
120 0;36,23 (2) 300 11;44,39
130 1;16,57 (1) 310 11;13,39 (1)
140 2;12,0 (1) 320 10;37,6
150 3;20,13 330 9;56,8 (1)
160 4;39,21 (1) 340 9;11,47
170 6;6,21 350 8;25,11
1. 11;39,35 in MS Tb, Iol. 80 r.
2. 0;49,47 in Ta (Iol. 26 r); 0;49,17 in Tb (Iol. 80 r).
3. 0;20,40 in MS Tb, Iol. 80 r.
4. 12;34,3 in MS Tb, Iol. 26 r; 12;35,3 in MS Ta, Iol. 80 r.
5. 0;10,13 in MS Tb, Iol. 80 r.
6. 12;10,9 in MS Ta, Iol. 26 r; 12;9,10 in MS Tb, Iol. 80 r.
Enter with o
v
the table oI the 'variation (ikhtilf ) and obtain A,.
Enter with the marka: (2f ) the frst table oI the 'minutes oI the anomalies
(daqiq al-hisas) iI 0 o
v
180 and obtain m
1
. II 180 o
v
360, then do the
same with the second table oI the 'minutes oI the anomalies and obtain m
2
.
Multiply mA,.
Add , mA, and you will obtain the 'corrected equation |oI anomaly| (tadl
muhkam).
The true longitude oI the Moon in its 'inclined sphere (falak mil) will be:
!
v
!
m
, mA,.
Instructions are given, aIter this, to calculate the equation oI time and to
obtain the true longitude oI the Moon on the ecliptic.
Table 10.6
IkhtilI al-Qamar ('Variation oI the Moon)
Second Table Ior the Computation oI the Iunar Equation oI Anomaly
0 0;0 90 2;35,59 (1)
10 0;22,21 (1) 100 2;39,39 (1)
102 2;39,49
20 0;44,21 (1) 110 2;38,20
30 1;5,41 (1) 120 2;31,34 (2)
40 1;26,1 130 2;18,54 (1)
50 1;44,54 140 2;0,18 (1)
60 2;1,55 (1) 150 1;36,4 (2)
70 2;16,29 (1) 160 1;7,3
80 2;28,3 (1) 170 0;34,28 (1)
Table 10.7
Daqiq al-Hissa ('Minutes oI the Anomaly)
Eirst Eunction oI Interpolation Ior the Computation oI the Iunar Anomaly
0/360 60;0' 90/270 33;12'
30/330 56;46 (1) 120/240 17;24 (1)
60/300 47;24 (1) 150/210 4;49 (2)
179/181 0;1 (1)
Table 10.8
Daqiq al-Hissa ('Minutes oI the Anomaly)
Second Eunction oI Interpolation Ior the Computation oI the Iunar Anomaly
0/360 0;0' 90/270 26;48'
30/330 3;14 (1) 120/240 42;36 (1)
60/300 12;36 (1) 150/210 55;11 (2)
179/181 59;59 (3)
Eollowing the logic oI the instructions I have just summarized, we can see
that the table oI the 'second equation (see table 10.6) computes two diIIerent
Iunctions:
Eor 0 o
v
180, the Iuncion involved is the lunar equation oI anomaly Ior
the minimum distance oI the center oI the lunar epicycle Irom the center oI the
Earth (,
R-e
). The table is displaced vertically 7;37,28 and, as the equation oI
anomaly is negative Ior 0 o 180, it gives:
7;37,28 ,
R-e
.
The table can be recomputed using a radius oI the epicycle oI 6;17,46,
31
an eccentricity oI 12;33,22
p
and a radius oI the deIerent oI 60
p
. The choice oI
the constant oI displacement is inadequate and it corresponds to an error in
the computation oI the maximum value oI the table, which should be 7;37,34
instead oI 7;37,28. The mistake was already present in Ulugh Beg`s :f and
7;37,28 is explicitely mentioned by Mram Chalab, in his commentary on
Ulugh Beg`s canons,
32
as the amount Ior the displacement and Ior the maximum
equation.
33
The table oI the 'variation (ikhtilf) is a standard table oI diIIerences
between the equation oI anomaly Ior minimum (,
R-e
) and Ior maximum (,
Re
)
distances:
,
R-e
,
Re
.
The interpolation Iunction involved Ior the aIorementioned values oI the
true anomaly (m
1
) decreases monotonically Irom 60 (Ior 2f 0) to 0 (Ior 2f
180). It can be recomputed using the expression:
m
1
(max ,
2f
max ,
R-e
) / (max ,
R-e
max ,
Re
)
in which max ,
R-e
7;37,28, and
max ,
Re
4;58,42.
It is easy to see the logic oI the system Ior the computation oI the lunar
equation oI anomaly at syzygies in which 2f 0, m
1
60 and ,(o
v
) ,
Re
:
7;37,28 ,(o) 7;37,28 ,
R-e
60(,
R-e
,
Re
) 7;37,28 ,
Re
.
At quadratures, 2f 180, m
1
0 and ,(o
v
) ,
R-e
:
7;37,28 ,(o) 7;37,28 ,
R-e
0(,
R-e
,
Re
) 7;37,28 ,
R-e
.
Eor 180 o
v
360, the Iuncion involved is the lunar equation oI anomaly
Ior the maximum distance oI the center oI the lunar epicycle Irom the center oI
the Earth (,
Re
). The table is also displaced vertically 7;37,28 and, as the equa-
tion oI anomaly is positive Ior 180 o 360, it gives:
7;37,28 ,
Re
.
The interpolation Iunction involved Ior the aIorementioned values oI the
true anomaly (m
2
) increases monotonically Irom 0 (Ior 2f 0) to 60 (Ior 2f
180). It can be recomputed using the expression:
m
2
(max ,
2f
max ,
Re
) / (max ,
R-e
max ,
Re
).
Again, at syzygies, 2f 0, m
2
0 and ,(o
v
) ,
Re
:
7;37,28 ,(o) 7;37,28 ,
Re
0(,
R-e
,
Re
) 7;37,28 ,
Re
.
At quadratures, 2f 180, m
2
60 and ,(o
v
) ,
R-e
:
7;37,28 ,(o) 7;37,28 ,
Re
60 (,
R-e
,
Re
) 7;37,28 ,
R-e
.
4 H:i+:q 1:iirs or +nr Coxiixri Iux:r Lqu:+ioxs
The Zf al-Sharf contains, as we have seen, a set oI double argument tables to
calculate the lunar equation oI anomaly (MS Ta, Iols. 11 r and v; MS Tb, Iols.
65 v and 66 r): see samples below, in tables 10.9 and 10.10. The arguments
are the marka: (double elongation), at intervals oI 30, and the mean anomaly
(at intervals oI 6): 720 values oI the equation oI anomaly are computed and I
have recalculated all oI them. On the whole, the calculator did a good job and
errors have usually an amount oI 1, with the only exception oI the equations
calculated Ior a double elongation oI 240, in which errors are systematic and
reach a maximum oI 8`.
34
It has been established that these habtaq tables have
been calculated using the lunar equation tables that I have just described and
that derive Irom Ulugh Beg`s Zf-i Sultn. Eor the recomputation oI the habtaq
tables I have used Benno van Dalen`s 1able Analysis and, especifcally, the sub-
programme called 1able calculator. The procedure used Ior the recomputation
has been the Iollowing one:
Table 10.9
Habtaq Table Ior the Iunar Equation oI the Anomaly Ior 0
o
o 180
o
Markaz (double elongation)
Mean 0 60 120 180 240 300 330
anomaly
0 6;35 (1) 5;44 4;53 6;5 7;37 7;13 6;54
12 5;40 (1) 4;47 3;45 4;44 6;23 (2) 6;11 5;57
24 4;49 3;55 2;45 3;28 5;10 (3) 5;12 5;3
36 4;4 3;11 1;54 2;20 4;2 (4) 4;17 4;15 (1)
48 3;27 2;37 1:16 1;23 3;1 (5) 3;30 3;33
60 2;59 2;15 0;53 0;39 2;10 (7) |2|;51
1
3;1
72 2;43 2;6 0;46 0;10 1;30 (7) 2;23 2;39
84 2;39 2;11 0;57 (1) 0;0 (5) 1;4 (8) 2;8 2;30
96 2;4|8|
2
2;31 1;28 0;10 0;54 (8) 2;6 (1) 2;35
108 3;11 3;6 2;17 (1) 0;41 1;2 (8) 2;20 |2|;53
3
120 3;47 3;56 3;25 1;34 (1) 1;28 (7) 2;49 3;24 (1)
132 4;34 (1) 4;57 4;47 2;47 (1) 2;13 (6) 3;32 (1) 4;9
144 5;32 6;8 6;20 4;17 3;16 (5) 4;29 5;5
156 6;36 (1) 7;24 7;56 (1) 6;0 4;34 (4) 5;36 6;9
168 7;44 (1) 8;41 9;33 (1) 7;49 6;2 (1) 6;50 7;49
1. MS Ta 3;51
2. MS Ta 2;43
3. MS Ta 3;53
. .
1. Eor 0 o
v
180.
1.1 Enter a table 11(c
m
) oI the lunar equation oI center j as a Iunction
oI the mean centrum, Ior e 12;33,22
p
, displaced vertically 13;15,34,
Ior arguments comprised between 0 and 330, with intervals oI 30. The
actual values entered were those that appear above in table 10.4.
Erom here on all the steps are given Ior one particular value oI c
m
.
m
) oI the Iunction oI interpolation m
1
Ior arguments
0, 30, 60...360. The values entered are those oI table 10.7 above.
1.3 Add
j o
m
o
v
to obtain the true lunar anomaly by adding, Ior each particular value oI c
m
,
o
m
11(c
m
), Ior o
m
0, 6, 12...174.
Table 10.10
Habtaq Table Ior the Iunar Equation oI the Anomaly Ior 180
o
o 360
Markaz (double elongation)
Mean 0 60 120 180 240 300 330
an.
180 8;52 9;54 11;3 9;38 7;37 (1) 8;7 8;29
192 9;55 10;59 12;19 11;|1|9
1
9;12 9;22 9;|3|6
2
(1) (1) (3) (1)
204 10;51 11;53 13;20 12;46 (1) 10;41 (5) 10;31 10;37
216 11;36 (1) 12;33 14;2 13;55 (1) 11;59 (6) 11;31 11;28
228 12;10 12;59 14;24 14;43 13;2 (7) 12;|1|7
3
12;7
240 12;30 13;9 14;28 15;9 13;47 (8) 12;50 12;32
(1) (1)
252 12;36 13;5 14;14 15;14 14;13 (8) 13;7 (1) 12;44
264 12;30 (1) 12;47 13;45 15;0 14;21 (8) 13;9 (1) 12;42
276 12;11 12;17 13;2 14;28 14;15
4
(3) 12;57 (1) 12;27
288 11;41 11:36 12;7 13;41 13;45 (7) 12;32 (1) 12;0
300 11;2 10;47 11;3 12;41 13;5 (6) 11;55 11;23
312 10;16 9:51 9;53 11:31 (1) 12:13 (5) 11;10 (1) 10:39
(1)
324 9;24 8;51 8;39 10;14 11;12 (4) 10;17 (1) 9;47
(1)
336 8;29 7;49 7;22 8;52 (1) 10;7 9;18 8;51
348 7;32 (1) 6;46 6;6 7;28 (1) 8;53 8;16 7;53
1. MS Ta 11;59
o
.
2. MS Ta 9;56.
3. 12;7 in MS Ta, in which 17` appears in a marginal correction.
4. 14;15
o
in MS Ta; 14;11
o
in MS Tb.
The result obtained will be table 13(o
m
).
1.4 Calculate a table 14(o
m
) oI the equation oI anomaly Ior minimum
distance (,
R-e
), the arguments being the true anomalies obtained Irom
13(o
m
), Ior each particular value oI c
m
. The table has to be displaced
7;37,28 and we must, thereIore, calculate
7;37,28 ,
R-e
.
The expression used is:
7;37,28(atan((6;17,46*sin(13)) / (47;26,37 6;17,46*cos(13)))),
in which 13 means 13(o
m
).
1.5 Calculate a table 15(o
m
) oI the equation oI anomaly Ior maximum
distance (,
Re
), also displaced 7;37,28, again using as arguments the
values oI 13(o
m
m
:
7;37,28 (atan((6;17,46*sin(13)) / (72;33,226;17,46*cos(13)))),
13 being 13(o
m
).
1.6 The fnal table Ior each value oI the mean anomaly, and still Ior our
particular value oI c
m
, will be:
14 ((15-14)*12),
in which
12 is 12(o
m
),
14 is 14(o
m
),
15 is 15(o
m
).
2. Eor 180 o
v
360
2.1 As in 1.1 above: 11(c
m
).
m
) oI the Iunction oI interpolation m
2
Ior arguments
0, 30, 60...360. The values entered are those oI table 10.8 above.
2.3 As in 1.3 above Ior o
m
180, 186, 192...354:
13(o
m
m
.
2.4 Calculate 7;37,28 ,
Re
Ior the values oI 13(o
m
), using the same
expression as above in 1.5:
14(o
m
).
2.5 Calculate 7;37,28 ,
R-e
Ior the values oI 13(o
m
), using the same
expression as above in 1.4:
15(o
m
).
2.6 As in 1.6 above, calculate
14 ((1514)*12).
Ackxovirixrx+s
The present chapter has been prepared within a research program on 'The cir-
culation oI astronomical ideas in the Mediterranean between the twelIth and
Iourteenth centuries, sponsored by the Direccion General de Investigacion
Cientifca y Tecnica oI the Spanish Ministry oI Education and Culture. Benno
van Dalen`s programs CAIH (Ior the conversion oI dates) and TA ('Table
Analysis), as well as Honorino Mielgo`s ATMM (Ior the analysis oI mean
motion tables) have been extensively used. InIormation about manuscripts in
the National Iibrary oI Tunis and the Sabhiyya Iibrary oI Sale (Morocco)
was given to me by Dr. Merce Comes and by my research graduate student
Hamid Berrani. Benno van Dalen read two successive draIts oI this chapter and
corrected some errors I had made. He also tried to improve my hopeless math-
ematical inconsistencies and sent me photocopies oI Sedillot`s publications on
Ulugh Beg`s :f. My gratitude to all oI them.
No+rs
1. See I. ex. David A. King, 'An Overview oI the Sources Ior the History oI Astronomy
in the Medieval Maghrib. Deuxieme Colloque Maghrebin sur l Histoire des Mathe-
matiques Arabes. Tunis, 1988, pp. 125157. See also a recent survey oI this manuscript
in A. Mestres, 'Maghrib Astronomy in the 13th Century: A Description oI Manuscript
Hyderabad Andra Pradesh State Iibrary 298, in J. Casulleras and J. Samso (eds.), From
Baghdad to Barcelona. Studies in the Islamic Exact Sciences in Honour of Prof. Juan
Jernet. Barcelona, 1996, I, 383443. A complete edition oI the canons oI the Hyderabad
compilation, together with a commentary and a partial edition oI the numerical tables
was presented by Angel Mestres as a Ph.D. dissertation (University oI Barcelona, Janu-
ary 2000).
2. The canons were edited by J. Vernet, Contribucion al estudio de la labor astronomica
de Ibn al-Bann. Tetuan, 1952. The solar tables were studied by J. Samso and E. Millas,
'Ibn al-Bann, Ibn Ishq and Ibn al-Zarqlluh`s Solar Theory in J. Samso, Islamic
Astronomy and Medieval Spain. Variorum. Aldershot, 1994, no X (35 pp.). Eor the plan-
etary tables see J. Samso and E. Millas, 'The Computation oI Planetary Iongitudes in
the Zf oI Ibn al-Bann, Arabic Sciences and Philosophy 8 (1998), 259286.
3. See E. S. Kennedy, 'The Astronomical Tables oI Ibn al-Raqqm, a Scientist oI
Granada, Zeitschrift fur Geschichte der Arabisch-Islamischen Wissenschaften 11 (1997),
3572. On these :fes a doctoral dissertation by Muhammad Abd al-Rahmn (Institute
Ior the History oI Arabic Science, Aleppo), Hisb atwl al-kawkib f l-Zf al-Shmil
f 1ahdhb al-Kmil li-Ibn al-Raqqm, was presented in the University oI Barcelona
(September 1996).
4. E. S. Kennedy and David A. King, 'Indian Astronomy in Eourteenth Century Eez: the
Versifed Zj oI al-Qusuntn, Journal for the History of Arabic Science 6 (1982), 345.
Reprint in D. A. King, Islamic Mathematical Astronomy, Variorum Reprints, Iondon,
1986, no. VIII (second revised edition by Variorum, Aldershot, 1993).
5. J. Samso, 'Andalusian Astronomy in 14th Century Eez: al-Zf al-Muwhq oI Ibn
Azzz al-Qusantn, Zeitschrift fur Geschichte der Arabisch-Islamischen Wissen-
schaften, 11 (1997), 73110. See also J. Samso, 'Horoscopes and History: Ibn Azzz
and his retrospective horoscopes related to the battle oI El Salado (1340), in Iodi Nauta
and Arjo Vanderjagt (eds.), Between Demonstration and Imagination. Essays in the His-
tory of Science and Philosophy Presented to John D. North, Brill, Ieiden-Boston-Kln,
1999, 101124.
6. See Kennedy and King, 'Indian Astronomy, p. 9; King, 'Overview, pp. 131132
and 144.
7. See Juan Vernet, 'Tradicion e innovacion en la ciencia medieval, Oriente e Occi-
dente nel Medioevo. Filosoha e Scien:e, Accademia dei Iincei, Roma, 1971, pp. 741
757; reprint in Vernet, Estudios sobre Historia de la Ciencia Medieval, Barcelona-
Bellaterra, 1979, pp. 173189 (see pp. 188189). See now Montse Diaz-Eajardo, La
teoria de la trepidacion en un astronomo marroqui del siglo XJ. Estudio y edicion
critica del Kitb al-adwr I tasyr al-anwr (parte primera) de Ab Abd Allh al-
Baqqr. Barcelona, 2001.
8. See J. Samso, 'An Outline oI the History oI Maghrib Zjes Irom the End oI the
Thirteenth Century. Journal for the History of Astronomy 29 (1998), 93102; see also
Samso, 'Astronomical Observations in the Maghrib in the Eourteenth and EiIteenth
Centuries, Science in Context 14 (2001), 165178.
9. See I. P. E. A. Sedillot, Prolegomenes des 1ables Astronomiques dOloug-Beg. 2
vols., Paris, 1847 and 1853.
10. Muhammad al-Arab al-Khattb, Fahris al-Khi:na al-Hasaniyya vol. III (Rabat,
1983), pp. 193195 (no. 227); see also D. A. King, A Survey of the Scientihc Manu-
scripts in the Egyptian National Library. Winona Iake, Indiana, 1986, p. 164 (G99).
MSS 11420 and 12616 oI the National Iibrary contain versions oI Ulugh Beg`s :f Ior
the coordinates oI Cairo (long. 55 Irom the western meridian).
11. Ms. Cairo DM814 is a copy oI this recension dated in 1325/19071908: cI. King,
Survey p. 142 (E46). A third (?), anonymous, Tunisian recension is preserved in two
mss. oI the Hasaniyya Iibrary in Rabat: 2650 (undated) and 2148 (dated in 1218/1803).
See Khattb, Fahris 335338 (no. 406408). Ekmeleddin Ihsanoglu et al., Osmanli
Astronomi Literaturu 1arihi (History of Astronomy Literature during the Ottoman
Period). Istanbul, 1997, p. 314 (no. 175) and pp. 347348 (no. 213).
12. Another copy oI the same :f is extant in MS 5990 oI the same library. The copy was
fnished in 1229 H./1814 by Muhammad b. Abd Allh al-Khayyr.
13. Copy dated in 1242/1826. Merce Comes obtained a microflm oI that manuscript
and called my attention to it.
14. See Abd al-HaIz Mansr, al-Fihris al-mm li l-makhttt. I. Rasd maktabat
Hasan Husn Abd al-Wahhb, Tunis, 1975, p. 398.
15. See also King, Survey, p. 143 (E53) who lists other manuscripts oI the same work:
ms. Cairo Talat Riyda 319,1 (pp. 116, ca. 1300 H.) contains the mean motion tables
oI this :f. A complete copy is in ms. Paris B.N. ar. 2536 (48 Iols., ca. 1150 H.).
16. The incomplete copy extant in MS Tb (Iol. 104 r) gives an example oI the computa-
tion oI the solar longitude and the date used is the 16th Jumd I 1099 H/ 19.3.1688.
17. Driss Iamrabet, Introduction a lhistoire des mathematiques maghrebines (Rabat,
1994), p. 143 (no. 506), mentions Ibn Mm and places him tentatively aIter the six-
teenth century. One copy (dated 1163 H/ 17491750) oI his work is extant in Cairo N.
I. (King, Survey p. 143, E54).
18. Tunis Nat. Iibrary, col. H. H. Abd al-Wahhb 18104, dated in 1301/18831884: see
A. H. Mansr, Fihris p. 397, and in Tunis Nat. Iibrary MS 2770. Another copy oI the
same work is extant in MS 1042 oI the Sabhiyya Iibrary in Sale (Morocco). Two copies
oI this work, dated in 1150/17371738 and in 1169/17551756, are also extant in Cairo
N. I.: see King, Survey p. 144, E55. I have another reIerence to a summary oI Sanjaq
Dr`s Mukhtasar, mixed with materials derived Irom Ibn al-Shtir`s :f, in MS Tunis
N. I. 5608. Einally, the Sabhiyya Iibrary also contains two summaries (Mukhtasars)
oI Ulugh Beg`s :f in MSS 1167 and 1168, the author being Muhammad b. Ab l-Eath
al-SI.
19. E. S. Kennedy and M. H. Kennedy, Geographical Coordinates of Localities from
Islamic Sources (ErankIurt, 1987), pp. 362363.
20. D. A. King, 'A Double Argument Table Ior the Iunar Equation Attributed to
Ibn Ynus, Centaurus 18 (1974), 129146. Reprinted in King, Islamic Mathemati-
cal Astronomy no. V. Eor other double argument tables Ior the computation oI lunar
and planetary equations see D. A. King, 'On the Astronomical Tables oI the Islamic
Middle Ages, Colloquia Copernicana 13 (1975), 3756 (cI. pp. 4445, 55), reprinted
in Islamic Mathematical Astronomy no. II; Claus Jensen, 'The Iunar Theories oI al-
Baghdd, Archive for History of Exact Sciences 8 (19711972), 321328; Mark J.
Tichenor, 'Iate Medieval Two-Argument Tables Ior Planetary Iongitudes, Journal
of Near Eastern Studies 26 (1967), 126128, reprinted in E. S. Kennedy, Colleagues
and Eormer Students, Studies in the Islamic Exact Sciences (Beirut, 1983), 126128;
G. Saliba, 'The Double Argument Iunar Tables oI Cyriacus, Journal for the History
of Astronomy 7 (1976), 4146 and 'The Planetary Tables oI Cyriacus, Journal for the
History of Arabic Science 2 (1978), 5365.
21. See a similar table in the Zf al-Sharf MS Ta Iol. 31 v; MS Tb, Iol. 85 v.
22. J. Chabas and B. R. Goldstein, 'Andalusian Astronomy: al-Zif al-Muqtabis oI Ibn al-
Kammd, Archive for History of Exact Sciences 48 (1994), 141 (see pp. 14 and 23).
23. I owe this inIormation to Angel Mestres.
24. This displacement is explicitely mentioned by Mahmd b. Muhammad b. Qdzdah
al-Rm, known as Mram Chalab (d. 1524) in his commentary to the :f oI Ulugh Beg:
cI. Sedillot II (1853), p. 140.
25. Mean motions have been calculated using Honorino Mielgo`s ATMM. See
H. Mielgo, 'A Method oI Analysis Ior Mean Motion Astronomical Tables, in J. Casulle-
ras and J. Samso (eds.), From Baghdad to Barcelona. Studies in the Islamic Exact Sci-
ences in Honour of Prof. Juan Jernet (Barcelona, 1966), I, 159179.
26. I have checked ms. Aleppo, Waqfyya 1307. A microflm oI this manuscript was
obtained Irom the Institute Ior the History oI Arabic Science (Aleppo), microflm no.
507.
27. I am Iollowing here Benno van Dalen`s notation: the two digits aIter the reIer to the
last two digits oI the parameter.
28. This eccentricity is mentioned by Mram Chalab in his commentary oI Ulugh Beg`s
canons: see Sedillot II (1853), p. 142.
29. Mram Chalab mentions an equivalent eccentricity oI 10;23
p
Ior a radius oI the
deIerent oI 49;37
p
. See Sedillot II (1853), p. 147.
30. The same set oI instructions appears in Ulugh Beg`s canons: see Sedillot II (1853),
pp. 138153.
31. Mram Chalab mentions a radius oI the epiycle oI 5;12
p
Ior a radius oI the 'inclined
sphere oI 60
p
. He seems to be making a mistake here: 5;12
p
is probably a rounded value
and it corresponds to a radius oI the deIerent oI 49;37
p
. 6;17,18
p
would be the corre-
sponding epicycle radius Ior a deIerent radius oI 60
p.
See Sedillot II (1853), p. 158.
32. See Sedillot II (1853), p. 151.
33. We have seen that Mram Chalab also ascribes to Ulugh Beg a radius oI the epicycle
oI 5;12
p
(Ior a deIerent radius oI 49;37
p
), equivalent to 6;17,18
p
(Ior R 60
p
), instead
oI our 6;17,46
p
. With this radius the maximum value oI the equation oI anomaly Ior
minimum distance would be:
,
max
atan (6;17,18/(6012;33,22)) 7;33,0.
34. Benno van Dalen, aIter reading a draIt oI this chapter, suggested that the errors
practically disappear iI one assumes that Sanjaq Dr made a mistake and used interpola-
tion coeIfcients 20;24 (instead oI 17;24) and 39;36 (instead oI 42;36). He is absolutely
right. AIter recalculating the column Ior mean anomaly 240, using these erroneous
coeIfcients, the only diIIerences (in the sample values) are:
12 6;23(1) 216 11;59(1)
168 6;2(1) 276 14;15(4)
180 7;37(1) 336 10;7(3)
192 9;12(1) 348 8;53(1)
204 10;41(1)
VI
Science and Medicine in the Maghrib and al-Andalus
A Survr. or +nr Prsr:rcn irrorr +nr +,:os
Eor the period 1834 to 1980 I have surveyed more than one hundred works
exclusively or partially devoted to the mathematics and astronomy in the
Muslim West. This number is modest considering the length oI the period, the
important historical role oI these regions, and the number oI works on the same
subjects written in the same period in other parts oI the Islamic world.
1
One
reason Ior this may be the political conditions oI the societies oI these regions
at the end oI the nineteenth and beginning oI the twentieth century, and the
consequent eIIects on their cultures.
Nevertheless, in spite oI their modest number, and the diIIerent aims and
levels oI these works, they can be said to have contributed as a whole to writ-
ing the frst draIts oI the history oI science in al-Andalus and the Maghrib in
the Middle Ages.
Two kinds oI works can be immediately identifed, each having a spe-
cifc goal. The frst is Iounded on the traditional criteria Ior scholarly research,
as developed in the nineteenth century, whose frst concern is to make known
the Iacts oI the subject in a disinterested way. The second, a product mainly
oI the Maghrib and Egypt during the frst halI oI the twentieth century, has a
clearly stated nationalistic goal: to nurture a sense oI pride in the heritage oI the
country and the importance oI its contribution to science.
2
Even this second kind, with its rhetorical Iorm and political aim, has also
contributed to a better knowledge oI the scientifc heritage oI the Muslim West.
Eor, in a feld that is almost virgin territory, any data can be helpIul, and, as
long as the researcher can separate the wheat Irom the chaII, he or she can fnd
in this nationalistic literature, iI not secure Iactual inIormation, at least some
hints (sometimes quite important) that may reveal new documents or suggest
new ways oI approach.
Eor this reason it has seemed useIul to make a rapid inventory oI the
signifcant publications in this feld beIore 1980, in order to put recent research
(including my own) in perspective. In addition, although Iocusing on the
11
A P:xor:x: or Prsr:rcn ox +nr His+or. or
M:+nrx:+ics ix :i-Axi:ius :xi +nr M:nrii
ir+vrrx +nr Nix+n :xi Six+rrx+n Crx+urirs
Ahmed Djebbar
Anxri DJrii:r 310 Prsr:rcn ox +nr His+or. or M:+nrx:+ics 311
mathematical aspects oI the scientifc tradition oI the Muslim West, my presen-
tation will also include the essential elements relating to the astronomical tradi-
tion oI this region, in both its theoretical and practical aspects. Eor astronomy
uses, implicitly or explicitly, mathematical concepts, objects, and instruments
belonging to various scientifc traditions (Greek, Indian, Eastern Muslim),
which Ied the science oI al-Andalus and the Maghrib. Moreover, some math-
ematicians in these regions, such as Maslama al-Majrt and Ibn al-Bann, pub-
lished astronomical writings, and it would seem unIair to neglect this aspect oI
their scientifc activities.
1nr Nixr+rrx+n Crx+ur.
Although the Muslim West was mentioned in the earliest works in the his-
tory oI mathematics or oI astronomy, by J. E. Montucla,
3
C. Bossut,
4
and J. B.
Delambre
5
in Erance, M. Pelayo in Spain,
6
and Hammer-Purgstall in Prussia,
7
the frst substantial investigations concerning its scientifc tradition were the
work oI J.-J.-E. and I.-A. Sedillot
8
in astronomy and oI E. Woepcke in math-
ematics.
9
These works Iormed part oI their authors` interest in the Arabic sci-
entifc heritage as a whole.
10
Astronomy
In astronomy, the translation oI the frst volume oI al-Marrkush`s work
(ca. 661/1262),
11
Jmi al-mabdi wa l-ghyt f ilm al-mqt |the Collec-
tion oI Principles and Goals in the Science oI the Determination oI Time|
12
appeared in 1834 and was Iollowed, ten years later, by an analysis oI the
whole treatise.
13
These publications stimulated interest in the history oI sci-
ences in the Maghrib, even though the contents oI the work in question refect
the scientifc tradition oI the East beIore the thirteenth century, rather than that
oI Marrakech, the city where al-Marrkush was born. However, one had to
wait until the second halI oI the nineteenth century to see a new interest in the
astronomy oI the Muslim West, which concerns two great fgures oI medieval
Spain: az-Zarqlluh ( az-Zarqiyl az-Zarql) (d. 493/1100) oI Toledo and
the Castilian king Alphonso X (12521284), who perpetuated the Arab scien-
tifc tradition oI Spain by his knowledge and his patronage. Steinschneider`s
14
works on az-Zarqlluh and those oI M. Rico y Sinobas
15
on the Libros del Saber
stimulated new research in the frst halI oI the twentieth century, which was
aided by the advance in the history oI astronomy as a whole in the nineteenth
century, together with the study oI the history oI mathematics.
Mathematics
The publication oI a translation oI the Muqaddima |Introduction| oI Ibn
Khaldn was the frst stimulus Ior research into the history oI Western Islamic
mathematics by E. Woepcke, A. Marre, M. Steinschneider, and H. Suter.
16
All
the investigations by these authors oI the Maghribi tradition in arithmetic and
algebra, were more or less related to passages on the classifcation oI the sci-
ences in the work oI Ibn Khaldn. Woepcke, Ior example, noticed the impor-
tance oI the chapter in the Muqaddima on mathematics. He published extracts
Irom it in 1854
17
and a complete translation two years later.
18
Since these passages also infuenced an important part oI the research
that will be mentioned below, I will also quote them and comment on their
contents in the light oI the research oI these Iour historians. Because diIIerent
interpretations oI these passages have arisen out oI the previous translations
by Quatremere,
19
de Slane,
20
Woepcke, Rosenthal,
21
and Monteil,
22
I will give
here my own translation, reIerring to a Iuture study Ior the justifcation oI my
readings.
23
AIter devoting section 19 oI chapter VI oI the Muqaddima to a general
presentation oI the rational sciences, Ibn Khaldn discusses, in section 20,
entitled Sciences of Number, the sciences oI arithmetic, algebra, commercial
transactions, and inheritances. Here are the main passages oI the book, which
is dated 779/1377:
24
The Sciences oI number: the frst oI them is arithmetic, and it is the knowl-
edge oI the properties oI numbers Irom the viewpoint oI composition, either
according to a successive 'progression or by duplication; as, Ior example,
iI the numbers Iollow one another while exceeding each other by the same
number, then the sum oI their two extremes is equal to the sum oI any pair
oI numbers whose distance Irom the two extremes is the same, and it is
equal to double oI the middle |term| iI these numbers are odd in number. Eor
example: the sequence oI the |integer| numbers, the sequence oI the even
numbers, the sequence oI the odd numbers. . . .
This science is the frst part oI mathematics and the best established, and it
is used in the prooIs oI calculation. The ancient and modern scientists have
publications on it, and the majority oI them integrate it into mathematics
without devoting |specifc| writings to it. This is what Ibn Sn did in his
books ash-Shif
25
and an-Naft,
26
as well as others among the ancient scien-
tists. As Iar as the moderns are concerned, they gave it up because it is not
practiced and because its useIulness is in the prooIs and not in the calcula-
tions; and it is Ior this reason that they abandoned it aIter they had extracted
the essence in arithmetical prooIs, as Ibn al-Bann did in the book Raf al-
hifb |The Raising oI the Veil|, as well as others. . . .
And among the branches oI the science oI number, |there is| the art oI calcu-
lation. It is a practical art |relating| to the calculation oI numbers by joining
and separation. Joining takes place in numbers in a single act |oI adding|,
and this is |then| addition, and by repetition |oI the same added portion|, i.e.,
one increases a number |by its own value| as many |times as there are| units
in another number; and this is multiplication. Separation also takes place Ior
numbers, either in a single act, like cutting oII oI a number Irom |another|
number and knowing oI the remainder, and this is then subtraction; or cut-
ting oI a number into a determined number oI equal parts; and this is divi-
sion. . . .
And among the best writings concerning the |subject| at this time in the
Maghrib, there is al-Hassr`s small book. Ibn al-Bann Irom Marrakech has
on the |subject| a useIul summary which establishes the rules oI its opera-
tions. Then he commented on it in a book, which he entitled Raf al-hifb,
which is obscure Ior a beginner because oI the frmly-structured prooIs it
contains. It is a book oI great value and we saw high-level proIessors giving
much consideration to it; it is a book which indeed deserves that.
The author, God bless him, Iollowed Ibn Munim`s book, Fiqh al-
hisb |The Science oI Calculation| and al-Ahdab`s |book| al-Kmil |The
Complete Book>|. He summarised their demonstrations and changed
thembecause oI the use oI |numerical| symbolsinto clear and abstract
justifcations which are the secret and the essence oI representation by
|numerical| symbols.
27
All these |matters| are obscure,
28
but the obscurity
arises |only| Irom the method oI demonstration |which| is specifc to the
mathematical sciences, since their problems and their operations are them-
selves quite clear. However iI one wanted to explain them, it would be
necessary to give the justifcations Ior these operations. And there are, in
understanding them, diIfculties which one does not fnd in the resolution oI
problems.
29
In these extracts oI the Muqaddima, European specialists in Arabic sci-
ences Iound a whole research program, namely the study oI the mathematical
works quoted by Ibn Khaldn. With one or two exceptions, their publications
were confned to the topics that were suggested to them by their own readings
oI the quoted passages.
An ambiguous passage in the section on calculation
30
directed Woepcke
toward his research oI symbolism in the mathematical writings oI the Muslim
West. Thus he discovered some symbols in the work oI al-Qalasd (d. 892/
1486), an Andalusi mathematician who studied in the Maghrib and taught in
Tlemcen and Tunis.
31
In 1854, Woepcke presented a note on his discovery to
the Academy oI Sciences.
32
In this note, Woepcke announced the discovery, in al-Qalasd`s hand-
book 'of a well developed algebraic notation among the Arabs of the West,
as well as a 'multiplication table of the algebraic powers in a Persian manu-
script.
A Iew years later, Woepcke published al-Qalasd`s book which con-
tained this symbolism and which is entitled Kashf al-asrr an ilm hurf al-
ghubr |Disclosure oI the Secrets oI the Science oI the Dust-Numerals|.
33
He
also translated an anonymous Maghribi handbook on calculation
34
as well as
extracts Irom several commentaries on 1alkhs aml al-hisb |Abridgement
oI the Operations oI Calculation| oI Ibn al-Bann (d. 721/1321).
35
In connec-
tion with al-Qalasd, we must also mention a note by M. A. Cherbonneau on
his writings and an article by G. Enestrm on a method oI approximation in
his Kashf al-asrr.
36
Although Ibn al-Bann`s 1alkhs was mentioned in the Muqaddima
and in almost all other mathematical Maghribi texts studied between 1850
and 1860, neither Woepcke nor any oI the other researchers mentioned above
devoted an article to the contents oI this handbook itselI. In 1864, Marre pub-
lished it with a somewhat literal and sometimes strange translation.
37
This
translation does not seem to have aroused much interest in Ibn al-Bann and
his work on the part oI the researchers oI this time, with the exception oI Stein-
schneider who devoted a brieI note to him in 1877.
38
An erroneous reading oI another passage oI the Muqaddima was the
origin oI research on al-Hassr (5th/12th c.), one oI the three other mathemati-
cians mentioned by Ibn Khaldn. In this passage, one reads 'Kitb al-Hassr
as-saghr which one can translate as 'the book of the small al-Hassr or 'the
book of the small saddle or 'al-Hassrs small book. On the basis oI this sen-
tence and its various interpretations, patient research was carried out by two
oI the most important historians oI sciences in Europe at the end oI nineteenth
century: H. Suter and M. Steinschneider. In 1874, the latter provided a satis-
Iactory interpretation oI the passage on the basis oI Hebrew texts. At the same
time, he revealed the existence oI a Hebrew translation oI the small book` oI
al-Hassr made by Moses Ibn Tibbon in 1271.
39
In 1893, the discovery oI an anonymous manuscript and its compari-
son with this Hebrew translation enabled him fnally to give the true title oI
the book as Kitb al-bayn wa t-tadhkr |Book oI the Demonstration and the
Recollection| and to provide new inIormation on the complete name oI its
author.
40
Suter investigated the biography oI al-Hassr and the remainder oI his
mathematical production. He conjectured the existence oI a second treatise by
Hassr, oI which the Bayn wa t-tadhkr was only an abridged version. AIter
long and unsuccessIul investigations to confrm this conjecture, he published a
German translation oI the Bayn.
41
Suter presented the results oI his research on the other mathematicians
mentioned by Ibn Khaldn in his biobibliographic work which he completed
at the turn oI the century.
42
This work can be regarded as the fnal result oI halI
a century oI research, edition and translation, in Europe and in Egypt, oI the
sources oI Arab Science in a wide sense.
Eor the primary sources oI the Muslim West, Suter initially used the frst
eight volumes oI the Iamous Bibliotheca Arabica Hispana
43
which were pub-
lished between 1883 and 1892, and al-Maqqar`s Nafh at-tb |DiIIusion oI the
PerIume|.
44
He also used the Eastern biobibliographic works Irom beIore the
twelIth century which had been edited by European scholars there.
45
It is natural to wonder whether, at the same time, one can detect in the
nineteenth-century intellectual circles oI the Maghrib, a similar interest Ior
the history oI the scientifc heritage oI the Muslim West in general. The most
recent biobibliographic publications by Maghribi researchers, especially al-
Mann, suggest a negative answer.
46
However, there are two phenomena that
deserve Iurther study.
The frst is the continuation oI the traditional scholarly activities oI pub-
lishing and teaching. In mathematics and astronomy, this included the continu-
ation oI curricula in the institutions oI higher education oI Qarawiyyn in Eez
and oI Zaytna in Tunis, which had not changed since the sixteenth century,
except in the reduction oI their volume and the progressive lightening oI their
contents.
47
The authors belonging to this traditionalist trend continued to write
handbooks prompted by the earlier writings oI the Maghribi tradition. In the
feld oI historical research, there were numerous publications on the political,
religious, and cultural history oI the area oI which none, to my knowledge, is
devoted to aspects oI the scientifc history oI the medieval period, or even oI
the Ottoman period (10th13th/16th19th c.). The only initiative concerning
the scientifc heritage was associated with the arrival, in 1865, oI lithography
(which was set up initially in Meknes, then defnitively in Eez).
48
1he Recen-
sion of Euclids Elements by Nasr ad-Dn at-Ts was lithographed,
49
not with
the aim oI putting it at the disposal oI researchers in the history oI science, but
rather to be used as a reIerence work Ior the teaching oI elementary geometry
at the Qarawiyyn in Eez.
50
The scientifc heritage continued to provide mate-
rial Ior teaching until the end oI the nineteenth century.
The second phenomenon is the publication oI translations, adaptations
or popularizations oI European writings about sciences (geometry, logarithms)
or about military technology. This is related to initiatives taken in the Otto-
man empire in order to acquire scientifc and technological knowledge Irom
Europe.
51
Among the translated works were Iegendre`s Elements of Geometry
and Ialande`s 1reatise of Astronomy.
52
Eor reasons that are still unclear, the
initiators oI this trend do not seem to have been interested in researches on the
history oI the scientifc heritage oI the Muslim West.
1nr 1vrx+ir+n Crx+ur. irrorr +,:o
In the twentieth century, research on the history oI mathematics and astronomy
in al-Andalus and in the Maghrib increased signifcantly but it continued along
the same lines as the work oI the pioneers oI the second halI oI the nineteenth
century which we mentioned above. In the biobibliography oI Andalusi mathe-
maticians, the works oI Suter, H. Hankel,
53
and C. Brockelmann
54
were supple-
mented by the publication oI Kitb ikhbr al-ulam bi-akhbr al-hukam
|The Book which InIorms the Scholars on the IiIe oI the Wise| oI Ibn al-QiIt
(d. 646/1248)
55
in an abridged version, and by the important discovery oI
the Kitb tabaqt al-umam |Book oI the Categories oI Nations| oI Sid al-
Andalus (d. 461/1068). This small work (on proIane sciences and philosophy)
was written by a specialist in astronomy and it contains invaluable inIormation
on the contemporaries oI the author and their works.
56
On the basis oI these
new sources, J. A. Sanchez-Perez compiled, in, 1921, a biobibliography on the
history oI mathematics and astronomy in al-Andalus.
57
This work contains data on 191 mathematicians and astronomers, but
in contrast to astronomical and astrological writings, Sanchez-Perez could
name very Iew extant copies oI mathematical works produced in al-Anda-
lus between the ninth and the fIteenth centuries, these being the Commentary
oI Ibn Zakariyy al-Gharnt (d. 809/1406) on Ibn al-Bann`s 1alkhs, Ibn
Badr`s Abridged Book in Algebra (7th/13th) and two writings oI Ibn Mudh
al-Jayyn (d. 460/1067): his Commentary on Book V oI Euclid`s Elements
and his treatise oI spherical Trigonometry, Kitb mafhlt qisiyy al-kura |The
Book oI the Unknown Arcs oI the Sphere|.
58
At the same time, Sanchez-Perez
gave a long list oI the lost mathematical works by the scientists who worked
in al-Andalus between the ninth and fIteenth centuries. This diIIerence is
even more striking because, except Ior Ibn Mudh`s book on trigonometry,
the mathematical works identifed by the time oI Sanchez-Perez are minor
writings, compared with the works oI the tenth to twelIth centuries. Thanks
to inIormation provided by Sid al-Andalus and by mathematicians oI the
thirteenth to Iourteenth centuries, such as Ibn Munim, Ibn al-Bann, and Ibn
Zakariyy al-Gharnt, we know that dozens oI mathematicians oI the tenth to
twelIth centuries, including Ibn as-Samh, az-Zahrw, and al-Mutaman, had
published works oI a high level on geometry, calculation and the theory oI
numbers. Ibn Khaldn also said, in his Muqaddima, with reIerence to the Book
of Algebra by the Egyptian Ab Kamil (d. 930), that it 'was commented on by
many Andalusis and that they 'excelled in their commentaries, whereas only
the rather average work by Ibn Badr has reached us.
60
No work similar to that oI Sanchez-Perez was written in this time, on
the scientifc tradition oI the Maghrib. Two contributions were made by H.
P.-J. Renaud (18811945) in the years 19321933: an appendix to Suter`s
book
61
and an article on the edition oI scientifc works in the Maghrib prior to
1880.
62
1nr Cox+rx+s or Prsr:rcn
Astronomy and Astrology
Eor these two important felds oI science in al-Andalus,
63
the most important
research undertaken in the frst halI oI the 20th century was that by Millas
Vallicrosa, in particular on az-Zarqlluh`s works.
64
This was Iollowed by con-
tributions by J. Vernet, G. J. Toomer, W. Hartner, B. Goldstein, J. Samso,
D. A. King, R. Iorch, and others
65
on certain aspects oI the astronomical pro-
duction oI az-Zarqlluh, and on other astronomers, such al-Majrt and al-
Bitrj (d. 581/1185).
The research in the frst halI oI the twentieth century showed that Anda-
lusi scientists designed new astronomical instruments, both instruments oI
observation
66
and universal astrolabes.
67
It also appeared that attempts were
made by Ibn Bjja (d. 533/1138), Ibn TuIayl (d. 581/1185), Ibn Rushd (d. 595/
1198) and al-Bitrj, to establish a new cosmology. On the basis oI the criti-
cism oI the Ptolemaic model by Ibn al-Haytham,
68
these scientists tried to
replace the Ptolemaic model by a model that was, in their opinion, more com-
patible with the physical world.
69
The originality oI the Andalusi mathemati-
cians was also revealed through their development oI astronomical tables, such
as the tables oI Ibn Mudh and the Toledan tables.
70
The research between 1900 and 1980 also produced Iour essential in-
sights. Eirst, it showed that a number oI scientifc works by Greek, Indian or
ninth-century Arabic authors, such as al-Khwrizm, Habash al-Hsib, and al-
Battn, were available in al-Andalus at an early stage.
71
Second, it revealed
that a Iatin astrological tradition infuenced the birth and the development oI
astrology in al-Andalus, which had been previously thought to depend only on
the traditions oI the Eastern Islamic world.
72
There are also connections with
the local tradition oI sundials prior to the advent oI Islam in Spain.
73
Third, it
showed the important role oI the group oI scientists associated with Alphonso
X in the transmission oI science Irom Arabic into Iatin.
74
This transmission
was studied with specifc examples, such as the universal astrolabe.
75
Eourth,
it demonstrated the important role oI mathematical methods, which are oIten
implicitly present in works on the theoretical aspects oI astronomy, astronomi-
cal instruments and sundials.
Mathematical methods must also have been used Ior the computation oI
the extant astronomical tables. Mathematical methods are discussed explicitly
in Ibn Mudh al-Jayyn`s trigonometrical treatise Kitb mafhlt qisiyy al-
kura. This book is a description oI trigonometrical procedures Ior the solution
oI astronomical problems.
76
His approach is independent oI that oI his imme-
diate predecessor in Asia, al-Brn, in his Maqld ilm al-haya |The Keys oI
Astronomy|.
77
Ibn Mudh develops a procedure Ior determining all the ele-
ments (sides and angles) oI an unspecifed spherical triangle Irom three given
elements. He uses the polar triangle Ior the resolution oI the case where three
angles are given, in a way independent oI Ab Nasr ibn lrq (4th/l0th c.) in the
Eastern Islamic world. He is, fnally, the frst known Andalusi mathematician
to have calculated a table oI tangents.
In the historical research oI the frst halI oI the twentieth century, the
place oI the astronomical tradition oI the Maghrib is quite modest by com-
parison with that oI the astronomy in al-Andalus. This astronomical tradition
could not compete with that oI al-Andalus, but nevertheless deserves some
attention.
The frst work in this tradition to be studied was Ibn al-Bann`s Minhf
at-tlib f tadl al-kawkib |The guide oI the student Ior the correction oI the
star movements|, oI which a partial Spanish translation was published by J.
Vernet,
78
who also listed, Iour years later, the astronomical manuscripts oI Ibn
al-Bann.
79
In the 1930s, H. P.-J. Renaud started to work on particular astro-
nomical subjects. He published biographical notes on two astronomers, Ibn al-
Bann
80
and Ab Miqra,
81
then he conducted research on the astronomical and
astrological traditions in the Western Maghrib as a whole,
82
on the obliquity
oI the ecliptic
83
and the appearance oI the lunar crescent.
84
He published the
only known calendric treatise oI the Maghribi tradition, whose contents also
show the close connection between the tradition in the Maghrib and that oI al-
Andalus.
85
An interest Ior the study oI the scientifc heritage oI the Muslim West
started relatively early in certain intellectual spheres in the Maghrib. Erom the
1950s on as we will see in detail in the Iollowing chapter devoted to mathe-
matics, the results oI this interest were modest in quantity and uneven in qual-
ity. The history oI astronomy does not seem to have benefted Irom the passion
Ior the Arabic sciences oI the Middle Ages which was aroused by the phenom-
enon oI Nahda |rebirth| preached by the Muslim reIormers oI the beginning oI
the century, such as Shaykh Abduh and Jaml ad-Dn al-AIghn.
86
I have not
Iound any trace oI original historical work devoted to a particular astronomer
oI al-Andalus or the Maghrib, other than notes in manuscript catalogues and
in general works on the cultural and scientifc history oI the Iormer Islamic
empire. However, some oI these general works contain important, and some-
times new, inIormation on particular astronomers oI the Muslim West or
on their contribution. Examples are Gannn`s book, which is limited to the
cultural history oI the Western Maghrib; Tqn`s book, which is devoted to
the entire history oI Arabic astronomy and mathematics; and the book oI al-
Azzw dealing with the history oI astronomy in the Islamic world, Irom the
arrival oI the Mongols until the second decade oI the twentieth century.
89
These books contributed in a double way to a better knowledge oI the history
oI Islamic science in general. The works by Gannn and Tqn were widely
distributed and went through several editions.
90
They changed the attitude
among the intellectuals oI the Arabic countries, and in the Islamic world as
a whole, toward the cultural and scientifc heritage oI the civilization oI their
ancestors.
91
In addition, the new inIormation in these works was oI interest to
the biobibliographers and the Western specialists in the history oI science, who
integrated it into their own publications. Eor example, Sarton said that the book
oI Tqn was indispensable Ior researchers in the history oI Arabic astronomy
and mathematics. He also published a positive review oI this work.
92
In 1942,
C. Brockelmann wrote to A. Gannn in connection with his book:
I received your invaluable book entitled the Moroccan genius in the Arabic
literature`; you have honoured me by sending it. I began to read it and I ben-
efted much Irom it in connection with what has escaped my research up to
now on the history oI the Moroccan literature. I hope to use its invaluable
contents Ior my proft and the proft oI my Orientalist colleagues, in the
Supplement to my frst book on the history oI Arabic literature which is now
published in the city oI Ieiden.
93
Einally, thanks to the research oI Slih Zak, Western specialists have
been inIormed about the existence oI the mathematician Ibn Hamza (16th cen-
tury), and his book 1uhfat al-add |The Ornament oI Numbers|, in which Zak
claims to have Iound some Iorm oI the concept oI the logarithm.
94
Mathematics
In the early twentieth century, research on the history oI mathematics in al-
Andalus and the Maghrib was limited to six authors and some specifc topics,
with an emphasis on biobibliographic research and editions oI texts.
Three oI these authors are Maghribi, namely al-Bn (d. 622/1225), Ibn
al-Bann (d. 721/1321) and Ibn al-Ysamn (d. 601/1204). The Iourth author,
al-Qalasd, is Andalusi but spent most oI his liIe in the Maghrib. The origin
oI the two remaining mathematicians, al-Hassr and Ibn Badr (7th/13th c.) is
uncertain. Judging Irom the contents oI their mathematical writings and Irom
their names (which are carried by a number oI Andalusi intellectuals men-
tioned in the biobibliographic works), one may guess that they were oI Anda-
lusi origin.
This historical research concentrated mainly on the science oI calcula-
tion, algebra, and, to a lesser extent, elementary geometry with plane fgures.
I will also mention astrology and the science oI inheritances as applications oI
mathematics, even though the research in these two subjects was more modest
in quantity.
Ibn al-Bann and his work were the subject oI the greatest number oI
studies. Renaud published two articles on him, frst a very concise article,
giving precise details on his date oI birth,
95
and secondly, a much longer one,
in which he reveals that Ibn al-Bann was the author oI many works in diIIer-
ent felds.
96
In his articles Renaud corrects certain errors according to which
Ibn al-Bann`s origin was Saragossa or Granada, and which were based on
conIusing
97
or unbased assertions, Ior which Casiri, the author oI the catalog
oI Arabic manuscripts oI the Escurial, seems to have been responsible.
98
Iater,
A. Gannn
99
and A. al-Es
100
published biobibliographies oI Ibn al-Bann, but
these did not bring in any new elements beyond what Renaud`s researches had
already revealed.
Only one oI the mathematical writings oI Ibn al-Bann was the subject oI
a detailed study, namely the 1alkhs, a handbook on the science oI calculation.
This work was frst translated (but not edited) by Marre, then edited, translated
a second time into Erench, and analyzed by Mohamed Souissi.
101
In this work,
Souissi used other Maghribi mathematical writings, such as Ibn al-Bann`s al-
Maqlt al-arba |The Eour Epistles|
102
and the commentaries on the 1alkhs
by Ibn al-Bann`s student al-Huwr (7th/13th centuries), and al-Qalasd.
103
This comparative study oI the sources themselves confrmed the existence oI a
certain continuity in the subjects and methods used in the mathematical teach-
ing in the Maghrib in the Iourteenth and fIteenth centuries.
104
We know today that these same topics and methods already occurred in
al-Hassr`s handbook al-Bayn wa t-tadhkr, oI which Suter had published
a detailed analysis in 1901. A comparison between this handbook and the
1alkhs in Marre`s translation could have revealed the relationships between
the mathematical traditions oI twelIth and the Iourteenth centuries. As Iar as
I know, no new research was done on this subject during the period which
interests us here.
A new opportunity Ior the realization oI this comparative study arose
in 1964 when A. Gannn revealed the existence oI a twelIth-century treatise,
written by Ibn al-Ysamn, with a rich mathematical content which is impor-
tant Ior a better knowledge oI the mathematical tradition oI the Muslim West as
a whole.
105
But the article, which was written in Arabic, apparently went unno-
ticed. Twenty years later, Ibn al-Ysamn continued to be known only Ior his
algebraic poem, al-urf:a al-ysamniyya |the Poem oI al-Ysamn|.
106
This
urf:a did not get a better Iate, even though a certain number oI manuscript
copies had been mentioned by Suter and Brockelmann. But in 1916, an impor-
tant algebraic work, Ibn Badr`s Mukhtasar f l-fabr, was edited and translated
into Spanish by Sanchez-Perez. As we will see in the next section, this text
throws light on the connections between the algebraic tradition oI the Muslim
West and the algebra oI al-Khwrizm and Ab Kmil (d. 318/930).
107
In geometry, the only contribution was the publication oI a Iacsimile
with English translation oI Ibn Mudh al-Jayyn`s treatise on the theory oI
proportions oI Euclid`s Elements Book V.
108
This text is important both Ior
the history oI this theory in the Islamic world and Ior the history oI transmis-
sion between the Eastern and the Western Islamic world. But its publication
did not inaugurate new research in the mathematical tradition oI the Maghrib
and al-Andalus.
Publications on applications oI mathematics are also rare and limited
to the presentation oI the contents oI a manuscript. Two texts in the science
oI inheritances drew the attention oI Sanchez-Perez, probably more Ior their
sociological interest than Ior the mathematical content.
109
The construction oI
magic squares was studied through al-Bn`s writings,
110
but the work oI al-
Bn, who lived in Egypt, is representative oI the Eastern Islamic tradition.
1he Contents ot the Published Mathematical 1ets
The research between 1900 and 1980 shows a predominant interest among
historians in the mathematical works themselves, through editions, transla-
tions and analyses oI texts. This has made it possible to make a preliminary
overview oI the mathematics that was taught or used in the Maghrib Irom the
twelIth century onwards, although nothing precise had yet been revealed on the
teaching oI geometry. Thanks to the handbooks analyzed by Suter, Woepcke
and Souissi, we know that the Iollowing subjects in arithmetic were taught in
the Maghrib, between the twelIth and Iourteenth centuries:
the positional decimal system and its representation by the ghubr symbols.
111
the seven traditional arithmetical operations, i.e., duplication, halving, addition,
subtraction, multiplication, division, and extraction oI the exact or approximate
square root oI a number, with various algorithms Ior integers and with extensions
oI some oI these operations to Iractions and to quadratic or biquadratic irrationals.
Arithmetic Iormulas Ior the summation oI fnite series oI natural, odd, or even
numbers, squares, cubes, as well as the fnite geometrical series.
Operations specifc to Iractions, such as reduction and conversion.
The second part oI these handbooks oI arithmetics is devoted to solu-
tions oI equations by arithmetical methods, that is, the rule oI three and the
method oI Ialse position (in the Muslim West called amal al-kafft |method oI
scales|), then by the six canonical equations oI al-Khwrizm.
112
In addition, irrational numbers are Iound in a poem by Ibn al-Ysamn,
entitled Urf:a f l-fudhr |Poem on the Roots|,
113
and in al-Qalasd`s letter
entitled Risla f dhawt al-asm wa l-munfasilt |Epistle on the Binomials
and the Apotomes|.
114
One fnds here the tradition oI Euclid`s Elements Book
X, rephrased arithmetically, as well as some extensions oI this tradition discov-
ered in the Eastern Islamic word between the ninth and the eleventh century.
The historical studies that were published during this period on the
numeral systems used in the Muslim West all deal with the question oI origin,
usually in respect to the positional decimal system. Probably Ior cultural rea-
sons, this subject inspired a series oI articles on the origin oI the symbols
which were used Ior this system in the Maghrib.
115
These articles contributed
nothing new to the history oI mathematics, because they were written Irom
cultural and ideological motivations by scholars who had no expertise in the
scientifc aspects oI the subject.
At the beginning oI this century, one also observes a revival oI a non-
positional numeral system which had been used Ior centuries, in some admin-
istrations oI the Western Maghrib. These were the so-called numbers of
Fe: using 27 diIIerent signs. The system was also known in Maghrib under
the name oI hurf a:-:imm |fgures oI the account book|. The calculation
which was associated with it was called hisb rm |Byzantine calculation|.
A number oI old texts, dealing with this system and its applications, by al-
Uqayl, al-Es, and Sakrj,
116
appeared in a lithographed edition. These publi-
cations might explain the interest around the 1920s Ior this numeral system on
the part oI researchers oI diIIerent backgrounds, such as G. S. Colin, C. Pellat,
and J. A. Sanchez-Perez.
117
This explanation needs to be supported by Iurther
research in the cultural history oI the Maghrib in the frst halI oI twentieth cen-
tury. It would also be interesting to know iI this revival represents a return to
the scientifc heritage similar to, but on a lower level than, the development in
China in the end oI eighteenth and throughout the nineteenth centuries.
118
The chapters on algebra in the arithmetical handbooks and the small
poem oI Ibn al-Ysamn which I have mentioned, are limited to the defnition
oI algebra, its objects, and its methods: the unknowns, the numbers, the Iour
arithmetical operations applied to these objects when they stand alone or are
combined in polynomial expressions, the six canonical equations, the opera-
tions oI balancing and, fnally, the method oI solution.
In the Muqaddima, Ibn Khaldn says on algebra:
the frst one who wrote on this science is Ab Abdallh al-Khwrizm and,
aIter him, Ab Kmil Shuj ibn Aslam; then people Iollowed his method.
His book on the six equations is among the best-written on the |subject|.
Many |are those|, among the people oI al-Andalus, |who| commented on it
well. And among the best commentaries is al-Qurash`s book. And we heard
that one oI the greatest masters in mathematics among the people oI the East,
was led to |a greater number| oI equations than these six species, bringing
them to more than twenty, and that he determined valid procedures Ior their
resolution, Iollowed by geometrical demonstrations.
119
This passage has been translated and analyzed many times by histo-
rians oI sciences since Woepcke. The passage showed that, in the Muslim
West, algebra did not remain in the condition in which al-Khwrizm had leIt
it, and as we fnd it in the Maghribi handbooks oI the Iourteenth to fIteenth
centuries analyzed by Woepcke, Suter, and Marre. But Ibn Khaldn`s asser-
tions were partially confrmed by the publication oI Ibn Badr`s Mukhtasar f
l-fabr |Concise Work on Algebra|.
120
This work contains all the material oI al-
Khwrizm`s book but in a diIIerent presentation and with extensions and new
applications. Ibn Badr treats the same topics as those studied in the Islamic
East beIore al-Karaj. He goes beyond al-Khwrizm because he generalizes
the rule oI the powers, discusses the necessary terminology, and introduces
irrational numbers. He also solves the same types oI problems which one
encounters in Ab Kmil`s work: problem oI 'tens, 'capitals, 'meetings,
'cereals, the 'pursuit problem, etc. But the number oI examples solved in
each category is less and Ibn Badr generally selected the problems which do
not involve too many technical complications.
121
EiIty years later, an English
translation oI Ab Kmil`s treatise on algebra was published on the basis oI the
Hebrew version oI Mordecha Einzi.
122
Ab Kmil`s work is important Ior the
transmission oI algebra Irom the East to the West, but its publication did not
lead to a re-examination oI the algebraic texts Irom the Muslim West.
In number theory, Ibn Khaldn was not oI great help Ior historians oI sci-
ence since he said merely that the mathematicians
had given it up aIter they had drawn the main points Ior arithmetic demon-
strations, as Ibn al-Bann did in the book Raf al-hifb, as well as others.
123
The last sentence revealed the title oI a work whose contents were prom-
ising, but this became available only in the 1980s. The last word oI this sen-
tence kept its secrets even aIter the publication oI Sid al-Andalus`s Kitb
tabaqt al-umam. These 'other mathematicians who wrote on number theory
seem to have been Andalusis. Sid al-Andalus, who designates them by
'scientists in the theory of numbers
124
seems to diIIerentiate them Irom the
'scientists in calculation. Knowledge oI the titles oI their works could have
made it possible to solve the question oI their origin but Sid seldom mentions
the works oI the scientists he introduces. Among the titles he quotes, only one
could be connected to an Andalusian tradition in arithmetic, namely the Kitb
tabat al-adad |Book on the Nature oI the Number|,
125
but up to now, we do
not know its contents.
Ibn Khaldn seems to reIer to an arithmetical tradition in the Maghrib
in his discussion oI certain types oI integers and oI the fgurate numbers oI the
Arithmetica oI Nicomachus oI Gerasa.
126
The existence oI such a tradition was
confrmed in 1976 by the publication oI an anonymous text, attributed to Ibn
al-Bann and dealing with the determination oI amicable numbers.
127
Orirx+:+ioxs :xi Prsui+s or Prsr:rcn :r+rr +,:o
BeIore 1980, research on the mathematical tradition oI al-Andalus and the
Maghrib Iocused on arithmetic and algebra, and was based on the known writ-
ings oI three authors: al-Hassr, Ibn al-Bann, and al-Qalasd (12th, l3th
14th, 15th century respectively).
In spite oI the major contributions oI the frst researchers in this feld,
not all questions raised by the publication oI Ibn Khaldn`s Muqaddima had
been answered, and some interpretations concerning the mathematicians or the
mathematical tradition oI the Muslim West were unsatisIactory. In addition,
the modest place oI mathematics in published sources suggested that proftable
research was still possible. Several questions were still open.
The frst question was the beginning oI mathematical activity in the
Muslim West. The second concerned the circulation oI scientists, ideas, and
techniques between al-Andalus and the Maghrib, in relation to the cultural and
political history oI the area.
A third question related to the contents oI the various handbooks Ior
teaching which had already been published and analyzed. Obviously these
contents did not refect the level oI mathematics in the Eastern Islamic world,
and they were well below the level oI some handbooks published in Baghdad a
Iew centuries earlier. The historians tried to fnd an explanation Ior this estab-
lished Iact, or to qualiIy it by discovering new sources.
In addition to these general questions, there were also more specifc ones.
The frst concerned the mathematical symbolism which Woepcke Iound in the
work oI al-Qalasd. What were the origins oI this symbolism, its develop-
ment, and, especially, its role in teaching? The second was whether other intel-
lectual activities in Western Islamic cities could have infuenced the thought oI
the mathematicians in relation to some mathematical concepts or defnitions.
In addition, the biobibliographic or historical works suggested two
equally important topics: the history oI the inIrastructure oI teaching, and the
place oI the community oI mathematicians and their activities in the context oI
the intellectual liIe and the cultural and religious practices oI the cities oI al-
Andalus and the Maghrib.
1nr M:+nrx:+ic:i 1orics 1n:+ Wrrr S+uiiri
Combinatorics
Research on this subject began with the discovery in 1978 oI Ibn al-Bann`s
treatise entitled 1anbh al-albb al masil al-hisb |Warning to Intelligent
People on the Problems oI Calculation|.
128
The analysis oI this treatise showed
the interests oI the author in combinatorics as well as his contributions to this
feld. This contribution seemed to be based on earlier work by another mathe-
matician, Ibn Munim, whose book Fiqh al-hisb |the Science oI Calculation|
had been mentioned by Ibn Khaldn in his Muqaddima.
Eurther progress was made by the analysis oI another previously unpub-
lished work oI Ibn al-Bann, Raf al-hifb. The chapter on fgurate numbers
and the summations oI fnite sequences oI integers, as well as some inIorma-
tion drawn Irom late commentaries on Ibn al-Bann`s 1alkhs, made it pos-
sible to supplement the combinatorial part oI 1anbh al-albb and to place
these parts in a context oI interests and practices which seem specifc to the
Maghrib.
129
The subsequent discovery oI Ibn Munim`s Fiqh al-hisb and the study
oI its important chapter 11 showed that the Maghribi contributions in combi-
natorics, prior to those oI Ibn al-Bann, were important both in quantity and in
quality. Not less than nineteen pages oI this work are exclusively devoted to
combinatorial defnitions, theorems, and techniques.
130
The Fiqh al-hisb showed the utility Ior a historian oI science oI a suI-
fcient knowledge oI the political, cultural, and ideological environment where
the mathematical activities took place. Thus it was possible to reIute several
assumptions about Ibn Munim`s liIe and activities which had prevailed since
the 1940s and which had been a source oI debate between European historians
oI sciences such as Steinschneider, Suter, Woepcke, and Renaud.
The technical contents oI the combinatorial practice in the two above-
mentioned sources can be summarized as Iollows. The combinatorial activities
oI the medieval Arabic tradition are not a direct continuation oI the Greek and
Indian traditions, but their origin is closely related to an Arabic cultural real-
ity.
131
Eor they originated not Irom the traditional mathematical themes, but in
linguistics and thus with the frst studies on the Arabic language and the lexical
morphological terminological discussion which they prompted.
At a later stage, one observes inside the mathematical tradition oI the
Islamic East the frst signs oI what one could call a combinatorial approach,
in connection with simple enumerations in the solution oI geometrical or alge-
braic problems, which did not require the establishment oI Iormulas or general
processes. But, paradoxically, the mathematicians oI the Maghrib who dealt
with combinatorics, were not inspired by this later tradition.
Ibn Munim tells us that his project is to mathematize a lexicographical
problem inherited Irom the period oI al-Khall Ibn Ahmad (d. 797). He initially
established a rule enabling him to determine all the possible combinations oI
n colors in groups oI p. Eor that, he has to construct the arithmetical triangle
according to both an inductive and a combinatorial approach. Then he demon-
strates, according to the same double approach, the Iormulas Ior the permuta-
tions oI a set oI letters, with or without repetitions, and the Iormulas giving,
by recurrence, the number oI possible readings oI the same word oI n written
letters, taking into account all the unwritten signs used in a given language
(e.g., vowels and sukns Ior Arabic). He also gives a Iormula Ior arrangements
without repetitions oI n letters in groups oI p, taking account oI the unwritten
signs (vowels, etc.) that can accompany these letters, and he solves a certain
number oI problems on the maximum number oI combinations with repeti-
tions. These results enable the author, by means oI a series oI tables, to solve
the initial problem oI the enumeration oI all the words Irom one to ten letters
which can be Iormed with the letters oI an alphabet, taking into account all the
possible repetitions oI the letters in a word and the (unwritten) signs which can
be placed on these letters.
The results in Ibn Munim`s work were new, and he also derived them by
a new method, which involved a purely combinatorial reasoning. In the history
oI combinatorics, the Fiqh al-hisb is, at the same time, the end oI a stage oI
calculation by means oI tables, and the beginning oI a new stage, the substitu-
tion oI arithmetic Iormulas Ior these tables. These Iormulas are subsequently
used to solve mathematical problems in various felds.
The contributions oI Ibn al-Bann develop the scope oI Ibn Munim. In
the 1anbh al-albb, Ibn al-Bann establishes the Iamous arithmetical Iormula
permitting a direct computation oI the number oI combinations oI n objects in
groups p without having to construct the arithmetical triangle, as his predeces-
sor had done.
In his Raf al-hifb, one fnds combinatorial theorems similar to those
in the 1anbh al-albb, but presented as complements to two chapters on the
theory oI numbers inherited Irom the arithmetical tradition oI Nicomachus: on
the summations oI integers and on the fgurate numbers respectively.
In later Maghribi mathematical writings, combinatorial approaches ap-
pear in very diIIerent felds, such as magic squares, inheritances, grammar, and
even religion.
132
The Iact that all later authors use the same combinatorial terminology,
and that none oI them claim authorship Ior results or applications, reinIorce the
continuity oI the combinatorial tradition since Ibn Munim.
133
Algebra
The frst research in the history oI Maghribi algebra aIter 1980 was related to
the diIIusion in the Muslim West oI the six canonical equations oI degree one
or two. The available texts were Ibn al-Ysamn`s Urf:a, Ibn Badr`s Ikhtisr
f l-fabr, both already published, and the still unpublished Kitb al-usl oI Ibn
al-Bann. To these texts, which are exclusively devoted to algebra, one can add
the algebraic chapters which usually conclude the handbooks oI calculation.
This research showed that the canonical equations were generalized and
that changes were made in the classifcation oI these equations, in close con-
nection with some developments oI the algebra oI polynomials.
134
We have
already mentioned specifc developments oI notational symbolism oI the
Muslim West discovered by Woepcke and Renaud.
135
Traces oI this sym-
bolism were Iound in Ibn al-Ysamn`s 1alqh al-afkr |The Eecondation oI
Spirits|,
136
and through specifc testimonies, it has been possible to Iollow the
evolution oI these symbols and abbreviations into an instrument Ior solving
arithmetical and algebraic problems.
137
Based on these specifc studies on equations and notational symbolism,
I wrote a general assessment oI the research on the available algebraic sources
Irom the Muslim West.
138
A Iew years later, Saidan published a critical edi-
tion and analysis oI the only work on algebra Irom the Maghrib which has
reached us, namely Ibn al-Bann`s Kitb al-usl wa l-muqaddamt f l-fabr
wa l-muqbala |the Book oI the Bases and oI the Preliminaries in Restoration
and Balancing|.
139
This is a work in the algebraic tradition oI Ab Kmil, and it
gives an idea oI the level oI teaching oI algebra in some cities in the Maghrib,
such as Bougie, Marrakesh, and Eez.
Since the publication oI Ibn Khaldn`s Muqaddima, we know that
there was an algebraic tradition in this area.
140
This inIormation was par-
tially confrmed by Ibn al-Ysamn`s Urf:a, Ibn Badr`s Ikhtisr, and fnally,
by some surviving citations oI the lost work oI the Andalusi mathematician,
Abu`l-Qsim al-Qurash, who taught algebra and the science oI inheritance in
Bougie, in the twelIth century. Two distinct aspects oI this tradition appear in
two other categories oI writings translated Irom Arabic or strongly related to
the Arabic algebraic tradition oI al-Andalus. The frst category consists oI the
chapters oI the Liber Mahameleth on Algebra
141
and some parts oI Ab Bakr`s
Liber Mensurationum, and it contains solutions oI practical problems using the
method oI the algebra oI al-Khwrizm.
142
The second category consists oI texts on the solution oI problems oI
(land) measurement, including Abrahm Br Hiyya`s Liber Embadorum,
143
parts oI Ab Bakr`s Liber Mensurationum and the Risla f at-taksr |Book on
Measurement| oI Ibn Abdn (d. aIter 976).
144
These texts show the existence
in al-Andalus oI a problem-solving tradition with algebraic aspects anterior to
al-Khwrizm, and related to the Babylonian tradition.
The analysis oI the available sources shows that the mathematical tra-
dition oI the Maghrib was not a IaithIul preserver oI various aspects oI the
algebraic activity in al-Andalus. In the Maghrib, one fnds neither direct or
indirect reIerences to three important traditions: algebra associated with the
geometry oI measurement such as it appears in Ibn Abdn`s handbook, study
oI the polynomials on the basis oI the work done by Sinn Ibn al-Eath in the
East
145
and, fnally, the geometrical solution oI cubic equations oI al-Khayym
(d. 526/1131).
146
The frst tradition existed in al-Andalus and its absence in the Maghribi
writings can be explained either by the loss oI the sources or by the lack oI
interest on the part oI Andalusi mathematicians Irom the twelIth century
onwards. Euture research may reveal which oI these alternatives is correct.
The second tradition is partially present in the Maghribi texts, but the alge-
braic writings oI Eastern authors, like Ibn al-Eath, al-Karaj (d. 414/1023) and
al-Samawal (d. 571/1175) seem to have been unavailable in the West. The
silence oI the mathematical sources and Ibn Khaldn`s testimony suggest that
the third tradition did not take root in al-Andalus and the Maghrib.
147
The contents oI Western algebraic writings prior to the thirteenth cen-
tury can be summarized as Iollows. The oldest witness in al-Andalus oI the
'algebraic tradition in measurement problems seems to be Ibn Abdn`s
handbook. Many problems are set out and solved according to the procedures
in Babylonian texts, without any reIerence to the six canonical equations oI
al-Khwrizm`s Algebra and without the algebraic terminology oI the ninth
century. Textual similarities to problems in the Liber Mensurationum and the
Liber Embadorum suggest that this tradition oI measuring survived in Andalus
until the beginning oI the twelIth century.
In the tradition oI al-Khwrizm and Ab Kmil, Ab Bakr`s Liber Men-
surationum contains problems oI measurement that are similar and sometimes
identical to those treated by Ibn Abdn and Br Hiyya.
148
The author oI the
Liber Mahamalet reIers explicitly to Ab Kmil`s Algebra, the Kitb al-kmil,
but no other work is quoted.
Three texts were infuenced by the tradition aIter Ab Kmil. The mathe-
matical poem oI Ibn al-Ysamn summarizes the algorithms Ior the solution
oI the six canonical equations and accompanies them with some operations
on quadratic irrationals. Ibn Badr`s Ikhtisr al-fabr wa l-muqbala |the Sum-
mary oI Restoration and Balancing| is a summary oI algebra in the tradition oI
al-Khwrizm and Ab Kmil, with some Iurther additions. However, an
analysis oI the methods and the problems treated by Ibn Badr reveals some
characteristics that are diIfcult to relate to what is known about the alge-
braic traditions oI the East prior to al-Khayym. Ibn al-Bann`s Kitb al-usl
|Book oI Elements| is the last important work on Algebra oI the Muslim West,
and later algebraic works, in the Maghrib and in Egypt, derive their inspira-
tion entirely Irom the problems and the methods oI this book, or reproduce
them with explicit reIerences.
149
The Book of Elements is in two parts. The
frst part deals with numbers; it is a summary oI Books VIIX oI Euclid`s Ele-
ments with additions such as the division by irrational expressions oI the Iorm
n p7m psp. The second part is about the solution oI various types oI problems
using algebraic methods. Eirst Ibn al-Bann treats problems whose solutions
are integers or Iractions, and he concludes with problems whose solutions are
irrational numbers. This part oI the book is inspired by Ab Kmil`s problems,
and sometimes use exactly the same coeIfcients, but Iollow a diIIerent Iorm oI
presentation.
150
The additions by Ibn al-Bann are problems oI algebra which
do not occur in Ab Kmil`s book but which ft into the same tradition, and
problems on the theory oI numbers, such as the representation oI an integer as
to the sum oI two squares oI integers or rational numbers.
Ibn al-Bann also dealt with algebra in two other works, the 1alkhs
and the Raf al-hifb. Although the 1alkhs contains only the rules and the
basic algorithms oI algebra, without demonstration and application, it was
the Arabic mathematical work that received the most commentaries in the
Maghrib, between the Iourteenth and the sixteenth century. The Raf al-hifb
contains some new elements, in particular purely algebraic methods Ior set-
ting out and calculating the solutions oI quadratic equations, without reIerence
to the geometrical demonstrations Iound in the Algebra oI al-Khwrizm and
Ab Kmil.
AIter Ibn al-Bann`s Kitb al-usl, three kinds oI algebraic texts were
written in the Maghrib: commentaries on earlier works, independent obser-
vations, and chapters in arithmetical works. Their contents are not always a
simple repetition oI the methods in the earlier works.
All the algebraic commentaries which have reached us concern Ibn al-
Ysamn`s algebraic poem. Some were written in the Maghrib.
151
The analysis
oI these texts does not disclose anything new, other than that algebraic symbol-
ism was intensively used in the mathematical teaching in the Maghrib during
the Iourteenth to fIteenth centuries.
The algebraic chapters oI two other texts are important Ior showing the
use oI this symbolism. The frst is Ibn QunIudh al-Qasantn`s Hatt an-niqb
an wufh aml al-hisb |the Iowering oI the Veil on the Various Opera-
tions oI Calculation| which contains, in particular, a symbolic expression oI an
equation whose second term is zero.
152
The second work is Rashf ar-rudb min
thughr aml al-hisb |Sucking the Nectar Irom the Mouths oI the Operations
oI Calculation| by al-Qatrawn, a mathematician who lived in Tunis probably
at the end the Iourteenth century or the beginning oI the fIteenth. This is the
only extant mathematical text oI the Maghrib in which the calculation oI the
square and cubic roots oI an abstract polynomial is explained. The interest oI
this Iact does not lie in the results themselves, since the extraction oI the square
root oI a polynomial oI arbitrary degree had been explained by as-Samawal
(d. 1175) in Baghdad, and since the extraction oI the cube root could be car-
ried out by any mathematician who understood the technique.
153
The inven-
tors oI the method in the Eastern Islamic world used the symbolism oI tables
to support oI the algorithm, but al-Qatrawn uses the Maghribi letter symbols
to write out the data, the operations and the results. Al-Qatrawn`s book is a
good illustration oI the circulation oI ideas and techniques among various sci-
entifc centers in the Islamic world.
Arithmetic
Since 1990, the history oI arithmetic in al-Andalus and the Maghrib has been
the subject oI intensive research in the Maghrib itselI. This has resulted in the
editions and studies oI the Iollowing works written between the twelIth and
Iourteenth century: Ibn al-Bann`s Raf al-hifb,
154
Ibn QunIudh`s Hatt an-
niqb,
155
Ibn al-Ysamn`s 1alqh al-afkr,
156
al-Uqbn`s Sharh at-1alkhs,
157
and al-Hassr`s Kitb al-bayn wa t-tadhkr.
158
In addition, studies have been
made oI the methods oI Ialse positions
159
and arithmetical algorithms
160
in the
tradition oI al-Andalus and the Maghrib. The relationship between arithmetic
and the science oI Islamic inheritances has been studied on the basis oI the
Kitb al-istiqs wa t-tafns f ilm al-hisb |the Book oI the Investigation and
the Classifcation in Calculation| oI al-Hubb (4th/10th c.),
161
al-Mukhtasar
|the Abridged book>| oI al-HI (d. 1192)
162
and Sharh Mukhtasar al-Hf
|Comment on al-HI`s abridged book>| oI al-Uqbn (d. 1408).
163
Another
research project has dealt with aspects oI arithmetic practiced in al-Andalus
between the ninth and the eleventh century and their presence in Maghribi
writings, or in Andalusi works written aIter the eleventh century that were
distributed and taught in the Maghrib. These investigations started with the
study oI the arithmetical parts oI three important works which had never
been analyzed beIore: in chronological order, the frst three chapters oI Ibn
al-Ysamn`s 1alqh al-afkr, chapters IVII oI the frst part and the entire
second part on Iractions oI Ibn Munim`s Fiqh al-hisb and, lastly, Book I oI
Ibn al-Bann`s Raf al-hifb.
164
This increased knowledge oI the practice oI
arithmetic in the Maghrib has led to the identifcation oI the frst volume oI
the Iull version oI al-Hassr`s work, al-Kitb al-kmil f ilm al-ghubr. Suter
conjectured the existence oI such a work and he distinguished it Irom the Kitb
al-bayn oI the same author.
165
Since research is still in its initial stage, it does not allow us to describe
the arithmetic tradition oI the Muslim West in detail, but it is already pos-
sible to sketch briefy the major outlines and the characteristic aspects oI this
tradition.
As in the East, the Indian positional decimal system, with modifed num-
ber symbols oI Indian origin, dominated Andalusi and Maghribi teaching. But
two other systems oI numeration and calculation were used in al-Andalus and
in the Maghrib. The Hisb al-yad (Einger Reckoning) does not seem to have
been marginal since mathematicians oI some importance devoted writings
to it, as they also did in the Eastern Islamic world.
166
The second system is
the Rm calculation (see above, p. 321). Evidence in some mathematical texts
shows that this system was used at least since the twelIth century. Authors oI
Andalusi origin or Andalusi Iormation, like al-Hassr, Ibn Munim and Ibn
al-Ysamn, discuss the system at length. In mathematical texts the system is
presented in the Iorm oI a defnition with comment, or in the Iorm oI an inde-
pendent chapter. Ibn al-Bann even devoted a whole treatise to it.
167
The arithmetical algorithms in the extant works are mostly the same as
those already used in the East. Presentation and usage change in the course oI
time. In the works prior to the thirteenth century, there are two separate sec-
tions on the operations oI doubling and halving respectively, multiplication
is treated beIore the addition, and an important place is given to the chapter
dealing with Iractions, which sometimes occupies up to halI oI the work.
168
Erom the thirteenth century onward, these characteristics disappear Irom the
handbooks. Erom now on Iractions are represented with the bar separating
numerator Irom denominator (as we do nowadays) but no author claims this
innovation.
1he 1heory ot Numbers
Ibn al-Bann`s Raf al-hifb was the frst work to show historians that number
theory was studied in the Maghrib. The work deals with problems oI the
Euclidean tradition, such as the study oI prime numbers and perIect numbers,
and problems Irom the arithmetical tradition oI Nicomachus, on fgurate num-
bers, summation oI a fnite series oI integer numbers, and amicable numbers.
169
An original contribution oI Ibn al-Bann is the integration oI combinatorial
theorems into the theory oI numbers. He expressed these theorems by means oI
a fnite series oI fgurate numbers in the tradition oI Nicomachus, or by means
oI the fnite series oI powers oI integers.
It was the discovery oI Ibn Munim`s Fiqh al-hisb that raised the veil
on some aspects oI the Andalusi tradition. In chapter VIII, Ibn Munim estab-
lishes the Iormula Ior the sum oI a series oI integers, such as the sum oI the frst
n even, odd, even-even or even-even-odd numbers, and the sum oI the frst n
squares and cubes. In this chapter, Ibn Munim also uses analysis and synthe-
sis to fnd results usually demonstrated by induction.
170
Chapter IX Iollows the
tradition oI Andalusi writings oI the eleventh to twelIth centuries on various
summations oI rows or columns oI the table oI the fgurate numbers.
171
Chap-
ter X is devoted to perIect and amicable numbers. The author gives here the
calculation oI Eermat`s couple oI amicable numbers (17296 and 18416), which
al-Hassr also gave in the lost second volume oI his Kitb al-kmil.
172
The second source oI inIormation on the Andalusi tradition in the theory
oI numbers is in the chapter on arithmetic oI the Kitb al-istikml, which was
identifed at the beginning oI the 1980s. It is an abridged version oI what was, in
the Arabic mathematical tradition aIter the ninth century, the core oI the theory
oI numbers, namely Books VII, VIII, and IX oI Euclid`s Elements,
173
the Arith-
metical Introduction oI Nicomachus
174
and the Risla f l-add al-mutahbba
|Ietter on the Amicable Numbers| oI Thbit Ibn Qurra (d. 901).
175
Geometry
What was known beIore 1980 about the geometrical tradition oI the Muslim
West was limited to the problems oI measurement and preliminaries Ior astron-
omy. The importance oI this discipline was confrmed by Sid al-Anda-
lus, the eleventh century historian oI sciences who discussed the activities
and works oI Andalusi authors who were specialists in geometry
176
and who
practiced this discipline beIore or in the eleventh century. Ibn Khaldn also
mentions important geometers who lived in the twelIth to thirteenth centuries.
177
But until the end oI the 1970s, no important geometrical work oI the Muslim
West had been identifed and studied, except al-Jayyn`s Kitb mafhlt qisiyy
al-kura.
Some Iatin and Hebrew sources are important evidence oI one aspect oI
the oI Arabic geometrical tradition oI al-Andalus, because they are translations
oI Arabic texts or compilations based on Arabic material. The feld covered by
these sources is limited, since they are mainly
178
concerned with topics related
to astronomy or land measurement.
179
The publications aIter 1980 Iall into three categories: identifcation oI
lost works, the transmission oI Greek and Arabic geometrical works Irom East
to West, and the teaching oI geometry.
In 1984, I identifed and analyzed an anonymous text containing inIorma-
tion on the original contributions oI the Andalusi mathematician Ibn Sayyid.
These are an extension oI the Greek material on conics, and concern what would
be called much later the study oI curves oI degree higher than 2. Ibn Sayyid
outlined a classifcation oI one category oI these plane curves and their use to
solve geometrical problems oI the Greek tradition which would be expressed in
modern terms by equations oI degree equal to or higher than 5.
180
The discovery, in 1984, oI most oI the geometrical chapters oI the Kitb
al-istikml oI al-Mutaman, by J. P. Hogendijk, answered the questions oI
Sarton in his Introduction to the History of Science.
181
The work throws new
light on the Andalusi geometrical tradition
182
and on the transmission oI the
geometrical writings oI the Greek corpus, such as Euclid`s Elements, Apollo-
nius`s Conics,
183
or Menelaus`s Spherics
184
in Arabic Irom the Muslim East to
the West. Al-Mutaman also used geometrical works oI the tenth century such
as Ibrhm Ibn Sinn`s treatise on the area oI the parabola, and even oI the
eleventh century, such as the works oI Ibn al-Haytham on optics and on analy-
sis and synthesis.
185
Al-Mutaman, or his Andalusi predecessors, also made
original contributions to geometry, such as the theorem oI Ceva
186
and the con-
struction oI two mean proportionals between two given lines.
187
However, the manuscripts oI the Kitb al-istikml are incomplete, and
they are a preliminary draIt by al-Mutaman oI a work in two parts, oI which
the second part was apparently never published. InIormation on the missing
sections and on the chapters oI the proposed second part oI al-Mutaman`s
project became available ten years later, thanks to the discovery oI two copies
oI a manuscript oI the thirteenth century. These manuscripts were written in
Asia by Ibn Sartq, a mathematician oI the East who had careIully studied the
published version oI al-Mutaman`s book and who had made a new version
that was not very diIIerent Irom the original one.
188
Another important problem in the transmission oI mathematical texts
relates to the Arabic manuscript copies oI Euclid`s Elements preserved in the
Escurial and in Rabat respectively. A comparative analysis showed that the
geometers oI al-Andalus and the Maghrib might have had at their disposal the
second al-Hajjj version oI the Arabic translation oI the Elements, which was
dedicated to the Caliph al-Mamn.
189
In the feld oI the elementary teaching oI geometry on the measurement
oI fgures several treatises (rasil) were published in the 1980s.
190
But much
work remains to be done beIore we can have an adequate idea oI this teaching.
To begin with, the published texts will have to be compared with those which
are not yet published, such as the geometrical chapter oI Ibn al-Ysamn`s
1alqh al-afkr
191
and Ibn Abdn`s Risla f t-taksr.
192
Conclusion
The Iollowing points that arise Irom this brieI survey, may indicate what
course to Iollow in investigating both material that has yet to be disclosed, and
texts which have been discovered, but not yet analyzed.
Eirst, there is the problem oI breaks oI continuities in the transmission
Irom the East to the West. Then, there are problems relating to the subject
matter oI calculation, geometry, and algebra.
In the science oI calculation, the questions are: Why did authors prior to
the thirteenth century give so much space to Iractions in their teaching manu-
als? What made Ibn al-Bann reduce this space, and why did he not hit upon
the idea oI reducing all Iractions to one Iorm? And why did the science oI
calculation become the main element oI mathematics (in all its aspects) in the
Maghrib in the post-Almohad period? The last question involves the position
oI mathematics in society as a whole, and the possible negative infuences oI
the environment on scientifc activity. But, in the light oI what we now know
on the cultural history oI Maghrib, we should also question the role oI this
environment in the protection oI contents and scientifc level, which one could
designate as minimal.
In geometry the survey has revealed various Arabic versions oI Euclid`s
Elements, which need to be careIully compared with each other, and Andalusi
and Maghribi geometrical texts which need to be ftted into the Euclidean tra-
dition.
193
In algebra, the most puzzling Iactor is the silence oI Andalusi biobibliog-
raphers concerning the presence oI the subject in the curricula oI the scientists
they mention. Moreover, they do not mention which works on algebra were
transmitted Irom the East to the West. It is only through the Iatin and Hebrew
translations made Irom the twelIth century onward that we know that al-
Khwrizm`s Mukhtasar and Ab Kmil`s Algebra must have been available
in al-Andalus. More research is needed to fnd these works, and their traces, in
Arabic sources in the region.
Einally, it has been shown that, in addition to research on the mathemati-
cal texts themselves, investigations using the biobibliographical writings Irom
the Muslim West have proved useIul. Eor they have resulted in several publica-
tions relevant to the history oI the Andalusi and Maghribi mathematical tradi-
tions, ranging Irom writings on the biobibliography oI mathematicians in the
Middle Ages,
194
through the mathematical activities in each oI the three main
geographical regions oI the Maghrib,
195
to the study oI the situation in a spe-
cifc period in a specifc region.
196
But these studies, while providing a useIul
starting point, also show how much more research remains to be done.
Ackxovirixrx+s
I should like to thank Jan P. Hogendijk Ior his helpIul advice and suggestions
during the preparation oI this paper, and C. Burnett Ior correcting the English
translation.
No+rs :xi Prrrrrxcrs
1. There is not yet, to my knowledge, any database oI the publications on the history oI
mathematics and astronomy in the Islamic world, but the existing partial bibliographies
confrm this disproportion. See D. A. King: The Exact Sciences in Medieval Islam:
some remarks on the Present State oI Research, M E. S. A., 4 (1980), 17 pp.; J. I. Berg-
gren: History oI Mathematics in the Islamic World: The Present State oI the Art, Middle
East Studies Association Bulletin, 19, 1 (1985), 933; D. A. King: An overview of the
sources for the history of Astronomy in the medieval Maghrib, 2
e
Colloque Maghrebin
sur l`Histoire des Mathematiques Arabes, Tunis, 13 Decembre 1988, Actes du Collo-
que: Tunis, Universite de Tunis I-I.S.E.E.C.-A.T.S.M., 1990, 125157; J. I. Berggren:
Mathematics and her Sisters in Medieval Islam: A Selective Review oI Work Done
Irom 19851995, Historia Mathematica 24 (1997), 407440.
2. A. Djebbar: Ies scientifques arabes Iace a leur patrimoine, Maghreb-Machrek,
Paris, Documentation Iranaise, 105 (1984), 4864.
3. J. E. Montucla: Histoire des Mathematiques, Paris, 17991802.
4. C. Bossut: Histoire generale des mathematiques, Paris, 1810.
5. J. B. Delambre: Histoire de lAstronomie du Moyen-age, Paris, 1819.
6. M. Pelayo: La ciencia espaola, Madrid, 18871889.
7. J. von Hammer-Purgstall: Literaturgeschichte der Araber, Vienna, 18531855.
8. J.-J.-E. Sedillot (17771832) graduated at the Ecole Polytechnique and then at the
Ecole des Iangues Orientales (Iounded in 1795). He was appointed as an associate
astronomer by the Bureau des Iongitudes (created especially Ior the study oI the his-
tory oI astronomy in Eastern civilizations). He completed the translation oI Ibn Ynus`s
Zf |Tables| started by Caussin de Pereval, beIore starting to work on other important
translations, namely Ulugh Beg`s Zf and the Book of the Principles and the Goals by
al-Hasan al-Marrkush. The last work was completed by his son I.-A. Sedillot.
9. AIter studying in mathematics in Berlin and Arabic in Bonn, Eranz Woepcke (1826
1864) was sent to Paris when he was 24 years old, by Alexander von Humbolt who had
directed him toward research on Arabic mathematics. He was the frst to publish the
contributions oI al-Karaj and al-Khayym in algebra and the use oI symbolic notation
in Maghribi mathematical writings.
10. It would be oI interest to study the motivations oI the frst publications in this feld
(which date back to the very frst year oI the nineteenth century)especially the trans-
lation in 1804 by Caussin de Perceval oI the frst chapters oI Ibn Ynus`s Hkemite
tablesand to consider the possible connections with the expedition to Egypt. See J.
Dhombres: Iìmage du monde arabe dans le bilan des activites scientifques dresse par
lÌnstitut de Erance sous lÈmpire, Sciences et 1echniques en Perspective, 20 (1992),
155163; J. Dhombres: Scientihc Motivation for and Mood From the Experience of the
Egyptian Expedition, XXth International Congress oI History oI Sciences, Symposium
'Science, Technology, and Industry in the Ottoman World, Iiege (Belgium), 2026
July 1997, Turnhout, Brepols Publishers, 2000, 9199; J. Dhombres, J. B. Bebert and J.
Eourier: La chaleur mathematisee, Paris, Berlin, 1998, chap. 5, LEgypte.
11. In this paper a notation such as 661/1262 means 661 in the (Islamic) Hijra era and
1262 in the Christian era.
12. J. J. Sedillot (trans.): 1raite des instruments astronomiques des Arabes compose
au trei:ieme siecle par Aboul Hassan Ali de Maroc, Paris, Imprimerie royale, 1834.
Reprinted: ErankIurt, Institut Ir Geschichte der arabisch-islamischen WissenschaIten,
1984.
13. I. A. Sedillot: Memoire sur les instruments astronomiques des Arabes, Paris, 1844.
Reprinted in: Al-Marrkush Ab Al al-Hasan ibn Ali ibn Umar (7th/13th.), Erank-
Iurt, Institut Ir Geschichte der arabisch-islamischen WissenschaIten, 1998, 45312.
14. Moritz Steinschneider was born March 30, 1816. He graduated in Prague and in
Berlin. AIter teaching in these two cities, he was given a post in the Berlin library. He
died in Berlin on January 24, 1907. Eor az-Zarqlluh, see his: Etudes sur Zarkali, astro-
nome arabe du Xl
e
siecle, et ses ouvrages, Bullettino di Bibliograha e di Storia delle
scien:e matematiche e hsiche, 14 (1881), 171182; 16 (1883), 493513; 17 (1884),
765794; 18 (1885), 343360; 20 (1887), 136, 574604.
15. M. Rico y Sinobas: Libros del Saber de Astronomia del Rey D. Alfonso X de
Castilla, Madrid, 18631867.
16. Heinrich Suter was born in 1848 near Zrich in Switzerland. He studied mathemat-
ics and received a doctorate in the history oI mathematics in 1871. At the age oI 40
years he began to learn Arabic and Ior the next thirty years he did research on Arabic
mathematics and astronomy. Eor a detailed biography oI H. Suter by J. Ruska, see Isis
5 (1923), 409417.
17. E. Woepcke: Notice sur des notations algebriques employees par les Arabes, Jour-
nal Asiatique, 5
e
serie, 4 (1854), 369372.
18. E. Woepcke: Traduction dùn chapitre des Prolegomenes dÌbn Khaldn relatiI aux
sciences mathematiques, Atti dellAccademia Pontihca dei Nuovi Lincei, 10 (185657),
236248. Reprint in E. Woepcke: Etudes sur les mathematiques arabo-islamiques,
ErankIurt, Institut Ir Geschichte der arabisch-islamischen WissenschaIten, 1986, vol.
1, 711723.
19. A. M. Quatremere: La Muqaddima, Paris, 1858.
20. M. de Slane: Prolegomenes, Paris, 18441862.
21. E. Rosenthal: 1he Muqaddimah, an Introduction to History, 3 vols, New York,
1958.
22. V. Monteil (trans.): Discours sur lHistoire universelle, Paris, Sindbad, 1978, 3
vols. All reIerences in this paper are to the Arabic edition: Ibn Khaldn: al-Muqaddima;
in Kitb al-ibar |The Book oI Examples|, Beirut, Dr al-Kitb al-Iubnn, Maktabat
al-Madrasa, 1983, vols. 12.
23. A. Djebbar: Ibn Khaldn et les Mathematiques, a travers la classihcation des sci-
ences de la Muqaddima, to appear.
24. This is the year in which Ibn Khaldn began writing the Muqaddima. He fnished
the frst version by 1382, when he oIIered a copy to the Prince oI Tunis. During the next
twenty years Ibn Khaldn introduced into his work additions and corrections, as he says
in the autograph manuscript now in Istanbul (Ms. Istanbul, AtiI EIendi, 1936).
25. Ibn Sn included in his compendious work on philosophy ash-Shif Iour sections
dealing with the Iour disciplines oI the quadrivium: number theory, geometry, astron-
omy, music. Eor geometry, he simply wrote a summary oI Euclid`s Elements (including
the three arithmetical books). His treatise on number theory is a work oI synthesis. See
Ibn Sn: Kitb ash-Shif |Book oI Cure|, geometry and arithmetic, ed. A. I. Sabra and
A. I. Mazhar, Cairo, 19751976.
26. Ibn Sn also included in his work Kitb an-Naft (which is a summary oI Kitb
ash-Shif) chapters on the Iour mathematical disciplines oI the quadrivium. These
chapters were later translated into Persian and enclosed in Ibn Sn`s Danesh-nameh
|the Book oI Science|. See Ibn Sn: Kitb an-Naft |Book oI Salvation|, Ie Caire,
1903; M. Achena and H. Masse (trans.): Avicenne, Le livre de science |Danesh-nameh|,
Paris, Ies Belles Iettres, 1958, 91270.
27. 'Lakhkhasa barhnah wa ghayyarah an ikhtilh 1-hurh fh il ilalin
manawiyyatin zhiratin hiya sirru 1-ishrati bi l-hurh wa :ubdatuh. Ibn Khaldn:
al-Muqaddima, op. cit., pp. 897898.
28. Ibn Khaldn uses the term mughlaq.
29. Op. cit., note 22, vol. 2, 896898.
30. Woepcke translates the passage oI note 27 as Iollows: 'Il resuma les demonstra-
tions de ces deux ouvrages et autre chose encore en fait de ce qui concerne lemploi
technique des signes dans ces demonstrations, servant a la fois pour le raisonnement
abstrait et pour la representation visible (hguree), ce qui est le secret et lessence de
lexplication (des procedes du calcul) au moyen des signes. See op. cit., note 18, 239.
31. On al-Qalasd, see M. Souissi: lim riyd andalus tnus al-Qalasd |al-
Qalasd, an Andalusi-Tunisian mathematical scientist|, Bulletin de lUniversite de
1unis, 9 (1972), 3349; A. S. Saidan: al Qalasd. In Dictionary of Scientihc Biog-
raphy, Ch. C. Gillispie (ed.), New York, Scribner`s Sons, 19701980, vol. 11, 229
230. Erom now on, the Dictionary of Scientihc Biography will be quoted as: D. S. B.;
A. Djebbar: al-Qalasd, un savant andalo-maghrebin du XV
e
siecle, Revue Arabe des
1echnologies, Paris, 9 (1990), 1223.
32. E. Woepcke: Note sur des notations algebriques employees par les Arabes, Comptes
Rendus de lAcademie des Sciences, 39 (1854), 162165. Reprinted in E. Woepcke:
Etudes sur les mathematiques arabo-islamiques, op. cit., vol. 1, 641644.
33. E. Woepcke: Traduction du traite dÀrithmetique dÀboul Haan Ali Ben Moham-
med Alkaladi, Extraits des Atti dellAccademia Pontihcia de Nuovi Lincei, 12 (1858
1859), 230275, 399438, in: Recherches sur plusieurs ouvrages de Leonard de Pise,
decouverts et publies par M. le Prince Balthasar Boncompagni, Rome, Imprimerie des
sciences mathematiques et physiques, 1859, 166.
34. E. Woepcke: Introduction au calcul gobr et hawa`, Atti dellAccademia Pontihcia
de Nuovi Lincei, 19 (1866), 365383. Reprint in E. Woepcke: Etudes sur les mathema-
tiques arabo-islamiques, op. cit., vol. 2, 541559.
35. E. Woepcke: Passages relatiIs a des sommations de series de cubes, extraits de trois
manuscrits arabes inedits de la bibliotheque imperiale, Annali di scien:e matematiche
e hsiche, compilati de Barnaba 1ortolini, 5 (1863), 147181. Reprint in E. Woepcke:
Etudes sur les mathematiques arabo-islamiques, op. cit., vol. 2, 476510.
36. M. A. Cherbonneau: Notice bibliographique sur Kalaadi mathematicien arabe
du XV
e
siecle, Journal Asiatique, 5
e
serie, XIV (1859), 437448. G. Enestrm: Sur
une Iormule dàpproximation des racines carrees donnee par Al Kalaadi, Bibliotheca
Mathematica, 236/9 (1886), 222229.
37. A. Marre: Ie Talkhys dÌbn al-Bann, Atti dellAccademia Pontihcia dei Nuovi
Lincei, 17 (1864), 289319. Eor example, the passage: 'al-gharadu f hdh l-kitb
talkhs aml al-hisb wa taqrbu abwbih wa manh wa dabti qawidih wa
mabnh |the goal, in this book, is to summarize the operations oI calculation, to
make understandable its chapters and its concepts and to fx its rules and its struc-
tures|, was translated by Marre as Iollows: 'Le but dans la composition de ce traite est
danalyser succinctement les operations du calcul, den rendre plus facilement accessi-
bles les portes et les vestibules, et den etablir solidement les fondements et la batisse.
M. Souissi explains the literal character oI this translation by the concerns oI Marre
to show that Ibn al-Bann, hls de maon, a employe la langage dun artisan. See
M. Souissi: Ibn al-Bann al-Marrkush, 1alkhs aml al-hisb |Ibn al-Bann Irom
Marrakech, Handbook on the Operations oI Calculation|, Erench edition, translation
and commentaries, Tunis, Publications de lÙniversite de Tunis, 1969, 10.
38. M. Steinschneider: Rectifcation de quelques erreurs relatives au mathematicien
arabe Ibn al-Bann, Bulletino di Bibliograha e di Storia delle Scien:e Matematiche e
Fisiche (Boncompagni) 10 (1877), 313314.
39. M. Steinschneider: Hebrasche Bibliographie, 80 (1874). Reprint: New York, Olms,
1972, vol. 3, 41.
40. M. Steinschneider: Die hebraischen berset:ungen des Mittelalters und die Juden
als Dolmetscher, Berlin, Bibliographisches Bureau, 1893, vol. 2, 557558.
41. H. Suter: Das Rechenbuch des Ab Zakary el-Hassr, Bibliotheca Mathematica,
serie 3, 2 (1901),1240.
42. H. Suter: Die Mathematiker und Astronomen der Araber und ihre Werke, Ieipzig,
Teubner, 1900.
43. The Bibliotheca Arabica Hispanica consists oI the editions made in Madrid and
Saragossa by E. Codera and his collaborators including: Ibn Bashkuwl`s Kitb al-sila
|the Book oI the continuation| (1883), al-Dabb`s Bughyat al-multamis f trkh rifl
ahl al-Andalus |the Desire oI the Researcher on the History oI the Men oI al-Andalus|
(1885), al-Mufam |the Iexicon|, Ibn al-Abbr`s at-1akmila li Kitb as-sila |the Com-
pletion to the Book oI the Continuation| (1885, 18861889) and Ibn al-Earad`s 1rkh
ulam al-Andalus |History oI the Scientists in al-Andalus| (18911892).
44. Maqqar (al-): Nafh at-tb min ghusn al-Andalus ar-ratb |Scent oI the PerIume oI a
Tender Branch oI al-Andalus|, ed. R. Dozy, G. Dugat, I. Krehl and W. Wright, Ieiden,
18551861.
45. Chapter 13 oI Ibn Ab Usaybia`s book is devoted to the medical doctors oI the
Muslim West. See Ibn Ab Usaybia: Uyn al-anb f tabaqt al-atibb |Sources oI
InIormation on the Categories oI the Physicians|, A. Mller (ed.), Cairo-Knigsberg,
18821884. Other equally important works by Ibn Khallikn and Hjj KhalIa were
also available. See Ibn Khallikn: Wafayt al-ayn |Necrology oI the Eamous|, trans.
M. De Slane, Paris-Iondon, 18431871; Hjj KhalIa: Kashf az-zunn an asmi al-
kutub wa l-funn |Revealing the Doubts about the Names oI Books and Arts|, trans. G.
Elgel, Ieipzig, 18351858.
46. M. Al-Mann: al-masdr al-arabiyya li trkh al-Maghrib |The Arabic Sources
oI Moroccan history|, Publications de la Eaculte des Iettres et des Sciences Humaines-
Rabat, Serie Etudes bibliographiques, 12; vol. 1, Casablanca, Muassasat Banshara li
t-tiba wa n-nashr, 1983; vol. 2, Mohammadia, Matbaat Eudal, 1989.
47. H. P.-J. Renaud: Iènseignement des sciences exactes et lèdition dòuvrages scien-
tifques au Maroc avant lòccupation europeenne, Hesperis, XVI (1933), 7889.
48. A. C. Binebine: Histoire des bibliotheques au Maroc, Publications de la Eaculte des
Iettres et des Sciences Humaines-Rabat, Serie Theses et Memoires, 17, Casablanca,
Imprimerie Najh al-jadda, 1992. Bencheneb and Ievi Provenal: Essai de repertoire
chronologique des editions de Fes, Alger, 1922.
49. J. Murdoch: Euclid. In D. S. B., op. cit., vol. 4, 414459.
50. Op. cit., note 47.
51. E. Ihsanoglu: Ottomans and European Science. In Science and Empires, P.
Petitjean, C. Jami and A. M. Moulin (edit.), Boston Studies in the Philosophy oI Sci-
ence, vol. 136, Dordrecht, Boston, Iondon, 1992, 3748. E. Gnergun: Introduction oI
the Metric System to the Ottoman State. In E. Ihsanoglu (ed.): 1ransfer of Modern Sci-
ence and 1echnology to the Muslim World, Istanbul, IRCICA, 1992, 297316.
52. I. Benjelloun-Iaroui: Les bibliotheques au Maroc, Paris, Maisonneuve et Iarose,
1990, 5556.
53. H. Hankel: Zur Geschichte der Mathematik im Altertum und Mittelalter, Ieipzig,
1874, pp. 247250.
54. C. Brockelmann: Geschichte der Arabischen Litteratur, Weimar, 18981902.
55. Ibn al-QiIt: Kitb Ikhbr al-ulam bi akhbr al-hukam |Book which inIorms
Scholars on the IiIe oI the Wise Men|, Iippert (ed.), Ieipzig, 1903.
56. Sid al-Andalus: Kitb Tabaqt al-umam |Book oI the Categories oI Nations|, I.
Cheikho (ed.), Beirut, Imprimerie Catholique, 1912. Translation: R. Blachere: Livre des
Categories des nations, Paris, 1935.
57. J. A. Sanchez-Perez: Biografias de matematicos arabes que orecieron en Espaa,
Madrid, Impr. E. Maestre, 1921.
58. We will return to these works below.
59. One oI these lost works in particular intrigued George Sarton, namely the Kitb
al-istikml oI al-Mutaman Ibn Hd. Sarton said about this book: 'It is strange that a
work believed to be so important and written by a king should be lost. G. Sarton: Intro-
duction to the History of Science, Baltimore, Williams and Wilkins, 1927, vol. 1, 759.
About the discovery oI this work and its importance, see below.
60. Op. cit., vol. 2, 899.
61. H. P.-J. Renaud: Additions et Corrections a Suter 'Die Mathematiker und Astrono-
men der Araber, Isis 18 (19321933), 166183.
63. In al-Andalus as well as in the Islamic East, astronomy had a more important posi-
tion than mathematics. Erom the very beginning oI the arabo-islamic civilization,
astronomy was related to problems oI religious practice, such as the determination
oI the azimuth oI Mecca, the calculation oI the times oI the daily prayers, and fxing
the frst day oI Ramadan. Astronomy was also used in astrology. The relative impor-
tance oI astronomy and astrology throughout the Arabic middle ages is refected by the
number oI pages devoted to them by the medieval Islamic and modern bibliographers.
Eor example, Ior the period prior to 1038, E. Sezgin devotes a volume oI 521 pages
to astronomy and another oI 486 pages to astrology, whereas mathematics occupies a
single volume oI 514 pages: E. Sezgin: Geschichte des arabischen Schrifttums, Band
V, Mathematik bis ca. 430 H. Ieiden, Brill. 1974; Band VI, Astronomie bis ca 430 H.,
1976; Band VII, Astrologie bis ca 430 H., 1978.
64. J. M. Vallicrosa: Estudios sobre A:arquiel, Madrid-Granada, 19431950.
65. Eor a detailed bibliography oI these contributions, see: J. Samso: 1he Exact
Sciences in al-Andalus; in S. K. Jayys (ed.): 1he Legacy of Muslim Spain, Ieiden,
Brill, 1994, vol. 2,952973.
66. J. Samso: 1res notas sobre astronomia hispanica en le siglo XIII. In J. Vernet:
Estudio sobre histoia de la ciencia arabe, Barcelona, 1980, 175177. R. Iorch: The
Astronomy oI Jbir ibn Afah, Centaurus, 19 (1976), 85107.
67. J. Samso and M. A. Catala: Un instrumento astronomico de raigambre zarql: el-
cuadrante shakkz de Ibn Tbugh, Memorias de la Real Academia de Buenas Lettras
de Barcelona, 13 (19711975), 531. J. Samso: A propos de quelques manuscrits
astronomiques des bibliotheques de 1unis. contribution a une histoire de lastrolabe
dans lEspagne musulmane, Actas del 11 Colloquio Hispano-Tunecino de Estudios
Historicos, Madrid, 1973, 171190; D. A. King: On the Early History oI the Universal
Astrolabe in Islamic Astronomy and the Origin oI the Term 'Shakkziya in Medieval
Scientifc Arabic, Journal for the History of Arabic Science, 3 (1979), 244257.
68. S. Pines: Ibn al-Haythams Critique of Ptolemy, Proceedings oI the Xth Interna-
tional Congress oI History oI Sciences, Ithaca, 1962. Paris, 1964, vol. 1, 547550.
69. E. J. Carmody: The Planetary Theory oI Ibn Rushd, Osiris, 10 (1952), 556586; B.
R. Goldstein: al-Bitrf. Principles of Astronomy, New Haven-Iondon, 1971.
70. H. Hermelink: Tabulae Jahen, Archive for History of Exact Sciences, 2 (1964), 108
112; G. J. Toomer: A Survey oI the Toledan Tables, Osiris, 15 (1968), 5174.
71. J. Vernet and M. A. Catala: Ias obras matematicas de Maslama de Madrid. In J.
Vernet (edit.): Estudios sobre historia de la ciencia medieval, Barcelona-Bellatera,
1979, 241271.
72. J. Samso: The early Development oI Astrology in al-Andalus, Journal for the His-
tory of Arabic Science, 3, 2 (1979), 228243. The author specifes there, in connection
with the loans Irom the Iatin astrological tradition: 'the aformentioned Arabic text is
based on the translation of a Latin astrological work which was known in al-Andalus
towards the end of the eighth or beginning of the ninth century, therefore being one
more item in the long series of contacts between Isidorian-Latin and Arabic culture in
Muslim Spain (p. 233).
73. D. A. King: Three Sundials Irom Islamic Andalusia, Journal for the History of
Arabic Science, 2 (1978), 358392; J. Samso: 1he Exact Sciences in al-Andalus, op.
cit., n. 65, vol. 2, p. 956.
74. J. Samso: Maslama al-Majrt and the Alphonsine Book on the Construction oI the
Astrolabe, Journal for the History of Arabic Science, 4 (1980), 38.
75. E. Poulle: Un instrument astronomique dans lÒccident latin: la Saphea,` Studi
Medievali, 10 (1969), 491510.
76. M. V. Villuendas: La trigonometria europea en el siglo XI. Estudio de la obra de
Ibn Mud, El Kitb mayhlt, Barcelona, Instituto de Historia de la Ciencia de la Real
Academia de Buenas Ietras, 1979.
77. Al-Brn: Kitb maqld ilm al-haya, La 1rigonometrie spherique che: les
Arabes de lEst a la hn du X
e
siecle, M. T. Debarnot (ed. and trans.), Damascus, Institut
Eranais de Damas, 1985.
78. Ibn al-Bann: Kitb Minhf at-tlib li tadl al-kawkib |The guide to the Student
Ior the Correction oI the Movements oI the Stars|, in J. Vernet: Contribucion al estudio
de la labor astronomica de Ibn al-Bann, Tetouan, 1952.
79. J. Vernet, 1956: Les manuscrits astronomiques dIbn al-Bann. In: Proceedings oI
the VIIIth International Congress oI History oI Sciences (Milano, 1956), Paris, Her-
mann, 1958, 297298.
80. H. P.-J. Renaud: Sur les dates de la vie du mathematicien arabe marocain Ibn al-
Bann` (XII
e
XIII
e
s. J.C.), Isis 37, 2 (1937), 216218; H. P.-J. Renaud: Ibn al-Bann`
de Marrakech, suf et mathematicien (XII
e
XIII
e
S. J. C.), Hesperis XXV (1938), 13
42.
81. H. P.-J. Renaud and J. S. Colin: Note sur le muwaqqit marocain Ab Muqri-ou
mieux Ab Miqra al-Battw, Hesperis XXV (1938), 9496.
82. H. P.-J. Renaud: Astronomie et Astrologie marocaine, Hesperis XXIX (1942),
4163.
83. H. P.-J. Renaud: Determinations marocaines de lòbliquite de 1ècliptique. Bulletin
de lenseignement public, 170 (1941), 321336.
84. H. P.-J. Renaud: Sur les lunes du Ramadan, Hesperis XXXII (1945), 5168.
85. H. P.-J. Renaud: Le calendrier dIbn al-Bann de Marrakech, Paris, 1948.
86. In this connection it would be interesting to study the writings oI these two Muslim
intellectuals, especially those which they published during their stay in Paris.
87. A. Gannn: an-Nubgh al-maghrib f l-adab al-arab |The Moroccan Genius in
Arabic Iiterature|, Eirst edition, Tetouan, 1938.
88. Q. H. Tqn: 1urth al-Arab al-ilm f r-riydiyyt wa l-falak |The Arabic Scien-
tifc Heritage in Mathematics and Astronomy|, Eirst edition, Beirut, Dr ash-shurq,
1941, 280 pp.
89. A. al-Azzw: 1rkh ilm al-falak f l-Irq wa alqatuh bi l-aqtr al-islmiyya
wa l-arabiyya f l-uhd at-tliyya li ayym al-Abbsiyyn (6561335/12581917)
|History oI Astronomy in Iraq and its Relationship to the Islamic and Arabic Regions
during the Periods oI Time Which Eollowed the Abbassid Period|, Baghdad, Matbaat
al-mafma al-ilm al-irq, 1958.
90. In 1954 the book oI Tqn appeared in a second edition which was enlarged Irom
280 to 450 pages. It was sold out within two years, and was Iollowed in 1963 by a third
edition which included corrections and additions. The book oI Gannn also appeared in
revised editions in 1961 and in 1975.
91. To these three works, written in Arabic, it is necessary to add Slih Zak`s book,
written in Turkish, the diIIusion oI which was more limited, but which also reached the
European researchers. See S. Zak: thr bqiyya |Remaining Relics|, Istanbul, 1911.
92. G. Sarton: Compte-rendu au Iivre de Q. H. Tqn, Isis 36 (1944), 140.
93. Op. cit., note 87, third edition, 1975, vol. 2, 6.
94. Op. cit., note 91, vol. 2, 286291. W. Hartner was the frst western specialist who
mentioned this mathematician while reIerring to S. Zak and to Q. H. Tqn. See W.
Hartner: Quant et comment sest arrete lessor de la culture scientihque dans lIslam?
In R. Brunschvig and G. E. Von Grunebaum: Classicisme et declin culturel dans 1his-
toire de lIslam, Actes du Symposium International dHistoire de la civilisation musul-
mane (Bordeaux, 2529 Juin 1956), Paris, Maisonneuve et Iarose, 1977, 334, note 4.
A manuscript oI Ibn Hamza`s work exists in Cairo (Ms. Cairo National Iibrary, Talat
riyda, turki 1), but to my knowledge, it has not yet been the subject oI any thorough
study. Since Ibn Hamza originated Irom the city oI Algiers, it would be interesting to
see how this work relates to the traditions in calculation in the Muslim West and that in
the Muslim East.
97. There existed another Abu`l-Abbs Ibn al-Bann originating Irom Saragossa who
wrote a mystical poem entitled: al-Mabhith al-asliyya an fumlat at-tarqa as-shyya
|Original Researches on the Entire Suf Way|. See Ms. Alger B. N., 2026, II. 214b
227b. A commentary on this poem was written by Ahmad Zarrq al-Barns, entitled
Sharh urf:at Ibn al-Bann as-Saraqust |Comment on the Poem oI Ibn al-Bann as-
Saraqust|. See Ms. Alger B. N., 2069 and 2295.
98. M. Casiri: Bibliotheca Arabico-Hispana Escurialensis, 2 vols., Madrid, 17601770.
H. P.-J. Renaud himselI says about Casiri`s errors: Les parties proprement litteraires
et historiques de luvre de Casiri sont peut-etre acceptables. On ne saurait en dire
autant des deux grandes sections qui ont trait a la medecine et aux sciences exactes.
Lucien Leclerc sen etait defa plaint tout au long de son Histoire de la medecine
arabe, apres avoir fait le voyage de lEscurial et constate par lui-meme les meprises
et les omissions de Casiri. Suter, malgre la prudence fustihee quil montre vis-a-vis des
assertions de cet ecrivain, na pas pu toufours les contrler et les reproduit souvent
dans son ouvrage. H. P.-J. Renaud: Additions et corrections a Suter 'Die Mathema-
tiker und Astronomen der Araber, Isis 18 (19321933), 167.
99. A. Gannn (no date): Dhikrayt mashhr rifl al-Maghrib |Memories oI the
Eamous Men oI Morocco|, Eirst edition (assembling Iorty biographies oI Moroccan
scientists), Tetouan, Manshrt mahad Mawly al-Hasan li l-abhth; second edition
(increased by ten biographies oI scientists), Beirut, Maktabat al-madrasa, Dr al-kitb
al-lubnn.
100. M. al-Es: Ibn al-Bann al-adad al-Marrkush |Ibn al-Bann the Arithmetician
Irom Marrakech|, Sahfat mahad ad-dirsat al-islmiyya, Madrid, 6, (12) (1958),
19.
101. Op. cit., note 37.
102. Edited by A. S. Saidan: Kitb al-maqlt f l-hisb li Ibn al-Bann |Ibn al-Bann`s
Book on the Epistles oI Calculation|, Amman, 1984.
103. Op. cit., note 37.
104. M. Souissi evaluates Ibn al-Bann`s handbook in these terms: On serait tente
dafhrmer que, che: Ibn al-Bann, larithmetique et certaines questions dalgebre ont
pris leur forme dehnitive, celle sous laquelle nous les rencontrons de nos fours. Nous
ne pouvons, pourtant, soutenir que tout, dans ces questions, est apport propre dIbn
al-Bann. Ce dernier, en effet, nous presente, en cette hn du XIIIe siecle et debut du
XIJe, une somme des connaissances arabes en arithmetique et en algebre. See op. cit.,
note 37, 3334.
105. A. Gannn: 1alqh al-afkr f l-amal bi rushm al-ghubr li Ibn al-Yasmn |The
Eecondation oI Spirits Ior the Use oI the Eigures oI Dust, oI Ibn al-Ysamn|, Mafallat
al-bahth al-ilm, Bagdad, 1 (1964), 181190.
106. B. Hims: Urf:at Ibn al-Ysamn f l-fabr wa l-muqbala |Ibn al-Ysamn`s
Poem on algebra|, Second International Symposium on the History oI Arabic Sciences.
Abstracts, Aleppo, I.H.A.S., 1979, 7678.
107. J. A. Sanchez-Perez: Compendio de Algebra de Abenbeder, edition, Spanish trans-
lation and mathematical analysis, Madrid, Apirica, 1916.
108. E. B. Plooij: Euclids Conception of Ratio and his Dehnition of Proportional Mag-
nitudes as Critici:ed by Arabian Commentators, Rotterdam, 1950.
109. J. A. Sanchez-Perez: Particion de herencias entre los musulmanes del rito Malequi,
con transcription anotada de dos manuscritos alfamiados, Madrid, 1914.
110. W. Ahrens: Die magischen Quadrate al-Bns, Der Islam 12 (1922), 157177.
B. Carra De Vaux: Une solution arabe du probleme des carres magiques. Revue
dHistoire des Sciences, 1 (1948), 206212. H. Hermelink: Die ltesten magischen
Quadrate hherer Ordnung und ihre Bildungsweise, Sudhoffs Archiv 42 (1958), 199
217.
111. These are our present 'Arabic number symbols except Ior the Iour and the fve,
which are slightly modifed. They were called in al-Andalus and in the Maghrib hurf
al-ghubr |letters or fgures oI dust|, and sometimes rushm al-ghubr |signs oI dust|.
112. Copies oI al-Khwrizm`s Algebra circulated in al-Andalus and were used Ior
the twelIth-century Iatin translations oI the book made by Gerard oI Cremona and
Robert oI Chester. See M. Steinschneider: Die Europaischen berset:ungen aus dem
Arabischen bis Mitte des 17. Jahrhunderts, Vienna, 19041905. Reprint, Graz, 1956,
24, 72. The contents oI al-Khwrizm`s Algebra are partially present in the hand-
books oI the Maghrib which have reached us. In his classifcation oI sciences, Ibn
Khaldn says about algebra: 'the hrst one who wrote about this art is Ab Abdallah
al-Khwri:m. See op. cit., note 22, vol. 2, 899. No copy oI al-Khwrizm`s Algebra
has been Iound in the Maghrib.
113. M. al-Es: Urjzat Ibn al-Ysamn I l-judhr |Ibn al-Ysamn`s Poem on the
Radicals|, Mafallat Rislat al-Maghrib, 1 (1942).
114. M. Souissi: Un mathematicien tuniso-andalou, al-Qalasd, in: Actas del II Col-
loquio Hispano-Tunecino de Estudios Historicos, Madrid, 1973, 147169; M. Souissi:
Rislat dhawt al-asm li Ab l-Hasan . . . al-Qalasd, |al-Qalasd`s Ietter on the
Binomial Numbers|, in Dawr al-ulm as-sahha f tanmiyyat al-buldn an-nmiya,
Dirst multaq Ali al-Qalasd, Beja (Tunisia), 2830 Mai 1976. Tunis, Ministere
des aIIaires culturelles, 1978, 161190.
115. A. at-Tz: al-arqm al-maghribiyya arqm arabiyya asla |The Moroccan Num-
ber Symbols are Authentic Arabic Number Symbols|, Mafallat al-Lisn al-arab, 2
(1965), 37; M. as-Sarrj: at-Tbi al-arab I l-arqm ar-riydiyya |The Arabic Infu-
ence on the Mathematical Number Symbols|, Mafallat al-Lisn al-arab, 3 (1965),
7064; Ab Eris: Dall fadd ala urbat al-arqm al-mustamala f l-Maghrib al-
arab |A New ProoI Concerning the Arabic Nature oI the Number Symbols Used in the
Arabic Maghrib| Mafallat al-Lisn al-arab, 10, 1 (1973), 231233; M. H. al-Yasn:
al-Arqm al-arabiyya f hlih wa tirhlih |Arabic Number Symbols, their Condition
and Circulation|, fq arabiyya, 12 (1980), 4249.
116. M. al-Uqayl: Silk farid al-yawqt f l-hisb wa l-farid wa l-mawqt |Pearl
Thread oI Sapphires on Arithmetic, Inheritance, and Timekeeping|, lithography, Eez,
1901; A. Sakrj: Irshd al-mutaallimn an-ns f sifat ashkl al-qalam al-fs |Guide
Ior EorgetIul Students on the Eorms oI Eez Writing Signs| (Commentary on the urf:a
oI A. al-Es), lithography (without date).
117. Ch. Pellat (trans.): Irshd al-mutaallim an-ns f sifat ashkl al-qalam al-fs de
Sakrf, Alger, 1917; G. S. Colin: De lòrigine grecque des 'chiIIres de Ees et de nos
'chiIIres arabes, Journal Asiatique, 222 (1933), 193215; J. A. Sanchez Perez: Sobre
las ciIras rmes, Andalus, 3 (1935), 97125; M. E. Viala: Le mecanisme du partage des
successions en Droit musulman, suivi de 1expose des signes de Fe:, Alger, 1917; A.
Guiraud: Jurisprudence et procedure musulmane, Casablanca, 1925, 109110.
118. J. C. MartzloII: Histoire des mathematiques chinoises, Paris, Masson, 1988, 31
32.
119. Op. cit., note 22, 898899.
120. Op. cit., note 107.
121. Op. cit., note 107, 2473.
122. M. Ievey: 1he Algebra of Ab Kmil, Kitb f l-fabr wa 1-muqbala, in a Com-
mentary by Mordecha Fin:i, Hebrew Text, Translation and Commentary with Special
ReIerence to the Arabic Text, Madison, Milwaukee and Iondon, University oI Wis-
consin Press, 1966.
123. Op. cit., note 22, 896.
124. Op. cit., note 56, 6465, 67, 68, sqq.
125. Op. cit., note 56, 69.
126. Op. cit., note 22, 896.
127. M. Souissi: Un texte dIbn al-Bann sur les nombres parfaits, abondants, deh-
cients et amiables, International Congress oI Mathematical Sciences, Karachi, 1420
July 1975; Arabic version in Bulletin de lUniversite de 1unis, 13 (1976), 193209.
128. M. Aballagh and A. Djebbar: Hayt wa muallaft Ibn al-Bann (maa nuss
ghayr manshra) |Ibn al-Bann`s IiIe and Works (with unpublished texts)|, Rabat,
Publications de la Eaculte des Iettres et des Sciences Humaines, 2001.
129. A. Djebbar: Enseignement et Recherche mathematiques dans le Maghreb des
XIIIeXIJe siecles. Paris, Publications Mathematiques dÒrsay, 1980, 81102, 55
112; hereaIter, the title oI this publication will be quoted as E. R. M.
130. A. Djebbar: Lanalyse combinatoire au Maghreb. lexemple dIbn Munim (Xll
e
XIII
e
siecles), Paris, Publications Mathematiques dÒrsay, 1985, 85101.
131. External infuences are, however, suggested by a remark oI al-Brn, on the met-
rics in Sanskrit poetry. See al-Brn: 1ahqq m li l-Hind, Beirut, lam al-kutub,
1983, 104. E. C. Sachau: Alberunis India, New Delhi, S. Chand & Co., 1888, vol. 1,
147.
132. Op. cit., note 129, 99112.
133. A. Djebbar: Materiaux pour lètude des pratiques combinatoires dans le Maghreb
medieval. To appear.
134. Op. cit., note 129, 640.
135. Op. cit., note 32, 162165; H. P.-J. Renaud: Sur un passage dÌbn Khaldn relatiI
a l`histoire des mathematiques, Hesperis 31 (1944), 3547.
136. A. Djebbar: Quelques aspects de lalgebre dans la tradition mathematique arabe
de lOccident musulman, Eirst Maghribin ConIerence oI Algiers on the History oI
Mathematics Arabic, 13 December 1986. Published in the Proceedings oI the ConIer-
ence, Algiers, Maison du Iivre, 1988, 99123. Eor the edition oI 1alqh al-afkr oI Ibn
al-Ysamn, see T. Zemouli: Muallaft Ibn al-Ysamn ar-riydiyya |Mathematical
Writings oI Ibn al-Ysamn|, M.Sc. thesis in History oI Mathematics, Algiers, E. N. S.,
1993.
137. Op. cit., note 129, 4154.
138. Op. cit., note 136, 106107, 111; M. Zerrouki: Ab l-Qsim al-Qurash, Haytuh
wa muallaftuh ar-riydiyya |Ab l-Qsim al-Qurash, His IiIe and His Mathemati-
cal Writings|, Cahier du Seminaire Ibn al-Haytham 5 (1995), Algiers, E. N. S., 1019.
139. A. S. Saidan: 1rkh ilm al-fabr f l-lam al-arab, II, Kuwait, 1985, 398613; A.
Djebbar: Mathematiques et Mathematiciens du Maghreb medieval (IX
e
XJI
e
siecles),
These de Doctorat, Nantes, 1990.
140. Op. cit., note 22, vol. 2, 898.
141. J. Sesiano: Le Liber Mahamalet, un traite mathematique latin compose au XII
e
siecle en Espagne, Premier Colloque Maghrebin sur l`Histoire des Mathematiques
Arabes, Algiers, 13 December 1986; in Proceedings oI the ConIerence, Algiers,
Maison du Iivre, 1988, 6998.
142. H. I. I. Busard: Iàlgebre au moyen-ge: le 'Iiber mensurationum dÀb Bakr,
Journal des savants, AprilJune 1968, 65124.
143. A. Bar Hiyya: Libre de Geometria, trans. J. Millas Vallicrosa, Barcelona, 1931.
144. A. Djebbar: Procedes algebriques et geometrie metrique dans lEspagne medie-
vale, a travers un manuel du X
e
siecle, Paris, University oI Paris-Sud, Prepublications
Mathematiques dÒrsay. In press.
145. Y. Atik: Lepitre dalgebre de Sinn Ibn al-Fath, Second Maghribin ConIerence
oI History oI Arabic Mathematics, Tunis, 13 December 1988; in: Proceedings oI the
ConIerence, Tunis, Tunis University II.S.E.E.C.-A.T.S.M., 1990, 519.
146. A. Djebbar and R. Rashed: Luvre algebrique dal-Khayym, Aleppo, Institute
Ior the History oI Arabic Sciences, 1981.
147. Op. cit., note 22, vol. 2, 899.
148. There are also Iatin texts translated Irom Arabic which have not yet been analyzed
and which could be connected with this tradition. See: op. cit., note 142, 71.
149. Ibn Al-Majd: Hw l-lubb wa sharh talkhs aml al-hisb |Collection oI the
Essence oI, and Commentary on, the Abridged Book on the Operations oI Calculation|,
Ms. Iondon, British Museum 7469, II. 144a245a.
150. J. Sesiano: La version latine medievale de lAlgebre dAb Kmil, in: M. Eolk-
erts and J. P. Hogendijk (ed.): Jestigia Mathematica, Studies in Medieval and Early
Modern Mathematics in Honour of H. L. L. Busard, Editions Rodopi, Amsterdam-
Atlanta, 1993, 315452.
151. See Ibn QunIudh al-Qasantn (d. 810/1407), Mabdi as-slikn f sharh urf:at
Ibn al-Ysamn |Principles Ior Those Who Undertake the Explanation oI Ibn al-
Ysamn`s Poem| and al-Qalasd (m. 892/1486), 1uhfat an-nshin al urf:at Ibn
al-Ysamn |The Ornament oI the Young People on the Poem oI Ibn al-Ysamn|.
152. Op. cit., note 136.
153. S. Ahmad and R. Rashed: As-Samawal, al-Bhir f l-Jabr, Damas, Matbaat
Jmiat Dimashq, 1972.
154. M. Aballagh: Raf al-hifb an wufh aml al-hisb li Ibn al-Bann al-Marrkush
|The Raising oI the Veil on the Eorms oI the Processes oI Calculation oI Ibn al-Bann
al-Marrkush|, Ees, Eaculte des Iettres et des Sciences Humaines, 1994.
155. Y. Guergour: al-Aml ar-riydiyya li Ibn Qunfudh |Mathematical Writings oI Ibn
QunIudh|, M.Sc. thesis in History oI Mathematics, Algiers, E. N. S., 1990.
156. T. Zemouli: Muallaft Ibn al-Ysamn ar-riydiyya |Mathematical Writ-
ings oI Ibn al-Ysamn|, M.Sc. thesis in History oI Mathematics, Algiers, E. N. S.,
1993.
157. A. Harbili: Lenseignement des mathematiques a 1lemcen au XIJ
e
siecle a trav-
ers le commentaire dal-Uqbn (m. 811/408), Magister in History oI Mathematics,
Algiers, E. N. S., 1997.
158. M. Zoubeidi: Le Kitb al-bayn wa t-tadhkr dal-Hassr, M.Sc. thesis in History
oI Mathematics, Algiers, E. N. S. In preparation.
159. K. Kouidri: Tarqat al-khataayn f t-taqld ar-riyd al-arab |Method oI Ealse
Position in the Arabic Mathematical Tradition|, M.Sc. thesis in History oI Mathemat-
ics, Algiers, E. N. S., 1999.
160. M. Abdelkader-Khaddaoui: al-Khawri:miyyt al-hisbiyyaf t-taqldar-riyd al-
maghrib |Arithmetic Algorithms in the Maghribi Mathematical Tradition|, M.Sc. thesis
in History oI Mathematics, Algiers, E. N. S., 2000.
161. E. Iaabid: Arithmetique et Algebre dheritage selon lIslam, deux exemples.
1raite dal-Hubb (X
e
XI
e
s.) et pratique actuelle au Maroc, Memoire de Matrise,
Montreal, Quebec University, 1990.
162. E. Iaabid: Lenseignement mathematique dans le Maghreb extreme a travers
lexemple des partages successoraux, These de Doctorat, Rabat, Universite Mohamed
V. In preparation.
163. M. Zerrouki: Ilm al-frid f l-Maghrib al-wast min khill sharh Mukhtasar al-
Hf li l-Uqbn |The Science oI Inheritances in the Medieval Maghrib Through the
Commentary oI al-Uqbn on the Mukhtasar oI al-HI|, M.Sc. thesis in History oI
Mathematics, Algiers, E. N. S., 2000.
164. A. Djebbar: Le traitement des fractions dans la tradition mathematique arabe du
Maghreb; in Paul Benoit, Karine Chemla et Jim Ritter: Histoire de fractions, fractions
dhistoire. Basel-Boston-Berlin, Birkhuser Verlag, 1992. Chapter XII, 223245.
165. M. Aballagh and A. Djebbar: Decouverte dùn ecrit mathematique dàl-Hassr
(XIIth c.), le Iivre I du Kmil, Historia Mathematica, 14 (1987), 147158.
166. J.G. Iemoine: Ies anciens procedes de calcul sur les doigts en Orient et en Occi-
dent (Note additionnelle sur le comput digital, texte inedit dÌbn Bundd), Revue des
Etudes Islamiques, Paris, 6 (1932), 158; J. S. Colin: Un texte inedit dÌbn Bundd sur
la dactylonomie, Revue des Etudes Islamiques, Paris, 6 (1932), 5960.
167. Y. Guergour: Sur les differentes numerations utilisees au Maghreb a lepoque
ottomane, XXth International Congress oI History oI Sciences, Symposium 'Science,
Technology, and Industry in the Ottoman World, Iiege (Belgium), 2026 July 1997,
Turnout, Brepols Publishers, 2000, 6774.
168. Op. cit., note 164.
169. A. Djebbar: Kitb al-usl wa l-muqaddimt f l-fabr wa l-muqbala |The Book oI
the Bases and the Preliminaries oI Algebra|, in: Op. cit., note 139, vol. 2.
170. A. Djebbar: La 1heorie des nombres dans le Fiqh al-hisb dIbn Munim, 5th
International Symposium on the History oI Arabic Sciences, Granada, 30 March4
April 1992; to be published.
171. A. Djebbar: Les nombres hgures dans la tradition mathematique de lAndalousie
et du Maghreb, in: Op. cit., note 139, vol. 2.
172. See Ibn Munim: Fiqh al-hisb, Ms. Rabat B. G., 416 Q, 320.
173. B. Vitrac: Euclide, Les Elements, Paris, Presses Universitaires de Erance, vol. 2,
1994, 247493.
174. W. Kutsch: Kitb al-Mudkhal il ilm al-adad, tarfamat 1hbit Ibn Qurra |The
Introduction to the Science oI Number, the Translation oI Thbit Ibn Qurra|, Beirut,
Catholic Press, 1958. Eor a Erench translation, see J. Bertier: Nicomaque de Gerase,
Introduction Arithmetique, Translation, notes and index, Paris, Vrin, 1978.
175. E. Woepcke: Notice sur une theorie ajoutee par Thbit Ben Korrah a làrithmetique
speculative des Grecs, Journal Asiatique, 4th series, 20 (1852), 420429.
176. Op. cit., note 56, 158181.
177. Op. cit., note 22, 901905.
178. An exception is: T. Ievy: Fragment dIbn al-Samh sur le cylindre et sur ses sec-
tions planes. Edition and Erench translation; in: R. Rashed (ed.): Les Mathematiques
inhnitesimales du IX
e
au XI
e
siecle, vol. 1, Fondateurs et commentateurs, Iondon, al-
Eurqn Islamic Heritage Eoundation, 1995, 928973.
179. In astronomy it is the geometry oI the instruments, which one fnds, Ior example,
in the Safha oI az-Zarqlluh and in the Astrolabe oI al-Majrt. Eor the geometry oI
measurement, it is Ab Bakr`s Liber Mensurationum and the Br Hiyya`s Liber Emba-
dorum mentioned above.
180. A. Djebbar: Ab Bakr Ibn Bffa et les mathematiques de son temps; in: Etudes
philosophiques et sociologiques dediees a Jamal ed-Dine Alaoui, Publications de la
Eaculte des Iettres et des Sciences Humaines, Dr El Mahraz n 14, Eez, 1998, 526.
181. J. P. Hogendijk: Discovery oI an eleventh century Geometrical Compilation: The
Istikml oI YsuI al-Mutaman Ibn Hd, King oI Saragossa. Historia Mathematica 13
(1986), 4352.
182. J. P. Hogendijk: The Geometrical Part oI the Istikml oI YsuI al-Mutaman ibn
Hd (eleventh century). An Analytical Table oI Contents, Archives Internationales
dHistoire des sciences 127 (1991), vol. 41, 207281.
183. On Apollonius, see G. J. Toomer: Apollonius; in D. S. B., op. cit., vol. 1, 179193.
184. J. P. Hogendijk: Which version oI Menelaus` Spherics Was Used by al-Mutaman
ibn Hd in his Istikml`? Proceedings oI the international conIerence 'Mathematische
Probleme im Mittelalter, der lateinische und arabische Sprachbereich, WolIenbt-
tel (Germany), 1822 June 1990, M. Eolkerts (ed.), Wiesbaden, Harrassowitz Verlag,
1996, 1744.
185. J. P. Hogendijk: al-Mutaman`s Simplifed Iemmas Ior Solving 'Alhazen`s Prob-
lem; in J. Casulleras and J. Samso (ed.): From Baghdad to Barcelona, Studies in the
Islamic Exact Sciences in Honour of Prof. Juan Jernet, Barcelona, Anuari de Eilologia
XIX-Instituto 'Millas Vallicrosa, 1996, vol. 1, 59101.
186. J. P. Hogendijk: Ie roi geometre al-Mutaman Ibn Hd et son livre de la perIec-
tion (Kitb al-Istikml),` Eirst Maghribin ConIerence oI History oI Arabic Mathemat-
ics, Algiers, 13 December 1986; in Proceedings oI the ConIerence, Algiers, Maison
du Iivre, 1988, 5166.
187. J. P. Hogendijk: Eour constructions oI two mean proportionals between two given
lines in the Book oI PerIection oI al-Mutaman Ibn Hd, Journal for the History of
Arabic Science, 10 (199219931994), 1329.
188. A. Djebbar: Ia redaction de lÌstikml dàl-Mutaman(Xl
e
s.) par Ibn Sartq un
mathematicien des XIII
e
XIV
e
siecles, Historia Mathematica, 24 (1997), 185192.
189. See J. W. EngroII: 1he Arabic 1radition of Euclids Elements. Book J, Disser-
tation Cambridge/Mass., Harvard University, 1980; G. De Young: 1he Arithmetical
Books of Euclids Elements, Dissertation Cambridge/Mass., Harvard University, 1981;
S. Brentjes: Varianten einer Haggag-version von Buch II der Elemente; in M. Eolk-
erts and J. P. Hogendijk (eds): Jestigia Mathematica. Studies in Medieval and Early
Modern Mathematics in Honour of H. L. L. Busard, Amsterdam, 1993, 4767; A. Djeb-
bar: Quelques commentaires sur les versions arabes des Elements dÈuclide et sur leur
transmission a lÒccident musulman. Proceedings du Colloque International 'Mathema-
tische Probleme im Mittelalter, der lateinische und arabische Sprachbereich, WolIen-
bttel (Germany), 1822 June 1990, M. Eolkerts (ed.), Wiesbaden, Harrassowitz:
Verlag, 1996, 104111.
190. M. I. al-Khattb: Sharh al-Iksr |Commentary on al-Iksr|, Dawat al-haqq,
Rabat, 258 (1987), 7787; M. I. al-Khattb: Rislatn f ilm al-misha li Ibn ar-
Raqqm wa Ibn al-Bann |Two Epistles oI Ibn ar-Raqqm and Ibn al-Bann on the
Science oI Measuring|, Dawat al-haqq, Rabat, 256 (1986), 3947.
191. A. Djebbar: Materiaux pour letude de la tradition geometrique dal-Andalus et
du Maghreb. In preparation.
192. Op. cit., note 144.
193. Y. Guergour: Les Elements et les Donnees dEuclide dans le Kitb al-istikml
dal-Mutaman Ibn Hd. These de Doctorat, in preparation.
194. D. Iamrabet: Introduction a lhistoire des mathematiques maghrebines, Rabat, al-
MariI al-Jadda Press, 1994.
195. A. Djebbar: Quelques elements nouveaux sur làctivite mathematique arabe
dans le Maghreb Oriental (IX
e
XVI
e
s.), 2nd Maghribin ConIerence on the History oI
Arabic Mathematics, Tunis, 13 December 1988; in Proceedings of the Conference,
Tunis, University oI Tunis-I.S.E.E.C.-G.E.H.M.A.-A.T.S.M., 1990, 5373; A. Djeb-
bar: Les activites mathematiques dans les villes du Maghreb Central (IX
e
XJI
e
s.), 3rd
Maghribin ConIerence on the History oI Arabic Mathematics, Tipaza (Algeria), 24
December 1990; in Proceedings oI the ConIerence, Algiers, OIfce des Publications
Universitaires, 1998, 73115; A. Djebbar: Quelques aspects de lactivite mathema-
tique dans le Maghreb Extreme (IX
e
XJI
e
), 4th Maghribin ConIerence on the History
oI Arabic Mathematics, Eez, 24 December 1992. To be published in the proceedings
oI the ConIerence.
196. M. al-Mann: Nasht ad-dirst ar-riydiyya I Maghrib al-asr al-wast ar-rbi
|Activity oI the Mathematical Studies in Morocco oI the Eourth Period oI the Middle
Ages|, al-Manhil, Rabat, 33 (1985), 77115.
A. DJrii:r 330
Ix+roiuc+iox
In the opening paragraph oI his al-Kulliyyt f al-Tibb, Ibn Rushd alerts his
readers that the book shall be controversial: 'We shall aspire to those doctrines
which conIorm to the truth, even iI this conficts with the views oI the people oI
the art.
1
In this chapter we shall be concerned with one controversy which Ibn
Rushd initiated in his medical textbook, namely a biting critique oI al-Kind`s
pharmacological computus. In the frst part oI this chapter we shall very briefy
sketch out al-Kind`s theory and, in much greater detail, look at the doctrines
advocated by Ibn Rushd and the criticisms which he directs at al-Kind. In the
second section we shall proceed to the central question oI this inquiry: to which
(iI any) historical context (or contexts), does this departure belong?
1nr Cox+rovrrs.
I Al-Kindi
Ieon Gauthier published a book-length study oI the theory oI al-Kind, includ-
ing the Arabic text (Irom the one surviving manuscript) oI the latter`s F
marifat al-adwiya al-murakkaba.
2
As the title oI his book readily indicates,
Gauthier approached his subject with the aim oI demonstrating that the work
oI al-Kind, and, to a much lesser extent, the critique oI al-Kind given by Ibn
Rushd, anticipate advances in European science that are associated with two
nineteenth-century fgures, E. H. Weber and T. G. Eechner. Needless to say, this
sort oI orientation in research is unacceptable in our own day. Al-Kind`s trea-
tise raises several issues that are oI great interest to the historian oI medieval
science, especially the early stages oI scientifc culture in Islamic civilization,
and surely warrants an in-depth study oI its own. That, however, lies Iar beyond
the purview oI this chapter.
The pharmacological computus with which we are concerned can be
summarized in a Iew sentences.
3
Classical pharmacology had classifed drugs
into Iour degreesmore precisely, it had characterized some as temperate, and
12
Axo+nrr Axi:iusi:x Prvoi+
Iix Pusni`s Cri+iqur or :i-Kixii`s
IHPMCCLCICL CCMIl
Y. 1zvi Iangermann
Y. 1zvi I:xrrx:xx 332 Axo+nrr Axi:iusi:x Prvoi+ 333
classifed those that were not so into Iour degrees. These degrees signiIy the
drug`s potency in terms oI the Iour elemental qualities: heat, cold, wetness,
dryness. Al-Kind`s contribution lies in the claim that the intensity oI drugs
increases geometrically with the increase in degree, according to the 'double
ratio; (nisbat al-dif ); thus a drug in the frst degree is twice as intense as a
temperate one, one in the second degree is Iour times so, one in the third degree
is eight times so, and, fnally, one in the Iourth degree is sixteen times so.
This is the distinguishing Ieature oI al-Kind`s theory, and it is the chieI
target oI Ibn Rushd`s criticism. It is the viewpoint associated with al-Kind in
the relatively meager medieval literature which carries on their debate.
4
This
summary suIfces Ior our purpose, and we shall not go into any more details.
Nevertheless, beIore moving on to Ibn Rushd, we should like to call attention
to these Ieatures oI al-Kind`s treatise:
1. Originality. Al-Kind is quite assertive concerning the original nature oI his
investigations, saying that they constitute an advance in scientifc knowledge.
5
Although earlier authorities spoke about the Iour degrees as applied to simples, none
explained how this is to be done with regard to compounds. Yet the determination oI
these degrees Ior compounds is a more urgent matter than it is Ior simples. Al-Kind
writes, 'I see that the attainment oI knowledge concerning the strengths oI com-
pound medications would |yield| tremendous benefts. As we shall see, the Andalu-
sians too are quite cognizant oI the Iact that they are undertaking new research.
2. Nicomachus. Near the beginning oI his treatise, al-Kind describes fve diI-
Ierent mathematical series; Irom among these he chooses the geometric series
noted above, since it takes 'natural precedence over the rest.
6
This notion is taken
directly Irom the Introduction to Arithmetic oI Nicomachus oI Gerasa. At the end
oI part one, chapter XVII, Nicomachus lists fve species oI 'the greater; and, at
the beginning oI the next chapter, he asserts that 'the multiple is the species oI the
greater frst and most original by nature.
7
However, it seems that this point is stated
a bit more emphatically in al-Kind`s version oI that treatise, which, we must note,
displays throughout a very diIIerent text than the Greek; it is not just a question oI
variants, but oI a substantial reworking oI the entire treatise. Al-Kind`s version is
preserved, along with notes to his lectures on the book, only in the Hebrew transla-
tion oI Qalonymos ben Qalonymos.
8
At this particular juncture oI the Introduction
to Arithmetic, there is a long explanatory note by al-Kind, which discusses the fve
series in cosmological context, Iurnishing additional evidence oI his interest in the
topic.
9
A Iull account oI al-Kind`s pharmacology, especially his attempt at math-
ematization, would pay close attention to the various maniIestations oI his interest
in Pythagoreanism.
10
3. Galen. Al-Kind cites some Galenic texts, and reIers to some diIIerence oI opin-
ion he has with other scholars concerning their interpretation.
11
All in all, how-
ever, the correct interpretation oI Galen`s pronouncements seems to have been a
relatively minor issue (and even less important to Ibn Rushd, though not entirely
neglected by him; see below) in this particular debate.
II Ibn Rushd
Ibn Rushd criticizes al-Kind and his teachings towards the end oI the fIth
book oI al-Kulliyyt, Kitb al-Adwiya wa-l-Aghdhiya (1he Book of Drugs
and Foodstuffs), in a chapter (the last in the fIth books) entitled 'Al-Qawl f
Qawnn al-1arkb, 'The Chapter on the Rules |or: Iaws| oI Composition,
that is, oI making compound medicines. We shall examine this chapter, Iollow-
ing Ibn Rushd`s exposition oI the principles and laws that govern this branch
oI pharmacology, paying close attention in particular to the critique oI al-Kind
with which it culminates. As we shall see, though, the chapter was never really
fnished. Several variants, especially at the end, testiIy to Ibn Rushd`s shiIts oI
thought concerning the issues at hand. His mature views are expressed not in
al-Kulliyyt, but in a separate monograph devoted to one oI the most Iamous
compound drugs, the theriac.
12
That essay is tightly organized and system-
atically developed. Signifcantly, Ibn Rushd makes no mention there at all oI
pharmacological calculations. As it seems to me, he came to realize that math-
ematics was not his Iorte, and he concentrated instead on an analysis in terms
oI natural philosophy. Our concern here, then, is with a relatively youthIul
venture on the part oI the great philosopher, an excursion into a subdiscipline
that he later abandoned. Nonetheless, this episode is instructive concerning
developments in Andalusian science and as well Ior the intellectual biography
oI Ibn Rushd.
Ibn Rushd begins the chapter with an explanation oI the 'necessity
(darra) oI mixing simples to Iorm compoundsthat is to say, why the physi-
cian, despite his preIerence Ior simples, may have no choice but to prescribe a
compound. According to Ibn Rushd, three 'things (ashy) Iorce the physician
to have recourse to compounds: a simple oI the required properties (quw) is
not available; there is available a simple that possesses the requisite proper-
ties, but not in the right quantities; the simple that is otherwise appropriate has
an additional, undesired property that must be neutralized. Having established
these three general classes, Ibn Rushd moves on to a detailed exposition oI
each one, including many examples. In the course oI this exposition, he reIers
several times to the qnn which is relevant to the case at hand. That term
reIers most oIten to rules that must be kept in mind when preparing medica-
tions; however, these rules are based upon aIfrmative statements concerning
the behavior oI natural substances, the human body, or the interaction between
the twostatements that oIten take on the appearance oI the natural laws.
Although a satisIactory analysis oI the issue is beyond the purview oI this chap-
ter, I would suggest that the term qnnspeaking oI how the term was used,
not necessarily how it was defnedoriginally reIerred to manipulative, proce-
dural rules, i.e. general instructions as to howto act, but evolved to mean natural
law, i.e. general observations concerning the natural world, which Iormthe basis
or justifcation Ior the procedures accepted by the medical proIession.
One reason Ior compounding drugs would be to reduce a substance hot
in the third degree to one hot in the second. In order to achieve this purpose we
have two options; we can either compound it with one cold in the frst degree,
or with one hot in the frst degree. Ibn Rushd appears to realize that the second
option is more problematical.
13
We are compounding two substances oI iden-
tical qualities, but the heat, rather than intensiIying, reaches an equilibrium
between the degrees oI the two simples. Ibn Rushd Ieels that this calls Ior an
explanation:
This law (al-qnn), I mean that a drug having less heat will reduce the heat
oI |a drug possessing| more |heat| was confrmed (yusahhihuhu) by Galen.
He took as evidence |the case oI| hot and tepid water. When the hot is mixed
with the tepid, its heat is necessarily reduced.
14
Since a hot drug, when mixed with a substance hotter than it, will result
in a net reduction oI the heat, one may suppose that a patient with a hot dis-
ease, say a burning Iever, could be treated with a hot drug oI a lower degree
than his illness; theoretically, this should reduce the patient`s Iever. Experience
shows, however, that a hot drug will clearly harm the patient. Eor example, a
person suIIering Irom a burning Iever who is given honey to drink will suIIer
greatly. In order to justiIy the law oI combination given in the preceding para-
graph, Ibn Rushd frst resorts to a computation. Pepper is hotter than nard. (We
learn elsewhere in the Kulliyyt that the Iormer is hot in third degree, the latter
in the second.) This means that it has a greater proportion oI heat, 'as iI you
were to say (ka-annaka qulta), a ratio oI heat to cold that is 5:1, whereas the
corresponding ratio in nard is 2:1. 'Thus, necessarily, when we mix a dirham
oI pepper with a dirham oI nard, the ratio oI cold to hot in the combination is
greater than its ratio in pepper. II you contemplate this, it is evident.
15
The
computation would seem to be the simple addition oI the parts, yielding a ratio
oI 2:7, which is indeed greater than 1:5.
It is noteworthy that Ibn Rushd does not help the reader along by explain-
ing anything at all about the mathematical operations employed in pharma-
ceutical computations. Some details relating to his method are broached only
at the end oI the chapter, in the course oI the critique oI al-Kind, and in a
manner that is not wholly satisIactory. In particular, Ibn Rushd does not exploit
his computational apparatus in order to solve the vexing problem oI the way
simples oI identical qualities but diIIerent degrees react; he does not develop
any computus that would show precisely by how much a simple, say, hot in the
frst degree will reduce another hot in the second. As we shall see, Ibn Rushd
has only the most general idea how to approach the problemthough this does
not stop him Irom castigating rival approaches. A Jewish writer, Mordecai ben
Joshua, also known as Viola de Rhodes (Provence, 14th? century), attempted
to fll in the gaps in Ibn Rushd`s exposition. In Iact, though, the computus that
he devises owes much to the theory oI al-Kind.
16
The simple example oI nard and pepper clearly does not in and oI itselI
provide a suIfcient explanation, and in the Iollowing paragraph Ibn Rushd
elaborates. Although his tone is not deIensive, it seems that he is trying to
anticipate objections to the rules that he has just stated. Thus he begins by
explaining that the 'parts oI hot and cold spoken about above occupy an inter-
mediate state oI being; 'they do not exist in pure actuality (bi-l-hl al-mahd),
but rather in some sort oI intermediate state (bi-darb min al-tawassut) between
potentiality and actuality.
17
He adds that, with the aid oI the concept oI 'inter-
mediate state, we can understand how homogeneous bodies can nonetheless
be analyzed into unequal numbers oI parts oI hot and cold. Another, unstated
advantage oI this characterization oI the 'parts is that it avoids any conIusion
between the quantity oI the substances and the analysis oI their qualities. In the
example given above, one dirham oI pepper has a total oI six parts, a dirham
oI nard has three, and a dirham oI the mixture has ninebut all weigh one
dirham. The division into parts is purely an analytical device Ior determining
the relative strengths oI qualities. This matter too will be clarifed presently.
Because these 'parts may also be viewed as potentialities, they repre-
sent the readiness oI the drug to be acted upon by the human body, more spe-
cifcally by the body`s innate heat. This Iact solves the puzzle alluded to earlier:
why a simple with a relatively low degree oI heat will harm a patient suIIer-
ing Irom Iever, i.e., increase his Iever, even though it will reduce the heat oI a
hotter simple when the two are mixed together. Even though (to return to the
example given above) honey has a relatively low degree oI heat, it is all the
same 'hot, and this heat will, under the agency oI the innate heat, become
'fery substance (fawhar nr). We may conclude, then, that there is an essen-
tial diIIerence between the way the 'parts react: when mixed with another
simple, they recombine according to a simple arithmetical procedure (as yet
unspecifed), but when consumed by the human, they are activated (though this
technical term is not employed here) by the innate heat.
Ibn Rushd speaks again oI a qnn in connection with the third cause Ior
compounding medicines, i.e., in order to mask one oI the qualities inherent in
a simple. This case is oI particular interest, Ior two reasons. Eirst, Ibn Rushd
emphasizes the generality and importance oI the qnn. 'It is as we have said
a general law (qnn fmi).
18
Second, he gives credit to the Ban Zuhr Ior
clariIying and sharpening a qnn that is only implicit in the writings oI the
ancients. 'It is Iound in the compounds oI the ancients even though they did
not reIer to it explicitly (bi-l-qawl), nor did they call attention to it. The people
who |provided| the best inIormation about this are none other than the Ban
Zuhr. Upon my liIe, they have many merits (mahsin) in this art.
19
This is one
oI several major accomplishments in the feld oI medicine that stand to their
credit. Ibn Rushd may have hoped that his own work, together with that oI the
Ban Zuhr, could produce an alternative to Ibn Sn`s monumental al-Qnn f
al-Tibb. We shall return to this in the second part oI our chapter.
The rule describes two synergetic eIIects:
1. The combination oI simples may produce in the resultant compound proper-
ties that were unpredictable on the basis oI the constituent ingredients. Ibn Rushd
observes that this particular rule applies only to secondary and tertiary qualities:
'On the whole (bi-l-fumlati) this rule applies only (innam) to secondary and ter-
tiary powers.
20
2. The reaction oI the human body to medication cannot be Iully predicted on the
basis oI an analysis oI the ingredients that make up the compound. Ibn Rushd
writes: 'The actions oI drugs upon |human| bodies are only a relative matter (amr
idf). In truth, this is not something that is consequent upon the parts oI the drug
itselI. It may happen that a drug that is itselI less hot will be, relative to the human
body, hotter than a drug that itselI possesses greater heat.
21
The reaction oI the human body to a given drugbe it simple or com-
poundis a strong variable in pharmacology, and one which can be estab-
lished only post facto, through experience, experiment, and observation. This
Iact was common knowledge, and it underlies, as Iar as I can tell, the disinter-
est (iI not disdain) which some authorities, most notably, Ibn Sn, display con-
cerning pharmacological theory. It is almost paradoxical that Ibn Rushd, who
invested more eIIort than most in the theoretical analysis oI the action oI medi-
cation should also have so strongly emphasized the essentially unpredictable
nature oI compound medications. His discussion oI the issues leads eventually
to the sweeping conclusion that any given combination oI drugs may have an
eIIect on the human body that is stronger than we might have predicted on the
basis oI an analysis oI the qualities oI the individual ingredients.
To return to the application oI this rule in the chapter under discussion
here. According to Ibn Rushd, 'this qnn is an important (muhimm) qnn
in medicine. Indeed, iI a person would only bear it in mind as he ought to, he
would hardly ever administer a cure with a simple drug. By my liIe, it is Iound
|implicitly| in the compounds oI the ancients. . . .
22
Ibn Rushd is going against
the grain oI a medical tradition, stated by Maimonides among others, that the
physician should administer a compound only when there is no appropriate
simple; and we have seen that, in the opening paragraph to the chapter, Ibn
Rushd appears to confrm this tradition. This apparent inconsistency is one oI
several Ieatures which, taken together, indicate that this chapter displays work
in progress, rather than Ibn Rushd`s mature views on the issue.
The next qnn addresses the matter oI the quantity oI simples that are
to be used in the preparation oI compounds. This qnn as well has several
Iacets (awfuh). Eor the most part they are general guidelines, almost selI-
evident. Unlike the preceding, they do not Ioreshadow any laws oI nature; nor
do they involve any mathematization oI pharmacology. One rule states that one
employs a smaller quantity oI a strong drug and a larger quantity oI a weak
drugthe exact quantities, oI course, depend upon the desired result oI the
fnal product. Another rule concerns compounds in which one particular ele-
ment is overwhelmingly dominant, the other ingredients being ancillary to it.
The exact quantities oI all oI the ingredients will be determined by the ultimate
strength that the medication should have, or other considerations, such as the
distance which the drug must traverse through the body in order to reach the
diseased organ. Ibn Rushd then speaks specifcally oI laxatives, oIIering again
a trivial example: iI Iour drugs are required Ior a potion, then the physician
mixes a quarter dosage oI each and has the patient drink the combination. He
then concludes:
These are all the rules and laws (fam al-dustrt wa-l-qawnn) which
are employed with regard to quantity. However, since the most important
thing Ior the doctor to know when compounding medicines is the degree oI
the primary, secondary, and tertiary powers, iI this is possible, we must say
something about it. We state: when someone wishes to determine the rank
(martaba), as Iar as primary qualities are concerned, oI a compound drug,
the way to do this is to take into consideration the degrees oI the primary
drugs which are in it.
23
With that statement, Ibn Rushd takes up the computus. He begins with some
trivial examples which, however, are important Ior establishing the rules that
will be employed in more complex cases: 'When you know the simple law
(al-qnn al-bast), you will necessarily know the complex law (al-qnn al-
murakkab) by means oI an investigative procedure (bi-wafh al-nazar).
24
Einally Ibn Rushd is ready to reveal his most complex computus, that
which applies to drugs compounded Irom simples oI opposing qualities and
varying degrees. Erom the preceding, however, the procedure is quite clear:
it is all a matter oI simple arithmetic, provided that we deal with comparable
units oI power, not oI weight: 'I mean by their equivalence, not the equiva-
lence in weight, but rather the equivalence in power.
25
The cold ingredient will
reduce the hot by the amount oI its degree; iI it is cold in the frst degree, it will
reduce the hot drug by one degree. Drugs oI opposing qualities will reduce
one another in accordance with the number oI degrees assigned to them. This
is in Iact an extension oI the second class; it is exactly the same computus by
which it has already been determined that drugs oI opposing qualities but equal
degrees neutralize each other`s eIIect.
The computus oI degrees can be combined with the notion oI specifc
quantitative units, yielding a more complete computus. Eor example, two units
oI a drug cold in the frst degree, when mixed with one unit oI a drug hot in
the third degree, will reduce it by two degrees, not one (the sum oI the degree
computus alone). Since the result will be proportional not only to the diIIer-
ences in degrees (which are always integers, when speaking oI simples) but
their relative quantities (in terms oI specifc units oI each substance), the fnal
computation may lead to Iractional degrees. Ibn Rushd illustrates with one
case in which the fnal result is 'an amount in the middle between the third and
the second,
26
but he does not go into any Iurther detail. The treatise oI Viola
de Rhodes alluded to above works out these Iractions in great detail. BeIore
ending his discussion on this matter, Ibn Rushd hints that this very same prin-
ciple explains why a double dosage oI a drug can be Iatal.
27
We may readily
supply the Iull explanation: a larger quantity oI a drug Iunctions (more pre-
cisely, multiple unit dosages Iunction) as unit dosages oI a higher degree.
The last and most troublesome class oI drugs is comprised oI those that
are compounded oI simples oI identical qualities but varying degrees. Ibn
Rushd has no complete solution to this problem. He does however, have some
strong opinions about solutions that others have proposed, Indeed, it is by way
oI this particular issue that Ibn Rushd takes on al-Kind and his computus. The
problem, as Ibn Rushd intimates, is simple, iI not 'selI-evident (bayyin bi-naf-
sihi).
28
Since a cold drug will reduce a hot drug by the arithmetical diIIerence
oI their respective degreesIor example, a drug cold in the frst will reduce a
drug hot in the second by one degree, yielding a compound hot in the frstit
stands to reason that a drug oI the same quality, e.g. one hot in the frst degree,
will reduce the same drug (hot in the second) by less; but by how much less?
Ibn Rushd can as yet provide no answer. He argues Ior a scheme in which drugs
oI the opposite quality will reduce by the greatest amount, temperate drugs will
reduce by less, and drugs oI identical qualities by even less; he speaks oI the
'proportion (nisba) by which these types oI substance will reduce. One might
have expected that his earlier remarks on the number oI 'parts in certain sub-
stances, as well as the clear cut method discussed in the previous section would
have presented some opportunities. But he oIIers no clear Iormulae Ior com-
puting the resultant powers oI the drug.
29
He does, however, have some defnite things to say about computi that
other physicians have proposed. It is here that Ibn Rushd frst mentions al-
Kind by name, placing the blame Ior introducing conIusion and irrelevancies
into the feld oI medical pharmacology squarely upon his shoulders:
The person who frst plunged them into this matter is none other than the
man known as al-Kind. That is because this man wrote a treatise in which he
sought to speak about the rules (al-qawnn) by which the nature oI a com-
pound drug may be known. But he went astray (kharafa) in speaking about
the art oI numbers and the art oI music, in the matter oI someone who looks
into something only incidentally.
30
This man adduced in that book senseless
and hideous things.
31
AIter this blistering introduction, Ibn Rushd is ready to display some
specifc points oI criticism. Iet us summarize the arguments which he raises:
1. Al-Kind`s computus contradicts the reasoning which underlies the system by
which drugs are graded by degrees. Ibn Rushd does not here say who designed
this system, preIerring instead to reIer to the inventors anonymously, in the third
person; in a later passage, however, at least according to the reading oI one oI
the manuscripts, the system is associated with Galen. A drug that causes sensible
heatapparently, the most barely sensible heat is intendedis assigned to the frst
degree. Then, drugs 'whose distance Irom that |frst degree| is |the same as| that
|frst drug`s| distance Irom temperance are classifed in the second degree. 'With-
out any doubt, it is double the frstnot Iour times, as in the theory oI al-Kind.
Similarly, drugs whose distance Irom the second degree is the same as the distance
oI the second Irom the frst are assigned to the third degree, 'and likewise Ior the
Iourth.
2. What could ever have Iorced the medical proIession to adopt the 'double ratio
advocated by al-Kind? It is not entirely clear whether Ibn Rushd rejects al-Kind`s
assertion (drawn as we have seen Irom Nicomachus) that the 'double ratio is the
most natural, or whether he is simply unaware oI it.
3. According to the system oI al-Kind, third-degree drugs would be already Iatal;
that is, iI we Iollow Ibn Rushd`s interpretation, whereby any drug Iour times as
intense as the frst degree is Iatal. What, then, are we to do with Iourth degree
drugs?
4. The increments between the degrees are unequal. 'What greater disorder
(ikhtill) could happen to the art?
32
Ibn Rushd elaborates, somewhat unclearly:
al-Kind has deIeated his own purpose. (Here again Ibn Rushd uses the third
person, and it is possible, though in my view unlikely, that he means to say that
al-Kind has missed the mark with regard to his` |Galen`s| purpose.) He wished to
preserve the |uniIorm| increase in degrees, but by doing so geometrically, rather
than arithmetically, he has caused the increments between each degree to increase
successively. 'So iI there were a fIth degree, it would be thirty-two parts`. . . This
is all delusion and drivel, and a discussion oI things which have no reality.
33
The last sentence, in which Ibn Rushd reinIorces his earlier denunciation
oI al-Kind, is a ftting fnale to his tirade. However, the manuscript tradition
indicates that Ibn Rushd continued to refect upon the problem; he was espe-
cially concerned with the proper understanding oI one oI Galen`s pronounce-
ments. His later thoughts are contained in this paragraph, which is Iound in one
oI the Arabic manuscripts oI al-Kulliyyt (St. Petersburg 124) and reproduced
in some oI the Hebrew and Iatin translations:
The way by which al-Kind came to err is that he made the frst degree double
the temperate with regard to the quality, hot or cold. This then required him
to maintain the double ratio. It may be said to him that Galen intended by the
frst degree that which adds one part in ten to the temperate. In this way, iI
the double ratio were to be compounded Ior |each| increase in the degrees, it
would not entail that the drug which is in the Iourth degree would be sixteen
times the temperate. That it is arranged (tarattaba) in this way |is indicated
by| Galen`s saying:
34
'By the frst degree I mean that which is evident to the
sense, when frst there appears a change in the body. Had he meant by the
frst degree double the temperate, then the change which is evident in the
body |upon the application oI a substance in the frst degree| would not be
the frst change. Contemplate this; it is quite clear. However, when a well-
known person commits an error, people by habit Iollow him, since the power
oI imitation (quwwat al-taqld) overwhelms their natures.
35
A somewhat abbreviated and slightly diIIerent version oI this paragraph
is Iound in the Hebrew translation oI Shlomo ben Avraham: 'But what Galen
said about this, when he stated, I intend by the frst degree, that change in the
body which is frst apparent Ior a drug.` Had he meant by the frst degree double
the temperate, then the change which is evident in the body upon |the appli-
cation oI| the drug would not be the frst change. Contemplate this; it is quite
clear. However, it is customary Ior people to Iollow the claims oI a Iamous
person, since they accept his opinion.
36
Yet a third version oI this paragraph is
Iound in the Iatin translatio antiqua.
37
In his later refections upon the problem, Ibn Rushd has arrived at three
additional arguments:
(P1) A quasi-mathematical argument, whose meaning is not entirely clear to me,
which implies a contradiction between a decimal system planned by Galen and al-
Kind`s geometrical series.
(P2) A textual argument, based upon a direct quotation Irom Galen. According to
Ibn Rushd, since the frst degree signifes the minimum sensual excitation, it cannot
possibly be double the temperate (which does not stimulate the senses at all).
(P3) A third argument Ior the supposed popularity oI al-Kind (though it could
equally be directed against Galen), namely people`s ingrained habit oI accepting
without question the pronouncements oI 'Iamous people.
Argument P2 leads us to raise the question oI Ibn Rushd`s Iamiliar-
ity with the entire Galenic corpus; and, even more to the point, it Iorces us
to wonder, as Gauthier did, whether Ibn Rushd read al-Kind`s treatise in its
entirety. Eor al-Kind cites two Galenic texts, neither oI which is reIerred
to by Ibn Rushd: 1arkb al-Adwiya (On Compound Drugs), in ten chapters,
better known in Arabic as Kitb al-Maymir, and the book on drugs writ-
ten Ior Andromachus, apparently a reIerence to the work known in Arabic as
Kitb al-1iryq il Bsn or Kitb al-1iryq il Fsun.
38
In addition, al-Kind
cites these sources as prooI that Galen had unambiguously rejected the very
computus advocated by Ibn Rushd: 'Moreover, Galen has already reIuted the
school which maintains that the strength (quwwa) oI the Iourth degree is Iour
times the frst, and the third three |times the frst|. . ..
39
Ibn Rushd, on the one
hand, gives no indication that he knows oI these passages, nor, in particular,
that they were cited by al-Kindand it seems almost inconceivable that he
could simply ignore them, iI he wishes to make a convincing critique. Yet, on
the other hand, he quotes a diIIerent Galenic dictum (not cited by al-Kind, at
least not in this context) whose misinterpretation, so he claims, is the source
oI al-Kind`s error.
Argument P3 is interesting Ior other reasons. As we have already sug-
gestedand we shall return to this point in the second part oI this chapter
al-Kind`s treatise seems to have generated very little interest. How, then, are
we to understand Ibn Rushd`s complaint concerning the uncritical acceptance
which al-Kind`s theory enjoys? It may very well be the case that, despite the
lack oI written evidence, al-Kind`s theoryor, at least, the basic idea that
with each higher degree, the intensity oI the drug doublesdid have wide cur-
rency. Alternatively, the appeal to habit as a source oI error may have become a
cliche; at the very least, Ibn Rushd`s contemporary, Moses Maimonides, oIIers
the very same explanation Ior persisting errors in both a medical and a philo-
sophical context.
40
Ibn Rushd has now completed his critique oI al-Kind. The two fnal
paragraphs oI the chapter do help to fll us in on the context oI the critique,
and we shall look at them briefy. Ibn Rushd briefy discusses the concept oI
khssa ('specifc).
41
This leads him to note a point oI disagreement with Ibn
Sn. According to the latter, most oI the active properties oI the theriac are
khawss, specifcs whose cause cannot be determined (la yumkin talluhu).
Eor that reason, Ibn Sn did not want to make any changes at all in the tradi-
tional recipe oI Andromachus. 'As Ior me, continues Ibn Rushd, 'I see it ft
to add many drugs to the theriac.
42
These are substances which may or may
have not been known to the ancients, such as aloe (d), ambergis (anbar),
and cloves (qaranful). Ibn Sn holds the view that the properties oI the theriac
are occult, and, thereIore, the physician should not tamper with the Iormula oI
the ancients, whose eIfcacy has been proven empirically over the centuries.
By contrast, Ibn Rushd, though not denying a role Ior empiricism, Ieels that
the properties oI the theriac, or some oI them at least, are explicable in terms
oI the predictable reactions oI the constituent simples; thereIore, there is no
reason why a physician ought not to experiment with diIIerent Iormulations.
Ibn Rushd thus closes the chapter with a strong, personal statement oI conf-
dence in rational pharmacology and in the advance oI science.
The next Iew lines are meant to introduce the discussion oI some specifc
compounds and their properties. This is in line with the plan oI this section oI
al-Kulliyyt (Kitb al-Adwiya wa-l-Aghdhiya), which is outlined in the open-
ing paragraph. There Ibn Rushd writes: 'AIterwards we shall move on to the
rules Ior the composition |oI drugs; qawnn al-tarkb|. We shall mention the
best known compound drugs, and we shall make known their natures, in accor-
dance with what the rules necessitate in that matter.
43
When that is completed,
the purpose oI this section will have been achieved.
44
The list itselI is not
Iound in the edition oI the Arabic text, nor in any oI the Hebrew manuscripts
which I have examined.
The very last sentence oI this chapter is relevant, in a strange way, to the
critique oI al-Kind. JustiIying the particular organization which he has chosen
Ior his materials, Ibn Rushd writes, 'Just like the authority on music (shib al-
msq) will speak oI the well-known instruments only aIter he has presented
the elements oI melodies and the ways in which they are combined, in order that
training should take place in that way (li-yaqaa bi-dhlika al-irtiyd), so also
is the case here.
45
Ibn Rushd could not have Iorgotten that, only a Iew pages
earlier, he had criticized al-Kind Ior appealing to music in order to justiIy a
decision made in pharmacological theory. Music, however, is not at all a major
Iactor in al-Kind`s treatise, and Ibn Rushd`s citation oI that particular point
while ignoring many other highly relevant arguments adduced by al-Kind
seems to be purely polemical. Why Ibn Rushd should then have chosen to make
a similar analogy in his own treatise is truly puzzling.
1nr Cox+rx+s
Ibn Rushd`s participation in the controversy surrounding models employed by
the astronomers oI his day has been quite adequately contextualized.
46
There is,
frst oI all, evidence oI dissatisIaction with the Ptolemaic models on the part oI
a number oI prominent thinkers, Irom all parts oI the Islamic world: Thbit ibn
Qurra and Ibn al-Haytham in the east, and Ibn Bjja and Ibn TuIayl in the west.
The westerners, so it seems, reexamined celestial physics as part oI a wide-rang-
ing reappraisal oI Aristotelian natural philosophy. Within this Andalusian trend,
Ibn Rushd stands out Ior his resolute commitment to Aristotlenot to Aristo-
telianism, by which I mean the body oI doctrines, including various accretions
and commentaries, not all oI them mutually compatible, built up around Aris-
totle`s teachings, but rather to a purifed, strict reading oI Aristotle`s own writ-
ings. Thus, within the Iramework oI the astronomical controversy, Ibn Rushd`s
attitude diIIers Irom that oI Ibn Bjja, whose dynamics, intended or not, contain
a radical reworking oI Aristotle, and Irom that oI al-Bitrj, who, it has been
shown, preIerred a number oI teachings, especially the theory oI impetus and
the identifcation oI tashawwuq ('desire, 'yearning) as the driving Iorce oI the
cosmos, which he may have discovered in the writings oI the maverick thinker
Abu`l-Barakt.
47
In short, Ibn Rushd`s interest in the problem situates itselI into
both the wider Islamic and the narrower Maghribian-Andalusian contexts; in
addition, there is a distinct twist to his views, which can only be explained by
taking into account his own personal inclinations and interests.
What about medicine? To be sure, Galen`s writings evoked widespread
criticism throughout the Islamic cultural orbit. Ab Bakr al-Rz wrote a
book length reIutation oI Galen.
48
Ibn Sn`s al-Qnn is replete with critical
remarks, including the standard accusation that, in attempting to philosophize,
Galen intruded into a feld in which he did not belong.
49
Moses Maimonides, a
native oI Cordova and Ibn Rushd`s contemporary, devoted the twenty-fIth sec-
tion oI his Fusl Ms to a stinging attack on Galen`s teachings in both medi-
cine and philosophy.
50
Ibn Rushd as well oIIers numerous strictures on Galenic
doctrines.
51
His critical posture, thereIore, conIorms to a general trend among
Arabic medical writers.
This point, however, is only oI limited value in approaching our particu-
lar problem. Ibn Rushd`s critique is directed at al-Kind, not Galen. To be sure,
there is evidence that the proper interpretation oI some Galenic texts was one
Iactor in the dispute between Ibn Rushd and al-Kind. Moreover, it is note-
worthy that Abu`l-Al ibn Zuhr, patriarch oI the Ban Zuhr Iamily oI physi-
cians, deIended Galen against the strictures oI al-Rz.
52
As we have seen, Ibn
Rushd`s work in medicine, particularly in pharmacology, is deeply indebted to
that oI the Ban Zuhr. II Ibn Rushd`s attitude towards Aristotle can be charac-
terized as unswerving admiration, his attitude towards Galen was more ambiv-
alent; his attack upon al-Kind`s fts neither into the critique oI Galen nor into
his deIense.
Two general observations are in order at this juncture, built upon remarks
oI Ibn Rushd concerning scientifc activity going on in his day. Eirst, an impor-
tant distinction between the work undertaken in pharmacology and in astron-
omy must be drawn. Research into pharmacology is part oI a comprehensive
reinvestigation into the science, one whose aim is to attain new inIormation. By
contrast, the program in astronomy was undertaken in order to recover the true
astronomy oI Aristotle; but pharmacology has no supreme authority, nor any
lost knowledge oI the ancients. True, the ancients had by experience arrived at
some eIIective medical prescriptions, but even then, they did not elaborate the
underlying theory. This important task was assumed only by later generations;
and it is only one oI several areas in pharmacology where the moderns can
make new and important contributions. Second, however, there exists a signif-
cant similarity between the works in pharmacology and astronomy. Credit Ior
advancing the new research programs belongs to the Andalusians, in particular,
the medical clan oI the Ban Zuhr.
Iet us return to the specifc issue oI the critique oI al-Kind. As Iar as I
can tell Irom the extant sources, at the time that Ibn Rushd launched his attack,
al-Kind`s monograph seems to have engendered hardly any interest at all.
Indeed, one searches in vain Ior reIerences to al-Kind`s computus in the two
most important medical encyclopedias, Ibn Sn`s al-Qann and al-Majs`s al-
Kmil. Al-Kind is cited about a dozen times in the Qnn, invariably in con-
nection with this or that particular remedy, but his computus is not mentioned.
Indeed, Ibn Sn does seem to have held much stock in pharmacological com-
putations at all. In the introductory 'scientifc tract (al-maqla al-ilmiyya) to
book fve oI the Qnn, Ibn Sn surveys the various reasons why it may be
necessary to employ compound medications: to combine or intensiIy various
secondary qualities, to delay the digestion oI the medication until it reaches its
destination, and so Iorth. Within this list we fnd a simple computus, not that oI
al-Kind, described as Iollows:
Perhaps we may need a drug which heats Iour parts (af:), but we can only fnd
one which heats three parts, and another which heats fve parts. We then com-
bine them, in the hope that the result oI the combination will heat Iour parts.
53
OI the various reasons which Ibn Sn gives Ior compounding medicines, it is
only in connection with this computation that he adds the disclaimer, 'in the
hope that (rfn an) the desired result will be obtained. Al-Majs also limits
himselI to a single trivial example. II the physician requires a drug hot in the
second degree, but no such simple is available, he may compound two other
simples, one hot in the third degree and the other hot in the frst. The resulting
compound will be hot in the second degree.
54
Even among the Andalusians, who, as we shall see, were especially
interested in pharmacology, the impact oI al-Kind`s book seems to have been
minimal. In the sources available to me, I have Iound only one clear trace. Ibn
Buklrish (f. Saragossa, eleventh century), author oI al-Mustan, exhibits a
computus which is surely based upon al-Kind, though the latter is not named.
55
Signifcantly, the only tract which directly addresses the theory oI al-Kind was
written by the same Abu`l-Al ibn Zuhr. UnIortunately, his 1reatise in the
Explication of the Letter of Yaqb bin Ishq al-Kind on Compounding Drugs
(Maqla f bastihi li-rislati Yaqb ibn Ishq al-Kind f 1arkb al-Adwiya) is
not extant.
56
Why was there so little interest in al-Kind`s book? Eor all practical pur-
poses, recipes Ior compound drugs were passed on by tradition. Medical prac-
titioners surely experimented with diIIerent Iormulas. Ibn Rushd, as we have
seen, is especially confdent oI his ability to improve existing Iormulae by the
application oI rational laws. On the whole, though, Ibn Sn`s approach (criti-
cized by Ibn Rushd) was the norm. Traditional recipes were passed on Irom
generation to generation; even when practitioners experimented with old or
entirely new Iormulae, their work was not guided by mathematical rules, such
as those worked out in the treatise oI al-Kind. Pharmacological theory was
oI little practical use, both beIore and aIter Ibn Rushd. This accounts both Ior
the weak interest in al-Kind`s treatise in the period preceding Ibn Rushd, as
well as the relatively low level oI interest in the controversy, once Ibn Rushd
published his critique.
57
Ibn Rushd`s work in pharmacology built upon earlier
advances oI the Ban Zuhr and, presumably, he was Iamiliar with the exposi-
tion oI al-Kind`s theory by Abu`l-Al. However, this very limited tradition is
only a small part oI our story.
Eor a Iuller appreciation oI the historical context oI this episode, it must
be considered against the background oI two larger developments. The frst
oI these is the conspicuous concern which Andalusian and Maghribi scien-
tists evinced in pharmacology; one ought to include as well related disciplines
such as botany and botanical lexicography. This special interest has oIten been
connected with the giIt oI a beautiIul copy oI Dioscorides` Materia Medica
by the Byzantine emperor Constantine VII to the Umayyad caliph Abd al-
Rahmn III. This event, which took place in 948/9, certainly had an immediate
impact. An impressive group oI Andalusian savants set about to study the text,
and the Byzantine emperor was persuaded to send a monk in order to assist
the Andalusians in understanding the Greek text. Three centuries later Ibn al-
Baytr still Iound cause to write a commentary on Kitb Diysqrds. All in
all, the corpus oI pharmacological writings produced in Spain and Morocco is
impressive indeed.
58
It seems especially noteworthy that, among the Andalusians who dis-
played a keen interest in medical pharmacology, we fnd two savants who
also fgure prominently in the reexamination oI astronomy: Ibn Bjja and Ibn
TuIayl. The Iormer wrote (in addition to some notes to part oI De plantiis) a
Discourse on Galens Book of Simples and, in collaboration with Abu`l-Hasan
SuIyn, a book recording experiences with the drugs described by Ibn Wfd.
59
Moreover, none other than al-shaykh al-akbarthe noted mystic Muhy al-
Dn ibn ArabinIorms us about a debate or contest or sorts between Ibn
Bjja and Ibn Zuhr regarding medicinal plants.
60
As Ior Ibn TuIayl: Ibn Rushd records correspondence which he had with
Ibn TuIayl concerning the proper defnitions oI the terms 'Iood (ghidh) and
'drug (daw), and especially the distinction between the two. Their disagree-
ment seems to have hinged upon the interpretation oI a statement made by
Galen.
61
As we have already stressed several times, Ibn Rushd`s most impor-
tant collaboration was that carried out with the Ban Zuhr. In an earlier study,
we took note oI the lack oI communication between some oI the major play-
ers in the Andalusian revolt against Ptolemaic astronomy.
62
Although scholars
can discern a community oI interest between Ibn Bjja, Ibn TuIayl, Ibn Rushd,
and al-Bitrj, there is very little evidence oI any joint eIIort or even discussion
between these indivduals. By contrast, there is abundant evidence oI lively and
Iecund exchanges in pharmacology. This may indicate that, oI the two sciences,
astronomy and pharmacology, it was the latter which aroused more intense
interest on the part oI the Andalusians.
The second context, and the one which promises to be oI greater signif-
cance Ior the history oI thought in Islamic culture, is the attemptperhaps one
ought to call it a programoI the Andalusians to construct an alternative to
the syntheses which were produced in the East. The most important target oI
this enterprise, in connection with both medicine and philosophy, was the work
oI Ibn Sn. Abu`l-Al ibn Zuhr`s negative opinion oI Ibn Sn`s al-Qnn
has been recorded by a number oI authors: 'Previously, he |Ibn Zuhr| had not
encountered this book |al-Qnn|, but when he examined it he condemned it
and discarded it, and did not include it in his private library. He kept tearing oII
the margins oI its leaves which he used Ior writing prescriptions Ior his own
patients.
63
In the fnal paragraph oI al-Kulliyyt, which as the title implies, deals
with generalities, Ibn Rushd notes that his book must be supplemented by
another treatise oI the type known as kannsh (singular: kunnsh), containing
instructions Ior the treatment oI specifc diseases. He hopes to have the oppor-
tunity to compose a work oI this sort himselI. In the meantime, however, he
warmly recommends 'the work known as al-1aysr, which was written in our
own time by Ab Marwn ibn Zuhr. It was I who requested oI him |to write|
this book. I made a copy, and this was the cause oI its publication (khurf).
64
In connection with this passage, A.Z. Iskandar has observed: 'The Kulliyyt oI
ibn Rushd and the 1aysr oI ibn Zuhr were meant to serve as a complete work
on medicine, possibly to serve instead oI K. al-Qnn which did not appeal to
Abu`l-Al Zuhr b. Abd al-Malik b. Marwn b. Zuhr and, very likely, his son
and disciple Ibn Zuhr, also the latter`s pupil, ibn Rushd.
65
Whether the com-
bined eIIorts oI Ibn Zuhr and Ibn Rushd were undertaken in response to the
work oI Ibn Sn, al-Majs, or al-Rz, it does seem to be the case that they
were striving to produce an alternative to the great compilations emanating
Irom the eastern reaches oI Islam. Ibn Rushd`s critique oI al-Kind may be
viewed as one small contribution to this program.
The point which we would like to stress is that the purpose oI this venture
was not tahfut, the 'destruction oI noxious ideas, but rather the construc-
tion oI an alternative. Moreover, the eIIort to compose a comprehensive work
on medicine was part oI a Iar-reaching program to create alternatives in other
felds oI science and philosophy. Earlier writers, especially Ibn Bjja, had pre-
pared the groundwork. Ibn Bjja was certainly an insightIul and penetrating
thinker, but he does not seem to have been interested in or capable oI produc-
ing a systematic work. Two relatively short works oI Ibn TuIayl are generically
identicalthey even bear the same titleas some treatises oI Ibn Sn: Hayy
ibn Yaqzn and al-Urf:a f al-Tibb.
66
These may, then, be viewed as Andalu-
sian alternatives to Avicennian writings. The central tasks oI this enterprise
were taken up by Ibn Rushd. We have already noted his collaboration with Ab
Marwn Ibn Zuhr in the feld oI medicine. Perhaps his series oI commentar-
ies to Aristotle were intended to serve as the Andalusian response to Ibn Sn`s
al-Shif. In philosophy, however, Ibn Rushd, had no collaborators; he did it all
himselI.
67
This leads us to our fnal observation. Although the broad program oI
pharmacological research as well as many oI the specifc statements made by
Ibn Rushd do fnd their place in the contexts which we have suggested above,
they do not exhaust the story. Ibn Rushd`s personality certainly stands out in
his bold confdence in the power oI rational inquiry. This, however, is just one
aspect oI the strong sense oI dissatisIaction with the work oI his colleagues
or most oI them I should say, excluding in particular the Ban Zuhrwhich
inIorms and inIects Ibn Rushd`s medical writings. Even aIter all oI the contexts
have been listed and described, one must take into account Ibn Rushd` strong
individuality, his personality, and his idiosyncrasies, iI one wishes to obtain a
satisIactory understanding oI the work oI this great thinker.
No+rs
1. S. Shaybn and U. al-Tlib, eds., Al-Kulliyyt f al-Tibb li-Ibn Rushd (Cairo, 1989),
p. 19. All oI my reIerences are to this edition; in my view, it supersedes the edition pre-
pared by J. M. Eorneas Besteiro and C. Alvarez de Morales, 2 vols. (Madrid, 1987).
2. I. Gauthier, Antecedents Greco-Arabes de la Psychophysique (Beirut, 1938).
3. I have borrowed the term computus Irom Iatin astronomy, where it reIers to a system
Ior carrying out certain computations, most notably Ior reckoning the date oI Easter.
4. The relatively low level oI interest in this issue will be discussed in the fnal section
oI this chapter. One writer who does express interest in the debate is Joshua Iorki, in his
(as yet unedited) Gerem ha-Maalot (see, e.g., Mss. Paris, BN heb 1143, II. 146ab; St
Petersburg, RNI, Heb I 334, II. 90b91a). He notes that the 'later Christian physicians,
such as Arnald de Villnova and Bernard de Gordon have decided in Iavor oI al-Kind,
and so shall he. Arnald`s Aphorismi de gradibus has been edited by Michael McVaugh
(Arnaldi de Jillanova Medica Omnia, vol. 2, 1975).
5. See Gauthier, pp. 12 oI the Arabic treatise.
6. Ibid., 5.
7. M. I. DÒoge, E. E. Robbins, and I. C. Karpinski, Nicomachus of Gerasa, Introduc-
tion to Arithmetic (Iondon, 1926), pp. 213214.
8. I discuss the Hebrew manuscripts, giving as well preliminary evidence Ior the nature
oI al-Kind`s version, in 'Studies in Medieval Hebrew Pythagoreanism: Translations
and Notes to Nicomachos; Arithmological Texts, to appear in Micrologus. See also
M. Steinschneider, Die hebraeischen Ueberset:ungen des Mittelalters (Berlin, 1892),
517519, and S. Brentjes, 'Untersuchungen zum Nicomachus Arabus, Centaurus
30 (1987), 212239. Note that the version oI Thbit ibn Qurra (published by Wilhem
Kutsch, Beirut, 1958), while in general Iollowing the Greek quite closely, also diIIers
signifcantly in a number oI places; these are not noted by Kutsch, despite the inclusion
oI occasional Greek terms in his notes to the text.
9. See, e.g., ms. Paris, BN heb 1029, II. 10b11a.
10. Eurther hints at al-Kind`s Iamiliarity with doctrines ascribed to Pythagoras are
evident in some oI the newly discovered manuscripts oI his Risla f l-qawl f l-nafs,
discussed by Charles Genequand, 'Platonism and Hermetism in al-Kind`s F al-Nafs,
Zeitschrift fur Geschichte der Arabisch-Islamischen Wissenschaften 4 (1987/88), 118,
esp. pp. 58; but see also p. 11 n. 40.
11. Gauthier, pp. 2425 oI the Arabic text.
12. Ibn Rushd`s Kitb al-1iryq was published by G. C. Anawati and S. Zayed in Rasil
Ibn Rushd al-Tibbiyya (Cairo, 1987), pp. 389422. The Hebrew translations are dis-
cussed by Steinschneider, Hebraeischen Ueberset:ungen, p. 676, to which should now
be added Ms. St Petersburg, Academy oI Sciences C67. A Spanish translation by Maria
Concepcion Vazquez de Benito is included in the volume Averroes. Obra Medica, pub-
lished as part oI VIII Centenario Averroes (Seville and Malaga, 1998), pp. 261277. I
am currently preparing an annotated English translation.
13. The problematics emerge later on (see below, p. 9II.); but Ibn Rushd oIIers no solu-
tion.
14. al-Kulliyyt, p. 304.
15. Ibid., 305.
16. I have studied the copy oI his treatise Iound in Ms Moscow, Ginzburg 462, II. 135b-
139b; I hope to discuss it in detail in a Iuture publication. See M. Steinschneider, Die
hebraischen Ueberset:ungen des Mittelaters (Berlin,1892), p. 799, no. 215.
17. al-Kulliyt, p. 305.
18. Ibid., 308.
19. Ibid., 308.
20. Ibid., 308. These terms are distinguished as Iollows (al-Kulliyyt, pp. 218219).
The Iour elemental qualities (hot, cold, wet, dry) are called primary. Secondary prop-
erties are quasi-elemental eIIects, e.g., hardening or soItening, which are induced
throughout the entire body. Tertiary properties are limited to one particular organ, Ior
example, a diuretic.
21. Ibid., 307.
22. Ibid., 308.
23. Ibid., 309.
24. Ibid., 309. In this context, bast and murakkab can only reIer to the simplicity or
complexity oI the computation.
25. Ibid., 310.
26. Ibid., 310.
27. Ibid., 311.
28. Ibid., 311.
29. Here too Viola de Rhodes attempts to work out the details according to the system
oI Ibn Rushd.
30. Iiterally: according to what happens to the person who looks into the thing with an
incidental investigation (al fihati m yarid li-man yanzur f l-shay al-nazara alldh
bi-l-arad).
32. Ibid., 313.
33. Ibid., 313.
34. The Hebrew translations (see Iollowing notes) exhibit: 'That which Galen said
shows this |i.e., that this is the correct interpretation|. So also the Iatin Antiqua trans-
latio, as translated by Gauthier, p. 97. Both oI these versions (which may not be inde-
pendent oI one another) clearly read tara instead oI tarattaba.
35. al-Kulliyyt, p. 313 n.1794. The St Petersburg manuscript is described by the editors
on p. 14, and, like the Madrid manuscript, it exhibits Ibn Rushd`s own later version oI
the book. This passage is Iound in the anonymous Hebrew translation, mss. Munich 29,
I. 179b, and St Petersburg, Academy B288, I. 157a.
36. Ms. St Petersburg, Academy C81, I. 165a. Concerning the two Hebrew translations,
see Steinschneider, Hebraeischen Ueberset:ungen, pp. 671675. The activity oI Shlo-
mon ben Avraham can now be dated to the early decades oI the thirteenth century; see
Y. T. Iangermann, 'Some New Medical Manuscripts Irom St Petersburg, Iorthcoming
in Korot.
37. It is translated by Gauthier, pp. 9697.
38. This work was the subject oI the unpublished doctoral dissertation oI Iutz Richter-
Bernburg, Eine arabische Jersion der peudogalenischen Schrift De 1heriaca ad Piso-
nem (Gttingen, 1969).
39. Gauthier, Arabic text, p. 24.
40. Guide of the Perplexed, part I, chapter 31 (trans. S. Pines, Chicago, 1963, vol. 1,
p.67), and al-Maqla f l-Rabw (1reatise on Asthma), soon to be available in an edition
and translation by Gerrit Bos.
41. There is special section on the khssa (or khssiyya) in al-Kulliyyt, 233II. It is
odd that Ibn Rushd does not reIer to it here. CI. Y. T. Iangermann, 'Gersonides on
the Magnet and the Heat oI the Sun, in G. Ereudenthal (ed.), Studies on Gersonides,
a Fourteenth Century Jewish Philosopher-Scientist (Ieiden,1992), 267284, esp. pp.
273274 concerning Ibn Rushd`s claim that all physical properties can be explained by
the proportionalities oI the Iour qualities, in eIIect denying the possibility oI any appeal
to specifc, occult properties.
43. Clearly by this statement Ibn Rushd intends the properties that are rationally learned
by the systematic application oI the qawnn.
46. A. I. Sabra, 'The Andalusian Revolt against Ptolemaic Astronomy: Averroes and
al-Bitrj, in E. Mendelsohn (ed.), 1ransformation and 1radition in the Sciences (Cam-
bridge, 1984), 133153; rprt., A. I. Sabra, Optics, Astronomy, and Logic. Studies in
Arabic Science and Philosophy (Variorum, 1994). CI. Y. T. Iangermann, 'Arabic Cos-
mology, Early Science and Medicine 2 (1997), 185213, esp. pp. 202206.
47. See now the important study oI J. Samso, 'On al-Bitrj and the Haya Tradition in
al-Andalus, in idem, Islamic Astronomy and Medieval Spain (Variorum, 1994), essay
XII. Note that Ibn Bjja wrote a short treatise F mhiyyat al-shawq al-tabiyy (On
natural desire and its quiddity), published by Jaml al-Dn al-Alaw, Rasil Falsa-
hyya li-Ab Bakr bin Bffa (Beirut, 1983), pp.97102. The treatise reported by Ibn Ab
Usaybia, Uyn al-Anb f Tabaqt al-Atibb, ed. N. Rid (Beirut, n.d.) (hereaIter:
IAU), p. 516, has tashawwuq instead oI shawq in the title. See Iurther S. Pines, 'Ia
Dynamique dÌbn Bjja, Melanges Alexandre Koyre (1964), pp. 442468; rprt., 1he
Collected Works of Shlomo Pines, vol. 2 (Jerusalem and Ieiden, 1986), pp. 440466.
Eor a very recent and thorough discussion oI the history oI the problem, see A. I. Sabra,
'Confguring the Universe: Aporetic, Problem Solving, and Kinematic Modeling as
Themes oI Arabic Astronomy, Perspectives on Science 6 (1998), 288330.
48. Kitb al-Shukuk ala Jalinus, ed. M. Mohaghegh (Tehran, 1993); cI. S. Pines, 'Razi
Critique de Galien, Actes du JIIIe Congres International dHistoire des Sciences
(Paris, 1953), pp. 480487; rprt., Collected Works, vol. 2, pp. 256263.
49. See, e.g, the end oI the very frst fasl (ed. Blq, vol. 1, p. 5): 'When Galen tried to
construct a prooI Ior the frst division, he did not wish to attempt that as a physician, but
rather insoIar as he wished to be a philosopher speaking about natural science.
50. The philosophical critique (but not the medical critique) was published by J. Schacht
and M. MeyerhoI, 'Maimonides against Galen, on Philosophy and Cosmogony, Bul-
letin of the Faculty of Arts, vol. 5, part 1 (Cairo, 1939). Some medical critiques are dis-
cussed by Y. T. Iangermann, 'Maimonides on the Synochous Eever, Israel Oriental
Studies 13 (1993), pp. 175198, esp. pp. 186188.
51. J. C. Brgel, 'Averroes contra Galenum`, Nachrichten der Akademie der Wissen-
schaften in Gottingen, Philologisch-Historische Klasse, no. 9 (1967), pp. 263340.
52. His Kitb Hall Shukk al-R: al Kutub Jlns is listed by IAU, p. 519. Mohagh-
ghegh, al-Shukk, p. 112n40, reports the existence oI a manuscript in Mashhad.
53. al-Qnn f l-Tibb, ed. Blq, vol. 3, p. 309.
54. al-Kmil min al-Sina l-tibbiyya, second section (amal), book ten, chapter ten.
I do not have access to the Blq edition, and consulted instead the ms. Bethesda,
National Iibrary oI Medicine WZ 225 A398k 1736. The same trivial example (but
substituting cold Ior hot) is reproduced by Ibn Rushd in the course oI his detailed and
extensive analysis oI compound medications; see al-Kulliyyt, p. 304. Although both
authorities employ the same simple arithmetical rule, there are some signifcant diIIer-
ences between their examples.
55. This section oI al-Mustan is available in M. al-Khattabi, Pharmacopee et regimes
alimentaires dans oeuvre (sic') des auteurs hispano-musulmans (Beirut,1990), pp.
312314. I have recently identifed a partial copy oI al-Mustan, Arabic in Hebrew let-
ters, in ms. Vatican Ebr. 530/1.
Y. 1zvi I:xrrx:xx 372
56. It is listed by IAU, p. 519.
57. See M. McVaugh, 'Quantifed Medical Theory and Practice at Eourteenth-Cen-
tury Montpellier, Bulletin of the History of Medicine 43 (1969), 397413, esp. p. 405:
'The course oI medical practice is, Ior him, still determined by the results oI the estab-
lished tradition. And in this he is Iollowed by all later writers at Montpellier, none oI
whom shows any signs oI using mathematical theory as other than a descriptive tool.
McVaugh`s remarks about Bernard de Gordon and the other physicians oI Montpellier
would, on the Iace oI it, seem to apply to most writers in the Arabic tradition as well.
58. M. MeyerhoI, 'Esquisse d`histoire de la Pharmacologie chez les Musulmans
dÈspagne, al-Andalus 3 (1935), 141; M. al-Khattabi, Al-Aghdhiya wa-l-adwiya inda
muallif l-gharb al-islm. madkhal wa-nuss |Erench title: Pharmacope et regimes
alimentaires dans oeuvre (') des auteurs hispano-musulmans| (Beirut, 1990).
59. Ibn Bjja`s treatises are listed by IAU, 516517. M. Asin Palacios, 'Avempace
Botanico, Al-Andalus 5 (1940), 255299, includes a text and translation oI Ibn Bjja`s
Kalm f l-Nabt (Discourse on Plants). Ibn Bjja`s essay on the nlfar, a heliotropic
plant, was published by al-Alaw, Rasil, pp. 106107.
60. Asin Palacios, 'Avempace Botanico, pp. 257258.
61. al-Kulliyyt, p. 215, n. 16. CI. IAU
62. Iangermann, 'Arabic Cosmology, pp. 202206.
63. A. Z. Iskandar, A Catalogue of the Arabic Manuscripts on Medicine and Science in
the Wellcome Historical Medical Library (Iondon,1967), p. 36, translating Irom a work
Ibn Jumay.
65. Iskandar, Wellcome Catalogue, p. 37. Iskandar adds, however, that Ibn Rushd did
write a commentary on Ibn Sn`s Urfu:a.
66. Eor the work oI Ibn TuIayl see the excellent collection oI essays edited by Iawrence
Conrad, 1he World of Ibn Tufayl. Interdisciplinary Perspectives on Hayy bin Yaqzn
(Ieiden, 1996). Concerning Hayy see Iurther D. Gutas, 'Ibn TuIayl on Ibn Sn`s East-
ern Philosophy, Oriens 34 (1994), 222241. Portions oI the Urf:a were published by
M. Muhammad, 'Qira I Urjzat Ibn TuIayl I l-Tibb, Mafallat Mahad al-Makhttt
al-Arabiyya 30 (1986), 4782.
67. According to the indications oI some Hebrew manuscripts, Ibn Rushd may have col-
laborated with his son, Ab Muhammad Abd Allh, in writing a Iew short philosophi-
cal monographs; this point still requires clarifcation. See Steinschneider, Hebraeischen
Ueberset:ungen, pp. 198199. On Ibn Rushd`s acquaintances, see J. Puig Montada,
'Materials on Averroes`s Circle, Journal of Near Eastern Studies 51 (1992), 241260.
b oI Kashgar (c. 600), 126
Abacus, 78
Abd al-Baqqr, Ab, 286
Abd Allh Husayn Qusa ibn
Muhammad ibn Husayn al-HanaI
al-Tnis, 286287
Abd al-Malik ibn Marwn ibn Zuhr,
366 367
Abd al-Mash oI Winchester, 24, 36, 39
40, 42
Abd al-Rahmn III, 365
Abfad numbers, 4. See also alpha-
numerical systems
Abrahismus, 38
Abridged Book in Algebra (Ibn Badr),
315
Abu`l-Al ibn Zuhr, 363, 365367
Ab Bakr, 327328
Abu`l-Barakt, 363
Abu`l-Hasan SuIyn, 366
Abu`l-Hasan Al ibn Mam al-HanaI
(known as Karbso, or Karbsa), 287
Ab JaIar Muhammad
See al-Khwrizm
Abu`l-Jd, 178, 180181
Ab Kmil (d. 318/930), 179, 315316,
320, 322, 326, 328, 334
Ab Marwn ibn Zuhr, 367
Ab Mashar (d. 886), 129
Ab Miqra, 317
Ab Sad ibn Sahl. See Ibn Sahl
Abu`l-WaI, 177, 190, 192, 200, 202,
204206, 221222, 238, 248
Accuracy oI vision (sidq al-ruya), 5759
Ad Jitellionem Paralipomena (Kepler), 97
Adelard oI Bath, 10
Adud al-Dawla 177, 181, 192
Aetius, 63
al-AIghn, Jaml al-Dn, 317
Ahmad ibn Thabt (d. 1234), 180, 182,
239, 241, 251-253
Akbar, Mirza, 255
Akbar scrolls, 255
Alexandria, school oI, 129, 133
Algebra, 322, 326329. See also
Mathematics, Arabic
Algebra (Ab Kmil), 328
Algebra (al-Khwrizm), 238, 327328
Algorithms, arithmetical, 330
Alhazen`s problem, 89
Almagest, 2336, 42, 8591, 160, 191
Almond (quadrilateral), 257
Alphanumerical systems, 2627, 35,
42
Alphonso X (Castilian king), 310
Amal al-kafft (method oI scales), 321
al-mir, Abu`l Hasan (d. 382/992),
61, 111112, 134
al-mul, Muhammad Bah al-Dn
(15471622), 250
Ammonius school, 129
Analytica posteriora. See Posterior
Analytics (Aristotle)
al-Andalus, 11, 286, 310, 316, 323,
332
al-Andalus, Sid (d. 1068), 912, 315,
323
Andalusian reaction, 148157
Andalusian restoration oI Aristotle`s
cosmos, 151157
Andalusian Revolt, 8687, 148
Andalusian school, 286, 323
al-Antk, Ahmad (d. 987), 200, 202,
206, 211, 217, 223232
Ixirx
)NDEX 374 )NDEX 373
Antioch, 35
Apices (numeral notations on abacus
counters), 11
Apogee
oI moon, 272
motion oI, 289
oI sun, 30, 272
Apollonius, 89, 125, 178179,
188190, 193, 332
Appearances (m yazhar), 5657, 69
Arabic numerals, 516. See also Hindu-
Arabic numerals
Arches (architectural), 237, 240247
Archimedes, 42, 178181, 183, 188,
191193, 248, 252
Architecture, surIace areas and volumes
in Islamic
arches and, 237, 240247
calculators and, 237240, 261
characteristics oI architecture, 237
238
defnition oI Islamic architecture and,
235237
fnancial compensation and, 260261
historical perspective, 235, 260261
mosque design and, 236237
muqarnas and, 237, 240, 254260
qubbas and, 237, 239, 248253
terms, defnitions oI, 257258
under Umayyad rule, 240
vaults and, 237, 239247
Aristotle, 3, 5758, 6061, 96, 105, 121
124, 129131, 135144, 146, 148,
149158, 363364, 367
Arithmetic, 142, 329331. See also
Mathematics, Arabic
Arithmetic (al-Khwrizm), 3, 1011, 13
Arithmetic texts, 238. See also specihc
titles
Arithmetica (Nicomachus oI Gerasa),
323
saI Khn, 269
Astrolabes, 8, 38, 188189
Astrology between ninth and sixteenth
centuries, contents oI research on,
316318
Astronomy, Arabic
Arabic optics and, 87
contents oI research on, between ninth
and sixteenth centuries, 316318
cross-cultural transmission oI
Liber Mamonis and, 2336
overview, 23
Habtaq tables and, 299301
Indian astronomical handbooks and, 3
lunar tables and
lunar equations and, 294298
lunar motion values and, 289293
mean motion values and, 289293
solar equations and, 294298
in nineteenth century, 310
Nitynanda and, 269283
Pisan Tables and, 3642
planets and
angular velocity oI, 155
distances oI, 155
latitudes oI, 3031
longitudes oI, 272273
motion oI, 3031, 149
order oI, 152
revolutions oI, 146
rotations oI, 270271
speed oI, 155
pre-Islamic, Hellenistic view oI world
and, 86
Sanskrit astronomy and, 269
spherical, 145146
:f and, 269270
al-Asturlb, Abu`l-Qsim, 188
Averroes, 150, 152157
Avicenna. See Ibn Sn
al-Azzw, 318
Bacon, Roger, 95
Baghdad, Great Mosque at, 236
al-Baghdd, Abd al-Qhir, 4, 7
Bah al-Dn, Muhmmad al-Amul
(15471622), 250
Balance of Wisdom, 1he (al-Khzin),
182
al-Balkh, Ab Zayd (m. 332/934), 134
Ban Zuhr, 356, 363365
Br Hiyya, Abrahm, 327328
Basra, Great Mosque at, 236
al-Battn, 10, 316
al-Bayn wa l-tadhkr, 313, 319
)NDEX 374 )NDEX 373
Bent (munkasira) rays, 64, 66
Berlin manuscript, 3839
Berlin, Staatsbibliothek, lat. Iol. 307,
3637
Berthold oI Moosberg, 23
Bibliotheca Arabia Hispana, 314
Bipeds, 258
al-Brn, 7, 160, 177, 181, 189190,
317
al-Bitrj (d. 581/1185), 151, 157, 366
Boethius, 12
Book of Algebra (Ibn Khaldn), 315316
Book of the Construction of the Circle
Divided into Seven Equal Parts, 180
181
Book of the Levels in the Explanation of
the Measurements (Qd Ab Bakr),
239
Book on Measurements (al-Hanbal), 242
'Book oI Optics (Kitb al-Manzir), 90
Book on What the Artisan Needs to
Know of Geometric Constructions, 1he
(Abu `l-WaI), 238
Bordered squares
Abu `l-WaI`s construction oI, 204206
defned, 199
direct construction oI, 206223
description oI, 223232
geometrical structure oI, 203204
oI odd orders, 202203
Bossut, C., 310
Brhmasphutasiddhnta, 270272
al-Bn (d. 610/1204), 318, 320
Byid Dynasty, 25, 177
al-Bzjn, see Abu`l-WaI
Chalab, Mram, 298, 305
Calculators, architectural, 237240, 261
Calculators Riches (Ibn Thabt), 239
Callippus, 157
Castile (Spain), 29
Catalogue of the Sciences (Ihs
al-ulm) (al-Erb), 5556, 68
Celestial revolutions, 146. See also
Planets, motion oI
Clarke, C. Purdon, 254255
Clay-plastered muqarnas, 258259
'Codex Emilianus, 11
'Codex Vigilanus, 11
Colin, G. S., 321
Collapsed vaults, 254
Color, visual, 91, 93
Combinatorics, 324326
Commentarium Magnum, 121
Commentary on Book V oI Euclid`s
Elements (Ibn Mudh al-Jayyn),
315
Commentary on Euclids Elements
(Proclus), 122
Commentary on Nicomachoss Arithmetic
(al-Antk), 200
Complete Compass (al-Kh), 189, 192
Conics (Apollonius), 89, 189190, 332
Constantine VII, 365
Constructing the Astrolabe with Proofs
(Ab Sahl), 189
Correction of the Almagest (Jbir ibn
Afah), 23
Crystalline humor, 97101, 103, 105,
107
Cupolas. See Qubbas
Curved muqarnas, 259
Curves, 187188
Damascus, mosque at, 240
al-Dn, Ab Amr Uthmn (d. 1053),
4, 12
Data (Euclid), 185186, 190
De aspectibus (al-Kind), 59, 88
De Caelo (Aristotle), 121, 152155
De Ia Hire, 283
De placitis Hippocratis et Platonis
(Galen), 105
Decimal system, Indian positional, 330
Defected (munatifa) rays, 6364
Demonstrations, 6061
Demonstrative science
mathematics as, 142148
philosophy as, 135142
Depth perception, 104109
Descartes, Rene, 95
De speculis (Pseudo-Euclid), 63, 65
Determining the Jolume of a Paraboloid
(al-Kh), 252253
Dionysius Areopagita, 126
Diophantus, 190
)NDEX 376 )NDEX 377
Dioscorides, 365
Direct (mustaqma) rays, 67
Discourse on Galens Book of Simples
(Ibn Bjja), 366
Discourse on the Measurement of a
Paraboloid (Ibn al-Haytham), 182
Distance, perception oI, 106109
'Division oI the Intellectual Sciences
(Ibn Sn), 141
Domes See Qubbas
Dominicans oI Cologne, 23
Double elongation, 289290, 299
Dresden Almagest, 2336, 42
Drkpaksasrini (Kevalarma), 283
Dust reckoning (hisb al-ghubr), 910
Eastern Arabic Iorms oI numerals, 8,
1015, 29
Eastern zj, 286
8 (numeral), 13, 15
Eisagge (Nicomachus oI Gerasa), 133
Element, 257
Elements (Euclid), 25, 85, 89, 128, 189
190, 315, 320321, 328, 331333
Elements of Geometry (Iegendre), 314
Elliptical arches, 253
Enumeration oI sciences (ihs al-ulm),
139140
Epagge, 144
Epicycles, 152, 157, 276279, 283
Epistemology, Iundamentals oI, 154
Epistle about the Characteristics of
Hyperbolic and Parabolic Domes
(al-Sijz), 251
Epistle on Music, 133
Epitome (Ibn Rushd), 149
Essence of Arithmetic (al-mul), 250
Etymologiae (Isidor), 11
Euclid, 25, 58, 6163, 65, 67, 70, 85, 87
90, 128, 178, 185186, 190, 315, 320
321, 328, 331333
Eudoxus, 145, 157
Evenly-even squares, 199
Eacet oI a cell, 257
Falsifa, 138, 157
Falsafa tradition, 127, 133, 140, 151
al-Erb (d. 950), 5666, 6869, 88,
129, 135, 138140, 144, 159-160
al-Earghn, 34
al-Eris, Kaml al-Dn (d. ca. 1318),
59, 67, 103
Ear Maghrib, 318, 321
al-Es, A., 319, 321
Eerrarius, Matheus, 35
Eez drawing, 255256
Eigure indice (Indian symbols), 29
Figure toletane ('Toledan symbols), 29
Fihrist (al-Nadm), 4, 7
F kayfyat sanat fam al-asturlbt
(al-Sijz), 8
Einger reckoning (hisb al-aqd), 45,
330
Fiqh al-hisb (the Science oI Calculation)
(Ibn Munim), 324, 330331
F marifat al-adwiya al-murakkaba
(al-Kind), 351
Eirst equation (tadl awwal), 294
5 (numeral), 1213, 15
Eorm oI objects, 9394, 9699, 109
Eractions, 330331
Fundamenta tabularum, 40
Fusl Ms (Maimonides), 363
Galen (d. ca. 214), 60, 8788, 97, 104
106, 111, 113, 123, 126, 135, 178,
352353, 360361, 363
Galenic theory oI Iour degrees oI
intensity or potency, 132
Gangrma oI Ksmra, 269
Geminus oI Rhodes, 23
Geometrical demonstrations, 60
Geometry. See also Mathematics, Arabic
halI-rhombus, 257
Hellenistic
methods oI, 182188
problems Irom, 178182
heptagon, 182, 186, 188
paraboloid, 182, 252253, 332
rhombus, 206207
Transversal Theorem and, 190
Gerard oI Cremona, 42
Germinus oI Rhodes, 23
al-Ghd, Abu`l-Hasan, 181
al-Ghazl, school oI, 159
)NDEX 376 )NDEX 377
Ghubr numerals, 710
Ghunya (Qusa), 286287
Ghunyat al-tlib f taqwm al-kawkib
(Qusa), 286
Givens (Euclid), 185
Gnostic tradition, 132134
Gravity
centers oI, 182
oI semicircle, 192
Great Book of Music (Nicomachus oI
Gerasa), 128
Great Mosque at Baghdad, 236
Great Mosques at Basra, 236
Greek Almagest, 2426
Greek mathematics, 136137
Greek science. See Philosophical
tradition in Greek science
Gur-i Amir dome, 247
Habash, 189
Habtaq method, 287
Habtaq tables, 299301
HalI-rhombus, 257
Hall al-aqd wa-bayn al-rasd (Ibn Ab
l-rijl al-Qayrawn), 285
Harley manuscript, 3839, 41
Harmonice Mundi libri J (Kepler), 122
Harmony, 133134
Harrn, philosopher-scientists Irom,
126127
Hasan b. Muhammad (known as Qd
Hasan al-Makk), 286
al-Hsib (in Bukhara), 181
al-Hsib, Habash, 316
al-Hassr, 313, 319, 323, 329330
Hatt an-niqb an wufh mal al-hisb
(the Iowering oI the Veil on the
Various Operations oI Calculations)
(Ibn QunIudh al-Qasantn), 329
Hayy ibn Yaqzn (Ibn Sn), 367
Height determination, 63
Hellenism, 126127, 129, 132, 135, 145,
159, 190, 193
Hellenistic geometry
methods oI, 182188
problems Irom, 178182
Heptagon, 182, 186, 188
Hermann oI Carinthia, 40, 42
Herod, hall oI, 236
Hindu reckoning system (al-hisb
al-hind), 3, 812, 15
Hindu-Arabic numbers, eastern Iorms
oI, 29, 3842.
Hisb al-aqd (fnger reckoning), 45,
330
Hisb al-hind method
Hisb rm (Byzantine calculation), 321
Homocentric spheres, 152, 156
al-Hubb (10th c.), 329330
Hugo oI Santalla, 40, 42
Hurf a:-:imm (fgures oI the account
book), 321
al-HI (d. 1192), 330
al-Huwr (7th/13th c.), 319
Hyginus, 40
Hyperbolic qubbas, 251
Iamblichus (4th century disciple oI
Porphyry), 123, 125, 128
Ibn Abdn (d. aIter 976), 327328,
333
Ibn Ab Al al-Qusantn, Abu`l-Hasan
Al, 285
Ibn Ab Jarda, 65
Ibn Ab Mansr, Yahy, 49
Ibn Ab l-Rijl al-Qayrawn, 285
Ibn Afah, Jbir, 23
Ibn Ahmad, al-Khall (d. 797), 325
Ibn Arab, Muhy al-Dn, 366
Ibn Azzz al-Qusantn, Abu `l-Qsim
(d. 1354), 18, 285
Ibn Badr (13th c.), 315, 319320, 322,
326328
Ibn Bjja (d. 1138), 148, 151, 158, 316,
363, 366367
Ibn al-Bann (d. 1321), 285286, 313,
315, 317319, 323326, 328331
Ibn al-Baytr, 288, 366
Ibn Bishr, Sahl, 38
Ibn Ezra, Abraham (1089/92-aIter 1160),
3640, 42
Ibn al-Eath, Sinn (10th century), 62, 71,
111, 327
Ibn al-Him (f. 1205), 285
)NDEX 378 )NDEX 379
Ibn Hamza (16th c.), 318
Ibn al-Hanbal (d. 1564), 242
Ibn al-Haytham, 59, 63, 65, 67, 85, 88
110, 140, 142150, 157, 177, 180,
182, 186, 190191, 252253, 316,
332, 363. See also Optics (Ibn al-
Haytham)
Ibn Hayyn, Jbir, 132133
Ibn Hunayn, Ishq, 24
Ibn Irq, Ab Nasr (4th/10th c.), 177, 317
Ibn s, Ahmad, 58, 6061, 65, 67, 70,
77, 88
Ibn Ishq, 287
Ibn Ishq, Hunayn (d. 264/877), 59, 63
Ibn Ishq al Tamm al-Tunis, Abu`l-
Abbs (f. Tunis and Marrakesh ca.
11931222), 285
Ibn al-Kammd (f. Cordova 111617),
285, 288
Ibn Khaldn, 251, 311315, 322324,
327, 331332
Ibn Iabbn, Kshyr (2nd halI oI 10th
century), 34
Ibn al-Iayth, Abu`l-Jd, 178, 180181
Ibn Iq, Qust (d. 300/912), 5961, 65,
70, 140
Ibn al-Majd, Shihb al-Dn (d. 1447), 10
Ibn Mudh al-Jayyn (d. 460/1067),
315317
Ibn Muhammad al-Mahall al-ShI,
Husayn (d. 1756), 15
Ibn Munim, 315, 324325, 330331
Ibn al-Nadm, 4, 127
Ibn al-QiIt (d. 646/1248), 9, 315
Ibn QunIudh al-Qasantn, 329
Ibn Qurra, Thbit, 24, 110, 127, 180,
182, 239, 241, 251253, 331, 363
Ibn al-Raqqm, Muhammad (d.1315),
285
Ibn Rushd (d. 595/1198), 121, 149, 151,
157158, 316, 351362
Ibn Sahl, Ab Sad al-Al, 77, 89, 177
178, 181182, 186, 188192
Ibn as-Samh, 315
Ibn al-Sar, Abu`l-Euth (d. 548/1153),
148
Ibn al-Shtir (d. 1375), 286
Ibn Sn, 65, 113, 135, 140142, 159,
160, 356, 361364, 366367
Ibn Sinn, Ibrhm (d. 946), 110, 177,
182183, 188, 192, 332
Ibn Sirtq, Muhammad, 177
Ibn Tamm, Ab Sahl Dunas, 89, 12
Ibn Thabt, 127, 180, 182, 239, 241,
251253, 331, 363
Ibn Tibbon, Moses, 313
Ibn TuIayl (d. 1185), 151, 316, 363,
366367
Ibn Tln mosque (Cairo), 240
Ibn Umar al-SI Abd-al-Rahmn
(d. 986), 36, 38
Ibn Wfd, 366
Ibn Yahy al-Sl (d. 986), 45
Ibn al-Ysamn (d. 1204), 810, 12, 46,
318, 320321, 326327, 329330, 333
Ibn Ynus, Kamal al-Dn, 180
Ibn Ynus (d. 1009), 287
Ibn Zakariyy al-Gharnt (d. 809/
1406), 315
Ibn al-Zarqlluh (d. 1100), 285, 310
Ibn Zuhr, Abu`l-Al, 363, 365367
al-j, Adud-al-Dn (d. 1355), 159
Ikhtisr al-fabr wa l-muqbala (the
Summary oI Restoration and
Balancing) (Ibn Badr), 327328
Ikhwn al-SaI, 132134, 141
Ilm f manqib al-Islm, 134
Image-oriented theory oI vision, 96105
Indian astronomical handbooks, 3
Indirect (ghayr al-mustaqma) rays,
6667
Induction (istiqr), 144145
Intellectual development in North AIrica
and Muslim Spain
contexts, 363368
controversy
Ibn Rushd, 353362
al-Kind, 351353
overview, 351
Introduction to Arithmetic (Nichomachus
oI Gerasa), 125, 128, 352
Introduction to the Phenomena
(Geminus oI Rhodes), 23
Ishq, Ab, 151
Isidor, 11
istiqr, 144
)NDEX 378 )NDEX 379
Iudicia (Pseudo-Ptolemy), 39
al-Jadhr, 286
al-Jhiz (d. 255/868-869), 4, 58
Jmi al-mabdi wa l-ghyat f ilm
al-mqt (the Collection oI the
Principles and the Goals in the Science
oI the Determination oI Time)
(al-Murrakush), 310
al-Jayyn, see Ibn Mudh
Jerusalem, 236
John oI Seville, 34
Karbso or Karbsa (beIore 17491750),
287
Kabsa, 288
Kalm school, 138, 159
Kalpa, 270, 272
al-Kamil (al-Majs), 364
al-Karaj, Ab Bakr Muhammad
(d. 1023), 235, 242243, 249, 322,
327329
Karna (hypotenuse), 283
Kashf al-asrr an ilm hurf al-ghubr
(Disclosure oI the Secrets oI the
Science oI the Dust-Numerals)
(al-Qalasd), 9, 313
Kashf al-haqiq f hisb al-daraf wa-l-
daqiq (Ibn al-Majd), 10
al-Ksh, Ghiyth al-Dn Jamshd ibn
Masd, 239241, 243244, 246248,
253, 256261
Kendra (anomaly), 283
Kepler, Johannes, 90, 9597, 103, 122
Kevalarma, 283
Key of Arithmetic (al-Ksh), 239, 247
al-Khayym, Umar, 177, 188, 192
al-Khzin, Ab JaIar, 182, 190
al-Khayym (d. 1131), 327
al-Khwrizim, Muhammad ibn Ahmad
(ca. 980), 4
al-Khwrizim, Muhammad ibn Ms
(ca. 830), 3, 7, 911, 13, 16, 238,
287, 316, 321322, 327328, 334
al-Kind, Ab YsuI Yaqb ibn Ishq
(d. ca. 257/870), 5960, 62, 65, 88,
126132, 134135, 141, 351353,
358362, 364365, 367
Kitb al-adwr f tasyr al-anwr
(al-Baqqr), 286
Kitb al-bath, 132
Kitb al-bayn wa t-tadhkr (Book oI
the Demonstration and the
Recollection) (al-Hassr), 313, 329
Kitb al-burhn (al-Erb), 139140,
159
Kitb Disqrds (Ibn al-Baytr),
366
Kitb al-hisb al-hind bi-l-takht ('Book
on Reckoning with the Board), 7
Kitb ikhbr al-ulam bi-akhbr al-
hukam (The Book which InIorms the
Scholars on the IiIe oI the Wise) (Ibn
al-QiIt), 315
Kitb al-ikhtisr f l-fabr (Ibn al-Bann),
326
Kitb Ikhwn al-saf, 132
Kitb al-istikml, 331332
Kitb al-istiqs wa t-tafns f ilm al-
hisb (the Book oI the Investigation
and the Classifcation in Calculation)
(al-Hubb), 329330
Kitb al-kmil, 328, 331, 334
Kitb mafhlt qisiyy al-kura (The Book
oI the Unknown Sphere), 315, 317,
332
Kitb al-maymir, 361
Kitb al-muallimn (al-Jhiz), 4
Kitb tabaqt al-umam (Book oI the
Categories oI Nations) (al-Andalus),
315, 323
Kitb tabat al-adad (Book on the
Nature oI Numbers), 323
Kitb al-tarb wal-tadwr (Book oI
Rectangularity and Circularity)
(al-Jhiz), 58
Kitb al-1iryq il Bsn, 361
Kitb al-1iryq il Fisun, 361
Kitb al-usl wa l-muqaddamt f l-fabr
wa l-muqbala (the Book oI the Bases
and oI the Preliminaries in Restoration
and Balancing) (Ibn al-Bann), 326,
328329
KuIa, 236
al-Kh, Ab Sahl Wayjan ibn Rustam,
177193, 252253
)NDEX 380 )NDEX 381
al-Kulliyyt f l-Tibb (Ibn Rushd), 351,
354, 360, 362, 366367
Iatin translations oI Almagest, 25
Iatitudes oI planets, 3031
Ieonardo oI Pisa, 16
Ieonardo da Vinci, 96
Liber Embadorum (Br Hiyya), 327328
Liber Mahamalet, 327, 328
Liber Mamonis, 2336, 39, 42
Liber Mensurationum (Ab Bakr), 327
328
Liber trium iudicum, 40
Libros del Saber (M. Rico y Sinobas),
310
Iight
accidental, 93
essential, 93
least, 94
physics oI, 9095
primary, 93
properties oI, 9293
secondary, 93
Locating Hidden Waters (al-Karaj), 239
Iongitudes oI planets, sun, and moon,
288290
Iucca (11421145), 38
Iunar anomaly, 289, 297
Iunar apogee, 272
Iunar epicycle, 276279, 283
Iunar equations, 294298, 299301
Iunar longitude, 288290
Iunar motion values, 289293
Iunar tables
Habtaq tables and, 299301
lunar equations and, 294298
lunar motion values and, 289293
mean motion values and, 289293
solar equations and, 294298
Iunar variation, 297
Macrobius, theories oI, 35
Maftih al-ulm, 4
Maghrib, intellectual circles oI, 10, 310,
314, 317, 320321, 323, 326, 331
Maghribi manuscripts, 1112, 15
al-Maghrib, Muhy al-Dn (d.1283), 286
Magic squares, 199233
Mahrja Jayasimha 27, 269
Maimonides, Moses, 151, 158, 160, 357,
361, 363
al-Majrt, 10, 310, 316
al-Majs, 35, 364365, 367
al-Makk, Qd Hasan (f. 17th c.), 286
al-Mamn (ruled 813833 A.D.), 34
35, 127, 238, 333
Manzir (optics) tradition, 5657, 62.
See also Optics Arabic
al-Manzir wal-mary al-muhriqa
(Ibn s), 65
Maqlt al-arba (The Eour Epistles)
(Ibn al-Bann), 319
Maqld ilm al-haya (The Keys oI
Astronomy) (al-Brn), 317
al-Maqqar, 314
Mary (catoptrics) tradition, 5657, 63
al-Mridn, Muhammad Sibt (d. 1527),
10
Marka: (double elongation), 289290,
299
al-Masd, 4
Materia Medica (Dioscorides), 365
Mathematics, Arabic.
Andalusi mathematicians and, 316
axiomatic, 136137
as demonstrative science, 142148
5 (numeral), 1213, 15
gnostic tradition and, 132134
Mathematics, Arabic (cont.)
al-Kh and, 177193
in nineteenth century, 311315
in ninth through sixteenth centuries
contents oI research on, 318320
eighteenth-century research on, 309
310
nineteenth-century research on, 310
315
orientation oI research on (aIter 1980),
323324
results oI research on, 323324
topics studied in research on, 324334
twentieth-century research (beIore
1980) on, 315316
optics and, Arabic, 150
philosophy and, in medieval Islam
Andalusian reaction and, 148157
)NDEX 380 )NDEX 381
Aristotle and, 135142
Ibn al-Haytham and, 142148
overview, 121122
philosophical tradition, 122135
primacy oI doctrine and, 148157
theology and, 157160
quadrivium oI, 130, 133
role oI, Ibn al-Haytham`s view oI, 149
7 (numeral), 15
6 (numeral), 13, 15
in tenth century
methods oI Hellenistic geometry and,
182188
overview, 177178, 192193
problems Irom Hellenistic geometry
and, 178182
3 (numeral), 5, 7, 15
2 (numeral), 5, 7, 15
vision and, 91
Mausoleums, 247248
Maya the Asura, 272
Mean motion values, 289293
Measurement of a Circle (Archimedes),
191192
Megale Syntasis (Ptolemy), 30
Megali Xintaxis, 24
Menelaus, 42, 332
Meno, 127, 135136
Metaphysica, 154
Metaphysics (Aristotle), 137, 152, 154
156
Metaphysics or Eirst Philosophy, 121
122
Metaphysics (Ibn Sina), 141
Metaphysics and philosophy, 154
Meteorologica (Aristotle), 149
Michael (bishop oI Tarazona), 40
Minhf at-tlib f tadl al-kawkib
(The Guide oI the Student Ior the
Correction oI the Star Movements)
(Ibn al-Bann), 285286, 317
Mirtiyya (Related to Mirrors), 63
Misha (surveying) tradition, 5657, 63
al-Misht al-manziriyya, 62
Miskawayh, Ab Al (d. 1030), 134
Module oI the muqarnas, 257
Moon. See also Iunar tables
apogee oI, 272
epicycle oI, 276279, 283
equations oI, 294298
longitude oI, 288290
motion values oI, 289293
variation oI, 297
Mordecai ben Joshua, 355
Mosque design, 236237
Muhammad (the Prophet), 235236
Mukhtasar, (al-Khwrizm) 330, 334
Mukhtasar f l-fabr (Concise Work on
Algebra) (Ibn Badr), 320, 322
Muqaddima (Ibn Khaldn), 311313,
315, 322324, 327
Muqarnas (stalactite vaults), 237, 240,
254261
al-Muqni f l-hisb al-hind (al-Nasaw),
7
Mrstus, 76
al-Murrkush, 310
Muslim Occident, 310, 314, 322323,
331, 334
al-Mutaman, 315, 332
Mutayyan muqarnas, 258259
Mutazilite discussions, 95
Nafh at-tb (DiIIusion oI the PerIume)
(al-Maqqar), 314
al-Nasaw, 7, 25
Natural philosophy (al-ilm al-tab), 121
'Nature oI Sight, 90
Neoplatonism, 123125, 127129
Neopythagoreanism, 124, 127
Neopythagoreans, 12
Newton, Sir Isaac, 94, 193
Nicomachus oI Gerasa (frst century
C.E.), 123, 125, 128, 130, 132133,
162, 190, 323, 331, 352
Nine Chapters on Arithmetical
1echniques (Chinese work), 235
Nine numerals (al-hurf al-tisa), 4, 7,
10, 15
Nitynanda, 269283
Numbers, theory oI, 331. See also
Mathematics, Arabic; specifc numbers
Numbers of Fe:, 321
Objects, determining size and distance
oI, 6264
)NDEX 382 )NDEX 383
Odd squares, 199
Oddly-even squares, 199
Ogival arch, 240241
On Centers of Gravity (Archimedes),
181
On Centers of 1angent Circles (al-Kh),
187
On Clarihcation of Finding Distances
between an Observe and the Centers of
Mountain Heights (Risla f /dh
wifdn abd m bayn al-nzir wa
marki: amidat al-fibl), 62
On the Common Mathematical Science
(Iamblichus), 128
On the Conhguration of the World
(f Hayat al-lam) (Ibn al-Haytham),
143
On the Light of the Moon (f Daw al-
qamar) (Ibn al-Haytham), 143144
On Measuring Paraboloids (Thbit), 251
On the Sphere and Cylinder
(Archimedes), 252
1000s, 29
Opposite point, 283
Opticks (Newton), 94
'Optics According to the Method oI
Ptolemy, 90
Optics, Arabic. See also Vision
accuracy oI vision, 5759
applications oI, versatility and Iallibility
oI, 6264
Arabic astronomy and, 87
elements oI vision, 6466
historical perspective, 5557
justifcation oI, 5961
mathematics and, 150
mechanisms oI vision, 6466
mediums oI operation, 6669
modes oI operation, 6669
paradox oI, 8590
Optics, Arabic (cont.)
scientifc character oI, 140
sources oI, primary, 7071
variety oI demonstrations, 5961
veracity oI vision, 5759
Optics (Euclid), 58, 6263, 65, 8788,
90
Optics (Kitb al-Manzir)
(Ibn al-Haytham)
Alhazen`s problem and, 89
crisis oI the science oI vision and,
9192
depth perception and, 104109
elements oI, 109110
Iundamental concept oI, 109
image-oriented theory oI vision and,
96103
paradox oI Arabic optics and, 8590
as phenomenalist theory, 93
physics oI light and, 9095
Ptolemy and, 8586, 109
refection and, 91, 9495
reIraction and, 91, 9495, 101103
Optics (Ptolemy), 6768, 88, 90
Ordered Iorm, 96
Pksa argument, 277, 283
Papyrus PERE 789, 5, 12
Parabolic qubbas, 251252
Paraboloid, 182, 252253, 332
ParvumAlmagestum, 23
Pecham, John, 95
Perception
depth, 104109
oI distance, 106109
visual, 96103
Peripatetic tradition, 96, 127, 132
Persepolis, 236
Pharmacological computus, 351352
Philosophical demonstrations, 60
Philosophical tradition in Greek science
background inIormation, 122123
gnostic tradition and, 132134
al-Kind and, 127132
Platonic heritage, 124125
school oI al-Kind and, 134135
transmission oI, through science, 126
127
Philosophy and mathematics in medieval
Islam
Andalusian reaction and, 148157
Aristotle and, 135142
Ibn al-Haytham and, 142148
metaphysics and, 154
)NDEX 382 )NDEX 383
overview, 121122
philosophical tradition, 122135
philosophy as demonstrative science
and, 135142
primacy oI doctrine and, 148157
spherical astronomy and, 145146
theology and, 157160
Physics
Aristotelian, 159
oI light, 9095
Pisan Tables, 3642
Placita philosophorum (Aetius), 63
Plane mirror, refection Irom, 63
Planetary Hypotheses (Ptolemy), 31,
160
Planets
angular velocity oI, 155
distances oI, 155
latitudes oI, 3031
longitudes oI, 272273
motion oI, 3031, 149
order oI, 152
revolutions oI, 146
rotations oI, 270271
speed oI, 155
Planisphaerium (Ptolemy), 10
Planispheric astrolabes, 189
Planudes, Maximus, 8
Plato, 91, 112, 122, 125, 127, 129, 132
133, 135136, 138, 146, 151, 157
Platonism, 123125, 127132, 146
Plotinus, 125, 127128
Point-Iorms oI light and color, 94, 96
97
Pointed arches, 246247
Polynomials, 326327
Porphyry, 125
Posterior Analytics (Aristotle), 6061,
124, 130131, 135, 138141, 144,
151, 159
Prior Analytics, 144
Probe experiment, 99100
Problemata Physica (Aristotle), 5758,
72, 140
Proclus, 122-123, 125, 128, 144, 146
Properties oI light, 9293
Proportion (nisba), 358359
Psephophoria katIndous (Planudes), 8
Pseudo-Aristotle, 140
Pseudo-Boethius, 12
Pseudo-Euclid, 63, 65
Pseudo-Ptolemy, 39
Ptolemaic astronomy, 191, 316
Ptolemy (d. ca. 170), 10, 2324, 28, 30
31, 34, 42, 6768, 8591, 105106,
108112, 123, 135, 146148, 151
152, 156, 160, 191, 274, 276
Pythagoras, 124
Pythagoreanism, 124125, 128, 130,
133, 142, 352
al-Qalasd (d. 1486), 9, 15, 312
313, 319, 321, 323324
al-Ql, Ab Abd Allh Muhammad,
287
Qalonymos ben Qalonymos, 352
Qnn, 353357
al-Qnn f l-Tibb (Ibn Sn), 356, 363
364, 366
Qnn al-Masd (al-Brni), 160
Qarawiyyn educational institutions,
314
al-Qatrawn, 329
Qawm al-Dn ibn Zayn al-Dn. See
al-Shrzi, Zayn al-Dn
Qazwn, 236
Qubbas (cupolas or domes), 237, 239,
248253
al-Qh, Ab Sahl, see al-Kh
al-Qurash, Abu`l-Qsim, 18, 327
Qusa, 286287
al-Qushj, 160
Radiation, 61
Raf al-hifb (the Raising oI the Veil)
(Ibn al-Bann), 324, 326, 328331
Rasil Ikhwn al-Saf, 133134
Rashf ar-rudb min thughr aml
al-hisb (Sucking the Nectar Irom the
Mouths oI the Operations oI
Calculation) (al-Qatrawn), 329
Ratio, 184185, 187
al-Rz, Eakr-al-Dn, 159, 367
Reasoning, process oI, 6061
)NDEX 384 )NDEX 383
Recension of Euclids Elements (al-Ts),
314
Rectilinear propagation, 9495
Rectilinear vision, 91
Refection (iniks), optical, 63, 6768,
91, 9495
ReIraction (initf), optical, 68, 91, 94
95, 101103
Regalis dispositio, 3536, 39
Reversed (munakisa) rays, 6364, 66
Rhetorica ad Herennium (Milan), 35, 39
Rhomboid, 257
Rhombus, 206207, 257
Risla f l-add al-mutahbba (Ietter on
the Amicable Numbers), 331
Risla f l-taksr (Book on
Measurement) (Ibn Abdn), 327, 333
Risla f dhawt al-asm wa l-munfasilt
(Epistle on the Binomials and the
Apotomes) (al-Qalasd), 321
Robert oI Ketton, 40, 42
Romakasiddhnta (Roman Zf), 270272,
282
RooI oI a cell, 257
Round arches, 240
Rudimenta (al-Earghn), 34
Rudolph oI Bruges, 40, 42
Rm calculation, 330
al-Sb, 185
Sbia, 126127
Sabian sources, 127
SaIavid Iran architects, 260261
al-Saghn, Ahmad, 177, 180181, 189
190
al-Sajgw (d. aIter 1592), 15
al-Samawal (d. 1175), 188, 327, 329
Samsam al-Dawla 177
Sanjaq Dr al-Tnis, 286287
Sarvasiddhntarfa (Nitynanda), 269
283
Science oI aspects (ilm al-manzir),
56, 69
Science oI the balance (ilm al-m:n),
132
Science discrimination, 5960
Science oI harmony (sinat al-talf),
134
Science oI number (ilm al-adad), 134
Science oI optics (ilm al-manzir), 56
57. See also Optics, Arabic
Science oI rays (ilm al-shu t), 60
61
Sciences of Number (Ibn Khaldn), 311
312
Sbkht, Severus, 3, 126
Second equation (tadl thn), 295296
Sefer ha-Mispar (Ibn Ezra), 39
Sefer Yesira (Ibn Tamm), 8
Sensible explanation/illustration, 60
Sergius oI Reshayna, 126
7 (numeral), 15
Shh Jahn, 269, 271
al-Shann, 178
SharaI al-Dawla 177, 181, 192
Sharh Mukhtasar al-Hf (Comment on
al-HI`s abridged book), 330
Sharh at-1alkhs (al-Uqbn), 329
al-SharI, Muhammad, 286
Shaykh Abduh, 317
al-Shif (Ibn Sn), 142, 367
Shrz muqarnas, 259260
al-Shrz, Qutb al-Dn (d. 711/1311),
67, 89, 110111
al-Shrz, Zayn al-Dn, 259260
Shlomo ben Avraham, 360
Shu (Persian work), 67
Shukk al Batlamys (Ibn al-Haytham),
143, 157
Siddhntakaustubha, 283
Siddhntasindhu, 269271
Sidq, MustI, 177
Sifr (zero), 4, 20
al-Sijz, Ahmad (f. 970), 5, 8, 177, 180
181, 183, 189, 251
Simple muqarnas, 258
Simplicius, 163
6 (numeral), 13, 15
Socrates, 136
Solar apogee, 30, 272
Solar equations, 294298
Solar longitude, 288
Solar marka:, 290
Solution of the Aporias in Euclids
Elements (Ibn al-Haytham), 144
Spandrels, 246
)NDEX 384 )NDEX 383
Sphere and Cylinder (al-Kh), 189
Spherics (Menelaus), 332
Spherical astronomy, 145146
Spiral motion, 156
Squares, 5758, 199. See also Bordered
squares
Stephen (at Benedictine monastery oI
St. Paul), 35
Stephen the Philosopher, 3536, 42
Sufhcient Arithmetic (al-Karaj), 235,
238239, 249
al-SI, Abd al-Rahmn, 23, 36, 39, 49,
51
Summa oI Asharite Kalm, 159
Sun
apogee oI, 30, 272
equations oI, 294298
longitude oI, 288
Sryasiddhnta, 270272, 282
Synaugeia, doctrine oI, 91
Syntaxis (Ptolemy), 3031
Syrianus, 125
al-Tabar, Badr al-Dn (ca. 824/1421), 63
Tables oI Pisa, 3642
Tabaqt al-umam (al-Andalus), 9
Tabiyyt, 154
1abulae astronomicae (de Ia Hire), 283
1adhkira, 159
1ahdhb al-akhlq (Ab Al Miskawayh),
134
1ahrr al-Manzir (Recension oI Euclid`s
Optics) (al-Ts), 6768
1akht, 78
1f al-a:yf (al-Maghrib), 286
Takht-i Suleiman vault in Iran, 254, 256,
260
1alkhs aml al-hisb (The Abridged
|book| oI the Operations oI
Calculation) (Ibn al-Bann), 313, 315,
319, 324, 328
1alqh al-afkr (The Eecundation oI
Spirits) (Ibn al-Ysamn), 326, 329
330, 333
1alqh al-afkr f amal rasm al-ghubr
(Ibn al-Ysamn), 9
1anbh al-albb al masil al-hisb
(Warning to Intelligent People on the
Problems oI Calculation) (Ibn
al-Bann), 324326
Tamerlane, mausoleum oI, 247
1aqwm al-khata wal-mushkilt allat
li-Uqldis f Kitbihi al-mawsm
bi-l-Manzir (Rectifcation oI
Euclid`s Optics) (al-Kind), 59, 65,
88
1rkh al-hukam (Ibn al-QiIt), 9
1arkb al-adwiya (On Compound Drugs)
(Ibn Rushd), 361
Tashkent scroll, 256
al-taysr, 367
10s and 10000s, 2627
1en 1reatises on the Eye (Ibn Ishq), 63
Theodosius, 42
Theology and science, 157160
Theon oI Alexandria, 88
Theorem oI Ceva, 332
Thierry oI Chartres, 40
3 (numeral), 5, 7, 15
Tier, 257
1imaeus (Plato), 132133
Timothy, Patriach, 162
Timurid architecture, 246247, 256, 259
Timurid scrolls, 255
Toledan tables, 316
Toledo (Spain), 29
Toletane fgure, 11
Topkapi scroll, 250251, 253254, 256
Transmission oI philosophy through
science, 126127. See also Cross-
cultural transmission
Transversal Theorem, 190
1reatise of Astronomy (Ialande), 314
1reatise on Geometrical Problem Solving
(al-Sijz), 183
1reatise on the Magic Disposition of
Numbers in Squares (Abu`l WaI),
200
1reatise on Measuring Paraboloids (Ibn
al-Haytham), 252253
1reatise on the Ratio Between 1hree
Lines (al-Kh), 181
Trepidation, 285286
1uhft al-add (The Ornament oI
Numbers) (Zak), 318
al-Tnis, Sanjaq Dr, 286287
)NDEX 386
al-Ts, Nasr al-Dn (d. 672/1274), 59,
63, 6768, 89, 110111, 159160, 314
2 (numeral), 5, 7, 15
Ulugh Beg, 239, 247, 269270, 282283,
286290, 294, 298299
Umayyad rule, architecture under, 240
al-Uqayl, 321
al-Uqbn (d.1408), 329330
al-Uqldis, 3, 8
al-Urd, Muayyad al-Dn, 112
Urf:a f l-fudhr (Poem on the Roots)
(Ibn al-Ysamn), 321, 326327
al-Urf:a f l-Tibb (Ibn Sn), 367
al-Urf:a al-Ysamniyya (the Poem oI
al-Ysamn) (Ibn al-Ysamn), 320
Vaults (architectural), 237, 239247.
See also Muqarnas
Visible image (khayl), 102103
Vision. See also Optics, Arabic
accuracy oI, 5759
color and, 91, 93
conditions oI, 92
crisis oI science oI, 9192
elements oI, 6466
image-oriented theory oI, 96105
mathematical treatment oI, 91
mechanisms oI, 6466
probe experiment and, 99100
problem oI, 95
properties oI, 92
rectilinear, 91
refected, 91
reIracted, 91
true, 91
untrue, 91
veracity oI, 5759
Visual cone, 96, 98, 103
Visual rays, 91
Visual-ray hypothesis, 65
Von Humboldt, Alexander, 3
Wafq al-add, 199. See also Quadratus
mirabilis
al-Wahhb, Ahmad ibn Abd, 58
al-Wahhb, Hasan Husn, 287
Western Arabic numerals, 7, 1115, 29
Witelo, 95
al-Yaqb, 4
Yavanacandracchedyakopayogin, 283
az-Zahrw, 315
Zak, Slih, 318
al-Zarqlluh, 51, 151, 310, 316
Zaytna educational institutions, 314
Zero (sifr), 4, 15
Zf, 192, 269270, 282, 285286, 298
Zf (al-Battn), 10
al-Zf al-Jadd (Ibn al-Shtir), 286
al-Zf al-Mukhtasar f ilm al-tadl wa
l-taqwm. See Zf al-Sharf
al-Zf al-Muwhq (Ibn Azzz al-
Qusantn), 285
Zf al-mumtahan, 34
Zf al-Sharf (Sanjaq Dr), 286, 288,
290291, 294, 299
Zf-i-fadd (Ulgh Beg), 269
Zf-i-Shh-Jhn, 269, 271
Zf-i Sultn (Ulgh Beg), 286, 289290,
292294
Zuhr, Ban, 356, 363365

The Enterprise of Science in Islam

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

The Enterprise of Science in Islam

Transféré par

Droits d'auteur :

Formats disponibles

1nr Lx+rrrrisr or Scirxcr ix Isi:x

Dibner Institute Studies in the History ot Science and 1echnology

is, once again,

Eigure 7.31 Eigure 7.32

Vous aimerez peut-être aussi