Vous êtes sur la page 1sur 133



The Genesis of Lachmann's Method

Edited a11d translated by Gle1111 \V. Most

The University of Chicago Press

SE BASTIANO TIM !'AN ARO (19l3-1000) was an Italian scholar in classics and
politics. He was a member of the British Academy and the Ac:cademia dei Lmcci and
the author of On MateTialism.

c LE NN w. Mo sT is professor of ;meient Greek at the Scuola Normale Supcriore in

Pisa and teaches at the University of Chicago in the Committee on Social Thought and
m the classics and comparative lirer:nure departments.

The University of Chicago Press, Chicago 60637

The University of Chicago Press, Ltd., London
(} 1005 by The University of Chicago IN MEMORY OF EUGENIO GRASSI
All rights reserved. Published :z.oos
Printed in the United States of America

14 IJ 11 11 10 09 08 07 06 OS l 2. 3 4 s

ISSN: o-1:1.6804054 (cloth)

Italian edition published by Liviana Edit rice, 198 1.

Library of Congress CatalogingIn-Publication Data

T1mpanaro, Sebastiano.
fGenesi del metodo del Lachmann. English)
The genesis of Lachmann's method I Sebastiano Timpan:uo ; edited and transl31ed
by Glenn W. Most.
p. cm.
Includes bibliographical references and index.
ISBN o u6- 80 405 4 (cloth : alk. paper) - ISBN o--:z.:z.6-80 406:z. (pbk.: alk. paper)
1. Lachmann, Karl, 17 93- 1851 . :z.. German philology. 3. Classical philology.
I. Most, Glenn W. ll. Title.

ro64 .L3T56 13 :z.005

430' .9'09 :z.-dcu

@The paper used in this public:mon meets the minimum requirements of the American
National Standard for Information Sciences- Permanence of Paper for Primed Library
Materials, ANSI z39.48 199:z..

Editor's /lltrod11ctio11 T

Editor's Note to the Reader

.. ..
Preface to the First Editio11 37
Preface to the Second Editio11 39

Introductio n 43

1 Eme11datio ope codic11111 from the Humanists to Bentley 45

l. The Need for a Systematic Rece11sio in the Eighteenth Century 58
3 The First Phase of Lachmann's Activity as a Textual Critic 75
Adde11d11m to Chapter 3 82
4 Lachmann as an Editor of the New Testament 84
5 Contributio ns of Lachmann's Contempor aries 90
Studies on rhe Text of Lucretius 102
7 What Really Belongs to Lachmann I I 5
8 Textual Criticism and Linguistics, and Their Crises at the End
of the Nineteenth and in the Twentieth Century I 19

Appendix A. Lachmann's First Attempt at a Mechanical Rece11sio

in I 81 7 139
Appendix B. Determinin g the Script of Lost Manuscript s 14 5
Appendix C. Bipartite Stemmas and Disturbances of the Manuscript
1 57
1 89
A. I Fi11a/ Remarks 011 Bipartite Stem mas/ 207

B. Editor's Notes: Diffcrmces among the Various Editions 216

C. Rece11t Bibliography 234

Index of Names 24 1
/lldex of Topics z.4 7
llldex of Manuscripts Discussed :z.5 1

Editor's Introduction

Sebastiano Timpan aro (1923 -2000) was one of the most importa nt
in twentie th-cenc ury Clas-
intellectuals of his generation and a major figure
as well known in
sics, history, and Marxis t and Freudia n theory. He is not
of the
England and America as he deserves to be, however, in part because
s, in part because of the very unconv ention-
seeming disparit y of his interest
it has
ality and originality of his views, and in part because traditionally
and German y, that
been from other European countries, especially France
because tra-
thinkers have more often been translat ed into English (if only
ditional academic Italian, a language in which Timpan aro fortunately
). I
not write, is singularly resistant to attempt s at rendering into English
only to facilitat e study of those cru-
hope the present edition will serve not
of tex-
cial aspects of the history of Classics and of the theory and practice
introdu ce to a
tual edition s that his book presents so clearly, but also to
somew hat wider audience the though t of a quirkily brilliant, soberly
sionate, and profoundly serious Europe an intellec tual.

Sebastiano Timpan aro was born into a family that combined a deep
ly to the history of science , with a no
cation to imellcctual pursuits, especial
. His father, Sebasti ano Timpan aro
less deep commi tment to leftist politics
ci, near Messin a in
Sr. (1888- 1949}, was born in the small town of Tortori
of Naples and
Sicily, but then went north to study physics at the universities
then Bologna. His promising career as an assistant in experimental
short in 1929 by his categor ical refusal
at the University of Parma was cut
governm ent require d of all Ital-
to take the oath of allegiance that the Fascist
school in Florenc e, where
ian state employees; he found work first in a private
n the
he taught mathematics and physics for a number of years, and then-i
the Do
meantime he had been obliged to swear the loyalty oath after all-at
institut e in the history of science in Pisa, which
mus Galilaeana, a research
I- -- _zs z LL &

Editor's Introduction 3
Edito r's Introduction
also taug ht at the elite Scuola Nor -
ship duri ng that period. In Pisa Pasquali
Sebastiano Tim pana ro's moth er, Ma family moved there he was intro duce d
he directed from 1942 until his death. male Superiore, and once Tim pana ro's
zo and took a degree in Greek lit of that uniq ue institution of schol-
ria Cardini (189 0- 1978 ), was born in Arez into the intense intellectual atmo sphe re
she studied briefly in Berlin with diminutive, provincial, but cultu r-
eratu re at the University of Naples; in r914 arly research and teaching and into the
centuries, Ulrich van Wila mow itz in which it is located. It was also at
two of the greatest Hellenists of the past ally and politically very anim ated town
r a brief period as a Dad aist poet (for ard Fraenkel, one of the greatest Ger-
Moellendorff and Herm ann Diels. Afte Pisa that Tim pana ro studied with Edu
with Tristan Tzar a), arou nd 1920 e to teach at the Scuola Norm ale
several years she corr espo nded frequently man Classicists of the century, who cam
rned to her Gree k studies. She ac- uali's death. Under their guidance
she aban done d poetry forever and retu regularly for a num ber of years afte r Pasq
in Italy, for her editions, transla~ ns for his later wor k on Classical lit-
quir ed considerable recognition, not only Tim pana ro laid the scholarly foundatio
sophy and science, especially on the ve all in the high ly technical disci-
tions, and studies of anci ent Gree k philo erat ure, especially on Latin poet ry (abo
ined for her whole career a teac her getical and lexical studies) and on
topic of the Pythagoreans, but she rema plines of texrual criticism and of microexe
re she met Sebastiano Tim pana ro ry duri ng antiquity. In both of these
in juni or high schools, first in Parm a (whe the history of scho larsh ip on Latin poet
Florence as her husb and, finally in ificant and lasting cont ribu tion s (col-
Sr.), then at the same private school in fields, Tim pana ro mad e numerous sign
ld War, she became actively involved a, and ioor a), thou gh he always pre-
Pisa. After the end of the Second Wor lected in Tim pana ro 1978 , 1986 , 1994
alist Party, cam paig ning especially erudite, highly condensed philo-
in local politics in Pisa for the Italian Soci ferred the form of the small, astonishingly
for the establishment of non- Chu rch nurs
ery schools in the city. literary mon ogra ph and thou gh he
logical note to that of the expansive
hed to his pare nts, even by Italian editions of such auth ors as Ennius
Sebastiano Tim pana ro was muc h attac never himself unde rtoo k the full-scale
his father's papers with a preface by he wou ld do.
stan dard s-he published a selection of and Virgil that his teachers had hoped
and the very last essay he wro te was ro was widely regarded, at hom e
himself in 19 52 (Tim pana ro Sr. 19 5i), On the basis of these studies, Tim pana
of his mother's works that he edited, t highly esteemed Italian Classical
the lengthy intro duct ion to a collection and abro ad, as one of the very few mos
dini 2001). From them he inher me a regular professo r at any Italian
which was published after his deat h (Car scholars of his time . Yet he never beca
ed all his wor k and his who le life: life, he taug ht a few times at the Uni-
ited interests and characteristics that mark university (tow ard the very end of his
lectual honesty and moral rigor; r): after som e years of teaching in
an uncompromising com mitm ent to intel versity of Florence as a visiting professo
rationality, as it is expressed for ex from 1960 until his retirement in
an unswerving dedication to an ideal of secondary schools near Pisa, he wor ked
ces, with at the same time a painful publishing house, La Nuo va Italia.
amp le in the progress of the natural scien 1983 as a proo frea der at a Florentine
of much of hum an history, especially complex set of partial reasons) was
recognition of the brut al irration:ility Wha t exactly the reason (or, likelier, the
:in exceptionally broa d multidiscipli- that was perh aps less extr aord inar y
in national and internation:il politics; for this seeming ano mal y-an anomaly
p:irticular fascination for French and r been confined to the universities
nary and multilingual cultu re, with a in Italy, where intellectual life has neve
German literature and intellectual histo
ry; a systematic preference for dis- r cou ntri es-w as always far fro~
than it wou ld have seemed in som e othe
s not of the cont emp orar y epigones . Some have suggested that the im-
cussing curr ent intellectual issues in term clear, even to those who knew him best
seminal thinkers who first set the not capa ble of recognizing, or of ac-
who set the passing fashions but of the perfect Italian university system was
l inability to compromise; an almo st he himself was not willing to com -
fundamental questions; a cons titut iona cepting, his merits, or alternatively that
ment to the poin t of irrep arab le rup- tutional ambiguities. It shou ld also
mor bid tendency to exac erba te disagree promise himself by entering into its insti
writings, above all strictly dea r and f from direct part icipa tion in the
ture ; an unm istak able tone in all his be born e in mind that, in remaining aloo
re, but with a sprinkling of collo- ds were university professors), he
precise, often austere, occasionally seve university (tho ugh man y of his closest frien
ashedly and sometimes rath er pon- set (though for different reasons) by
quialisms and, rarely, meta phor s, unab was able to cont inue a family trad ition
rical path os and of even the slight- ely to scholarly research and publi-
derously didactic, free of any false rheto his pare nts, to devo te his free time entir
est trace of wit or hum or for their own
sake. university adm inist ratio n and exam -
catio n with out beco ming emb roile d in
ed with his pare nts to Florence. activity to friends and to anon ymo us
Tim pana ro was born in Parm a and mov inations, and to direct his pedagogical
good fortune to be able to study ent, always unpredictable, some-
In the university of that city he had the readers rath er than to the physically pres
52), the greatest Italian Classicist seemed to inspire in him a degree of
Classics with Giorgio Pasquali (188 5-19 times rather unruly students who often
figure in the sometimes difficult me-
of the twentieth century and a crucial diffidence bord erin g on drea d.
an national trad ition s of scholar-
diati on between the Germ an and the Itali

Editor's Introduction s
Editor's Introduction
4 ro developed a sophisticated yet
deepest affinity. In parti cula r, Tim pana
aspects of Classical scholarship reached back to such ancient
Tim pana ro's focus on the most technical highly personal philosophical position that
tific discipline com para ble, at etius but that wou ld have been
and his confidence that Classics was a scien prec urso rs as Epicurus and, above all, Lucr
ces did not mislead him into sup- is no accident that he chose to em-
least in certain regards, to the natu ral scien mos t at hom e in the eighteenth century. It
a direct, unmediated relation to collection of essays titled On Ma-
posing that we might be capable of having body his views not only in the form of a
Classical scholarship as thou gh it b; English trans latio n, Tim pana ro
antiq uity or into neglecting the history of terialism (Tim pana ro 1970 , 1975 a, 1997
of superseded errors. On the con- translation of P. Thir y d'Ho lbac h's
were nothing more than an accu mula tion 1975b, r98o b, 1996) but also in that of a
paired the scrupulous investiga- lbac h 1985 ), to which heap -
trary, from the beginning of his studies he 1772 treatise Le bo11 sens (Goo d sense; d'Ho
less scrupulous investigation of and also provided an extensive
tion of anci ent literary texts with the no pended Voltaire's observations on the text
before him, both in antiq uity and the text in both historical and sys-
those scholars who had investigated them intro duct ion of his own , which explored
tions of a single text alon g the line in degree of oversimplification, we
in modern times: the interpretative varia tematic perspectives. At the cost of a certa
h it with unexpected and not al- unde r the headings of materialism,
of the history of its reception helped enric may summarize Tim pana ro's philosophy
stability of that text provided a materialist, Tim pana ro was firmly
ways absu rd new meanings; the relative hedonism, atheism, and pessimism. As a
variety of its attested understand- supposed auto nom y and freedom
fixed poin t that permitted the historical opposed to any Idealistic idolatry of the
l. Tim pana ro's first book , dedi- the tot~I physical and biological
ings to be orde red and rendered meaningfu of the hum an agent. Instead he insisted on
omo Leopardi's cont ribut ions to severely criticizing on this acco unt,
cated to his pare nts, was a stud y of Giac dete rmin ation of all hum an phen ome na,
a, 1997a): in it he dem onst rated pana ro 1974, 1975 c, 199:z.; En-
Classical studies (Tim pana ro 1955 , 1977 amo ng many others, the later Freud (Tim
being nothing more than a self- st all twentieth-century Marx ists,
once and for all that Leopardi, so for from glish trans latio n, Tim pana ro 1976 a), almo
his personal tribulations, was of lapsing frequently into Ideal
absor bed Rom antic poet obsessed only with and even Mar x himself, who m he accused
was not only deeply inspired by Engels. Against the various forms
a Classical scho lar of Euro pean rank who istic hum anism and co who m he preferred
also capa ble of interpreting and detto Croc e (1866-195:z.), domi-
his read ing of Gree k and Latin texts but of Idealism that, in such figures as Bene
of his cont emp orari es. This first the twentieth century, Tim pa naro
emen ding them alon g with the very best nated Italian philosophy throu gh mos t of
work on the subject, was followed an enlightened materialism - and
book , which has remained the stan dard leaped back to Leopardi as the model for
ngs (Leopardi 1969 ) and by a h it was intimately linked. For,
by an edition of Leopardi's philological writi also for an enlightened pessimism with whic
ry and intellectual history, mostly gical appa ratus that makes up
number of othe r studies in mod ern litera in this view, an essential part of the biolo
rs (collected in Tim pana ro i96 5, s is the unremirring search for
devoted to nineteenth-century Italian write hum an beings as well as all othe r organism
ough Tim pana ro modestly defined (a conclusion he derived from che
1969 , 198o a, 198:z., 1984 a, 1994b). Alth pleasure. But given that there is no God
lar of nineteenth-cen tury cultural in that natu re has nor been con
himself once as noth ing more than "a scho failure of all attem pted theodicies), ic is certa
logy" (Bruscagli and Tellini 1996: fulfill our ends, and hence the he-
history who comes from Classical philo struc ted with a view towa rd us, so as to
in which Italians unde rstan d their desires must inevitably be often
1 5}, in fact his cont ribut ion to the way donistic desire for the satisfaction of our
influential. In particular, Tim - ediably. Thos e who knew Tim-
own intellectual history has been extremely frustrated - systematically, painfully, irrem
pone nt of much early nineteenth- ds between irony and despair; his
pana ro's insistence on the Classical com pana ro wel l-th e oscillations in his moo
inence the continuities between ures of friendship or of schol-
cenrury Italian writing has brou ght ro prom inter mitte nt burs ts of intense joy at the pleas
htenment think ing and a num ber never doub ted that his pessimism
a liberal strain of eighteenthcentury Enlig arship; his unfailing, angu lar cour tesy-
iously been superficially pigeon- ully argued philosophical position
of nineteenth-cen tury write rs who had prev was not only a well considered and caref
unted as conservatives or anti-
holed as Rom antic and were often disco but also a deeply ingrained way of life.
ablish an impo rtant link of con- whole life a passionately en
nationalists. In so doin g, he has helped reest Nonetheless, Tim pana ro was also for his
een the eighteenth and nineteenth n politics. He began, soon after
tinuity in European cultural history betw gaged milit ant on the far left wing of Italia
ber of the PSI (Partito Socialista
centuries. the end of the Second World War, as a mem
h-century figures shou ld not both his parents; he shar ed with
Tim pana ro's conc entra tion on nineteent Italiano, or the Italian Socialist Parry), like
of the eighteenth century, espe- Democrats, who , overdy Cath -
mislead: it was rathe r with certa in aspects them a lifelong hostility to both the Christian
poin t to positive future develop- Church, governed Italy uninter-
cially with those that seemed to him to olic and with the supp ort of the Cath olic
that he himself evidently felt his
ments or that had been unju stly neglected,
.. Sll!J

Editor's Introduction 7
Editor's Introduction
same kinds of mechanical processes as result so often in mistakes in tran-
ruptedly for the first two decades after the war, and the Italian Commun ist
scribing manuscri pts (Timpana ro 1976a). A few articles, mostly on related
Party, which he regarded as covertly theological in structure , ideology, and
themes, have also been translated (Timpana ro 1976b, 1977b, 1979, 1984b,
appeal-o pium for the masses. But it has never been easy, in Italy or else-
1988). The result is that Timpana ro is known in the English-speaking world
where, to be both left wing and principled, and at the same time to wish to
above all as a materialist philosopher, Leftist theoretician, and critic of psy-
attain actual political power. When the Italian Socialists decided in 1964 to
choanalysis. These are, of course, importan t aspects of his vast and diverse
abandon the oppositio n and to join the government in a coalition with the
producti on-but it can be argued that they are the ones of least permanen t
Christian Democrats, Timpana ro and his mother, like many other Socialists
value, in part because they were tied closely to what appear now to have
who reacted against what they considered a betrayal of their founding prin-
been largely ephemeral modes of Leftist politics and thought (already in the
ciples, became active members in the newly founded PSIUP (Parti~o Soc~a
1990s Timpana ro ruefully confessed that On Materialism had come to seem
lista Italiano di Unita Proletaria, the original name of the PSI), which splin-
to him like "a fossil"; Timpana ro 1994b: xi), in part because philosophy
tered off from the left wing of the PSI. The PSIUP dissolved when it failed
was in fact not Timpana ro's greatest strength. Instead it is his historical and
dismaUy in the general elections of 1972., and Timpana ro joined a successor
philological studies that seem to me to represent his most lasting contribu-
party, the PDUP (Partlto di Unita Proletarla h but the PDUP, despite coa-
tion to Western culture. The present translatio n of Tlte Genesis of Lach~
lescing with other small left-wing groups, was crushed in its turn in the elec-
mann's Method will give a wider circle of readers the opportunity to enrich
tions of 1976. Timpana ro resigned from the PDUP in that year and never
their understan ding of this profound and enigmatic figure.
ag:iin joined any other political party on offer in Italy's alphabe~ soup of left-
wing acronyms. But until the end of his life he continued to write frequently
in newspapers, journals, md private letters about specific political issues and
about the tensions and contradic tions in the Italian left-an ample, indeed
Timpana ro's book examines the historical development and the systematic
an inexhaustible subject for reflection, which, in his last years, he seasoned
limits of a particularly significant moment -the attempt during the nine-
with a cautious rapproche ment with environmentalist positions (Timpa
teenth century to increase the degree of rationalization, standardi zation, and
naro 2001b). Timpana ros gaunt, well-dressed figure, with his high fore-
professionalization, within the evolution of a process that for millennia has
head and piercing eyes, was a familiar sight not only at scholarly confer
been central to human culture- the transmission of written texts. It is only
ences and seminars, where he was received by colleagues and students with a
if his argument is placed within this larger horizon that its full import can
respect bordering on deference, but also at political demonstr ations, where
be understoo d.
a vague sense of his scholarly renown and institutional marginality sur-
A written record has this advantage over an oral utteranc e-that it lasts
rounded him with an aura of ascetic purity, indeed almost of saintliness. He
in time beyond the moment of expression, in a physical form independ ent
must have been a formidable political opponent , erudite, eloquent, drasti
of the speaker's and listeners memories. Yet this physical form too has its
cally and unrelentingly polemical, limited only by an intransigent incapac-
limitations, for it is restricted to a single spatial location and must be en-
ity to comprom ise and perhaps a certain remoteness from the lived realities
trusted to an ultimately perishable medium to bear it. For the one reason or
(as opposed to the theoretica l study) of politics and economics.
the other-e ither because the existing copy no longer suffices for the new,
Oddly, the only books of Timpana ros that have been translated into En-
spatially dispersed uses to which it is now to be put (usually, new readers),
glish before now are his collection of essays 011 Materialism, in ~hich he
or because it has become damaged over time (by overuse, inadequate mate-
attacks European Marxism for havi ng betrayed the legacy of genuine mate-
rials, or simple old age)-it may become desirable to produce new copies of
rialism bequeathed by Engels, and French Structuralism for having mis-
written texts. Before the age of photogra phs, photocopies, and scanners,
understoo d and misapplied the linguistic theories of Saussure (Timpana ro
which copy texts by purely mechanical processes simply on the basis of the
1 975b, 198ob,
1996), and The Freudian Slip, his sustained polemic against
contrast between lighter areas and darker ones, the only way to produce
the later Freud and Freudian psychoanalysis in general for having aban-
new copies was to transcribe them from old ones, element for clement, most
doned the materialist insights of Freud's early writings, and against the psy-
often semantic unit for semantic unit. If greater accuracy of transmission
choanalytic explanati on of so-called Freudian slips in particular, which in
was required, this could be done visually, by a scribe copying onto one new
most cases, based on his conceptio n of materialism and his experience of
medium the text he saw before his eyes (but the disadvantage was the smaller
textual criticism (and proofreading~, he attributes instead to exactly the
Editor's lnrroducuon
Editor 's Jnrroduction
one that was closest to
the same time from a that derived from an authoritative provenance; the
number of copies that could thereby be produced at hand; and so forth. And of course even then the copyi
st (or his corrector)
er of copies was sough t
single exemplar). If on the other hand a large numb was still free as he saw fit to make whatever he thoug
ht were corrections and
by the exemplar was
after, an acoustic procedure could be preferred where other improvements.
ed to it and copied down , each
read out before a group of scribes, who listen The advent of printing in the fifteenth century altere
d various parameters
had heard (at the cost of
onto his own medium, what they thoug ht they of the process of textual transmission but at first had no effect whatsoever
and other forms of
greater inaccuracy, due to homonyms, distraction, noise, on these methodological issues. Printing vastly multip
lied the number of tar-
one, that for most
interference). It is only a guess, but probably a good get texts that could be made on the basis of a single
source text and created
was one that began
of the history of huma n cultur e the normal situation a greater degree of textual identity (though, especially
at first, by no means
sourc e text) and ended up, as result
with a single exemplar to be copied (the complete identity} among those target texts. But printi
ng did not in itself re-
and usually as purpose, with more than one copy of the text (the source text ds of textual trans~ission,
normally entailed quire people to change fundamentally the metho
plus the target text, or multiple target texts): transmission and indeed for several centuries printers, correctors,
and editors continued
med neither by ma-
multiplication. And given that the procedure was perfor to use most of the same rule-of-thumb criteria that
their predecessors, the
ns err, transmission al-
chines nor by gods but by humans, and that huma ancient and medieval copyists, had developed. What
changed matters most
y entailed prolifera-
ways entailed error, and multiplication of copies usuall was instead the concatenation of three factors during
the period from the fif-
se in the number of manu -
tion of errors.
copyist's task: he (it teenth to the nineteenth centuries: the vast increa
To have only a single source greatly simplified the scripts, Greek and Latin, that became available throu
ghout Europe durin g
it as faithfully as he
was of course usually a he) could attempt to transcribe and after the Renaissance; the gradual concentration
of the holdings of li-
he saw fit, so as to cor-
or his supervisors wished, intervening into the text ns braries no longer in a large number of small collections
(each of which might
consid ered ro be improvements of
rect obvious errors or to effect what he have one manuscript of Cicero) but, more and more, in a small number of
availa ble two source
various sorts. But what was he to do when he had large collections (each of which might have dozen s); and the general in-
to differ from one
texts? Given the proliferation of errors, these were bound crease in the ease of communication and travel over
the cours e of the early
were of any consider
anoth er in their readings, at least occasionally, if they modern period . The result was that eventually there
was no longe r a scar-
reading to put into the
able length. On what basis was he to choose which city of potentially available source texts from which
further copies could be
ion occu rred- and presumably for
target text? However rarely such a situat derived but an impressive, indeed intimidating overa
bund ance- Mont fau-
in the largest scriptoria,
many centuries it did not occur frequently except con's Palaeographia graeca ( 1708) already lists more than 11,00 0 Greek
regularly enough for a
monasteries, and libraries- it must have happened manuscripts. Since the fifteenth century, printers and
editors had tended by
op: whichever seemed
certain set of rule-of-thumb criteria of choice to devel force of inertia to copy their own rexes from those
printed by their prede-
better reading would be
to be the grammatically or semantically or logically cessors, correcting them by the lights of their ingen
uity where they seemed
imported into the tar
preferred from case to case, or both readings could be in error or adduc ing for comp arison some one manu
script (or a very few
rization for one of
get text with or witho ut an expression of greater autho manuscripts), almos t always one (or ones) that had the advantage of prox-
to give a general pref-
them. The next step methodologically will have been imity or the appearance of old age. As the number of
available manuscripts
one whenever possible,
erence to the one source text over the other available proliferated, further criteria for preferring one readin
g over anoth er were
the latter or indicating
either suppressing apparently cquipollent readings in developed- the better reading was the one to be found
in the most manu -
simpl ified the copyist's work,
them as inferior alternatives . This will have scripts, or in the oldest ones, or in the oldest one of
all- but a rational jus-
to choos e among variants
freeing him from the obligation to use his brain tification for such criteria was neither provided nor availa ble. At least from
the numb er of source
from case to case and, in effect, reducing once again the lofty perspective of a nineteenthc entur y Classical
philol ogy anxious to
basis was the copyist
texts. But at this point a new question arose: on what establish its credentials as a serious science, this was
obviously an unsatis
to prefer? Over the cen-
to choose which one of the available sources he was factory state of affairs.
develo ped, each with its own par-
turies, various contradictory criteria were The first attempt to provide a thoroughly mechanical
and systematic pro-
; the most legible man-
tial and specious justification: the oldest manuscript cedure for rationalizing and standardizing the choic
e amon g manuscripts,
readings; the one that
uscript; the one that appeared to have the most good and hence amon g readings, was developed during
the ninete enth century
corrections; the one
had the fewest corrections; the one that had the most
Editor's Introduction 1[

Ediror's Introduction If
10 racterizes the met h od 1tse
Lach cal investigatio n to Lachmann's met hod that cha .
ntieth century has been kno wn as f T' with exemplary clarity in the
and since the beginning of the twe ( 179 3- Th e aim s o 1mpana ro's ana lysi s are set out
association with Karl Lachmann n, and its results are summarized
mann's met hod because of its texts in second ~ara~raph of his intr odu ctio
pro duc ed celebrated edit ions of clusions are the followin g: tha t man y of
185 1 ), a Ger man Classicist who is ge and luctd~y in cha pter 7. These con
ern German . Lachmann's met hod associated with the met hod wer e w II
Latin, Greek, and medieval and mod g a stan dard ized , the techniques and presuppositions foll o~
nealogical and mechanical in natu
re and aims at providin aissance Hum anis ts or dur ing the
man uscr ipts established either amo ng the Ren the eigh-
s on the basis of multiple the New Testament scholars of
rati ona l procedure for editing text ord er to ing centuries, especially amo ng
or use his personal judgment in ar formulation of the met hod asso
wit hou t requiring that the edit of man - teenth century; that the part icul
Its goal is to determine the filiation Lucretius (L<1chmann 18s o) w<1s an-
choose amo ng variant readings. ch oth er ~i~h Lac~mann's celebrated edition of
ch one s have been copied from whi such oth er scholars as Ore lli, Zum pr,
uscripts, tha t is, to ascertain whi odu ce new erro rs, t1~1pated m mos t ~f i_ts essentials by k and
ones: given that every act of tran
scri ptio n is likely to intr s before the publication of this wor
R1tschl, <1nd Madv1g m the decade
a manuscript B, if it has been cop
ied mechanically from a man usc ript
A, will
it has in part icul ar alm ost in its enti rety
in Bernays's wor k on Lucretius
of ~8
it does not have all of them, r 846 ); and that Lachmann's own
have all the erro rs tha t A had (if ce is likely (a~d ~o a lesser exte nt in Purmann's of t
dur ing the transcription and hen earlier editions bur even in his grea
pro bab ly corr ecte d som e of chem ly to have pltcat1~n oft~~ method, not only in his da-
ically afte r all), and it is also like inconsistent and was mar ked by fun
not co have been copied mechan be the case , then B Lucretius ed1t1on, was generally all, both
at least one new erro r of its own
. If this can be sho wn to aro dem ons trat ed, once and for
text it sha res menral errors. In sho rt, Tim pan he did
of the constitution of the not in fact Lac hma nn's met hod (for
can be discarded for the purposes is not er that "La chmann's met hod " was chm ann 's
brings no new information that nn's met hod was not in fact "La
with A, since B, com pare d with A, it mus t nor invent it} and tha t Lac hma
mechanical, bot h in the sense that consistently). Since the first pub lica tion of
roneous. Lachmann's met hod is app lied met hod " (for he d id not app ly it
ption of manuscript s if it is co be use the term "La chm ann 's met hod " with -
presuppose the unth inki ng transcri tion s of filia tion is Tim pan aro' s study, scholars who
determination of rela so at thei r peril.
to them and in che sense chat the ility . Idea lly, out quo tati on mar ks have don e line of
s and calculations of pro bab sketches a brief hist ory of the dec
achieved on the basis of simple rule be rati ona l (2) In cha pter 8, Tim pan aro the be
ings based on this met hod will second half of the nineteenth and
choices of manuscripts and of read on ob Lachmann's met hod between the
taste of the individual scholar but es that wer e per
gi~ning of_rhe twenrie~h century. His focu
s is on the ana logi
in that they will depend not on the d; hen ce they will be ual ed-
atized and eval uate the developments in Classic al text
jective evidence tha t can be mathem you ng or old, in ce1ved dur mg that period between on the
, because any scholar, in comparative historical linguist
cap able of becoming stan dard ized exa ctly the sam e iting on the one han d and those a ver-
experienced or exp ert, sho uld on
principle com e up with nineteenth cenrury in which
hma nn's ~ther: a first period in the middle of the
rmation . We may interpret Lac applied, with euphoric hopes for
results if he is given the same info texts, s10n of Lachmann's met hod was
the proliferation of possible sou rce tions within the lnd o-E uro pea n fam ily of
met hod as a defensive reaction to tify it to the search for genealogical rela
e manageable number, and can iden s of that cen tury by a wav e of dis
intended to reduce them to a mor
professionalization of Cla ssic s dur ing the ~ang~ages was followed in the last decade iplines. The discovery tha t the
as one imp orta nt element in the hed to be rec- 1llus1onment and skeptic ism in
both disc
s led
hed rules tha t all who wis in a relatively small number of case
nineteenth century, since it establis so as met hod was fully applicable only
ipline cou ld be expected to follow it by artificially reducing the num
ber of wit
ognized as full members of the disc som e textual critics to misapply the ba-
erally acceptable results. on and hyp othesizin g filia tion s on
to pro duc e uniform and hence gen rds: nesses they too k into con side rati
hmann's met hod in three rega e at all, and othe rs to emp hasi ze instead
Tim pan aro' s boo k examines Lac inve nt it? ( 1) Wh at sis ~f inadequate evidence or non not be
icul ar, did Lac hma nn atio ns in which the met hod cou ld
( r) Wh at was its origin, and , in part vali d is it? The se the imp orta nce of thos e man y situ ex-
was its subsequent development and
outcome? and (3) How not been copied mechanically and
app~ied safely b:ca use targ et texts had tury, the
three questions determine the stru
ctur e of his boo k. the beginning of the twentieth cen
rs the devel clus 1vely from smgle sources. By led to a
( 1) In cha pter s 1 to 7 and
Appendix A, Tim pan aro conside sm in all the hum an scie nces had
thro ugh general reaction aga inst Positivi
edition from the fifteenth cen tury s met hod in such Classicists Pas quali,
widespread discr~st of L~chmann'
opm ent of the met hod s of textual the orig in of the var-
tury and inquires into cha pter is dedicated.
the middle of the nineteenth cen of the gen ius of to who m the closing secu on of the t are
Lachmann's method. Par t e of Lachmann's met hod , and wha
ious features that go to mak e up gen ealo gi- (3) Wh at then is the lasting valu
lies the very sam e technique of
Tim pan aro 's stud y is that he app
Editor's Introduction I)
Editor's Introduction
1986). As is indicated by Timpanaro's posthumous paper published here, he
the limits of its validity? Ir is to these questions that chapter 8 is partly and
himself believed that Reeve had definitively proven him wrong in several im-
Appendices Band C are fully addressed (as is also the posthumou~ essay,
portant points (though not in all), and this judgment of his, though perhaps
which I have tided "final Remarks on Bipartite Stemmas" and which ap-
roo harsh, seems substantially correct.
pears here for the first time). They largely take the form of a long drawn-out
But the issues at stake are of great enough interest from a methodologi-
argument with Paul Maas (1880-1964) , who provided the most authorita-
cal point of view that the value of Timpanaro's arguments is not thereby
tive statement of the principles of Lachmann's method in his lapidary Text
simply vitiated: anyone interested in questions of textual transmission can
kritik (Maas 19 58 [19271). Against Maas's almost mathematical formula-
learn from the debate as a whole. And beyond the specific question of bi-
tion of the principles of mechanical textual edition, Timpanaro insists on
partite stemmas, the disagreement between Maas and Timpanaro retains a
the infinite diversity of human error and in particular on three well-attested
permanent significance. For Maas's concluding aphorism, "No spe,cific has
and universally recognized ways in which manuscripts can come to differ
yet been discovered against contaminatio n" (Maas 1958 [1927): 49), in set
from one another without the differences being explicable in terms of me-
ting narrow limits to the scope of rationality in textual edition, had seemed
chanical copying: contaminatio n (when scribes make use of more than one
to assign everything o utside them to the sway of caprice and blind chance:
source text and on the basis of their own judgment mix readings from the
Timpanaro's sustained effort was directed not only to extending those lim
one with readings from the other); scribal conjecture (when copyists deliber-
its as far as he could, but also to demonstrating that even inside of them mat-
ately make corrections of their own in the text they are copying because, for
ters were not at all so geometric and predictable as Maas seemed to suggest,
one reason or another, they are dissatisfied with the source text's readings);
Ulcimately, what was at stake in the debate between them were different
and polygenesis (when the same errors are produced entirely independently
views of what was to count as rationality in textual edition-for Maas, only
in different transmission processes, either by chance or because under cer-
pure algorithms; for Timpanaro, statistical probabilities and stochastic pro-
tain circumstances certain kinds of errors occur with greater probability).
cedures as well. If only for this reason, the disagreement between them,
As Timpanaro's investigation of the limits of Lachmann's method pro
which reflected nor only their personalities but also larger differences in the
ceeds it becomes more and more focused on one highly technical but very
conditions of the sciences when the two men were being educated, could
impo;tant issue: why is it that so many textual traditions seem to take a bi
never have known 3 final winner. This makes their debate not less interest-
partite form, dividing nearly into two and only two branches or families of
ing and significant but all the more so.
witnesses? The problem was first posed by the French medievalist Joseph
Timpanaro's demonstratio n in the present srudy is compact, pointed,
Bedier (1 9 13, 1928); his suggestion that the explanation for this striking
stringent, and, especially fo r the first of its three goals, entirely convincing.
prevalence was to be sought not in the intrinsic nature of textual transmis-
Yee it is remarkable not only for these strengths, which it shares with other
sion itself but in the faulty methods of textual editors provoked an intense
great works of scholarship, but also in three ocher regards, in which it is un-
discussion among many scholars, including Pasquali and Maas. Where Pas-
usual, indeed anomalous.
quali denied the truth of Bedier's observation, arguing that bipartite ste'.11
First, when it was first published, in the form of two articles that ap
mas were far less frequent than Bcdier thought, Timpanaro agreed with
peared in 1959 and 1960 (Timpanaro 1959, 1960), it was che work of an
Bedier about the fact of their overwhelming dominance; and where Maas
astonishingly young man: Timpanaro was only thirty-si x years old at the
accepted that apparent fact but attempted to provide a statistical explana
time, yet he was already prodigiously learned, assured in his judgments, ma-
tion for it in terms of the possible outcomes of various scenarios involving
ture in his formulations, authoritative in his to ne.
the copying of manuscripts, Timpanaro attacked both Maas's statistics and
Second, even more strikingly, at the time he wrote these 3rticles Timpa-
his whole methodology. Timpanaro himself seeks a cautious middle way,
naro had never himself performed a critical edition of any text, ancient or
accepting the predominance of bipartite stemmas as an objective fact due to
modern. He was best known in this period for his work on Leopardi's philo
the modes of textual propagation in the Middle Ages, but suggesting that
logical writings, yet he h3d not edited these and did not do so until a decade
certain erroneous editorial techniques increase the number of apparent cases
later, and then in coll3boration with Giuseppe Pacella (Leopardi 1969). To
of biparticism beyond what ought to be the case. His own arguments (in-
be sure, Timpanaro had already published a number of articles that demo n-
cluding most recently another posthumous text, Timpanaro 2001c) did not
strated his interest and competence in questions of textual criticism of Larin
conclude the discussion once and for all, but gave rise to a series of contri-
poetry, especially a series of three articles presented as preparato ry toward
butions by other scholars, particularly Michael Reeve (especially Reeve
Editor's Introduction
Editor's Introduction
not be more evident. The histories of Classical scholarship tend to be dia
a new edition of the fragments of Ennius(Timp anaro 1946, 1947, 1948-49).
chronic in sequence, biographical in format, and sometimes unmistakably
But that edicion of Ennius had not yet appeared at the time, ten years later,
hagiographic in tone: along the line of a more or less simple temporal axis
and in fact it never did appear. Indeed, by die end of his life the only edi-
they string together the biographies of the great scholars of the past, indi-
tions that Timpanaro published were of writings by modern Italian authors
cating their personal vicissitudes, strengths, and foibles, listing their great
which he prepared, almost always in collaboration with f~iends, on the b~-
works, and inviting the student to admire their contributions to scholarship,
sis of autograph manuscripts or early editions (e.g., Ascol'. 19 59; Bartoletti~
to try to understand them, and to imitate them as best he can. The manuals
Bornmann-Manfredi-Timpanaro 1961, 1970; Leopardi 1969; Pasquah
of textual criticism are usually systematic in orientation, synchronic in struc-
86a, 1986b), and two popular editions of foreign ce~ts with lt~lian trans-
19 ture, entirely unbiographical in format, and usually cool and reasoned in
lation (d'Holbach 19 8 5; Cicero 1988, 1998, the latter indeed. quite remark-
tone: they attempt to construct a series of rules to explain how errors come
able but for its extraordinarily rich commentary, not for its text). Thus
about and to indicate to the future editor what techniques and lines of rea-
Tim~anaro never published during his entire lifetime a single criti~al edition
soning he should apply in order to construct the best possible edition out of
of an ancient author on the basis of the very procedures of collat10n, recen-
the materials at his disposal. The biographies of individual scholars are usu-
sion, and emendation of manuscripts that he had analyzed with such ex-
ally closely focused on the life and works of the particular person who forms
pertise and penetration in this book. He also planned for man~ year~ to
their object, often employ archival material, and arc usually more interested
write a handbook of textual criticism, but nothing ever came of this pro1ect.
in synchronic and local, rather than in diachronic and large-scale contextu~
We should bear this in mind when we read what are almost the closing
alization. Finally the introductory manuals are most often encyclopedic in
words of the body of this book: "And the practical exigency remains, that
character within the terms of the specific discipline that is being presented;
certain critical editions not be postponed forever for the sake of studying the
only rarely do they succeed in combining perceptive analysis of individual
history of the tradition in all its smallest details, that scholars not bury them-
figures with the larger development of the field and of cultural history as oi
selves so deeply in the study of medieval and Humanist culture that they for-
whole. What unites all these kinds of studies is that with very few exceptions
get to return to textual criticism" (below, p. 13~)". .
indeed, their author is legitimated in the eyes of his readers by the first-class
Third The Genesis of Lac/Jmann's Method 1s m a certam sense a book
work he does in separate publications as an editor of Classical texts-an ap-
without ~ genre. Those who work on the modern study of ancient Greek and
parent generic requirement of such studies that makes the fact that Timpa
Latin literature tend in general ro produce one or another of various kinds
naro never did edit a Classical text himself seem even more anomalous.
of sharply differenciated studies of its techniques. There ~ave been (1) nu-
The Genesis of Lachmann's Method, on the contrary, is an attempt to
merous histories of Classical scholarship, such as Bursian 1885, Sandys
historicize the techniques of textual criticism and edition such as had not
( 903-8) and 1915, Gudemann 1909 (1907), Kroll 1919 (1908),
1908-:z.i 1 really ever been ventured before. It fits into none of the categories of schol-
Wilamowitz 19 8 :z. (19 2.1 ), or, more recently, rfeiffer 1968 and 1976,
arship on the history of Classical scholarship we have just distinguished.
Reynolds-Wilson 1991 (1968), Kenney 1974, and Briggs and Calder 1990;
(1) Timpanaro is indeed writing a history and for that purpose does make
(:z.) many manuals of textual criticism, of which the best-known ones of the
use of biography as one of the important strategies with which he organizes
twentieth century are Havet i911, Kantorowicz 19:z.1, Maas i958 (19:z.7),
his material, but he employs the accounts of individual lives only tactically,
Dain 1975 ( 1949 ), van Groningen i963, and more recently. West 1973;
instrumentally, in order to explain and make these historical developments
( ) various biographical studies of individual scholars, ~htch ha;e a~
concrete, never as an end in itself. Besides, he makes no claim to present a
tempted to set their work in the context of their personal hves a~~ h1ston-
survey of the whole of Classical scholarship in the period he considers, but
cal periods, such as Bernays i8 55 on Scaliger, Mahly 1864 on Polttian, Jebb
only a single technique, Lachmann's method. (:z.) Moreover, he considers ed-
i88:z. on Bentley, Pattison 1892 on Casaubon, Clarke 1937 on Porson, and
itorial procedures not as a timeless set of universally valid rules that were
Grafton 1983-93 on Scaliger; and finally (4) a few introductory textbo~ks
always waiting to be discovered and, once revealed, could now be formal-
discussing methods in an individual scholarly discipline as a whole, which
ized once and for all in a permanent and perfectly systematic arrangement ,
have included important technical histories of that discipline, most notably
but rather as the product of specific human needs as these developed over
Wachsmuth 1895 on ancient history and Traube 1965 (1909-:z.o): 1.1-8o
the course of centuries in frequent contact with other cultural domains.
on paleography. T he differences between these various kinds of works could
Editor's Introduction 17
Ednor's Introduction
reconstitution of which
and it provides a pen- only dispensable witnesses to ancie nr texts, for the
(.3) His book does bear Lachmann's name in its title, they could be sacrificed witho ut loss, but also a precio
us testimony to later
it does not really fit into
etrati ng analysis of many of Lachmann's works , yet understandings of earlier writings, documents of recep
tion that had not only
ical focus is far too wide (for
the category of biographical studies: its histor a hermeneutic value but also a cultur al and historical
dignity of their own.
the present), while
it considers the history of a metho d from ancie nt times
to stimu lus and the model for
It is evident that Pasquali's book provided the
many aspects of Lach-
its thematic focus is far too narro w (for it ignores Timp anaro 's-ind eed, Timp anaro himself acknowledg
es as much when he
question of Lachmann's
mann's life and work s and concentrates only on the writes in the preface to the first edition that his book
is "an investigation
ys similarities in certain
metho d). (4 ) And finally, Timp anaro 's work displa born, one might say, in the margins of Pasquali 1952
( 1934), of which it pre-
Pasqu ali and Timp anaro
respects to Traube's history of paleography, which supposes the re~der 's familiarity and to which it const antly refers" (below,
lon's De re diplomatica
much admired, and Traub e's analysis of Jean Mabil p. 37). Pasquah posed many of the basic questi ons for which Timp anaro 's
closest analogue
/ibri VI (Traube 1965 (1909 -20): i..z.0- 30) is perhaps the study went on to try to provi de more detailed and
satisf acrory answers.
Lachmann; yet Timp a
within Classical studies to Timp anaro 's accou nt of Moreover, the very strucr ure of Pasquali's book bears
a striki ng affinity to
y of any discipline in
naro's mono graph is certainly not a study of the histor Timp:maro's: both books begin with a historical exam
inatio n of the origins
l to emph asize that its objec t is
its entirety, and Timp anaro himself is carefu of Lachmann's metho d and of Lachmann's own use
of it, and then go on to
but only a single and
not even textual criticism as a whole (below, p. 37), assess the limitations of its validity, thoug h of cours e the propo rtions be-
very specific editorial procedure. two cases. Indeed, Pas-
doub t some are due tween the two parts are completely inverted in the
How are we co explain the oddities of this book? No quali's text has even influenced a numb er of the verba
l formulations in Tim-
less no set of expla-
to the individual peculiarities of its autho r, and doubt panar o's-f or exam ple, comp are the following text
of Timpanaro:
ther, or perha ps even to reduce
nations will ever suffice to clear up altoge
unfair to Timp anaro not
significantly, its mysteries. Nonetheless it would be Later -in the nineteenth century, as we shall see, and
unfortunately even to
metho d to the very same
to attem pt to subject his own book on Lachmann's day-t his procedure, which has received the technical
name of climinatio
success to that very
kind of genealogical analysis that he applied with such codicmn dcscriptorum (elimination of derivative manus
cripts), has often be
ion as follows: What are
method. We may therefore rephrase the above quest come a convenient expedient for saving the Classical
philologist time and
ions within which it makes most ation that there is a
the personal influences and scholarly tradit trouble: insufficient evidence, or even the simple observ
sense to situat e The Gc11csis of Lachma1111's Method? mass of rcccntiores (more recent witnesses! alongs ide a manuscript of con-
ali, Timp anaro 's
The most obvio us answe r is of course Giorgio Pasqu siderable antiquity, has too easily suggested that the more
recent ones derived
to the eighteen highly
teach er at Florence and Pisa. Pasquali had responded from the older one. (Below, p. 47 )
(Maa s 1958 [191.7])
condensed and abstra ct pages of Maas's Textkritik
filled more than twice as many
with a detailed and highly critical review that with that of Pasquali:
organ (Pasquali 19i.9) ;
pages in the most prestigious Germ an disciplinary
y mono graph , ten times
five years later appea red the first edition of a length And one must answer that every time there was an ancien
t manuscript on the
longer, which presented a full version of his views,
Storia de/la tradizione e same contents on the
criticism; Pasquali one side and a certain number of recent ones with the
critica del testo (History of the tradit ion and textual other, that generation tended ro derive these latter ones
from that former one
1951. [19341) . Wher e Maas had striven to formu late in its most essential means of proof, or, t~
forma lizing a mechanical and did not hesitate unscrupulously to abuse technical
form an extreme version of Lachmann's metho d, speak more clearly, was satisfied with demonstrat ions devoid of any value.
g manuscripts in or-
set of rules for determining relations of filiation amon {Pasquali 19521 1934) : 26)
sake of textual constitu-
der to permi t mere copies to be discarded for the
to the validity of any
tion, Pasquali insisted on all the factors that set limits It is only if this relation of openly professed depen
dence is borne in mind
copyi sts could be not only fallible
such metho d, reminding scholars that that we can fully under stand the justice of Timp anaro 's decision to conclude
writer s, that medieval
machines but also creative and intelligent readers and the final chapt er of his book with a length y evalu ation of Pasquali's schol-
ity of ancie nt editions,
varian ts might well go back in certai n cases to a plural arship , which , first published less than a decad e
after his master's death,
script copies were not
and above all that medieval and Renaissance manu
Editor's Introduction 19
Editor's Introduction
of the libraries and minds of Timpanaro's parents during his childhood and
moves well beyond the limits that considerations of mere argumentative rel
youth, and they have left an evident trace on all of his own work. Sebastiano
evance might have imposed in order to assume the dimensions and charac
Timpanaro Sr. had already been passionately engaged in the history of sci-
ter of a formal eulogy. ence even before he was compelled to abandon his acrive participation in re-
Yet the evident similarities and filiation between the two books should
search in experimental physics, and afterward he devoted his considerable
not mislead us into mistaking the fundamental difference in their orienta
energies and intelligence to this field above all: he always regarded the his-
tions. It is not accidental that Pasquali's investigation into the origins and
tory of science as a crucial link between science and civil society, a link that
development of Lachmann's method occupies only the first, brief chapter of
had to be vigorously nourished and strengthened in order to prevent a po
his book (Pasquali 195:. \1934}: 3-i:z.), for his principal object is not so
tentially quite dangerous cultural fragmentation, particularly in a country
much the historical evolution of a set of procedures as rather what he con
with a strong humanistic tradition like Italy. And Timpanaro's m;>ther de-
siders to be the facts of the transmission of a large number of mostly ancient
voted her work on ancient Greek philosophy above all to the connections
texts, which he takes to be such that those procedures cannot be applied to
becween philosophical and scientific thought in the period from the pre-
them appropriately and successfully. Ultimately, Pasquali's is a manual of
Socratics through Aristotle, for example, in the case of the Pythagoreans,
editorial technique, addressed above all to this question: given the true na-
insisting against many of her colleagues on those elements in their doctrines
ture of textual transmission, how best are we to edit texts? Its principal dif-
that were more mathematical and naturalistic, less mystical and shamanistic.
ference from other such manuals is that, so far from setting itself the task of
Timpanaro himself cultivated an active interest in the discipline of the
promulgating rules that are to be applied more or less mechanically, it in
history of science, for example, maintaining a fairly close personal relation
sists on the many factors that set narrow limits to the validity of all such
with the prominent Italian Communist philosopher of science Ludovico
rules; the extraordinar y richness of its documentatio n ends up performing a
Geymonat (despite the absolute incompatibility of their political views), col
deconstruction of traditional manuals of textual criticism, but from within
lecting, reading, and annotating his books and publishing two important re-
the genre. Even Pasquali's introductory chapter on Lachmann's method is
views of his works (Timpanaro 1973, 1977c). And his quasi-professional fa.
intended not so much to provide a dispassionate investigation of its history
miliarity with the scientific discipline of historical linguistics is evident on
as rather to deprive it of the authority that, for so many scholars, derived
many pages of the present book. For this presence of a strong scientific com.
from the prestige of his name. Timpanaro's book, by contrast, is the hisrory
ponent in T impanaro's scholarship, it seems evident that his parents are an
of the gradual discovery of a particular scholarly method over the course of
important explanatory factor. To those who have studied their works (ed
centuries; the considerations of that method's degree of validity, which oc
ited expertly by Timpanaro himself), much in the present study will have a
cupy the final sections of his book, are appended to that historical investi-
familiar ring. Thus Timpanaro takes care here to link the historiography of
gation and are not at all indispensable to its argument. Ultimately, Timpa
science with that of society as a whole-but already his mother had writ
naro's is a history of Classical scholarship, differing most strikingly from
ten, "the historian of science is a scientist with a critic's and a historian's
traditional examples of that genre in being organized not as a series of biog
mentality, who feels the need to illuminate scientific creation and to insert it
raphies but rather in terms of a single and highly specific scholarly method,
into the process of scientific thought and of the general vision of the world
to whose discovery a large number of individuals contributed in varying
of its period, with which it stands in a relation of such intimate necessity that
ways and degrees over the co urse of hundreds of years.
if it is abstracted from it that vision ends up not only being distorted, but
In a larger sense, the closest p:irallels to Timpanaro's study of Lach-
also falsified" (Cardini 1951). We can see throughout these pages how fruit
mann's method are to be found not in the field of Classical scholarship after
ful a cautious use of biography can be for the history of science, given that
all , not in histories of the discipline nor in manuals of textual criticism nor
there is no science without scientists-b ut already his father and mother
in biographies of scholars nor in introductory textbooks on method, but m-
had written the history of science in precisely the same way, in terms of in-
stead in the histo ry of science: for Timpanaro's investigation into the devcl
dividuals like Leonardo and Galileo, Alcmaeon and Melissus, and his father
opment of a single method, the most striking analogies arc the sorts of stud-
had declared programmatically, "the history of science ... must present to
ies of the development of an individual scientific technique or concept-the
us the physicists, the chemists, the naturalists, the mathematicians, the as-
integral calculus, the heliocentric theory, the theory of relativity, and the
tronomers, living and working, in such a way that they become as familiar
like-which have been standard fare in the history of science since the nine
to us as we are ourselves. We must live their triumphs, their investigations,
tecnth century. These were the kinds of studies that filled a considerable part
Ednor's Introduction 11
Editor's lncroduction
of his Classical scholarship and the almost inextricable mixture of tradition
their hypotheses in all their details, in all their nuances, in all their energy"
ality and originality in his methodology (his papers are collected in Perosa
(Timpanaro Sr. 1951: 13).
2000). Perosa's catalogue of the 1954 exhibit on Politian at the Laurentian
Between this very narrow context provided by Pasquali and Timpanaro's
Library in Florence (Perosa 1954) provided one extraordinary example of
parents, on the one hand, and the very broad context supplied by the his-
how the methods of a long-dead scholar could be analyzed tersely, techni-
tory of science on the other, it is very difficult indeed to identify a specific,
cally, and nonetheless quite interestingly; and Perosa's seminars at Pisa must
middle-range institutional context within Classical studies proper that can
have supplied many more-Timp anaro himself refers to them with un-
adequately explain the genesis of The Genesis of Lachmann's Method. In-
mixed admiration (below, p. 46). And some of Timpanaro's own earliest
deed, Timpanaro's book is not really very much at all like any of the stan-
and already most polished work belongs very dearly to this line of research
dard forms of the history of Classical scholarship that preceded it. Yet there
(e.g., Timpanaro 1951).
is one field of scholarship that flourished in Italy in the 1940s and 1950s,
In the 1950s, when Timpanaro's Genesis of Lachmann's Method was in
and with which Timpanaro was demonstrably familiar, in which the kinds
the course of gestation, Classical studies in Italy had not yet fully recovered
of questions he raised here and the kinds of strategies he deployed in trying
from Pasquali's death and were languishing to a certain extent. Bue in those
to answer them did have an important place: the study of the scholarly prac
same years Humanist studies were burgeoning there, and it does not seem
tices of Renaissance Humanism. Of course, the Italian Humanists have Imm
at all unlikely that it was from this lively, fresh, and challenging field that
been a favorite object for study by their modern compatriots, but such study
Timpanaro derived some of his deepest inspiration and at least some of the
has usually not been focused on the kinds of precise technical issues that
models and standards for his own highly personal brand of scholarship.
are so characteristic of Timpanaro's book but on other kinds of questions:
This might well help to explain why his book on Lachmann begins, perhaps
biographical, political, literary, archival. Yet already in the 19i.os Remigio
somewhat oddly, not with the ancient grammarians who first invented some
Sabbadini had examined, acutely if only briefly, "the method of the Hu-
of the philological practices that Timpanaro investigates, nor with the great
manists" (Sabbadini 192.2), and even before the Second World War so too
eighteenth-century biblical scholars who first explicitly formulated them,
did a few other scholars, most notably Berthold L. Ullman and P. 0 . Kris-
nor with the nineteenth-century German Classicists who were Lachmann's
teller, who were not themselves Italian but were closely associated with Ital-
most immediate predecessors, but rather with the Humanists of the Italian
ian scholars and institutions (thus Kristeller taught German at the Scuola
Renaissance, and above all with Politian. And it represents a point of some
Normale Superiore di Pisa from 1935 ro 1938, before the Fascist racial laws
difference between Timpanaro and Pasquali, for the latter, for all his love
forced him to emigrate to America). But it was not until after 1945 that
and knowledge of Italian literature and of the history of the reception of the
these scudles of philological method really took off, especially in Italy and
Classics, never quite achieved a degree of familiarity with the Renaissance
England, perhaps in part as a sober and dignified reaction against the ex-
Humanists co mparable to that of his student Timpanaro, preferring instead
cesses of the rhetorical culture of Fascism. These were the years of such
to concentrate on such traditionally literary Italian Classics as Dante, Pe-
pathbreaking studies as those by Jose Ruysschaert on Lipsius and Tacitus
trarch, and Boccaccio.
(Ruysschaert 1949), by Giuseppe Billanovich on Petrarch (Billanovich
Finally, it is not entirely impossible that some role in shaping the young
1951), and by Carlo Dionisotti on Filetico and Virgil (Dionisotti 1958), to
Timpanaro's sense of the history of scholarship~ the nature and impor
mention only a few particularly remarkable examples. For Timpanaro, the
tance of the field, the kinds of questions to ask, the strategies available for
crucial figure in this scholarly movement was certainly Alessandro Perosa
answering them-was also played by the historian Arnaldo Momigliano
(1910-98), who studied atthe Scuola Normale Superiore di Pisa from 1928
(1908-87). Momigliano had left Italy for England in 1939, but after the
to 1932. and then was the administrative secretary there (a post also involv-
war he reestablished intense contacts with Italian colleagues, especially
ing teaching duties) from 1933 to 1953, and was professor at the University
younger ones; in 1964 he accepted a position at the Scuola Normale Supe-
of Florence from 1959 to 1980. Much influenced by his teacher Pasquali,
riore di Pisa, while retaining his chair at the University of London. A close
Perosa turned away in the 1940s from the scudy of Greek and Latin litera-
intellectual and personal friendship developed between the older historian
ture themselves and devoted himself henceforth to research into the Italian
and the younger Classicist during the 1950s and lasted until Momigliano's
Renaissance Humanists- research conducted with the learning and inter-
death; it is documented both in the books and articles of each to be found
ests of a trained Classical scholar. He investigated above all Politian as a Hu-
in the personal libraries of the other (now both part of the library of the
manist philologist and demonstrated convincingly the range and brilliance
Editor's Introduction 2.)
Editor's Introduction
any significant way, that the
Momigliano did indeed influence Timp anaro in
Scuola Norm ale Superiore di Pisa) and in the letter
s exchanged by the two given his tendency scrupu~
decad es, scores of which have latter never acknowle.d~ed this in his writings,
scholars over the course of more than three lously to record all his intellectual debt s-ev en
Timp anaro 's keen aware-
Ricca rdo Di Dona to (Di
been preserved and will be published shortly by ness of his disagr~ements with Momigliano in other
matters would scarcely
anaro only mentions Mo
Dona to 2005). To be sure, in this book itself Timp have prevented him from recognizing such a debt
to him in this one. Viewed
ical footnotes added in
migliano three times, tangentially, in bibliograph from above and outside, the fundamental affini
ties in meth od between the
moreover, nothi ng in the
later editions (chap. 6, n. 6; chap . 8, nn. 37, 38); two scholars are unmistakable, and Momiglian
o is proba bly the most im-
goes beyo nd discussions of spe l figure active in the field of
surviving correspondence between the two porta nt Italian, or indeed European, intellectua
s of intell ectual affinity and
cific scholarly and political issues, and expression the history of Classical scholarship in the late 1950 s who could have exerted
iously took Momigliano
personal cordiality, to prove that Timp anaro consc any degree of formative influence on Timp anaro
. But the likeliest proxi mate
of the 1950s. Nonerheless,
as the model for his own work in the latter pare expla nator y conte xt for Timp anaro 's work rema
ins not Classical scholar-
ived his own work witho ut
it is difficult to imagine that he might have conce ship at all in the narro w sense but the history of
Italian Humanism .
anaro studied closely the
any reference to Momigliano's whatsoever. Timp In any case it should cause no surprise that Timp
anaro 's work on the his-
shed in 1955 (Momigliano
first volume of Momigliano's Contributi, publi to~y of Clas~i~al scho.larship is here associated so closel y with the histo ry of
ing and now classic articles
19 55}, which contained a number of path break science, for 1t is only m a narro w and historically
provi ncial view of the his
history and antiq uaria n re
on such topics as the relation between ancie nt tory of science that this discipline can be thoug
ht to be restricted to the
aphy on the Rom an Empire
search (67-1 06), the development of historiogr ~tudy of the physical and biological sciences. "Science"
and its equivalents
(107 -64), the genesis and
from the seventeenth to nineteenth cenruries m many languages designate any discip lined and instit ution alized effort to
and such impo rtant indi
function of the concept of Hellenism (165 -94), apply the sustained exercise of reason to the study of man and his world
ury historiography as Gib-
vidual figures of eighteenth- and nineteenth-cent and precisely this was certainly Timp anaro 's deepl
y felt and often asserted
-48), Nieb uhr (:z.49-62),
bon (195 -z.u) , Grote (213 -31), Creuzer (:z.33 understanding of the natur e of Classical philology.
Of cours e there are many
the later ones he was writ-
and Oroysen (26 3 - 7 4 ). In all these essays, and in other comp onen ts to the professional activity
of the Classical scholar (as,
formu lating the first version of
ing durin g the same time as Timp anaro was for that matter, of the natur al scientist as well)
besides purely rational
their conta ct became ever
his own study, and in the following years, when ones -tast e, intuition, empathy, experience, to name only these. But in the
of scholarly method and
more intense, Momigliano focused on questions present book Timp anaro concentrates as far as possi
ble on some of the tech-
genesis of concepts and pro
endeavored not only to unravel the complicated nical ~roce.dures of Cla~sical scholarship that are
most clearly an expression
tical interchange between
cedures, deploying for this purpose a subtle dialec of rationality and considers their sources and scope
. After all, there can be
but also to determine the
individual biography and larger social contexts, no doub t that the meth od of examining the filiati
on of manuscripts is a tri-
could renew the methods
exten t to which these concepts and procedures umph of. reason, for it is quite irrational to accep
t rhe readings found in a
ordin ary importance of
that historians use even today. And given the extra manu scrip t merely because it happ ens to be near
ro hand or is old or because
the nineteenth century, it is
German Classical scholarship in the first half of other scholars have accepted it in the past.
to the period and coun try
not surprising that Momigliano kept returning Thus the story Timp anaro tells in this book is that
of the gradual triumph
book -thou gh it is also not sur
that Timp anaro studied most closely in chis ~f reason over the forces of habit, laziness, religious intole
rance, and stupid-
concerned with historical
prising, given that Momigliano was much more ity. For all his ":'aterialism, Timp anaro seems to
ackno wledg e a teleological
rarely touched on Lach-
questions than with philological ones, that he only force th~t certamly does not opera te in the wider
field of huma n history as
mann himself in his own work. do so in chis narro wer tech-
Genesis of Lacbmann's a whole but at least seems ro a certain extent to
It would be tempting co suggest that in The r and endo~ing it
ian answ er co a Pasqualian ques ni~al domain, driving his story forward as a kind of moto
Method Timp anaro provided a Momiglian with a forceful internal dynamic. This may be why
he emphasizes that some
ficati on. Much separated the
tion, but to do so would be a drastic oversimpli figures, such as Gottfried Herm ann, lag behin d the point at which other s
disciplines: beyond their
two men, and not only their ages and academic have. already. ar~i~ed, and why he always takes
pains to isolate as clearly as
ences, Momigliano still
evident and mutually acknowledged political differ possible the md1V1dual steps forward made by this
schol ar or that one be it
a version of the Crocean
retained a certa in measure of comm itmen t to Le Clerc or Madvig or many other s. An expec
tation, if not of the inevi
e all it is extremely odd, if
Idealism that was anath ema to Timp anaro . Abov
Editor's lmroduction
Editor's Introduction
ly worth y of ad
tability, then at least of the strong probab ility of rational
progress within the of human reason, not perhap s the greatest one but certain
ered this metho d and to have
limited confines of a partic ular science seems to underl ie his accou nt and mirati on and investigation, to have discov
bly leads some people
makes it d ifficult for Timpa naro to accou nt satisfactorily
for two kinds of learne d to apply it. Yet the success of a metho d inevita
applying it to cases
exceptions to this progress, which he nonetheless takes
care insistently and to attem pt irrationally to overextend it and to abuse it by
perhap s the true vil-
repeatedly to point out: that progress is not contin uous
but encou nters re- to which it is not in fact applic able-s o already Cobet ,
doing to cast dis-
lapses and regressions, as in Germ any at the beginn ing of the nineteenth lain o f Timpa naro's story, and so too Maas -and in so
d appro priatel y, would have led
century; and that schola rs' theoretical precepts often do not square fully credit irrationally on a metho d that, applie
y, Ernesti, and to entirely accept able results.
with their actual practice (for example, in the cases of Bentle naro's study is
In a certain sense, then, one underlying purpose of Timpa
Orelli). errors and agains t
But the story Timpa naro tells is not a purely intern al one,
in the sense of co rescue Lachm ann's metho d agains t Lachmann's own
suggested that that
involying nothin g more than increasing degree s of ration alizati on of sci- his teache r Pasquali's overst ated polemics. Pasquali had
to discre dit it further by driv-
entific procedures, determ ined only by strictly scienti fic consid eratio ns. It is metho d was not Lachm ann's, buc only in order
aro proves that the
also extern al, for the process he describes is partly imped
ed, but a' so partly ing a wedge between it and his celebr ated name. T impan
it wns not merely
suppo rted and encou raged by the fact that the techniques
of text editio n do metho d is not Lachm ann's, but so as to demonstrate that
a culmin ation of
not exist ' n isolation from the other regions in the larger
cultur e aroun d the invention of a single person , however gifted, but rather
What this book
them but are instead profou ndly influen ced by them. After all, one of the many centur ies of philological insights and hard work.
a shared d iscove ry in which many
texts whose editio n was at issue in this period was the New Testam ent, and shows is that Lachm ann's metho d was
lines, and interes ts all had
questi ons of which readin g to choos e in that text could
easily acquir e not schola rs of different genera tions, countr ies, discip
vered -quite the
only a scholarly weight but also a doctri na l one. T impan
aro takes care, fol- a hand. It was no t inevitable that it should have been disco
were not always
lowing Pasquali, co point to the impor tance of the divisio
n becween Catho - contra ry-an d once it was discovered, its limits and value
reason that the
lics and Protestants, and within Protes tantism to the influen
ce of heretical assessed correc tly-ag ain, quite the contra ry. All the more
here, howev er auster ely, has its own
curren ts, for the field of textua l criticism of the New Testam ent in the sev- story of its discovery, as it is recoun ted
cho ly pathos .
rs were oddly un- deeply human fascination and rather melan
enteen th and eighteenth centuries (altho ugh both schola
ent criticism during
interested in the no-les s-impo rtant role of Old Testam
che same period). So too, Timpa naro explores at length
the relations be
Indo-E uropc an linguis tics, and points
tween Classical textua l criticism and was intran sigent in
to the impac t of the general rise of irratio nalist tenden cies at the end of the Timpa naro was as scrupu lous in his scho larship as he
ntly revised , correc ted,
nineteenth centur y on the value attribu ted to Lachm ann's
metho d. Despite his politics. One result was that this study was consta
began its life in the
che narrow ness of his focus, Timpa naro is careful to
situate his theme as and update d over the course of almost four decad es: it
and went throug h
broad ly as possible: his story is one that moves from the
fifteenth centur y to form of two lengthy articles (Timpana ro 1959, 1960)
and one in Germa n
the twenti eth, from Italy to France to Hollan d to Engla nd to Germa ny and three editio ns as a book in Italian (1 9 63, 19 81, 1985 )
ished, the originally lithe and
back once more co Italy and France, from Huma nists to theolo gians to Clas transla tion ( J 971 ). Each time it was republ
enrich ed {or should
sical scholars. austere line of argum entati on becam e more and more
of the second ary
In his schola rship as in his policie s, Timpa naro may
have been a pro- one say encrus ted ?) by fu rther referen ces to and discussion
tion. For Timpa -
gressivist, but he was very far from being triump halisti
c-qui te the con literat ure that had appea red since the preceding p ublica
l. In genera l it may
crary. He saw rationality as difficult to achiev e and as alway s endan gered by naro 's work was widely discussed and broad ly infl uentia
menta lly contra dicted the find-
the threat ening forces of various kinds of irratio nality -in chis regard his be said that recent schola rship has never funda
confir med and deepened
and enviro nment al- ings of the historical portio n of the book but only
study of Lachm ann's metho d is ac one with his political system atic sections
ist writings. It is after all quite ration al to apply Lachm
ann's metho d to t hem, whereas the positions T impan aro adopt ed in the
hod have sparke d
those cases in which it is appro priate to do so-w here
there is one arche- concer ning the limits o f the validity o f Lachmann's met
conta minat ion-a nd it would be consid erable discussion, contro versy, and disagreemen
type, mechanical transm ission , and no point, somet ime aroun d 1986,
cases, and it is a triumph Anoth er result was that after a certain
quite irratio nal not to do so: there are indeed such
Editor's Introduction Edicor's lncroduccion

Timpanaro, always a perfectionist, but now aged, fatigued, and depressed, edition or translation of reference, followed in parentheses by the original
finally abandoned the attempt to revise his book yet again. A moving testi- year of publication of the work; colon; volume and page number, separated
mony to this decision is the hitherto unpublished essay, a rejoinder to Mi- by a period, if the work comprises more than one volume, otherwise only
chael Reeve, one of his critics, which was presumably written around t986 the page number. The full bibliographical details can be found in the bibli
and appears posthumously in this edition in Additional Materials A, "[fi- ography of works referred to by Timpanaro.
nal Remarks on Bipartite Stemmas]"-and no less moving testimony is pro I have also added a selective bibliography of scholarly works on the var
vided by the fact that Timpanaro chose not to publish it. The reader of this ious topics treated by Tirnpanaro that have appeared since 1986, the date of
volume should be aware that in the years after 1986, Tirnpanaro refused re- the last Italian edition of this book. This bibliography is of course far from
peatedly to allow his book to be translated into English, claiming that it was complete, but I hope it provides at least a starting point for readers who
by now obsolete and that he no longer had the energy required to update it. wish to deepen their acquaintance with these topics and to find out how
It is translated now by the generous permission of his wife and literary ex Timpanaro's views are regarded te day.
ecutor, Maria Augusta Timpanaro Morelli. My own view is that the histo~
ical portion of this work remains almost entirely valid, that the systematic ACKNOWL E DGM ENTS
sections can be enormously stimulating to students and scholars engaged in
the problems of textual edition, and that the fact that Timpanaro himself no At the conclusion of what has been an intense and very agreeable labor, ir is
longer felt able to correct and revise his work as he wished to should not a great pleasure to thank all those who have made it possible.
prevent it from being made available to the many readers who will be able First of all I thank Timpanaro's wife and literary executor, Maria Au~
to benefit from it-and who will be able to decide on their own whether he gusta Timpanaro Morelli, who has supported this project from the begin~
was right, or whether I was. ning, for giving me permission to translate this book and for making avail
The basis for this translation was the last Italian edition published dur- able to me Timpanaro's own copy of the 1986 edition and the unpublished
ing Timpanaro's lifetime; I was able to make use of his personal copy, with text that is included in an appendix. Her hope, and mine, is that with this
a number of marginal annotations in his hand. Considering the importance translation her late husband will continue to find and inspire readers not
of this book and its evolution through a series of different versions, it has only in Italy but throughout the world.
seemed appropriate to indicate that development as tactfully but as clearly I am also very grateful to Mario Teto, an advanced student at the Scuola
as possible. Hence while the numbered footnotes are all Timpanaro's own, Normale Superiore di Pisa, who has assisted this publication in many ways.
the notes signaled by letters have been inserted by myself in order to indi- In particular, he collated the various editions of the work and thereby fur-
cate textual divergences between the final Italian edition and the earlier Ital- nished the materials indispensable for the genetic critical apparatus; he sys
ian and German ones; these divergences are presented in a genetic critical tematized and checked all of Timpanaro's references and prepared the bib-
apparatus in Additional Materials B, "Differences among the Various Edi- liography of works cited by Timpanaro; he helped me to put together the
tions." All divergences in the text that are nor merely stylistic have been in bibliography of items that have appeared since r986 (Additional Materials C,
dicaced; on the other hand, no divergences in the notes have been indicated, " Recent Bibliography" ); he prepared a first. transcript of the very difficult,
as this would have swollen the apparatus beyond any degree of usefulness. and at times illegible, manuscript of Timpanaro's unpublished essay; and he
The reader who chooses to ignore these notes can be confident that the text checked and corrected my English translation with care and elegance. It is a
he reads represents the fully considered views of Timpanaro toward the end source of great satisfaction to me, and would certainly have been to Timpa-
of his life, but whoever dips into their riches will be able to trace out in de naro as well, to know chat Italy is still producing young Classical scholars
tail the development of his thought over a span of more than four decades of the very highest quality.
and to admire all the more his scrupulous honesty and generosity of spirit. Riccardo Di Donato (Pisa) has improved this edition by his sugges-
Timpanaro was generally quite careful in his citations of primary and tions for the introduction and by his generosity in making available to me a
secondary sources, but the method of citation he used, although well estab preliminary transcription of the correspondence between Momigliano and
lished in Italy, may seem haphazard to American readers. I have system Timpanaro, which he has edited and which will appear in :z.005 . Antonio
atized his references in the following way: author's surname (with first ini- Carlini (Pisa) helped with the decipherment of Timpanaro's unpublished
tial or other indication in case of homonyms); the year of publication of the essay.
Editor's Introdu ction
:z.8 Editor's Introdu ction
F. Bornma nn, M. Manfred i,
took place Banolet tiBornm ann Manfrcd iTimpa naro 1970: V. Banolett1,
At a conference on Sebastiano Timpa naro and his parent s that Grassi," Ate11c c Roma, sc:r. 5, r5: 20-1 4
a in Sicily, and S. Timpan aro, eds., " lnediri di Eugenio
on u-13 August i.003 in Tortorici, in the province of Messin BCdier r9r 3: J. BCdier, preface to J. Rcnan, Le lai de /'Ombrc , xxv-xli v. Paris, 1913 .
and his family and presen ted a first ver-
I learned much about Timpa naro BCdier r9:z.H: J. BCd1er, "La tradit1on m:musc rrtc du Lai de /'Ombre , n Romania H ( 1918):
. My thanks to the organiz-
sion of part of the introdu ction to this volume 16r-96 .
Silvia Rizzo, and to
ers, Michel e Feo, Vincenzo Fera, Giacom o Ferrau, and Bernays 18ss: J. Bernays, Joseph Justus Scaliger. Berlin, 1855.
di Lettere e Fi- n of Livy,n Journal of
the Centro lnterdi partim entale di Studi Umanistici, Facolta Billanovich 1951 : G. Billanovich, "Petrarc h and the Textual Traditio
losofia, Universita degH Studi di Messina, for inviting me to this the Warburg and Co11rta11/d I nst1tmes 14: 13 7-108.
eds., Classical Scholarship:
and to all those who participated in the lively discuss ion after my lecture. Briggs and Calder 1990: W.W. Briggs and W. M . Calder Ill,
not only to A Biographical f.11cyclopedio. New York and London , 1990.
Michael Reeve (Cambridge) had the genero sity and kindne ss
t the first Bruscagli and Tdlini t996: R. Bruscagli and G. Tellini, eds.,
Bibliografia dcgli scritti d1
lend his suppor t to this project at an early stage but also to subjec
Lanfran co Caretti. Rome, 1996.
ng criticism
version of the translation of the opening chapte rs to a searchi Bursian 188 5: C. Bursian, Geschichtc dcr classische11 Philolog
ie in De11tschla11d: Vo11 den
Zetzel (Co-
from which, I hope, it has emerged much improved. J. E. G. Anfiinge11 bis zur Cegenwart. Munich , 1885 (reprint: New York,
transla tion in its penulti mate form and L'Eco de/la Scuola
lumbia) read throug h the whole Clrdini r951: Maria Timpan aro Cardini , "La storia delta scienza ,"
made many suggestions for its improvement. Nuova (Turin), Suppl. 3, :z.o March 19 5 1.
h), and c impegno civ1le, ed. Se-
Franz Martin Scherer (Heidelberg), Jan-Di rk Mueller (Munic Card ini :z.001: Maria T1mpan aro Cardini , Tra at1tichita classica
naro's ref- 1.
Antje Wessels (Berlin) were very helpful in trackin g down Timpa bastiano Timpan aro. Pisa, :z.oo
in Tuscany. Cicero 1988: Marco Tullio Ciceron e, Della divi11azio11e, ed.
S. Timpan aro. Milan, 19 88.
erences to Germa n works of scholarship that could not be traced 11e, 4th rev. and updated ed., ed.
introdu ction was read by Giusep pe Cicero 1998: Marco Tullio Ciceron e, Della divinazio
The penultimate version of the whole
Ginzbu rg (UCLA ), Filippo maria S. Timpan aro. Milan, 1998.
Cambi ano (Pisa), Anton io Carlin i, Carlo Essay. Cambridge, 1937.
Claudi a Wassm ann Clarke 1937: M. L. Clarke, Richard Porsot1: A Biographical
Pontan i (Pisa), Lucia Prauscello (Pisa), Mario Telo, Dain 1975 (1949): A. Dain, Les 111an11scrits, Paris, 1975 (1st
ed. Paris, 1949).
greatly from
(Chicago), and Isabell e Wiena nd (Fribourg) and has benefited d'Holba ch I 98 5: r. Thiry d' Holbac h, II buon se11so, ed. and s.
trans. T1mpan aro. Mila n,
n (Princeton)
their observations and criticisms. Above all, Antho ny Grafto r985.
e, which has
subjected it to an incisive, erudite , and deeply generous critiqu Di Donato :z.005: R. Di Donato , ed., Amaldo Momig/iano-
Scbastiano Timpanaro: Car
enormously improved it. teggio. Pisa, :z.00 5.
and Susan virgiliana di M. Filetico, "
My editors at the University of Chicago Press, Alan Thoma s Dioniso tti 19 J 8: C. Dionisorti, '"l.avin i;i venit litora ': Po lcmic;i
and (what is Medioe vale e Uma11is tica iSJ - 315.
Bielstein, were always supportive, imaginative, resourceful, Italia 1 :
History of Classical Schol
Grafron 1983-9 3: A. Grafton , Joseph Scaliger: A Study i11 the
even rarer) patient.
arsbip, :z. vols. Oxford , 1983-9 3.
My heartfelt thanks to all. hte der klassische11 Philo
Gudema nn 1909 (r907): A. Gudcma nn, Gnmdriss der Gcschic
logie, :tnd ed. enlarged. Leipzig and Berlin, 1909 (rn ed., 1907).
Glenn W. Most Ha vet 1911: L. Ha vet, Marmcl de critiq11c verbale. Paris, 1911.
Baratti Florence-Pisa Jebb 188:z.: R. C.Jebb, Bc11tlcy. New York, 188:1..
}tme 2004 Ka ntorowicz 19:z.1: H. Kantoro wicz, f.infiihr1111g i11 die Textkrit
ik. Leipzig, 19:z.1.
Kenney 1974: E. J. Kenney, Tlte Classica l Text: Aspects of f.ditirtg i11 the Age ofthe Primed
Book. Berkeley, Los Angeles, and London , 1974.
Kroll 1919 (1908): W. Kroll, Ceschichteder klassiscl1e11 Philolog
ie, ind improved ed. Ber
ii viaggio nclla Vcnc lin and Leipzig, 1919 (1st ed., 1908 ).
Ascoli 1959: G. I. Ascoli, wNorc lencrari o anisrich c minori durantc VI, trans. and ed. C. Lach-
zia, nella Lombar dia, ncl Picmontc, nella Liguria, nel Parmigi
ano, Modcncse c Pont1fi- l.lchma nn 11150: K. l.achma nn, Lucrctii De rerum natura /ibri
de/la Scuo/a Normale Superiore m:annus. Berlin, 1850.
cio. MaggioGiugno 18 5:z., n ed. S. Timpan aro, A1111ali ed. G. Pacell:a and S. Timpa-
di Pisa, scr. :z., :z.11: 151-9r . Leopardi 1969: G. Leopardi, Scritti filologici (181 7-1832 ),
i, F. Bornma nn, M. Man- naro. Florence , r 969.
Bartole niBorn mannM anfrediTimp:m aro 1961: V. Barrolcn rhe German by B. Flower.
fredi, and S. T impana ro, eds., wlnediti di Eugenio Grassi,n Atc11e c Roma, scr. 5, 6: Maas 1958 (r9:1.7): P. Ma:as, Textual Criticism, trans. from
Oxford, 1958 ( Ti!xtkritik, 3rd ed. Leipz1s , 1957; rn ed. Leipzig, 19:z.7).
1:1.9 - 65.
30 Editor's Introduction Editor's lnrroduction JI

Mahly 1864: J. Miihly, Angelus Politia11us: Ein Culturbild aus der Renaissance. Leipzig, Timpanaro 1965: S. Timpanaro, Classicismo e illuminismo nell'Ottocemo italiano. Pisa,
1864. 1965.
Momigliano 19 55: A. Momigliano, Contributo alla storia degli studi classici. Rome, 19 55. Timpanaro J 969: S. Timp11n3ro, Classicismo e illuminismo nell'Ottoce11to italiano, l.nd ed.
Pasquali 1919: G. Pasquali, review of P. Maas, Textkritik (Leipzig and Berlin, 1917), enlarged. Pisa, 1969.
Gnomo11 5 ( 1919): 417-35, 498-51.1. limp;m3ro 1970: S. T1mpanaro, S11I materialismo. Pisa, 1970.
Pasquali 1951 (1934): G. Pasquali, Storia della traditione e critica del testo. Florence, Timpanaro 1973: S. Timpanaro, review of L. Geymonat, Storia de/ pensiero filosofico e
1951 (1st ed. Florence, 1934). sciemifico, vols. 1-6 (Milan, 1970-71), Belfagor 18: 371-78.
Pasquali 1986a: G. Pasquali, Serini filologici, ed. F. Bornmann, G. Pascucci, and S. Tim Timpanaro l 97 4: S. Timpanaro, II lapsus fre11dia110: Psicanalisi e critica test11alc. Flor-
panaro. Florence, 1986. ence, 1974.
Pasquali 1986b: G. Pasquali, Rapsodia sul classico: Co11tributi all'Endclopedia italiana, Timpanaro 1975a: S. T1mpan3ro, S11I materialismo, ind ed. rev. and enlarged. Pisa, 1975.
ed. F. Bornmann, G. Pascucci, and S. Timpanaro. Rome, 1986. Timpanaro 197 5b: S. Timpanaro, On Materialism, trans. L. G3rncr. London, 1~7 5 .
Pattison 1891: M. Pattison, Isaac Casaubon, IJJ9-1614. Oxford, 1891 (reprint: Geneva, Timpana ro 1975c: S. Timpanaro, 11 lapsus freudia110: Psicanalisi e critica testuale, 1nd ed.
1970). rev. Florence, 197 5.
Perosa 1954: A. Perosa, ed., Mostra de/ Poli:.iano 11ella Biblioteca Mediceo Laurelltiana: Timpana ro 1976a: S. Timpan;uo, The Frc11dia11 Slip: Psychoanalysis a11d Textual Criti
Ma11oscritti, libri rari, autografi e documcnt1: Firenze, 13 scttembrc-30 novembre cism, trans. K. Soper. London, 1976.
1954. Florence, 1954. limpanaro 1976b: S. Timpanaro, " Freudian Slips and Slips of the Freudians," New Left
Perosa 1000: A. Perosa, Studi di filologia umamstica, 3 vols., ed. P. Viti. Rome, 1000. Review 95: 45-54.
rfeiffer 1968: R. Pfeiffer, History of Classical Sclrnlarship from tl1c Beginning to the End Timpanaro 1977a: S. T1mpanaro, La (ilologia di Giacomo Leopardi, 1nd ed. rev. and en
of the Hellenistic Age. Oxford, 1968. larged. Rome and Bari, 1977.
rfeiffer 1976: R. Pfeiffer, History of Classical Scholarship from 13 oo to I 8so. Oxford, Timpanaro 1977b: S. Timpanaro, "Friedrich Schlegel and the Beginnings oflndo-European
1976. Linguistics in Germany," trans. j . P. Maher, in Ober die Sprache und die Weisheit der
Reeve 1986: M. D. Reeve, "Stemmatic Method: 'Qualcosa che non funziona?' " in The lndier: in Beitrag wr Begriind1111g der Altert11mskunde, by F. Schlegel, ed. E. F. K.
Role of the Book in Medieval Culture: Proceedings of the Oxford 1"ternational Sym Koerner. Amsterdam, 1977.
posium, 26 Septembcr-10 October, ed. P. Ganz, Bibliologia 3: 57-69. Timp3naro 1977c: S. Timpan3ro, review of L. Geymonat, Storia de/ pe11siero filosofico e
Reynolds-Wilson 1991 (1968) : L. D. Reynolds and N . G. Wilson, Scribes and Scholars. scicntifico, vol. 7 (Milan, 1976), Belfagor 31: 713-15.
Oxford, 1991 ( ISt ed. Oxford, 1968). Timparutro 1978: S. Timpanaro, Co11trib11ti di filologia c di storia de/la lingua latin.i.
Ruysschaert 1949: J. Ruysschaert, Juste L1psc ct les Arma/es de Tadte: Une methode de Rome, 19 78.
critique tcxtuelle au XVlc sit-de. Turnhout, 1949. Timpanaro 197 9: S. Timpanaro, "The Pessimistic Materialism of Giacomo Leopard i,"
Sabbadini 19u: R. Sabbadini, II metodo degli umanisti. Florence, 19u. New Left Review 116: 19 - 50.
Sandys 1908- u ( 1903- 8): J. E. Sandys, A History of Classical Scholarship, 3 vols. Cam Timparutro 198oa: S. Timpanaro, Aspetti e fig11re della c11lt11ra ottocentesca. Pisa, 1980.
bridge, 1908-u (1st ed., 1903-8). Timpanaro 198ob: S. T imp01naro, On Materialism, trans. L. G01rner, :z.nd ed. enl;irgcd.
Sandys 19 15: J. E. Sandys, A Short History of Classical Scholarship from the Sixth Cen London, 1980.
tury B.C. to the Present Day. Cambridge, 1915. Timp;inaro 1981: S. Timpanaro, Ant1leopardiani e 11comoderati nella sinistra italiana.
Timpanaro 1946: S. Timpanaro, "Per una nuova edizione critica d1 Ennio," St11di ltalia11i P4sa, 1981.
di Filologia Classica, n. s., 11: 41 ~ 81. Timpanaro 1984a: S. Timpanaro, II socialismo di Edmo11do De Amicis: Lctt11ra de/
T1mpanaro 1947: S. Timpanaro, "Per una nuov3 edizione criti~ di Ennio," Studi ltalia11i "Primo maggio. Verona, 1984.
di F1/ologia Classica, n. s., u : n - 77, 179-107. limpanaro 1984b: S. T impa naro, "Freud's 'Roman Phobia,'" trans. K. Soper and M. H.
Timpanaro 1948-49: S. Timpanaro, "Per una nuova edizione critica di Ennio," St11di lta Ryle, New Left Review 147: 4 ~ 31 .
liani d1 Filologia Classica, n. s., 13: 5-58 , 135. Timpanaro 1986: S. Timpanaro, Perla storia dctla filologia virgilia11a antica. Rome, 1986.
Timpanaro 1951 : S. Timpanaro, " Atlas cum compare gibbo (Maniale VI 77)," Rinasci- Timpanaro 1988: 5. Timpanaro, "Otto Skutsch's Ennius," trans. N. M. Horsfall, in y,,
mento 1: 311-18 (reprinted in Timp3naro 1978: 333-43). bo,,11s d1sce11di perit11s: Studies in Celebration of Otto Sk11tsch s Eightieth Birthday, ed.
Timp:maro 1955: S. Timpan3ro, La (ilologia di Giacomo Leopardi. Florence, 1955. N. M. Horsfall, 1- 5. London, 1988.
Timpanaro 1959: S. Timpan3ro, " La gencsi del metodo del L:u:hmann, pt. 1," St11di /ta T1mpanaro 1991: 5. Timpanaro, La "fobia romana" e altri scritti su Freud e Mcri11gcr.
liDni di Filologia Classica 31 (1959): 181- u8. Pisa, 1991.
Timpanaro 1960: S. Timp3naro, "La genesi del metodo dcl Lachmann, pt. 1," Studi lta- T1mpanaro 1994a: S. Timp01naro, N11ovi cor1trib11ti di filologia c storia della li11g11a lati11a.
liani di Filologia Classica 32 (1960): ) 8- 63. Bologna, 1994.
32 Editor's Introduction

T1mpanaro 1994b: S. Timpanaro, Nuovi studi sul 11ostro Ottocento. Pisa, 1994.
Timpanaro 1996: S. Timpan3ro, 011 Materialism, trnns. L. Garner, 3rd ed. London, 1996.
Timpanaro 1997a: S. T1mp3naro, La filologia di Giacomo Lt!opardi, J rd ed. rev. 3nd with Editor's Note to the Reader
addenda. Bari, 1997.
Timpanaro t997b: S. Timp3naro, Sul materialismo, 3rd ed. rev. and enlarged. Milan,
Timpanaro 1001a: S. Timpan:iro, Virgilianisti antic/Ji e tradizione indiretta. Florence,
T imp:1n:1ro 1001b: S. Timpan:uo, II Verde e ii Rosso: Scritti militami, 1966-2000, ed.
L. Conesi. Rome, 1001.
T1mpan:1ro 1001c: S. T1mpan:1ro, "Stemmi trip:inite e lapsus (anrichi e modcmi)," ed.
G. Magn3ld1, IJ Po11te 57, nos. 10-11(October-Novem ber1001): 32.3-30.
Timpanaro Sr. 1951: Seb. Timpanaro (Sr.I, Scritti di storia c critrca de/la scienza, (ed. Se-
basti3no Timpanaro). Florence, 19 51. There are two separate systems of annotation in the present edition:
Tr3ube 1965 (1 909 - 20): L. Traube, Vorlesungc111111d Abbandlrmgcn, J vols., ed. F. Boll.
Vol. 1: Zur Palaograp/Jic u11d Ha11dscl1ri(tcmkimdc, ed. P. Lehmann. Munich, 1965.
1. Notes indicated by superscript numbers or asterisks formed part of the
van Groningen 196 J: B. A. van Groningcn, Traitc d 'bistoire et de critiq11e des textcs grecs.
Amsterdam, 1963.
original Italian edition of Timpanaro's work and are keyed to footnotes
Wachsmuth 189 5: C. Wachsmuth, Ei11lcit1111g i11 da1 St11di11m dcr alt cm Gescl1icbtc. Lcip at the bottom of the page.
zig, 1895. 2. Notes indicated by superscript letters have been added by the presented-
West 1913: M. L. West, Textual Criticism a11d Editorial Tcclmiq11c. Stuttgart, 1973. itor in order to mark differences among the various editions of the Ital-
Wilamowitz 1981 ( 1921): U. von Wilamowirz-Moellendorff, History of Classical Scbol- ian original and are keyed to Additional Materials B, .. Differences among
arsbip, trans. from the German by A. Harris, ed. with ;rn introduction and notes by the Various Editions," on pages 216- 33 below.
H. Lloyd-Jones. London, 1982 ( ~ Gescbicl1tc dcr P/Jilologie, Leipzig and Berlin, 1921;
rev. ed., 1917).
Preface to the First Edition

A first version of this study appeared earlier as Timpanaro 1959, 1960. That
it can now reappear in a revised and enlarged form is due to my friend Gian-
franco Folena and to the director of the Bibliotechina del Saggiatore, Prof.
Bruno Migliorini. To both of them I express my deep gratitude.
The modifications and additions have been more numerous than I antic-
ipated. Nonetheless, the aims and limits of my work remain the same ones
as are indicated in irs brief inrroduction. This is not a history of textual criti-
cism-that would demand a much greater breadth of treatment-but merely
an investigation that aims to clarify the gradual formation and then the cri-
sis of that "genealogical method" that goes under Lachmann's name: an in-
vestigation born, one might say, in the margins of Pasquali 1952a ( 1934), of
which it presupposes the reader's familiarity and to which it constantly
refers. And because many of the methodological problems faced by Lach-
mann, his contemporaries, and his followers are still unresolved, it seemed
to me that it might be useful if I made some contribution to the discussion
concerning them: see in particular chaprers 6 and 8 and the second and third
appendices (appendices B and CJ.
In the course of revising this essay, I received valuable contributions from
Konrad Muller {a textual critic who knows the history of his discipline as
few others do), Fritz Bornmann, and E. J. Kenney. Thanks to them, chap-
ter 5 and the first two appendices appear here in a more complete and cor-
rect version. The third appendix, which is published here for the first time,
owes much to the friendly collaboration of Vincenzo di Benedetto and
Alfredo Stussi. I have also taken account of suggestions and criticisms by
Eduard Fraenkel, Antonio Carlini, Antonio La Penna, Manfredo Manfredi,
and Jean Panvini.
What Eugenio Grassi's friendship meant to me I cannot express ade
quately. An incomparable expert on Greek language and style, animated by
the aspiration to submit all hypotheses to rigorous verification and to clear
the field of the many ambitious and arbitrary constructions in which Clas-
Preface to the First Edition

sical philology abounds, he was a unique guide and judge, for me as for his
other friends: now that we can no longer avail ourselves of his illuminating
criticism, our work has become all the more difficult and uncertain. Besides Preface to the Second Edition
the interpretation of ancient texts (the activity for which he was best titted
by nature), he also made extremely acute contributions to the methodology
of textual criticism: among the posthumous writings published in Grassi
1961, see his observations on the concept of an archetype and the difficulties
it presents. It is to be hoped that those observations, together with others he
has left us, will be reconsidered and further developed by future scholars.

The first edition of this licrle volume, which appeared in 1963 (= Timpa-
naro 1963a), has been out of print for many years. In 1971 a German edi-
tion, Die Entsteh1mg der Lachmannschen Mcthode, appeared in Hamburg,
published by Helmut Buske(= Timpanaro 1971), with additions and cor-
rections due in part to myself, in part to my friend the translator, Dieter
lrmer, who is also an important scholar on Demosthenes and the Greek
medical writers, and to his colleague Volkmar Schmidt. But this second Ital
ian edition is not simply a retranslation of that German edition into the
Italian language. From 1971 until now I have felt the need to deepen and
broaden my research on various points, even without aiming at a degree of
completeness that would be impossible anyw3y. Moreover, two important
works have appeared that, more than all others, have induced me to add
corrections and additions to my work: Rizzo 1973 (an exemplary book,
whose wealth of historical sense and interpretative acumen is far fuller than
the title would suggest) and Kenney 1974, of which a revised and enlarged
Italian edition will appear shortly. Kenney's book, notable for its lucidity of
exposition and sureness of judgment, is a history of textual criticism from
the Renaissance to the present-something never even attempted until now
(this fact was neglected by certain reviewers, learned and acute, but too
sarcastic and inclined to criticize before reading attentively), comprising a
much broader range of material than the present study. As is only obvious,
Kenney has reconsidered many of the arguments and problems I had already
discussed, and has often done so in greater depth. Hence I have derived great
profit for this new edition from his book, as from Rizro's.
It would take too long to list here other recent works that I have found
useful for this new edition: the reader will find them cited in the course of
the volume. One danger I ran, paradoxically, was to take too much account
of all these new contributions-not that they did not deserve it, but I did
not wish my little book, even if corrected and enlarged, to become "bloated"
just in order to repeat badly everything that Kenney and other future schol-
Preface to the Second Edirion Preface ro rhe Second Edicion

ars had said and will say so well: it was supposed to stay within the limits clarify the influences and relations of collaboration and antithesis, to show
of its specific subject matter, which were already explained dearly in the how the "crisis" of the method (a crisis that has deprived it of its absolute
preface to the first edition, reprinted above; and the first two chapters in par validity but has not at all vitiated its usefulness) was manifested not only in
ticular were supposed to retain their introductory character. It was not easy the period after Lachmann, which I discuss in my last chapter, but was al-
for me, in those first two chapters, to reconcile the need to be "streamlined" ready found in tmce in philologists' studies long before Lachmann, and even
with the need to be not too incompletely informative: 1 hope that 1have not in Lachmann himself in certain oscillations and contradictions beneath his
succeeded too badly. Upon certain figures and problems of seventeenth and invariably self-confident and peremptory tone. I had no desire to make liti-
eighteenth-century philology, about which interest has been particularly gious and antihistorical claims for priority, to hunt for "precursors"-even
lively in recent years, I did linger, perhaps a bit too much with respect to the tho ugh it is nor without interest to determine from whom Lachmann de
general economy of the book, even if too little with respect to their impor rived certain ideas and even certain technical terms, and to present, even if
tance: l would like to return to deal with them elsewhere soon, although l only in brief hints, an image of "the man Lachmann" that is free not only of
recognize that the history of the philology of that period is indissolu bly moralistic censure but also of that hagiography of which some of his overly
linked to the history of the movements of religious reform, a field in which zealous disciples made him the object. In my view, it is not at all useless,
it will not be easy for me to acquire an adequate preparation. even in the history of a rigorously technical d iscipline, to take account of bio
To counterbalance my many revisions and addicions, I have performed graphical elements, to seek to "make the characters come alive," without of
only a few little cuts in comparison with the German editio n. Both my friends course making any concessions to an episodic and gossipy biographism.
lrmer and Schmidt, whom l have already mentioned, but especially the fo r My revision and enlargement of this work have been greatly assisted not
mer, had inserted into that edition some notes briefly indicating the results only by works published in these last years but also by personal or episto
of their own research on particular problems of textual criticism. Although lary exchanges of ideas with various scholars. Here too I have preferred to
these notes were very interesting in themselves, they ran the danger of re declare my debts as l go along, in the course of my exposition. But here l
maining somewhat extraneous to the continuity of my exposition; and any wish to mention in particular Giovan Battista Alberti, Severino Caprioli,
way, lrmer and Schmidt themselves late"r developed those ideas in separate Antonio La Penna, Scevola Mariotti, Dante Nardo, Antonio Enzo Quaglio,
works and will go on, I believe, to develop them further. For that reason I Wolfgang Schmid, Alfonso Traina, and Gian Piero Zarri. To Dante Nardo l
have left out some of their contributions and condensed others into a few am deeply grateful not only for suggestions of considerable importance, de
words. To Irmer I have a further debt of gratitude: he was the first to men rived from his experience as a textual critic and student of the history of
tion to me, and then to supply, Lutz-Hensel 1975. philology (the essays that he will publish soon on Giulio Pontedera and
Appendix C, on the other hand, has been much reworked to take acco unt Pietro Canal will show these two Latinists, who are not known well enough,
of recent studies. Among the many reasons that lead to disproportionately under an a lmost entirely new light ), but also for the friendship he showed
increasing the number of bipartite stemmata, I have now emphasized more me by supporting the publication of this edition of my work with the Li
forcefully the one that from the very beginning had seemed to me the most viana publishing house. And an expression of thanks just as warm and af
important one, and I have tried to explain its mechanism better. I wish I fectionate is due to the director of the series, Sergio Romagnoli, for accept
could have been briefer: I do not believe that exclusively and abstractly ing this little book.
stemmatological problems should be overvalued in textual criticism, but, Twenty years have passed since an early and cruel death extinguished at
precisely in order to reduce them to their correct dimensions, it was neces o nly thirty-three years the man to whose memory this work was dedicated,
sary to examine in a bit more depth certain arguments and to unmask cer and is still dedicated. But the memory of his lucid intellect and of his melan
tain specious sophisms. choly, concealed under irony, still remains present in his friends, as do ad
Whoever has seen in this wo rk of mine a desire to diminish the figure of
Lachmann has misinterpreted my intentions. I wanted instead to show how On Ponrcdcra see now Nardo 1981; on Pierro Canal, a distinguished Venetian philol
'' Lachmann's method " was the result of a collective effort in which ocher ex ogisr of rhc nineteenth century, whom rhc character of his studies excludes from rhe scope
cellenr philologists besides Lachmann participated, each o ne with his own of chis liule volume, imporranr studie' edited by Nardo and his students have already been
intellectual character. What mattered to me was to define each one's role, to published and others will follow.
Preface to the Second Edition

miration for what he achieved in such a brief time and mourning for what
he was not destined to accomplish.
S.T Introduction

This edition has sold out in a short time; I have prepared only a corrected
reprinc. I have rectified typographical errors and my own mistakes in the
text (I am grateful to G. B. Alberti, R. Fuhrer, A. Golzio, M. D. Reeve,
A. Rotondo, and A. Stussi for poincing out several to me); I have made a
minimum of additions at the back of the volume, but I have left the treat-
ment substantially unchanged without subjecting it to further revisions that
would have ended up deforming it and destroying its original physiognomy. Of the two parts into which Lachmann divided textual criticism-recensio
I believe that this little book can still serve generally to orient readers and [recension] and emendatio [emendation]-the second had been practiced
that it remains valid in many essential points. In others it is outdated; but since antiquity. In the seventeenth and eighteenth centuries it had also been
perhaps it has contributed toward stimulating new ideas and new research the object of good methodological discussions, to the degree that this was
in a field of studies in which much work still remains to be done in depth. possible considering that by its nature it is an "art" rather than a "science." 1
But other scholars will continue that work better than I can, bringing fresh In the nineteenth century, methods of emendation were refined further (this
energy and new ideas to it.b was especially due to progress in the study of the language and style of the
various epochs and authors): but they were not transformed in a revolu-
tionary way; nor can one say that, as far as divinatory talent isb concerned,
even the best conjectural critics of that century were superior to a Turnebus
or a Bentleyc or a Reiske.
Instead, the great novelty of nineteenth-century textual criticism was the
scientific foundation of recensio. But how this was attained; how much of
"Lachmann's method" should really be attributed to Lachmann, and how
much should be claimed for his predecessors and conremporaries instead;
through what phases Lachmann himself passed in the course of developing
his method-all this still remains to be clarified. The histories of Classical
scholarship say almost nothing about it. Valuable hints concerning these

1. Among these discussions of the art of conjecture before the nineteenth century, the

two best are the Ars critica of Le Clerc t73 0 (1697) and Morel 1766 (= Quantin 1846:
969-1116). The work of Le Clerc also treats some questions thattoday we would assign
to recmsio. On chis see below, pp. 61-63, 68n2.9. On Morel, see Kenney 1974: .44 - 46.
Robortello 1s57, a short treatise, is worth recalling without too much anachronistic sever-
ity, since it represents the 6rst attempt at this kind of discussion (although obviously the
attempt is imperfect, in certain aspects it anticipates the future); on this see urlini t 967:
65-70; Kenney 1974: 19-36. Butthe Middle Ages (twelfth century) had already possessed
in Nicola Maniacutia of Rome an isolated scholar-in several respects an interesting
one-who enunciated certain theoretical princi ples of textual criticism. He is 1hc subject
of a series of studies that are valuable (although occasionally tending toward exaggera
tion): Peri 1967 and 1977.

hisrorica[ matters can be found in some discussions of textual criticism: I

Quentin 19;z.6: 2.7-38 and passim; Dain 1975 (1949): 160-86; Giarratano
1951: 106 - 2.3; and ;ibove all, Pasquali 1952.a (1934), in which, as in all his Emendatio ope codicum from
writings, philology and the history of philology arc closely united (see espe
the Humanists to Bentley
cially chap. 1, "Lachmann's Method," which has already appeared sepa
rarely as Pasquali 1931).? Nonetheless the need is felt for a broader study,
such as Joseph Bcdier called for as early as 192.8.3 I wished to make an at-
tempt; others, after me, will do better. I only wane to warn readers that the
first two chapters have a purely introducrory character; they do not aim at
all.i to trace out a history of textual criticism from the Humanists until the
end of the eighteenth century, but only to isolate some historical presuppo
sitions and partial anticipations of "Lachmann's method."
In the vast majority of cases, the editiones principes [first primed editions]
:z.. In the following pages other writings by Pasquali are cited. Kenney 197 '4 has re- made by the Humanists were based on recent manuscripts, since these were
cently dealt with the subject more fully. The biography of Lachmann (Hern 18 5 1) offers easier to get hold of and more comfortable for the typesetters to rcad. 1
very little on the questions that interest us here. Hence these editions for the most part reproduced a text that had been ad-
3 BCdier 19l.8 . As is well known, the ideas BCdicr proposed in this anicle are highly justed and "prettified" by copyist-interpolators. This text, propagated from
debat:ible (see below, chap. 3, n. :z.3), but he was right to deplore the lack of a study on the
one edition to another, constituted the "vulgate."
genesis of "l.:ichmann'5 method" ( 19:z.8: 163n:z.).
In order to improve and correct the vulgate where it did not seem satis-
factory, one could have recourse either to conjectures or to collation of man-
uscripts considered more authoritative. Classical philologists followed both
chese paths from the Humanist period until the end of the eighteenth cen
tury, some expressing their preference for the one, some for the other. The
two approaches are described by Ruhnken at the beginning of his famous
Elogi111n Tiberii Hemstcrhusii [Eulogy for Tiberius Hemsterhuisl: 2 "There-
fore scholars embarked upon one method of doing Criticism or the other,
depending on the differences among their native talents. Some rashly up
rooted things that were solidly established and should not have been dis-
turbed at all, and harassed things that were certain by means of uncertain
conjectures; others did nothing more than gather together the materials pro
vided by manuscripts." It is easy to see that Ruhnken is indicating the
harmful exaggerations of the two approaches here rather than their positive
aspects: evaluating his hero's achievement in a way that was certainly exag
gerated, he wanted to show that Hemsterhuis had been the first scholar to
harmonize the two rcquiremencs and to establish the true, balanced ars cri

t . Cf. Dain l 97 r 9-49 ): 160- 6 r, who points out 11ery well the dose affinity between
Humanist manuscripts and che first printed editions, in borh external appearance and ed-
itorial technique; Pasquali 195:1.a (I 9Hh 49-50 , 78, and most of chap. 4; Kenney 197 4:
4-5; Rizzo 1n3: 69-75.
:z.. Ruhnken 1875 (1789): 1.
CHAPTER ONE E.nrendatio ope codic11m from the Humanists to Bentley 47

tica [art of textual criticism]. But what must be particularly emphasized is able copies]-recent manuscripts or printed editions-with the genuine
that even the champions of the manuscripts tended not to use as a constant reading of a codex pervetustus [very ancient manuscript) that he has found
basis for their editions the very manuscripts they considered superior; in- in the Laurentian Library or to which some other Humanist has drawn his
stead, they kept che vulgate as the basis, and only had recourse co the man attention.
uscripts where the vulgate was not satisfactory. Thus whac they practiced Up to this point there is nothing substantially new in comparison with
was nor recensio and emendatio, bur rather, according to their own con- the ancient grammarians: for example, controversies regarding textual crit-
ception and terminology, two different types of emcndatio of the vulgate: icism in Aulus Gellius's Noctes Atticae (which constitute the principal model
emendatio ope codicum [emendation with the help of manuscripts] and of the Miscellanea in their compositional structured too) are often resolved
emendatio ope ingenii [emendation with the help of native wit] or coniec by recourse to manuscripts of venerable antiquity (sometimes too venerable
turae [of conjecture).3 to be believable !}.-~ But Polirian often goes on to add a consideration of a
In the Humanist period, Politian was the most rigorous proponent of genealogical nature in order to reinforce his preference for the older manu-
emendatio ope codirnm.~ He has recourse quite rarely to conjecture in his script: recent manuscripts are copies of the older one; hence they do not
first and secondb Miscellanea (nor are his conjectures comparable in bril have the value of an independent tradition.6
liance co Valla's in Livy or Marullo's in Lucretius): for the most part he con- Later-in the nineteenth century, as we shall see, and unfortunately even
trasts the corrupt reading of the exemplaria quae sunt in manibus [avail- today-this procedure, which has received the technical name of elimlnatio
codicum descriptomm [elimination of derivative manuscripts], has often be-
3. As is well known, ancient grammarians already used the term cmendare in the gen- come a convenient expedient for saving the Classical philologist time and
eral sense of "to co rrect," also including in it therefore corrections made as a consequence
trouble: insufficient evidence, or even the simple observation that there is a
of (for the most pan occasional) collation of manuscripts: cf. the words emcndatio and
mass of recentiores [more recent witnesses] alongside a manuscript of con
emendare in the Thcsa11ms, and Pascal l 918; further b1bliograph1cal references in Rii.zo
siderable anciquity, has too easily suggested that the more recent ones de
1973: 15001. So too, the corresponding Greek term 6wp0oiiv and other analogous ones
(era ypalj>E1v, 11eroneEvat, etc.) were used in the same ambiguous sense: cf. Ludwich rived from the older one. Did Politian too sometimes work in this way? In
1885: 93, 104-5. Among ltalianists, even ones of high standing in textual criticism, the the first edition of this book l suggested that he had done so, and even now
wide sense of emendatio has l01sted until recent times: cf. Pasquali 1941: u6 ( = Pasquali I cannot bring myself to exclude this possibility altogether. And yet even
1968: 1.159}, in which a pass01ge of Michele Barbi is translated, as it were, into the mod- then I cited a well-known example of an eliminatio that Politian based on
ern terminology of Classical philology. The distinction between the rwo types of emenda- solid evidence; and now l am inclined to believe, though with some reser
tio is often enunciated in Vettori I 569 (e.g., 30.:u., tit.: "A passage: corrected [. . .}in part vations, that Rizzo is right ( t 97 3: 31502.) to maintain that such cases make
with the help of 01n old manuscript, in part by conjecture"; 36.6; etc.); when he simply says
up if not the totality, at least the great majority.I In chapter 25 of his first
emcndarc or corrigcre without any addition, the: context shows that he generally means
Miscellanea, Politian demonstrates that the Laurentian manuscript 49,7 of
correcting ope codic1m1, in conformiry with his conservative tendency, to which we shall
refer in ;a moment. Among the many other passages from Classical philologists of the
Cicero's Epistulae {amiliares, which had one q uire out of order by an error
fifteenth to eighteenth centuries that could be cited, see also N. Heinsius in the p reface to
his edition of the works of Ovid (Heinsius 1661: without page numeration): "which I h01ve 5. Cf., e.g., Aulus Gellius t.7; t.11.2; 1.3.5; 9.14; 11.10.6; 13.11. 16; 18. 5.11. More
already emended in many pass01ges, in pan on the authority of old manuscripts, in part by generally on the use of manuscripts by Greek and Latin grammarians, cf. Lehrs 1882:
relying upon the guidance of my native wit alone." Sabbadini 1910: 56 - 6o is still useful 344- 49. Oscil101tions between age and large numbers of manuscripts as 01 criterion fo r
for providing a prdimin:iry orientation on the champions of conjectures and of manu- judging were not lacking; on Galen's ideas on this subject cf., e.g., Bri>ckcr 1885: 417.
scripts in the Humanist period; some more: detailed observations now in Kenney 1974: 6. E.g., Misc., chap. 5 ("an extremely old manuscript of the Argonautica of Valeri u~
16- 17 and chap. 1 p01ssim. Fl01ccus I. ..) from which I think all the other available ones are derived"; cf. the passage
4. During the l:ist ten years, studies on Politian and editions of his unpublished writ of the second Miscella11ea cited below, n. 8), 15 (to which we shall refer shonly), 41, 89,
ings have flourished intensely; this tendency is still in full swing. This is not the right place 93, 95. Cf. the famous subscriptio to the incunable containing the Srlvae of Statius inter
for a full bibliogr:iphy; I therefore limit myself to referring to the lively synthesis of Graf alia: "I have come upon a text of the Silvae of Statius [... j from which alone all the other
ton 1977a (with an upto-date bibliography, and rich in references to earlier and contcm available ma nuscripts seem to have emanated, though it is full of errors and corrupt and,
porary Hum01nists). I also wish to recall with undiminished gratitude the seminars of Ales what is more, as I think, reduced to half its o riginal extent" (cf. Perosa 1955: l 5, and be-
sandro Perosa at Pisa on the first Miscella11ea, which I had the good fortune to attend at low, n. 9; this is not the place to linger on the very controversial problem of just which
the beginning of the 1950s. In what follows I shall cite other scholarly contributions. manuscript Politia n s01w); and the analogous s11bscriptio to Apicius (Rizzo 197 3: 3 t 5n2).
CHAPTER ONE mc11datro 011e codic11m from the Humanists to Bentley 49

of binding, is the ancestor of a group of more recent Laurentian manuscripts with clarity ' the concept of an archetype, which explains the corruptions
in which the same disturbance in the order of the letters is found without shared by a whole tradition. In fact, despite certain professions of an ex
this being explicable by a displacement of the quires. 7 He eliminated the tremcly conservative creed ("I prefer to be in error with old manuscripts
apographs of an old manuscript of Valerius Flaccus by analogous reason rather than out of excessive fondness for my own ideas"),12 Vettori is any
ing, as the second Miscellanea now prove.11-8 Thus Politian not only vener thing but uncritical in his actual philological practice: his defense of the
ates h the oldest manuscripts in general: he also has the beginnings of a his manuscript tradition is almost always the result of a deeper interpretation,
torical understanding of manuscript traditions. He is also aware that when a better understanding of the writer's style. This is true especially of Cicero,
a conjecture is necessary, it must find its starting point in the oldest stage the author whom Vettori knew best; it is somewhat less true of the Greek
of the tradition that we can reach, not in the deceptive patchings-up that texts he published, in which he sometimes really did fall into excesses of
corruptions have undergone in the more recent manuscripts (Misc. chap. 57, conservatism. l l
cf. chap. :z.o)-a criterion that will not be fully recognized until the age of But, even if only in passing, Erasmus had already arrived at the use of m
Lachmann. the concept of archetype for the purposes of emendatio. In his Adagia 14 he
What is more, Politian already understood char the manuscripts {at least proposed a correction to a proverbial expression cited in Aristotle's Meta-
rhe oldest and most valuable ones) had to be collated not occasionally but physics, u and observed, "The agreement of the manuscripts will not seem
systematically, registering all the readings that diverged from the vulgate at all astonishing to those who have even a modicum of experience in as
text, induding those that were certainly erroneous but that might turn out scssing and collating manuscripts. For it very often happens that an error of
to be useful for restoring the text. This is the criterion he asserted in the sub- one archetype, so long as it has some specious appearance of the truth, goes
scriptio11cs to the writers De re mstica, to Pliny, Starius, Pelagonius, and on to propagate itself in all the books that form as it were its descendants,
Terence; he had a full and justified awareness of its methodological novelty, 'and the children of its children and those who are born later."' 1 ~
even if earlier Humanists and, probably, medieval scribes had already begun We must linger for a moment on this passage, which I discussed inade-
to apply it.~ In this regard he was a precursor of Ernesti and Wolf {see be- quately in the first edition of this work, in part because of a banal error of
low, pp. 71-74) and was already beginning to overcome the erroneous con- interpretation (corrected by Rizzo 1973 : 316). Scholars used to think that
cept of eme11datio ope codicum, which implies that collations arc made not
constantly but only occasionally. 12.. Vertori 1571: 166. C f. 7 1: "Since by narure I was always shy a bout changing any

Politian's orientation toward conservative textual criticism, his polemic thing r ashly in someo ne else's wr irings."
against the copyist-interpolarors,i and his tendency to bclittlck recent man 13 . On his edition o f the Agamem11011, cf. Fraenkcl 1950: 1.34- 35. On Vcrtor i's per
uscripts as copies of a vctustissimus still preserved recur in Pier Vettori, with sonality in general, which even now is no t su fficien tly known (some indicarions m Gm fton
19 75: 162.-79 ), we expccr much new informarion from Lucia Cesarini Marrinclh's re
greater support from arguments and examples. 111 Always disinclined to make
search. O n a group of imporranr students and followers of Vecrori, who distinguished
conjectures, he is especially hesitant when the old manuscripts unanimously themselves i11tcr alia in climi11atio11l!$ dcscriptorum following in the footsteps of Polman,
attest a reading: "I scarcely think that all the old manuscripts can be subject see Grafton 1977b: 17 5-76; and, with particular reference to rhe Corpus iuris, before and
to the same error. " 11 This was a legitimate hesitation until scholars attained after Vettori, Caprioli 1969 (Caprioli's fundamcnral study extends into the sixteenth ccn
tury as well), and Troje 1971.
7. In rurn Laur. 491 7 derives from Laur. 49. 9 of the mnth century. Politian noticed 14 Erasmus 1538 ( 1500 ): 209 (Chilias I, century VI, adage 3 6). In the first edition
this too !see the beginning of the same chap. 2.5), but he did not take the time to supply (Pam, 1500) this ad:igc is not yet to be found; ir appears for the first time in the Vcnii;c:
the evidence: "The fact that this one is copied from that one is d ear from many proofs edirion of 1 508 1 as N. G. Wilso n informs me.
which I shall omit now." The whole question is well clarified by Kirner 190 1: 400-406. 15. Mctaph. a 993b5: TiS' civ lh)pas- ci1ui pTm; i.e., "Who would err rrying to hit (a rar
8. Politi:m 1972: 4.6-7 (chap. 1); cf. 1,23 and n. 45; Branca t973 =347-51. get as btg as) a door?" The expression, Alex:inder of Aphrodisias explains, Ei X1J1TTm om)
9. Cf. the passages cited by Perosa 1955: 15, :n, 16, 38, 66. ,..;;,, nitoTwv Twv i rri oKom'w To{EU<i1rrwv. Erasmus wanted to correcr OupaS' 10 OqpoS',
10. Among the many passage~ that could be cited, see his preface to the Epistulac fa wrongly: cf. Leutsch 18 5 1: 6 78.
miliares of Cicero (Vettori I 586 [ t ssSI: 69-701. Vettori writes illter alia: "These perverse 16. So (or else, certainly erroneously, Ka( Tm I the ed itions of the Adagia available to
correctors have inflicted u pon them (sc, writers) wounds no fewer than those dealt them me. The Homeric texr (I/. 10.308) has Kai rro(b,:;,v rrail>ES', rni KEii, ere. The accusative
by time itself and the ignorance of earlier i;enturies. " noiOoS' is an ad:iptation required by the conrext in Erasmus; the o ther changes will have
1 1. Vettori 1540: S14. been errors of citation from memory.
so <; HAPTER ONE Eme11datio ope codicum from the Humanists to Bentley

the Humanists (like the ancients before them: d. Cicero, Ad Att. 16.3.1) was an insidious error, one with an appearance of truth (fucus), so that it
meant by the term archetypum or codex archetypus only the "official text" did not occur to them to correct it. Erasmus was certainly thinking of an an
checked by the author and intended to be published afterward in further d ent archetype in that passage from the Adagia, since he is not embarrassed
copies. A wider and deeper examination (Rizzo 1973: 308-17) has made it by the fact {which he notes expressly) that Alexander of Aphrodisias had al-
clear that alongside that meaning (perhaps the prevailing one) the term also ready read the presumed error 0upas in his copy of Aristotle. But, in gen-
has many other usages in the Humanist age, among them the one that will eral, he shows that he believes it to be impossible for "crude" errors such as
go on to prevail later, namely, that of a manuscript-even if it is later than meaningless expressions or lacunas m spread through the entire manuscript
the author by many centuries, even if it has been preserved by chance and is tradition: subsequent copyists would have noticed these and would have
devoid of any "official" quality or authority, even if it is disfigured by errors tried to heal them. Erasmus's conception of textual transmission was almost
or lacunas-from which all the others are derived. It is in this sense that too unmechanical-it was appropriate only for certain traditions.n
Merula, in the preface of his Plautus edition of 1472, applies the term arche The great French Classical philologists of the sixteenth cenmry, from
typum to the lost unus fiber [single book] from which the copies still extant Turnebus to Lambinus, to Daniel, to Pithou, all felt the requirement to seek
of the Plautine comedies are derived {Rizzo 1973: 314), even if he introduces out old manuscripts and to use them for their editions, but Joseph Scaliger felt
the word with a velut (as it were], which gives it an almost metaphorkal it more than anyone else. A conjectural critic much less gifted and spontane
meaning; and Politian uses the same term in his commentary on Statius's Sil- ous than Turnebus, hostile to collections of occasional adversaria, rending
vae for Poggio's manuscript, which he had seen and judged mendoS11s [full to compose organic works already in the first period of his activity, 19 ani-
of errors] and dimidiatus [reduced to half its original extent].> 7 What is still mated by a historical spirit far more than by a taste for the interpretation of
lacking, it seems to me, before we arrive at the "Lachmannian" usage of individual passages, he was the first to set himself the problem of recon-
archetypum (or of the adjective archetypus ), is the limitation of the term to structing a medieval archetype (even if he did not use the term archetype to
lost ancestors alone and, what is more, to ones distinct from the original or designate it~. In his Castigationes in Catullum he believed that on the basis of
official text. Politian himself applies the term archetype more than once to the corruptions in the apographs he could establish that their ancestor was
the Pandects first preserved in Pisa, then in Florence, that is, to a manuscript written in pre-Caroline minuscule.20 His demonstration must be considered
that was still extant and that he thought was one of the official rexes Jusrin- a failure,1 1 but in any case the very attempt is quite interesting. Scaliger too,
ian circulared to various dties. 18 following in the footsteps of Politian and Vettori, cook issue with fifteenth
In any case, the importance of the passage of Erasmus we have cired con-
sists not in his application of the term archetypum to a lost common copy- 19. See his letter to Janus Dousa of 1594 (Scaliger t6z7: p.I: "We have m<1de many
as we have seen, at least Merula had preceded Erasmus in giving this mean- observa[ions on authors in both languages, which could give birth to a monstrous prog
eny of Variants, Old Readings, Miscellanies and other things of this type, with which the
ing to the term, even if with a somewhat cautious wording; and probably
ambition of Philologists is nowadays wont to run riot.. . But in order fo r our sleepless
Erasmus too considered legitimate the other meanings commonly given it in
nights to bear fruit, we undertook to interpret and purify ::1uthors ::1s wholes." This de
the Humanist period-but rather in his energetic affirmation of the right mand for wholeness was later developed fully in the great historical and ' hronological
to correct a reading that appears erroneous without allowing oneself to be works of his more mature years. Cf. Bernays 1855: 46- ..p.
intimidated by the comensus codirnm (consensus of the manuscripts] (as, i.o. Scaliger 1582. (1577 ): 4 of the Castigationes: "Moreover, I suspctt that that Gal
even after Erasmus, Pier Vettori allowed himself to be, as we saw just now): lic copy was written in Langob<1rdian letters, since the errors that have been disseminated
it was not the case that each and every copyist committed the same error in- in later manuscripts by inexperienced scribes seem to have arisen precisely from 1hose
dependently of the others, by an improbable phenomenon of polygenesis crabby characters, as we shall indicate carefully in the proper place." By ~ Langobardian
spreading through the whole tradition; instead, it was a single copyist who letters" we must understand here pre-Caroline minuscule and not Beneveman script, as is
made clear by the fact that on pp. 2.3 and 7 3 Scaliger hypothesizes errors due to the con
was responsible for the error, and subsequent copyists repeated it because it
fusion between a and u: "'because lhere is no difference between these letters in the Lan
gobardian script" IP i.3); "'11 and a are 1he same in Langobardian characters" (p. 73 ). On
17. Politi3n 1978: n.16; cf. 10.13-17; <1nd above, n. 6. the rather broad usage of the term Langobardian letters in the Humanist period, cf. Casa
18. See the pass3ges cited by Rizzo 1973: 313 and 313ni.; and, for other Humanists massima 1964: 566-67; Rizzo 1973: 12.2.-2.J. For other confusions of letters that Scaliger
who were in contact with Politian or underwent his influence, Caprioli 1969: 393-404 considered significant, cf. below, Appendix B, n. 3.
and passim. 2.l. See again Appendix B, n. 3.
s:z. CHAl'TER ONE Eme11datio ope cod1c11m from the Humanists to Bentley

century interpolators, who gave particular offense to this enemy of stylistic Knowledge about manuscript material was greatly increased by the work
allurements and lover of the austerities of archaic Latin; 22 but he understood of the Dutch Classical philologists of the seventeenth century, in particular
that the old manuscripts too were contaminated by corruptions that had by Nicolaas Heinsius, who, as is well known, examined an enormous num-
to be healed by conjecture: "Just as more recent editions must be weighed ber of manuscripts during his travels throughout Europe and made colla-
against old copies, like gold at an assay, so too the manuscripts must be cor- tions that even today are admired for their exactitude. 26 He was also able to
rectly weighed in the scales of judgment." 2 ' And on the basis of the perfectly indicate the best manuscripts for many texts; and not only did he have a
legitimate hypothesis of an archetype in rather bad shape, he felt himself all clear concept of what a medieval archetype was, as Scaliger already did, but
too authorized to transpose sections of poems, especially of Tibullus, so as to he was also able to distinguish two families of manuscripts in the manu-
provide them with a logical order. 24 Caution in recensio, excessive boldness script traditions of Curtius Rufus and Prudentius, though the method he ap-
in ememiatio: we shall find the same contrast once again in Lachmann.""' plied in doing so was not entirely rigorous. Yet the fact that, as Kenney has
From his Italian predecessors and contemporaries, especially from Vet observed, he took as the basis for his own editions of Ovid the text that his
tori, Scaliger derived and developed further the demand for a complete col- father, Daniel Heinsius, had prepared for his edition of 1629, indicates that
lation of the manuscripts in his edition of Catullus and in the edition of Va his point of view was somewhat obsolete.'.27 And his love for facile and el-
lerius Flaccus he scarcely sketched out; later, in his edition of Manilius, he egant Latin versification, more Ovidian than Ovid himself, led him not only
returned to the practice of merely occasional collations.q,H often to propose conjectures for rhe sole purpose of beautifying the text but
also more than once to prefer brilliant but specious readings found in recent
:z.:z.. See, e.g., the Castigatio11es (Scaliger 1581 Ir 577 j: 105}. For Scaliger's fond ness for manuscripts. Later he worked with a much more cautious methodology on
archaism one may recall his Latin version of the O rphic hymn$ and his famous judgment the text of Petronius: he understood that many "anomalies" were due not
on Enn ius (quoted by Bernays 1855: i.8i.).
to corruptions but to that author's unique style. 2H But here we find a limit-
13. In the Prolegome11a to Scal1ger 1600 (1579): 8.
ing c;ase: no one could have dreamed of "Ciceronianizing" the Latin of the
14. Heyne 1817 (r 755 I: xviii- xix protested against these transpositions even before
Satyricon, so enormous was its difference from so-called Classical Latin, and
Haupt 1875- 76: 3.30- 41 did.
On Scaliger, A. T. Grafton (of whom I cited a preparatory work, Grafton 1975) has not just in the Cena. Probably the scholarly disagreements about N. Hein-
now published volume 1 of a large-scale complete study: Grafton 1983 (important for sius and the divergent assessments of him for all the general recognition of
Scaliger's precursors too). The second volume: is eagerly awaited. To the disagreement that his greatness, now as tn the past, derive from the fact that he was a transi-
I had expressed in n. i. 5 below, Grafton replies couneo usly, repeating his thesis ( r 98 3: tional figure, perhaps more conspicuously than others-half a Humanist in
3 i.o), and points out to me that the chronological computations Scaliger had to perform the restrictive sense of the term, half a Classical philologist aware of new re
for the great works of his second period of activiry demanded just as much patience and quirernents.29 The defects of Heinsius the "Humanist" were magnified in
attention of him as had been required for doing complete co llation5 of manuscripts. This
j5 quite true; but that patience and attention were a conditio si11c qua non for the great
chronological works to which Scaliger devo ted hi mself, whereas a critical edition could che reason for this methodological regression to the worsening of relations between
also be prepared by relying solely upon conjectural criticism and neglecting 10 do a sys- Scaliger and Italian Classical philologists, a5 Grafton suggests. On the other hand, it is true
rematic reccnsio first, even despire an awa reness of the risks to which such a procedure chat Scaliger's taste for textual criticism had not weakened at the time of his Manilius edi-
would give rise (see below, p. 75). It is beyond dou bt that Scaliger's relations wirh Vettori cion (Grafton 1975: 17 ;; and already Housman in a letter dted by Grafton, ibid., n. 71 I.
and his followers were ruined both by personal reasons and by Vetrori 's conservatism as Instead, I would think of a decrease in the patiem:e and powers of concentration necessary
a textual critic (cf. Grafton 1983: 184-86), but it continues to seem impro bab le to me that for complece collations, which can become weaker even when a scholar is still very inter-
this was the reason why Scaliger regressed methodologica lly and became convinced that ested in problems of textual criticism.
sysrcmatic collations of rhe manuscripts were superfluous (to rhe co ntra ry of what he had 16. Cf., e.g., Munari r95oand 1957; further bibliography in Kenney 1974: 59n51; the
believed when he had edited Catullus). To be sure, as far as I know he never provided a article by M. D. Reeve announced by Kenney has appeared, Reeve 1974; cf. Reeve 1976.
cheorerical justification for this regression. Cf. also what I point out in chap. :i., nn. 3 9 and '2.7. Kenney r974: 6:1.-63. Cf. the objections of Grafton 1977b: r73, and rhe reply of
47, concerning Ernesti and Wolf. Neicher of rhem repudiated the principles they had ex Kenney 1980. See also the formulation, directed substantially toward the past and nor the
pounded with such lucidiry, yet they did not have the self-consistency to apply them in all future, which we cited above, n. 3, from Heinsius's preface to his Ovid edition.
cheir editions. Cf. also Jocelyn 1984: 60,P :i.8. Cf. Blo k 1949: '2.46-53, esp. 147-51.
15. For Catullus, cf. Grafton 1975: r58-61; for Valerius Flaccus, Waszink 1979: 81 '2.9. Such contradictions are well brought out by Kenney 1974: 57-63 {and they are
and n. 11; for Manilius, Grafton r975: 174-76. To me it seems implausible to attribute repeated in Kenney 1980). The characrcrization of N. Heinsius given by L Miiller 18691
Eme11datio ope codic1m1 from the Humanists ro Bentley

other Dutchmen, for example, in Jan van Broekhuizcn (Broukhusius). And stitution of the text": this judgment of Lachmann's, repeated by other first-
in general in the seventeenth century, and to some extent still in the eigh- rate scholars,x32 is substantially true, even if it runs the risk of giving too
teenth, it was more in breadth than in depth that the examination of manu- simplistic a picture of Bentley's mode of operation, in which the recourse to
scripts made progress. An extreme example of this tendency is the Jesuit manuscripts (in the case of Horace, often recentiores) and the work of con-
Girolamo Lagomarsini, a pure and simple collector of variants drawn from jectural emendation did not follow first the one and then the other, but were
innumerable manuscripts and printed editions of CicerolO (whereas an- interconnected, and indeed for the most part the latter preceded the for-
other Italian student of Cicero, Gaspare Garatoni, was later to demonstrate mer.r.n But Bentley's healthy distrust for the vulgate was illuminated almost
a much superior natural talent for Classical philology). With only a few a hundred years after his death by none other than Lachmann, who discov
exceptions, which the progress of research may well reveal to have been ercd it by starting out from the field in which it was most visible, New Tes-
more numerous,u the hints at a history of the manuscript tradition which we tament criticism (see below, pp. 63f., 85f.). For the most part, this aspect of
have noted in Politian, Erasmus, and above all Scaliger, were not developed Bentley's activity as a textual critic remained concealed from the Classical
very much. 31 philologists who were his contemporaries or immediate successors by an-
Richard Bentley, the most brilliant Classical philologist berween the end other, more conspicuous one: his conjectural criticism, extraordinarily in
of the seventeenth century and the first decades of the eighteenth, and one genious (just think of his emendations to Callimachus and Manilius),' but
of the most brilliant Classical philologists of all times,w possessed a great ca- often rash.34 He himself contributed to this impression with some of his vig-
pacity for evaluating manuscripts (even if not for reconstructing their ge- orous pronouncements, like the celebrated "For us, reason and the facts
nealogy) and for distinguishing genuine readings from interpolated ones. are worth more than a hundred manuscripts," 35 or like the perhaps even
"No one who knows Bentley well will doubt that a new editor of Horace, more characteristic passage in the preface to his edition of Horace in which
once he has eliminated most of Bentley's conjectures (what in fact is not dif- he maintains that conjecture, precisely because in it the Classical philolo-
ficult), will find that after him he has almost nothing left to do for the con- gist's personal responsibility is entirely at stake, ends up yielding more se-
cure results than accepting the transmitted reading or choosing berween
variants does.hh,Jli Against a lazy and uncritical adherence to the vulgate or
S 1-54 is incomplete, but charming for its liveliness :ind perceptiveness, :ind hence still
worth reading. Blok 1949 is fund:iment;il for its richness of documentation and for hav-
3:1.. lachm;mn 1876: :1..:1.53nt (= Lachmann 1830: 8:1.ont). Cf. Wilamowitz 198:2.
ing brought to light new aspects of Heinsius's personality; the title, N. Heinsius in Dienst
(19:1.1): 8t; Kenney 1974: TL.
11a11Christina11a11 Zwede11, is infelicitous in;ismuch as it suggests primarily a biogr;iphy of
H On this point cf. now, with gre:ic precision, Brink t978: 1141-48, esp. 1147-48.
Hcinsius as a diplomat; the book is wrinen in Durch, so my consultation of it w:is labori
Bur I believe that Brink exagger:ites in trying to reh:ibilitate Bentley's Hor:itian conjectures
ous and incomplete. The m:iin results of Blok's book :ire summarized by Waszink 1979:
and in maintaining, in the footsteps of Paul M:ias, that the transmitted text of Hor:ice is
7:1.-73 , 8:1.-83, and arc sh:ircd by Gr:ifton 1977b. I cannor get rid of m y i mpres~ion that
extensively corrupt (see :ilso below, n. 37 ). Besides, it is not true th:ir Wilamowin (as Brink
its tendency is excessively :ipologetic, even if in :my case, I repear, HeinsiU5's gre:itness is
reports 1978: 1144) said that the text of Hor:ice "has no need" of conjectures but rather
not in question.
th3t it has "very little need" ("verschwindend wenig": cf. Wilamowin 19:1.7: 36): the d1f
30. Nardo 1970: 147-48, H-S8 demonstrated that the majority of the so-called
ference (due perhaps to the ltali:in translator of Brink's English text?) is not negligible.
Lagom:irsini:in manuscriprs are in re:1lity old printed editions, for the most part derived in
H On Callimachus, cf. Pfeiffer 1976: 153, 3nd Pfeiffer 1949-53: :1..xliv-xlvi (sec :ilso
turn from inforior manuscripts.
31. One of these exceptions, pointed out to me by D:intc Nardo (cf. now Nardo Hemmerdinger 1977: 490-9:.). On Manilius, Housman 1937 ( 1903): xvi-xix. This is not
1981 ),is constituted by Giulio Pontedcr:i, a distinguished bot:inisc of Vicenia who dedi the place ro linger on the splendid conjectural contributions of Bentley to these authors
cared himself to the study of :intiquity and m:idc excellent conrributions to the texts of the and to many others.
L:itin agricultural writers (C:ito, Varro, Columcll:i, P:ill:id1us), for the most part defending 35. In his edition of Horace (Bentley 1711), note on Carm. 3.:1.7.1 s (for ;inalogous
the readings of the m:inuscripts against the vulg:itc :ind thereby developing further that pronouncements by other Classical philologists, cf. Kenney 1974: 41n:1., 99). In point of
fact, to those words Bentley added, "cspeci31ly with the further vote of the old Vatican
hnc of research, which had been inaugurated by Politian :ind Pier Venori, and in these very
manuscript." Bur in any case the vetat he supported against the vetet of the better tradi
same authors. The importance of Ponteder:i's studies (see esp. Pontedera 1740) was no-
ticed by Johann Gottlob Schneider, who in his edition of the Scriptores rei ntsticae :ilso re tion is miscaken.
36. Bentley 1711: :1. (preface): "In these Hor:itian labors, then, we offer more readings
produced his posthumous Epistulae ac dissertatio11es ISchneidcr 1794-96: 4.:1., d. 1.vii-
viii); m:iny of his contributions, mediated by Schneider, have been :iccepted by more by means of conjecture than with the aid of manuscripts, and, unless I :im entirely mis-
taken, for the most p:irt more certain ones: for when there are variant readings, :iuthority
recent editors, but until now historians of classical scholarship have neglected him.
CllAPTER ONE 111e11datio ope codirnm from rhe Humanists to Bcorlcy 57

to the first reading offered by just any manuscript, this argument had a cer- The English Classical philologists of the second half of the eighteenth
tain degree of truth (cf. Kenney 1974: 71-73); but it tended to set up as the century and the first half of the nineteenth who were inferior to Bentley in
goal of a critical edition not the historically most probable text but the best ralent and breadth of horizon but nevertheless followed his powerful ex
text that the editor's taste and mentality could imagine. I do not believe that, ample in textual criticism (Musgrave, Porson, Dobree, Elmsley) were above
as some have said, Bentley's edition of Milton's Paradise Lost, which is full all conjectural critics, endowed with a refined knowledge of linguistic and
of arbitrary conjectures, is evidence of his senile decline or of his lesser fa- metrical usage, especially regarding the recitative parts of Greek tragedy and
miliarity with English poetry than with Latin and Greek poetry (even though comedy. But they also fd t the need to check the manuscripts. If Porson " was
there may well be some truth in this latter explanation); instead, I think we conditioned by the focr that he never stirred from England, where only re
should seriously consider Brink's somewhat paradoxical suggestion (Brink centiores manuscripts" of Euripides1x and the other tragedians existed, be-
1978: n61-64) that Bentley's Miltonic conjectures, misguided on the level fore him Samuel Musgrave ventured as far as Paris and collared two impor-
of textual criticism, were an indirect form of literary criticism, opposing a tant manuscripts of Euripides there, and after him Peter Elmsley went to
different taste to Milton's and to the corresponding poetic language. But, in haly, studied the Laurentian manuscript of Sophocles (he was the first to rec-
my opinion, something similar also happens in Bentley's edition of Horace, ognize clearly its superiority),-1'" and collated and evaluated Vatican manu-
even if to a lesser degree: in the overwhelming majority of cases, Bentley's scripts of Euripides, for the most part correctly.~ 0 To Elmsley we also owe
hundreds of conjectures on Horace are "corrections" not of the transmitted the suggestion that all the manuscripts of Aeschylus derive "from rhe same
reading but of the poet; and many of them betray a lack of understanding of copy, which appears to have survived alone the general wreck of anc ient lit
that element irreducible to pure rationality in the strict sense, which is in- erature." ~ 1 We already know that the concept of an archetype goes back
herent in any poetic language, in quite different forms and degrees.
17 three centuries before Elmsley. All the same, his stricdy "medieval" concep
tion of the archetype of Aeschylus, as the sole manuscript to have escaped a
"shipwreck" that befell civilization, is interesting because it anticipates
itself often deludes people, and encourages the deplorable itch ro emend; but when con
Madvig's and Lachmann's formulations. Ir is altogether another matter that
jcctures are proposed against the testimony of all the manuscripts, not only do fear and a
one can no longer think today of an archetype of this kind for the text of
sense of shame tweak one's car, but reason alone and the clarity of the meanings and ne-
cessity itself dominate. Furthermore, if you produce a variant reading from one manu-
script or another, you achieve nothing by claiming authority for one or two witnesses Bur m y wish not to separate his English followers from Bentley has led
against a hundred, unless you bolsrer it with enough arguments to settle the matter on me to jump too far ahead in my story. We must now take a step back, in or
rhcir own almost without the ll!~timony of a manuscript. So don't worship scribes alone: dcr to show how New Testament philology gave rise to great progress in the
no, venture your own wisdom, so that it is only when you have tested on their own the in methodology of texrual criticism.JJ
dividual points against the general drifr of the discourse and the character of rhe language
that you pronounce your opinion and deliver your verdict." On the value assumed in the
Enlightenment by Horace's sapcrc a11dc, see Venturi 1959; but rhe pious Bentley, a fero- mancc as a conjccrural critic not with the text of Horace hut with the tragedies of Seneca;
cious adversary not only of atheists but also of deists (see below, p. 6,lf.), limircd his own if one skims rhe appararus cnticus of the Paravian edition of Mo ricca 1946- 47, o ne will
"Enlightenment" to textual criticism. find only very few conjccrures of Bentley's worthy of attention and innumerable orhcr ones
37. I have lingered a bit on this poinr because some writings about Bentley (Goold that are entirely useless, uviolcnt," even tasteless. One almost fails 10 recognize rhc bnl
196,}; Shackleton Bailey 1963; Brink t918: 1087-t 164) that add new points of view and hanr emender of C.111imachus.
very intelligent considerations nonetheless tend row;srd an indiscriminate exaltation of all 38. Di Benedetto 1965: 10.
of llcntley, as though his greatness would be diminished by any recognition of the limits 39. Cf. Jebb 1900 [1886): hv.
possessed by every scholar, even the greatest. This is particularly noriceable in rhe essay by 4 0. Di Benedetto 1965: 11- u.
Brink, who in another respect is the very one who has dug mosr deeply into Bentley's per .p. Elmsley 1810: l.1 9. The passage is cited by Dindorf 1876: 40 5. Its correct inter-
sonaliry. Many critics of Bentley's conjectural boldness can be accused of a myopic con prcration (derivation of rhe surviving manuscripts from an archetype losr today, not from
servatism, but it would be difficult to nullify the observations of a Housman in this way: the Mc:diccan manuscript, as Dindorf had understood it) is due to Wilamowitz 19 14: xx111.
but sec whar he says in the already cited preface to his edition of Manilius (Housman 1937
I l 903 I: xvi i11fra-xvii) concerning the "faults" of Bentley's Man iiius, which are "the foul rs
of Bentley's other critical works"; and he repeats analogous reservations in the preface to
his edition of Lucan, Housman t9l.7 ( 19:z.6): xxxii-xxxiii. Bentley gave his worst pcrfor
The Need for a Systematic Rccc11sio 59

2 indeed permitted to amass variants at the foot of the page-John Mill col-
lected more of them in his Oxford edition of 1707 than anyone else did-
The Need for a Systematic Recensio but every attempt to introduce modifications into the text, even if on the
authority of the oldest manuscripts, encountered the theologians' fierce op-
in the Eighteenth Century position: "If someone [ ... ) would dare to change even a little word or a single
letter or one stroke of a letter by the application of critical judgment, then
at once with their cries of protest they rear him apart as impious, and with
great fierceness they accuse him of heresy," wrote Wettstein ( 1730: 158)-
and he himself had had firsthand experience of these theological furors.
Such intolerance was stronger in Protestant countries than in Catholic
As we have suggested, it was above all the study of the Greek New Testa- ones: "For the Reformation, in contrast to Catholicism, Sacred Scripture is
ment that made the technique of recensio progress beyond the point it had the only source of truth, and at the same time, quite otherwise than in Ca&
reached with Scaiiger. This was observed by Giorgio Pasquali 1952a (1934): tholicism, it is the only book that the whole populace reads. What will hap-
8: "with regard to recensio, philologia profana [... )is still, without know pen if that first certainty, from which all others well forth, becomes uncerA
ing it, a tributary of philologia sacra"; he also indicated the reasons for it. rain?" 1 All the same we must add that a similar error, which allo wed the
The New Testament has an extremely rich manuscript tradition; conjectural textus receptus to be venerated as the tradition and made a return to the an&
criticism can achieve little or nothing: hence the problems of choosing among cient manuscripts seem a rash innovation, was not limited to theologians
the innumerable variants and assessing the different degrees of authority of alone but was widely diffused among Classical philologists as well. It de-
the manuscripts moved to the forefront. And here every question of textual rived from that error of method we have already described, namely, to take
criticism aroused a particularly lively interest, since it went beyond pure the vulgate as the basis and emend it aherward by recourse to manuscripts
philology and implied, or at least could imply, questions of theology. and conjectures: in this way recurring to manuscripts seemed to be a break
The editio pri11ceps of the Greek New Testament, edited by Erasmus, was with tradition rather than a return to it. Exceptionally fierce struggles were
one of that great Humanist's least successful editions, for he prepared it in necessary in the field of New Testament studies to defeat this illogical form
haste and based it on Byzantine manuscripts of little value. 1 But here too of conservatism; but so too in Classical philology Ernesti, Reiske, and Ruhn-
that phenomenon occurred that we described at the beginning of chapter r: ken had to work hard~-just like, in Dante studies, Bartolomeo Perazzini,-~
most of the subsequent editions reproduced the text of the editio pri11ceps,
with some contamination. One of these editions, the so-called textus recep- what mattered to Erasmus was making a LHin translidon of rhe New Tesrament, nor an
ed ition of the Greek text. Even if this were true (and ic is not, at leilst nor entirely: many
tus, published by Elzevier of Leiden (1 614, 1633), had an enormous dif-
of de Jonge's arguments ilre strained), che fact remains that in char ctse Erasmus ought to
fusion and was adopted by the Protestant churches. 2 From then on it was
have published only the Larin translation: whoever p u bl i sh~ rhe Greek text as well (and
this was the first printed Greek text!), and publishes it .is extraordinarily badly as Erasmus
1. Waszink 1979: 7 5- 77 has obscned that for the most part Erasmus's o riginal contri- d id, cannot be excused by his lack of inreresr in thii: work. It is not even fair to accuse
butions to textual criticism should not be sought in his editions of Classical texts, prepared Wettstein of being a detractor of Erasmus: Wettstein passes a severe judgment on the
hastily for the printers' use, buc must he gleaned from his Latin translations of Greek texts. Greek text published and republished by Erasmus, but he also defr:ndcd it generously, in
:i.. Gregory 1900- 1909: :i..937- 4:1. (still fundamental). There is a more concise but the first and especially in 1he second Prolegomena, against a tracks by intolerant theo lo-
very clear exposirion in Hundhausen in Werzer-Welre r88:z.-1903: :i..608 - ~. Metzger gians. More generally, let me say that R. Pfeiffer and Dutch scholars have contributed to
r968 (1964): chapters 3 and 4 jscc also the addenda at the end of the second edition) are the development of a "cult" of Erasmus, a defensive attitude regarding every aspect of his
rich m information and very up to dare, but rhe author does not charactcriic the individ- personality, of which that great man has no need.
ual personalities of the New Testament critics of the eighteenth century distinctly enough, 3. Pasquali 1952.a ( 1934): 9; and sec all of his fine tribute to Wettstein, which con
and, as chapters 6 and 7 demonstrate, he does not always have a clear understanding of eludes: "Even in a field as technical as textual criticism, the greatest discoveries arc for the
the princ ipies and methods of more recent textual criticism; to a certain extent his hisior- most part the work of men of noble spirit."
ica I exposition is thereby also impaired. 4. For Reiske and Ernesti, see below, pp. 70-73, For Ruhnken, cf. Ruhnken 1875
On the textus receptus, de Jonge 1978 is fundamental. But I cannot agree with the (1789):2.I.
same scholar's recent defense of Erasmus's edition: de Jonge 1984. Acco rding to de Jonge, 5. See below, n. 4;.
60 CHAl'TER TWO The Need for :1 Systematic Recensio 61

a remarkable philologist from Verona who remained almost completely iso We have named Richard Simon and Jean Le Clerc. Strictly speaking,
lated," so that in Italy that" prejudice lasted even into the nineteenth cemury: neither one still belongs to the history of attempts to edit the New Testa
many of the readings in Plautus that Tommaso Vallauri defended were vul- ment. Simon worked especially on problems of authenticity, stratification,
gate readings of very little or no documentary authority, or even quite recent and the historical criticism of the Old and New Testaments (as did, for the
conjectures, to which Ritschl often opposed not his own conjecrures but Old Testament, Baruch Spinoza, a contemporary of Simon's, another .. her
readings from the Ambrosian palimpsest (Milan, Biblioteca Ambrosiana etic" greater than all those we have named so far, hated equally by Jewish
G82 sup.).6 The most stubborn defenders of the receptus were Protestants, theologians and by Christian ones), but he only provided a few illuminating
but the very spirit of the Reformation encouragedJ the textual criticism of hints concerning textual criticism in the strict sense.9 Le Clerc worked on
the New Testament and consequently that of Classical texts too. Wettstein these very same problems, with a strongly rationalistic criticism like that of
always appealed against his persecutors to the principles that had inspired Simon, but dissenting from him in many points. As a philologist he dealt
the Reformation. 7 He understood that those principles remained alive not with "profane" Latin and Greek texts and prepared a number of editions of
in the restrictive dogmatism of the great Protestant churches (Lutheran, chem, none of them excellent. But his methodological treatise, the Ars cri-
Calvinist, Anglican) in the coumries where they had won the victory and tica, which we mentioned at the beginning of this study (Introduction, n. 1 ),
been recognized by the political powers or were even identified with them, demonstrates that it was his heterodox theology that provided the impulse
but rather within the "heretical" currents of Protestantism itself, in devel- to his philology. He applies his principles in the same way to Latin, Greek,
opments that were rationalistic or, often, simultaneously rationalistic and and Hebrew philology. From the second to the fourth edition of his work,
mystical, but of a mysticism that was not contemplative and inert but sub he adds after its two volumes a third one, Epistolac criticae et ecclesiasticae,
versive. It was to such currents that the principal New Tescament critics be- itt quibus ostenditur 11sus artis criticae, cujus poss1mt haberi vo/1mte11 ter-
longed. Jean Le Clerc was an Arminian, as were already two other men tiwn (Critical and ecclesiastical epistles, in which is demonstrated the use
whose characters were stronger than his and whose interests were wider, of the critical art, of which they can be considered the third volume]; and in
Gerhard Johannes Vossius and Grotius, who both found the time to make the Ars critica he often has recourse to examples from the Old Testament,
distinguished contributions to various kinds of philology, including Classi and even more from the New Testament. His Epistola de editione Milliana
cal philology (but this was especially true of Grotius, despite his dedication [Epistle on Mill's edition] is directed even more exclusively to the textual
to law, to theology, to active politics}; Wettstein was a Socinian, or at least criticism of the New Testament; it is found (without page numeration) after
was suspected of Socinianism; Semler was strongly rationalistic; even the the preface of Ludolf Kiistcr's edition of the Novum Testamcntum Graewm
more timorous Lutheran Bengel was a Pietist and adhered to millenarian (Kuster 1710). Here Le Clerc asserts a restriction on the criterion of the lee-
tendencies in his commentary on the Apocalypse, exerting an influence in
this direction on the English Methodists.e In comparison,' the Catholics,
even n:imed, while Wenstein is merely alluded to once, insignificantly, :ind Le Clerc is in :1
except for one distinguished heterodox, Richard Simon,~ contributed very certain sense restored ro the mnmstream of Catholicism thanks ro his edition of the works
little to the criticism of the Greek New Testament in the seventeenth and of Erasmus ( 137).
eighteenth centuries; and textual studies of the Latin Vulgata as well ceased 9. See, e.g., below, n. :1.8. After so much neglect of Simon as a textual critic of the New
almost entirely once the Sistine and Clcmenrine editions had established a Testament, it is plcns:inr to read the claims made for him by Reynolds-Wilson 1991 (1968):
text with eclectic criteria." 188; yet, with regard to the methodology of rcxru:il criticism in the strict sense, they go
too for. Chapters l.9 - 3 :z. of Simon I 689: 3 36- P 6 contain nor so much innovative meth
6. Cf. Virelli 1962.: 130-31. odologic:il criteria as rather important individual observations (e.g., 150 on the value of
7. Weusrein 17 30: 158, :ind esp. 1734: uo. the so-called codex Dc::.ae of Cambridge) combined with other assertions that are still quite
8. On the Sistine :ind Clementine editions, sec Quentin 19:1.6: 18- 10. On attempts un old fashioned: e.g., Simon- who had indeed observed the p roc~ of b:inaliz:irion in the
derraken in C:uholic milieus ro base editions of r.urisric texts upon systematic cotl:itions transmission of texts, as we shall see in a passage cited late r-declares that he prefers the
of m:inuscripts, cf. Peti1mengin 1966 (accur:irc :ind intelligent, bur with some :1pologetic "simpler" rending compared ro the one rh:it ~contains an expression that seems more
cxc11ggcrnrions). Rudolf Pfeiffer's C:itholic viewpoint prevented this distinguished and much forceful" (:z.77: that 1s, he prefers rhe lcctio fad/ior), :ind he still believes in the criterion of
lamented scholar not only from undersr:inding the development of New Testament texrual the majority of manuscripts (ibid.). It ls intolerable to claim thar Bentley's Proposals (be-
criticism, but even from narrating it, even after its importance for the history of Classical low, p. 63f.) "scarcely mark nny advance" beyond the work of Simon (Reynolds-Wilson
philology had been pointed our: in Pfeiffer ~976, Bengel, Semler, and Griesbach arc not r 99r lr 968) 187)!
6.z. CHArTER TWO The Need for a Systematic Rcccnsio

tio brevior (shorter reading], which we will quote shortly (p. 69n30) and word for word, but ..whole phrases, or in order to save time they even read
assigns great importance to the indirect tradition (which Mill had largely whole sentences and only write them down afterward" ( 1730 [ 1697]: :z.. 5).
neglected), though he docs not overestimate its value and distinguishes pas- He requires that a conjecture be able to explain the genesis of the corrup-
sages in which the divergences of the indirect tradition are due to errors of tion (1730 (1697): 2.177), but he does not establish this as an absolute re-
citation by memory from ones that instead probably preserve better read- quirement ("if it can be done"); indeed, he admits (1730 [1697): :z..278 and
ings. This must be taken into account if we are to arrive at a fair evaluation esp. 9) the existence of "inexplicable" corruptions, since the copyist or, what
of his philological personality-not a great one, to be sure, but still quite a amounts to the same thing, the person who reads aloud to him may even
notable one. Certainly, if he is judged only on the basis of his notorious have substituted for the text of the model completely different words refer
polemic with Bentley about the fragments of Menander and Philemon, it ring to thoughts that occupied his mind at that moment-we are not far
is easy to arrive at a dismissive judgment of him: Le Clerc's knowledge of from the "Freudian slip," indeed, we have gone too far beyond it in acer
Greek meter was below the average of his time, Bentley's was well in ad tain respect, since such a substitution would not have been facilitated by
vance of his age; Bentley's victory was clear. But this does not authorize us similarities of sound or sense, by what Freud will call BegiinstigtmgetJ [fa-
to speak of "the mental aberration of a vain and vainglorious man" (Brink voring conditions]. He has no interest in the genealogy of manuscripts but
1978: 1140), as though that unhappy outcome indicated everything about limits himself in general to preferring the oldest manuscripts (1 7 30 (1697):
Le Clerc, as though there were no Ars critica. The conservative approach 2.190); but we should not forget that it was only later that the rehabilitation
that Le Clerc advocates more than once in this work, his polemic against of the recentiores became justified and fruitful, and that the first, necessary
conjectures "invented so that the writer, who had not expressed his mean- stage had to be one of distrust for the all-toooftcn interpolated manuscripts
ing so badly after all, would only speak more elegantly or wittily," 10 is aimed of the Humanist age. We shall refer shortly to Le Clerc's contributions re-
against Scaliger, whom Le Clerc sometimes also cites explicitly, and cer- garding other editorial criteria (mtts scribe11di [the author's habitual style],
tainly against most Dutch Classical philologists too (perhaps not yet at lectio di{ficilior [the more difficult reading)): in this way we shall see even
Bentley). But it would be mistaken to see here merely a mediocre scholar, in- more dearly how the Ars critica (which, let us recall, had many editions and
capable of making conjectures, venting his spleen against the great emenders hence a broad diffusion ) paved the way in large measure for the immediately
of texts. In the age and environment in which Le Clerc was operating, his subsequent development of New Testament textual criticism.h
polemic was justified. Le Clerc knows well that the ancient texts are corrupt But history follows winding roads, even the history of a limited problem;
in many places; he knows that many corruptions are not medieval but an- and so the first project of editing the New Testament which overcame t he
cient, and he proves this with a rich documentation (1730 [1697): z..i:z.- general conception of the textus receptus was due not to one of the religious
27); perhaps he even assigns too little importance to so-called paleographia reformers to whom we have referred and w whom we shall return later, but
cal corruptions, upon which later scholars will insist too much, and he pays to Richard Bentley, a man as brilliant and bold in philology as he was or-
intelligent attention to psychological corruptions, from more mechanical thodox in matters of religion. And it was, precisely, a religiously orthodox
ones such as what will later be called "saut du meme au meme" [leap from aim that inspired him, besides the philological problem itself: he wished to
the same to the same] (1730 (1697]: 2.48-56, though his examples are defend the authority of the biblical text against the "free thinkers" (in real-
inconclusive) and similar errors due to vernacular pronunciation (1730 ity not atheists but deists) led by Anthony Collins. For free thinkers, the ex
I 1697): 2.56-78), to substitutions of synonyms and analogous phenomena istence of so many variants in the New Testament manuscript tradition, such
( 1730 [ t 697 ]: 2. 5 and elsewhere). He notices, and here too he is ahead of as had been amassed by Mill in his edition of 170 7, already mentioned
his times, that most errors arise from the fact that copyists transcribe not (above, p. 59), was an argument against the Gospels' authenticity and truth;
for Protestant theologians of strict observance, it was a reason to fear a simi
lar polemical use of the text's "uncertainty" and hence to refuse to part com
lo . Le Clerc l7JO (1697): 1.169. Le Clerc adds: "For wh nt writer has ever polished
pany with the receptus; for Bentley, it was an incentive to establish the text
senrences so pcrfecdy that the subject matter can never be expressed better?" Wilamowitz
1981 ( 1911 )~ 173 - 74 will make an analogous observation; he regards this as an achieve
more solidly and thereby to defeat skepticism. 11 So he planned an edition
mcnr of recent textual criticism, which has overcome the "demand for absolute perfection
implicitin the canonical authority of antiquity." Cf. also Le Clerc r730 (1697): i . lO- Il, 11. Sec the polemic against Collins in Bentley 1713 = Bentley 1836-)8: 3.:i-87-3 68
.z.59, and elsewhere. (esp. H7- 6 t ).
CHAPTER TWO The Need for a System:1tic Rece11sm

based on a comparison of the oldest Greek manuscripts with the Latin Vul- this was much riskier than in Anglican England), they were more acute re-
gata and the citations in the Patristic texts, which ought to have restored for garding theo retical qucstions.t
us the state of the tradition as it had been at the time of the Council of Each felt k needs unknown to the other: hence their polemics and their
Nicaca. 12 He recognized that rcccnsio had to take precedence before con- mutual incomprehension. Bengel has the merit of having been the first to try
jectural criticism in a textual tradition that was so rich and ancient (Bentley to determine the relations of kinship among manuscripts: "Manuscripts are
1836-38: 3.488). But even though, as we have seen, his project was dictated closely related to one another if they have the same ancient arrangements of
by intentions that were anything but subversive in religious matters, it none- text on the page, subscriptions, and other subsidiary features ... 1" Besides
theless encountered the theologians' opposition. And so Bentley ended up these kinds of evidence he also argued on the basis of shared readings,
giving up this plan, also because of his commitment to other projects and though he did not yet go so far as to distinguish between shared corruptions,
because of the very difficulty of completing so enormous a task.ll the only truly probative evidence, and shared correct readings. He imagined
There were cerrainly more practicable ways to improve the recept11s. O ne that in the distant future the whole history of the New Testament tradition
could draw upon Mill's apparatus and introduce more reliable individual could be summarized in a tabula genealogica, that is, in what will later come
readings into the text; in certain cases, one could also have recourse to con- to be called a stemma codicum. 18 Whal is more, Bengel also saw lucidly that
jectures. The first procedure was already followed in 1709-19, that is, such a genealogical classification would furnish a secure criterion for choos-
before Bentley's Proposals, by the theologian and mathematician Edward ing among the variants, thereby allowing editors to o vercome the old and
Wells; 14 the first and second ones together, by the Presbyterian Daniel
Mace. 1 ~ Both improved the recepttts very notably; they performed a coura-
r6. Bengel 1763:1 (r 734): 1S. T he 1763 volume also cont:1ins (pp. 62.5-95:.) Bengcl's
geous and philologically valuable deed. Nonetheless this was still just occa
other \'l'ritings on New Tcst:1ment texrual criticism, including the Prodromus Novi Test/I
sional corrections, ope codicttm and, more rarely, i11ge11ii. Bentley's project mcllti Cracci rcctc ca11tcq11c adoma11di (already published at the beginning of Bengel
was methodologically more innovative, because he intended to set aside the 17 i5). Bengel also edited Cicero's Epist11/ac fa1111liarcs and some Patristic texts besides the
rcccptus altogether and to refer constantly to the manuscripts. It was along one just cited. On him cf. Nestle r S93 (useful, despite the irritatingly apologetic tone);
this path that Johann Albrecht Bengel and Johann Jacob Wettstein, the two Nolte 1913; and now Miilzer 1970, an ample monograph that, however, discusses Bengel
greatest New Testament critics of the eighteenth century (whom we have as a theologian much more fully than :1s a textu:1I emic {in any case see ch:1p. 6, useful for
already mentioned for their religious positions), moved and made further various cpis1ol:1ry testimoni:t ). On the expression apparatus criltc11s (used by Bengel, per-
haps for 1he first time) and o n his use of symbols (Greek letters, which however md1catc
progress. More cautious than Wells or Mace regarding interventions into
nor manuscriprs bur Mdcgrees of v:iluc" of the various readings), cf. Kenney 1974: 156 and
the text (not least because in continental, Calvinist and Lutheran Europe,
n. 4; for the symbols, also Metzger 1968 ( 1964): 11 J ;ind Malzer r 970: 16i. Wettstein
will be 1hc firsr one who uses symbols 10 indicau: the m:1nuscripts (capiial lctrers for MSS
ll.. Bentley 17:.1 = Bentley 1836-3 8: 3.477-86. Cf. Gregory 1900-1909: i.949 - in uncials, Arabic numerals for the MSS in minuscules): cf. Metzger 1968 ( 1964): 1r4 and,
50; Fox 1954: 10 5- 8; Metzger 196 8 (1 9 64): 109-10; Kenney 19 74: 100 (quoting the cs for a precursor in this usage (Savilc 161i), Kenney 1974: 1 S7 i11fra. In general, on the de-
sential part of a lcrrerwritrcn by Bcnrley in 1716 to Archbishop William Wake, which con velopment of the technique of critical editions, sec Kenney 1974: 15i- 57. Bur sec :1lso, for
rains che fundamental o ucli nes of ihc projccr); Brink 197 8: 1150-5i (he rightly cmpha the Humamsl :1ge, Rizzo 1973: 301-i3 and the pass:1ges listed in the index (p. 3901 un-
sizes chc methodological awareness implied in Bentley's refusal to try to go back ro an der Msigle per indicarc mss."
epoch hcforc rhe Council of Nicaca). For rhc "gcograph1eal criterion" already adumbrated r7. Bengel r 763a ( 1734): 18: " bur if the readings themselves :1rc collared, they 1cmlto
by Bentley in the Remarks cited above, see below, p. 8 5f. go together"; he goes on to cite various groupings of m:1nuscripts.
13. r.m of the material he left unp ublished, now preserved in 1hc library of Trimly 18. Bengel 176p (1734): 10. I will not cite the whole passage, already quoted by Greg
College, Cambridge, w:1s l:11cr published by Ellis 186i . ory 1900-1909: i.908 and by Pasquali 19 5ia ( 1934): 9. Bengel added: uMagn:1m comcc-
14. Metzger 1968 (1964): 109 :1nd n. 1. I have not been able to inspect the cdiuon t:1nca nostra silvam habent: scd m:1num de t:1hula, ne risu;;'m pcriculo cxpon:1tur vcri1:1s"
of Wells. [Our con1cctures arc based on :1 lot of m:1teri:1I: but hands off, lest 1he truth be exposed IO
15. [Mace! 17:.9. Cf. Gregory 1900-1909: :..950-51 (and 3.136o); Mclachlan 1he risk of laughter). Hence he believed that the attempt had to be postponed 10 better
1938-39 (this article, which cannot he found in lraly, was made available 10 me by E.J. times. Ma1111m de tabula, with aufcr understood, is a well-known proverbial expression in
Kenney); Men ger 1968 ( 1964): 110- 1:.. On other Engl ish New Testament critics of rhe Latin ro say MStop!" "Enough!"; Bengel uses it with the traditional meaning, bur ar the
second half of the eighteenth century, sec Metzger 1968 (1964): 115-17. same rime he alludes jokingly to 1he talmla gem!alogica of which he has just spoken.
66 CHAPTER TWO The Need for a Systematic Rece11sio

deceitful criterion of the majority of manuscripts. "Two or more groups, of accepted by some earlier editor": 1 1 so his edition turned out to be much in
ten agreeing, are worth the same as one: two or more manuscripts of a single ferior to the methodological principles he expounded in the Apparatus.
group are worth the same as one when they agree with one another. But Wettstein, a scholar of a much more combative character, pursued the
when they disagree with one another, a group or a manuscript agreeing with polemic against the recept11s initiated by Bentley and other Englishmen;n.2l
many does away with the present error of its comrades (i.e., with the error and he had no trouble demolishing the arguments that Bengel had adduced
of its present comrades}": 19 hence the important thing is not that a reading to justify his own prudence.21 So too, Wettstein attacked the apparently rea
be attested by the majority of the manuscripts but by the majority of the sonable criterion, followed by some, of retaining the receptus where there
families; only within each family does the majority of the manuscripts have were no reasons to abandon it, for he saw clearly that in this way the rc-
a value for reconstructing its ancestor's reading. This is already the proce ceptus rather than the manuscripts continued to serve as the basis of the
dure that Lachmann will later develop, and that Paul Maas (1958 (192.7]: edition, and chat editors were thereby led to accept a large number of ba
p. 6, sec. 8) will call elimiuatio lcctionum singularium lelimination of unique nalizations, if not of outright corruptions (1730: r67). In che edition he
readings]-an infelicitous expression, but we too will use it for lack of a bet completed shortly before his death, in general he too ended up limiting him
ter one: 1 a procedure to follow whenever the tradition is not too contami self to expounding his own disagreements with the recept11s in the notes, but
nated. Further on Bengel repeats even more explicitly that it is the consensus we can explain this if we think of the persecutions he had had to undergo
of manuscripts bclongillg to different families that guarantees the antiquity (accusations that he was aiming to deny the divinity of Christ by altering the
of a reading.10 Obviously, in a cradition as contaminated as that of the New text of the Gospels; removal from his office as pastor; exile from Basel until
Testament Bengel could not apply these criteria immediately (nor could more he found refuge in Amsterdam)." But Wettstein showed no interest in the cri
recent scholars do so); they became fruitful only when they were applied to teria of the genealogical classification of the manuscripts which Bengel had
simpler and more mechanical traditions. What is more, Bengel's fear of con formulated with such intelligence; instead he remained attached to the cri-
troversy (which in fact broke out immediately}m and persecution led him to terion of majority rule, and he never understood the arguments with which
refuse to accept into the tl!'Xt "even one syllable that had not already been Bengel had demonstrated its fallaciousness.H.

19. Bengel 1763a (1734): 2.1 : "che present error of ics comradesft must mean a differ 2.t. Bengel t 763a ( r 734): 6o7. The only largely new ediuon he made was of che
encc of reading rhac is found in a group of m:muscripcs existing now bur chat was not Apocalypse.
found in their model. 2.2.. Wetmcin 17 30: 166-67. On the relatio ns between Bentley and Wettstein, cf. Jebb
10. Bengel 1763a (1734): 65: "Bue a difference among che witnesses closesr to the r8B9 : 159; Bercheau 1908: 199.
source, the first hand, and most distant from one another, does have value; in this way they 2.3 Wettstein 1734: 2.t8-3 1, and 1751-p: i . r 56-70.
reveal che genuine reading by their agreement.ft Ibid., 68: " If chat agreement embraces a 24 . Wccmein 1730: 195, and even more r734: u6-::.8; Wecrstcin 1751 - 51: t. 166-
divcrsiry of manuscripts, all doubr will be annulled. ft It is apparent from che whole con 67. Even before Wettstein, chc criterion of che majority of manuscripts had been codified
texc char Bengel understands th is "dista nee" or "diversiry" more in the sense of belonging by Mastrichc c71 1: r 3. Wettstein had no particular foich in the oldest manuscripts, agree
to different famil ies than as geographical distance. So too later, Griesbach 1796 (1774): ing with Bengel in chis point (i:ce below, n. 33); this was the principal reason for his grad
c.lxxii: "And yet if those wimesses which can really be considered different agree with one ual separation from Bentley.
another in a friendly way, that should finally be considered :m agreement which lends au On Wensrein there is only one point I am anxious co emphasize here. As is well
tho rity to chem. ft Nonetheless, since Bengel distinguished a natio Asiatica and a natio Afri known, the edition of 17 5 1-5 2. represents a refusal (perhaps a forced refusall to insert into
cana in the New Testament manuscript tradition, che "distance" came to assume a geo che text the readings differing from che recept11s. which Wettstein considered bccrer. Bue the
graphical meaning as well, as already in a reference by Bentley and later, more explicitly. Prolegome11a of char edition are, so to speak, felicitously inconsistent with chat refusal: the
m Lachmann: see below, p. 85f. The geographical meaning becomes more explicit in a auack on the recept11s is repeated and developed there, che polemic against Bengcl's timid
later writing of Bcngel's, Bengel 1763b ( r742.): sec. viii, regula v: "Bue these manuscripts iry is repeated wich a forceful, sometimes even excessive, cone, the accusations of "im
were diffused through the churches of all che ages and climates (et climat11ml, and in spite pieryft arc rejected with undiminished energy. My citation on chis page ac n. :.3 already in
of the multitude of variants they come so close to che original text that they show the gen- dicated this, but did so coo hastily. On Wercscein's precursors regarding the formulation of
uine reading all together." Herc clima, as already in postClassical Latin. cnn only mean mternal criteria for che selection of variants, editorial technique (che use of signs), and many
"zone:," .. region.,. other methodological principles, Armando Golzio has collected a considerable amount of
61! CHAl'TER TWO The Need for a Sysremaric Rccensio

Nonetheless, Wettstein's Prolegomena of 1730, which represent the most terrain already prepared for them; but it was left to them, and particularly
interesting phase of his thought, assigned the first place in the choice of read to Wettstein, to develop more fully the theoretical assertion and practical
ings to internal criteria: usus scribe11di and lectio difficilior. He and Bengel application of these two norms.10 It was only later that Wettstein, preoccu-
were in agreement on this question. It is only when two readings are equiv- pied by occusations of subjectivism in the choice among variants, ended up
alent in themselves, declares Bengel, "that the decision is referred to a more adhering above all to the criterion of the majority of manuscripts, but with
accurate examination of the manuscripts": " a position opposed to that of out ever repudiating internal criteria.'1 31
Lachmann, who will recur to i11dicium [judgment] only when two readings In the second half of the eighteenth century another New Testament
have the same external authority. critic, Johann Salomo Semler, distinguished between iiusserliches and i1111er-
One of these internal criteria, the 11s11s scribendi, was already well known liches Alter {external age and internal age], that is, between the antiquity of
to the ancient grammarians; 26 the philologists of the fifteenth to seventeenth a manuscript and the antiquity of the readings attested by it: a manuscript
centuries then made ample use of it, even if they employed it perhaps more that is more recent than another one can preserve readings that are more an
for conjectural eme11datio than for choice among varianrs.27 Sporadic an cient. 11 Bengel had already noticed this, as others had even earlier, but not
ticipations of the criterion of the lectio difficilior can also be found from an-
tiquity until the seventeenth century; 28 as far as I know, the first to formu- or clarior, which for Le Clerc always had its origin in a marginal gloss rhat had intruded
late it precisely was Jean Le Clerc. 2 ~ Thus Wettstein and Bengel found the inro the rexr and substiruted for the original reading, or at least in co,,scio11s b:inalizations;
yet ar least as often, if nor more ofren, the origin is an unconscious banalization.
material and has evaluated ir imelligemly: I hope that the results of his researches will he 30. Bengel 176p ( 1734): 17: "Where the one [sc. reading) is more easy, the otherless
published soon. Another field in which Golzio has enriched, modified, and also corrected so, rhc one that is old, weighty, brief, is preferred; the one that charms us by its grearer pcr-
many of the matters discussed in this work of mine 1s the conceptual and methodological spicac1ry and fullness, as though n had been inrroduced deliberately, is generally set aside."
contribution made by Orclli and M:1dv1g, which now turns out to be more significant than A fulle r discussion is in Wettstein 1730 : 179, 184 (on lectio d1ffici/ior) and 188 (on us11s
whar I had indicated (see below, JlJl 9of., 97 f., 10::.f.), even if the importance assigned to scr1bcnd1, from which he rightly disringuishcs the repetition of a passage with identical
Madvig was a point to wh ich I was particularly committed.r words, which is suspected of leveling and is therefore to be rejected in favor of the "varied"
:z.5. Bengel 1763a (1734): 18. This position is now reaffirmed by Waszink 1979: 87, expression). I shall not cite in their entirety the passages of Wenstein, which arc already
with good reason. quoted by Pasquali 195:z.a (1934): ro-u. Ir should be noted that already in Bengel, and
16. Especially to Aristarchus: Lehrs 188::.: 354-56; Pasquali 1951a ( 1934): ::.33, ::.40- then in Wemtcin and Griesbach, ;ind even in recent manuals, the lcctio brevior appears a$
41; rfeiffer 1968: 117- 18 (but one mighr still have expected to find something more here). a subspecies of the lectio difficil1or-bur in face the lectio '1rcvior is a m11ch more uncer-
Aristarchus certainly npplicd rhe idea of 11s11s scribc11di excessively, hypernnalogic;:illy. tain criterion, since if rhe fuller reading can derive from rhe desire to n:rnke the text clearer
z.7. Ir is asserted as a cricerion for eonjeccural emcndacions, e.g., by Le Clerc 1730 or from interpolations of various kinds, the briefer reading can be caused by omissions
(1697): z..:z.70-81. (Dain 1975 (19491: l.O), especially by unconmous elimination of words not strictly nec-
:i.8. E.g., in Galen, Medici Cracci, ed. Kiihn, 18.1.1005; 17.1.98, 101, 110 ("chis was essary to the contexr yet stsll present in the authentic text: cf. Timpanaro 1976: 35 - 40;
the ancient reading, but it was altered by many interpreters in order to make it clearer"). other ex:implcs in Rizzo 1977: 104-5. In this point, Le Clerc proved himself co be more
Bur Galen uses chis internal criterion only so as co confirm the aurhority of the oldest man- caurious, for in the epistle mserred into Kiisrer's edition (cf. above, p. 61) he had main
uscripts, which remain rhe fundamental principle for him. Somt: furrher details on Galen's taincd the authenticity of rwo words that :ire nor strictly nec1:5sary in pare of the tradir1on
procedure arc furnished by D. lrmer in the German edirion of my book, Timpanaro 1971: of Man. 3: 1 t and that arc a bsent in other witnesses, "for there was no reason why rhese
19. Probus (in Scrvius auctus on Virgil, Ae11cid 1::..605) followed the criterion of the most words should have been added, for they are obscure ;ind add nothing to clarify the meaq.-
archaic reading (which is a special case of the lectio difficilior) in order ro prefer floras to ing of rhe passage: on rhe contrary, for these very reasons they could have been cl1mmatcd
flauos in one passage of Virgil. I c:innot bring myself to believe chat this is an archaism that as obscure and useless."' As already in the case of rhe lectio faci/1or, Le Clerc speaks of an
Probus ;irbitrarily introduced; I hope ro discuss this passage in more detail elsewhere. In intentional alteration, which 1s not the most frequent case; but in itself his argumentation
the Middle Ages a reference ro the lectio difficilior is found in lrnerius: cf. Kantorowicz is entirely correct and demonnrates thar the lcctio longior can even be the /cctio d1fficil1or.
19:z.1: 31. In the seventeenth century, Simon 1689: 37 5-76 too observed that copyists tend 3 1. The Animadvcrsio11cs et ca11tio11es for choice among variants arc no longer to be
ro banalize, but he did not deduce from this observation an explicit criterion for choice found in the much fuller Prolegome11a co Wettstein 1751-51, but they are republished sep
among variants. aratcly at the back of the edition.
19. Le Clerc 1730 (1697): 1.193: "If one ofthcm (sc. the readings] is more obscure and 3::.. Semler 1765: 88-89. Semler 1765: 396 also polemicizes against the criterion of
the ochers clearer, then the more obscure one is likely ro be true, the others glosses." The the majority of manuscripts but does so only very briefly. In this regard he makes no prog
only defect of this formulation i5 the too restrictive character of the concept /cctio facilior rcss beyond Bengel, but a little later Emestt will go another step m the criticism of "pro
CHAl'TER TWO The Nc:c:d for a Systc:m:iric Recensio 71

with such clarity.rn In the end Johann Jacob Griesbach summarized the re- could still point to the critics of profane texts as an example that students of
sults of earlier criticism in a didactically perfect form in the Prolegomena to the New Testament should follow; but by 1770 the roles had been reversed
his second edition;l but although by now he fully recognized the inconsis after the works of Wettstein himself, Bengel, and Semler, and Johann Jacob
tency of the receptus, he too did not free himself from it courageously Reiske wrote, "We should not treat profane authors with less scrupulous
enough..U. veneration than the New Testament. For the very same reason that we care-
fully collate manuscripts of the New Testament, it is only fair that we inspect
the manuscripts of Demosthenes and all the other ancient authors too and
dig our and publish their readings. For this is the only way to demonstrate
In the meantime, Classical philologists had noticed that they had fallen be- the historical rrurh of any text, be it sacred or profane, on the basis of the
hind the theologians in textual criticism. In 1730 Wettstein (1730: x66J consensus of many ancient manuscripts of approved reliability." 3'
In fact, Reiske did more for the text of the Attic orators and of Atticists
like Libanius" with his splendid conjectures than by investigating the man-
fane texts" (cf. below, n. 43). On the classification of manuscripts proposed by Semler, de-
veloping Bengel's classific::nion, see below, chap. 4, n. s. uscript tradition (many of his conjectures were later confirmed by manu-
33. The fact that reccn1 manuscripts could have good readings had been observed, scripts of which he had no knowledge).37 But the need to use the manu-
e.g., by Nicolaas Heinsius (1661: 1.195: " in the Arundel manuscript, 1hough it is recent, scripts as the text's constant foundation instead of only making occasional
but nonetheless offering very correct readings"), but withou1 wondering whether this co llations was reaffirmed a little later with great clarity by Ernesti in the
might be due to felici1ous conjectures. As for Bengel, he had limited himself 10 observing preface to his edition of Tacitus, and once again by Friedrich August Wolf
that "almost the whole variety of readings was crea1ed a long rime before 1he Greek man at the beginning of his Prolegomena ad Homcrtttn.3" Each of these scholars
uscrip1s extant today" (1763a 117341: 11) and that therefore ancient manuscripts are al observed correctly that to follow the old method of having recourse to the
ready not less corrupt than more recent ones: a claim that is doubtless exaggerated
manuscripts only where the vulgate was not satisfactory resulted in leaving
(though shared also by Wettstein), although it is true that the variants and corruptions of
in the text a large number of small corruptions and lcctio11es faciliores that,
greatest importance often go back to very ancient times, as Le Clerc had noticed.
34 . Griesbach 1796 (17741. As for methodological cri1eria, 1here is nothing in Gnes- for better or worse, made some sense and hence did nor arouse suspicion.
b:ich th:it is not already found in his predecessors, and hence it is mis1aken to place him at Wolf writes: "A true, continuous, and systematic recension differs greatly
the beginning of the "modern cri1ic:1l period fl of New Testament texrual criticism, :is Metz from this frivolous and desultory method. In the latter we want only to cure
ger 1968(r964): l19 does: he honor:ibly concluded one import:int period, but he did not indiscriminately the wounds that are conspicuous or are revealed by some
begin a new one. So too, 1he cri1erion of the lectio "'cdia (Griesbach l 796 [1774 [: Ixiii), to manuscript or other. We pass over more [readings] which are good and
which P:isqu:ili ( 1951a [ 1934]: 1 l) poin1s as a novdty, had already been formula1ed by Ben- passable as regards sense, but no better than the worst as regards authority.
gel 17633 (17.34): 17 ("Where there arc not only two readings, but m:iny readings, the ,,,;J.
Bur a true recension, attended by the full complement of useful instruments,
die one is the best. For from 1his one, :is though from a center, the 01hers dispersedfl): this
seeks out the author's true handiwork ac every point. It examines in o r-
is wh:it h:is been called "diffracrion" in our times by Contini 19 55: r J4, and then elsewhere.
None1hc:lcss, Gricsbach's formul:itions arc clearer and more eleg:int than his predecessors'.
35 Nonetheless he introduced into his own 1ext some re:idings 1ha1 diverged from the 36. Reiske 1770: lxxvi.
rccept11s. 3 7. Nonetheless, :is Dieter lrmer has pointed our 10 me, Reiskc had the merit of being
Jn the 1981 edition I modified certain details of this first part of chapter 1 concern- the first to recognize the value for the m:1nuscrip1 1radition of Demos1henes of m:inuscript
ing the great New Testament critics of the eighteenth century; but i1 would need a more: A (Augusr:inus, now Monacensis 485) and to make use of it. In his own edition of Demos
drastic revision th:in ;my of the 01hers, even though only :1 few years have: p:issed (and even 1hc:nes (H. Wolf 157:z.), Hieronymus Wolf had already cited some readings of this manu
recalling th:it ch:ips. l and 1 have a merely introductory character). I learned much 1h:1t I script, communicated to him by Simon Fabricius, hut he had not introduced them into his
had not known previously by taking p:irt in the seminar on Wettstein conduc1ed by my text (cf. Voemel 1857: 183, 193-94; I also owe to D. lrmer my knowledge of this work).
friend Antonio Ro1ondo at 1hc lstituro di Storia of the Facolt3 di Letrere (1983-85); I am 38. Ernesti 1801 ( 1771}: vi. F. A. Wolf 1985 (1795): 43- 55. Wolf was not only cle:irly
profoundly grateful to Rotondo and his students for this experience:. One: fruit of this sem- influenced by Ernesti but :ilso probably by Semler; on the intimate friendship between Wolf
inar will be the edition with commentary of Wcttstein's Prolcgonuma (1730 and 1751-51) and Semler, cf. Koerre 1833: 1.119 (brought to my attention by Konrad Milllc:r, who also
prepared by Cecilia Asso, a student of Rotondo. On Wettstein the "heretic," who is linked points our 10 me Wolf's reference, in 1he preface to his Homer edition [F. A. Wolf 1 !104:
with Wettstein the 1extual critic even more closely 1han is usu:illy thought, we: await :1 book I F. W. Wolf 1869: 1.15:z.], to Gricsb:ich, "eminent founder of S:icred criticism," with
by Rotondo himself, which will be innovative and rich in unpublished documents.' reg:ird 10 1he various degrees of external and internal probability of 1ransmitted readings).
72. CHAPTEk TWO The Need for a Systcm:nic Rcc1msio

der rhe wirnesses for every reading, not only for those that are suspecr. It manuscripts [are] not the worse ones], which, as we know, is in itself not
changes, only for rhe mosr serious reasons, readings that all of these ap- new, but it is worth the trouble to quote Wolf's own formulation (F. A. Wolf
prove. It accepts, only when rhey are supporred by witnesses, others that are 1985 [ 1795 )+ 46): .. For newness in manuscripts is no more a vice than youth
worthy in themselves of the author and accurate and elegant in their form. in men. In this case, too, old age docs not always bring wisdom. Insofar as
Not uncommonly, then, when the witnesses require it, a true recension re- each follows an old and good authority well, it is a good witness." The com-
places attractive readings with less attractive ones. It rakes off bandages and parison between a manuscript's "youth" and a human being's is little more
lays bare the sores. Finally, it cures not only manifest ills, as bad doctors do, than a graceful witticism, but the final phrase explains well why a recentior
but hidden ones too." 39 Only a procedure of this sorr deserves the name of can on principle be not at all a deterior: the recenr copyist may have copied
recensio and not of mere recognitio (F. A. Wolf 198 5 [1795}: 45); and only an ancient and good manuscript well (and, Wolf seems to imply, directly). 42v
after a systematic rece11sio will one be able to go on ro conjectural emenda- So too, during the last three decades of the eighteenth century, the need
tio, for which Wolf did not feel much sympathy anyway. 4 1 Thus it was that for a genealogical study of the manuscripts, which we have seen Bengel as-
rhe old concept of emendatio ope codiwm was completely overcome. scrr in the field of New Testamenr studies, gradually spread among Classical
But Wolf's recognition of the need for a systematic recensio and for a re philologists. Ernesti clarified even better than Bengel had done the principle
pudiation of the vulgate is not accompanied by a too exclusive faith in the that several manuscripts deriving from the same ancestor had the same value
most ancienr manuscriprs. It will have been Semler 41 who positively infltt as only one; 4 l Christian Gottlob Heyne and the Alsatian jean Schweig
enced him in this recognition of recentiores 11011 deteriores [the more recenr

39. F. A. Wolf r985 I r795): 43-44. Cf. also Wolf's preface to his edition of Pla1o's 4z.. Cf. Kamorowicz 192.1: 2.1-u; Pasquali 19pa ( 1934): 46. Their argumcms arc
Sy111posi11111 (F. A. Wolf r782.: v-vi ~ F. A. Wolf 1869: 1. t J 5-;61: KBuc this cannoc hap formulated more clearly and fully than Wolf's bur are nor substantially different from his.
pen if manuscripts and old edition ~ arc only checked and compared occasionally, for in 43. Ernesti 180 1 (1 772.): xxviii: "When manuscrip1s arc involved, one should make
dividual obscure or apparemly erroneous passages." But so alrt'ady Ernesti 1801 ( 1772.): sure that we have not as many in number as possible, bur as many of those that possess as
vi: "In previous ages, those who set about to edit ancient writers thought it was enough it were the legal right to give an opinion.... For if you have a hundred manuscripts of the
to check manuscripts and primed editions in those passages where they got stuck( ... 1. In same book, but it is certain that they are derived from a single apograph, then all together
this way( ... ] they ended up leaving many things untouched that could have been emended they only have the right and force of a single book." Griesbach 1796 (1774): 1.Jxxi will
from those same manuscripts and printed editions. And yet no one is so sharp-eyed that express this concept with very similar words. But although Italy was a philologically quite
he can see all the faults of the vulgate reading on his own, and not sometimes approve cor "depressed" milieu (in which, nonetheless, Verona constituted an exception), Domenico
ruptions as though they were correct." In Erncsti, however, these fine theoretical pro Vallarsi had already enuncia1ed this principle there in the preface co his edition of Saint
noun,ements almost never found any practical application: both his edition of Tacitus and Jerome (Vallarsi 1766 ( 17341, reprinted in Migne 1845: p. xxxix, para. 35): manuscripts
his (better) edition of Cicero are based in substance upon preceding edition ~, not upon thnt agree in errors or in arbitrary changes (criticomm ausis) are worth "not more than
manuscripts, as C. G. Zumpt 1831: 1.xxiv-xxviii rightly observed regarding Cicero. From one manuscript": evidently Vallarsi wrongly neglected the possibility of contamination
this point of view, his edition of Callim~chus (Ernc$ti 1761 ) is better; in the preface when he considered the arbitrary acts of critici. And above all, textual criticism or Dante
(fol. 5b>Ernesti asserts that all the nonintcrpolatcd manuscripts of the Hy11111s arc derived in the second half of the eighteenth century possessed at Verona itself a philologist or Eu-
from a single lost model, on the basis of their agreement "in lacunas and in readings" ropean standing in the figure of Bartolomeo Perazzini (Perazzini 1775), 10 whom Folena
(cf. Pfeiffer 1949-53: :i..lvl. 1965: 67-69 has justly called our attention. Perazzini, besides repeating the aforcmcn
40. KA pleasant pastime" is what he calls it at the beginning of the 1'ro/cgo111c11a; F. A. tioncd genealogical principle (probably under Vallarsi's influence), also wro1e passionate
Wolf 19 8 5 ( 1795): 45 (cf. 44; and F. A. Wolf 1869: 1.2.4:1.). All the same, Wolf is not an un polemics against the defenders of the vulgate that have much in common with those of the
'riti,al wnservativc: c'hoing Bentley's famous phra ~c (above, p. .H I. he admits that one New Testament philologists who were his c;ontcmporaries or a little earlier than he. This
must prefer "talem" to "treasure 'hests full of parchment" (1795: vi), but he maintains similarity is increased by the fact that the Divi11a Commedia was considered a "sacred
that emendation shou Id not prc:,ede recension nor, even less, ta kc its place; and, hke Ernesti, text" too for a complicated series of religious and artistic reasons, though to a muc;h lesser
he insists on the importan'e of emendation for healing "latent errors" (sec above). But we degree than the New Testament. Perazzini speaks explicitly of a "prejudice of a sacred
should bear in mind that although Wolf is speaking in general terms in the Prolegomena, text" (1775: 56), and he repeats: ~nor should any text be sacred, unless it has first been
he is thinking above all of the Homeric text, one of the very few from antiquity that have perfectly emended"; and again: "Hence it is not I who should be called an innovator, bur
no need of conje,tural criticism (except for the prohlem of the interpolated verses>. He will rather those who altered the text which was once received" (i.e., in this case, not 1he rc-
later judge emendation more favorably, in F. A. Wolf 1807: 40 I'"' F. A. Wolf 1!169: :i..llpl. cepws in the sense of the New Testament, but the oldest stage of the tradition, the antiq11a
41. Sec above, p. 69, and n. 38 for the relations between Wolf and Semler. lcctio, as he says immediately before).

hiiuser tried to reconstruct genealogies, the former for the manuscripts of 3

Tibullus, the latter for the Manual of Epictetus.44 To be sure, these scholars'
attempts turned out quite imperfectly, not only because of their inexperi- The First Phase of Lachmann's
ence and incomplete knowledge of the manuscript material but above all for Activity as a Textual Critic
the objective reason that even today makes it impossible in so many cases
to trace a stemma codicum: contamination. Heyne-a philologist whose
greatest originality certainly did not consist in textual criticism but who
even as a textual critic is more valuable than is generally said 45-was well
aware of this phenomenon of contamination; for the New Testament, Gries-
bach was too.46
So too, the manuscript tradition of Homer was too contaminated to per-
mit the fulfillment of Wolf's proposal that the manuscripts be organized
"into classes and families." 47 But thanks to his use of the Venetian scholia After the considerable progress achieved by the method of textual criticism
discovered by Villoison, Wolf was able to achieve something else in his Pro- during the eighteenth century, we wicness a return to old positions in the
legomena: the history of a text in antiquity.48 In this way he prepared the first fifteen years of the nineteenth century. Gottfried Hermann and Im-
way not for Lachmann, but rather for Jahn's and Wilamowitz's concept of manuel Bekker, the two greatest textual critics of the generation after Wolf,
Textgeschichte [history of a textJ and for all the nineteenth- and cwencieth- differed greatly from one another in many ways, but both remained quite
century studies on "ancient variants and ancient editions" (to repeat the unaffected by the need for a systematic recensio as it had been adumbrated
title of one of Pasqual i's chapters). For Wolf, the Homeric question itself was by the great New Testament critics and by Ernesti and Wolf. Hermann was
nothing more than the first phase, oral and popular, of the history of the text an admirable expert on Greek language and style and supplied contribu
of the Iliad and Odyssey: in the Prolegomena it is only discussed in these tions of decisive importance to the study of meter, but he had no interest in
terms and not as a problem in literary history.4!1 manuscript tradition: his editions are based not on manuscripts but on pre-
ceding editions, and the improvements he contributed to the text of the
44. Heyne 181 7 (1755 1: xiii-lxxviii: bur whac Heyne provides is in the fim inst:rnce Greek poets are the frui ts of conjecture or, when he docs choose between
a genealogy of the cditio11s of Tibullus, and only secondarily one of the m:rnuscripts. variants, are based solely on internal criteria. To be sure, he succeeded very
Schweighiiuser 1798: prefoce: to classify the manuscripts, Schweighiiuser relied on shared often in resolving once and for all textual difficulties that had remained un-
corruptions less than on the reciprocal arrangement of the text of Epictetus's Ma1111al and solved until then: for after all, a thorough knowledge of an author's language
ofSimplicius's commentary on it (cf. Timpnnaro 1955: 70). Schweighiiuser h:id more mer-
and style always remains the first and essential condition for restoring his
its in the cli111i11atio codicum descriptomm: see below, p. 99.
text. And yet his complete indifference with regard to the documentary foun-
45. Heyne 18 17 ( t755): xv, xxxvi (where he speaks of "apographs, perh:ips prepared
with others, or m:ide our of them"). Heyne's fame as a textual critic, and :is a Classical
dation of the classical texts represents not only one aspect of his lack of un-
philologist in general, w:is impaired by the scornful tone :idopred toward him by his stu derstanding for the new Classical philology of Wolf and Boeckh bur also a
dents Wolf :ind Lachmann (and, in a less technic:il field, Friedrich Schlegel), in great mea- step backward compared to the textual criticism of the eighteenth century.'"
sure unjustly; fo r L.ichm:inn m p:imcular, sec Lachmann 1876: i. . 106. Bekker, on the other hand, was an indefatigable explorer of manuscripts:
46. Griesb:ich t 796 ( 1774): lxxvtii: "The readings of the one recension have been in
troduced mto the m:inuscripts of the other family," etc. Already Semler ( l 76 5) had often ob- 1. Cf. Jahn l 849: zo. Sauppe l 841 : 5 Sauppe 1896: 8:1. writes, "Whoever wishes to

served that in different passages the very same manuscript belongs to different "recensions." perform the an of criticism properly must first of all examine rhc manuscripts and seek our
47. F. A. Wolf t985 (1795): 44. In his edttions of other authors (Plato, Cicero, ere.), and mvemgace their characteriscics :is carefully :is possible. I recall chat you Ii.e., G. Her-
which in certain cases would have: allowed an :1pphcat1on of the genealogical method, mann I gave rhis advice very often." It remains uncertain whether S:iuppc captandae bc11e-
Wolf ltmitcd himself ro hasty recog11itio11es in contrast with his principles. 110/entiae ca11sa is attributing here to his reacher something rh:ic the l:itrer had in reality never
48. Simon 1689 had alre:idy arrived at this concept of a "history of the text," but he said, or whether Hermann really did give chis advice to his students wirhout going on to
had had no followers among Classical philologists. Cf., for now, rfeiffer 1976: t;o. apply it himself; we h;ive already noticed :i similar contrast between theory and pr:ictacc
49. On this point cf. Timpanaro 1980: ll.5 and n. :1.8. in Ernesri and in Wolf himself; cf. chap. z, nn. 39, '47 Cf. the addendum to this ch;iprer.
CHArTER T HREE The First Phase of Lachrnann's Activity as a Textual Critic 77

we are indebted to him for the rediscovery or appreciation of first-rate tex- refers the reader for everything regarding interpretation to the commentary
tual sources such as the Urbinas of lsocrates (Vatican, Urbin. gr. 111 ) , the on the poet that J. G. Huschke was planning to publish, just as in the pref-
Parisinus manuscripts of Demosthenes (Paris. gr. 2.934) and of Theognis ace to his Lucretius ( 185oa: 15) he will refer to the future exegetical works
(Paris. suppl. gr. 2.88), the Ravennas of Aristophanes (Raven. gr. 42.9 ), and of Steinhart and Reisacker. 7 The critical apparatus, which in his first edition
many manuscripts of Plato. Yet despite the fact that Bekker was Wolf's fa- of Propertius is still "selective," becomes "scant" in his editions of Catullus
vorite pupil, not even he ever thought of doing a systematic recensio: as Wi- and Tibullus and in his second edition of Propertius.
lamowitz rightly notes, .. in his choice of manuscripts and readings Bekker With regard to the manuscript tradition, Lachmann insists above all on
essentially relied upon his sense of language and style," z. which, for Attic the d istinction between interpolated manuscripts and uninterpolated ones.
prose, was certainly very experienced. Manuscripts that contain Humanist interpolations must be set aside: you
This step backward in the methodology of textual criticism at the begin- arc in trouble if you yield to the attraction of their specious readings! In this
ning of the nineteenth centuryJ helps to explain the impression of great orig- polemic favoring truth over elegance, which assumes a tone of particular ve-
inality that Lachmann's first works make even when they do nothing more hemence in his preface to Propertius,~ in this distrust of the docti Jtaliani
than reaffirm, sometimes with less refinement and caution,h principles that [learned Italians],' Lachmann feels himself particularly close to Scaliger but
were well known to the Classical philology of the eighreenth cenrury. has harsh words for most of the Dutch Classical philologistsY1 Lachmann's
In the field of rhe criticism of Classical texts (we shall refer to Germanic language, like Scaliger's, reveals not only the pure requirement of documen~
texts shortly), Lachmann began his activity with an edition of Propcrtius in tary truth but also a motivation typical of aesthetic criticism: anyone caA
18r6, followed by editions of Catullus and Tibullus and an editio minor of pable of judging can tell that the original reading is more beautiful than the
Propertius (all 182.9). His review of Gottfried Hermann's edition of Sopho- interpolated one; Classical harshness is preferable to Classicist tinsel.
cles' Aiax 4 and two reviews concerning Tibullus 5 are also important. Lachmann sympathizes more with Bekker than with Gottfried Hermann
In this first phase of his thought, Lachmann claims, polemically and par among the Classical philologists nearest to him in age, and this is only
adoxically, that the most urgent cask is to supply rigorously diplomatic edi- natural. 1 It is strange that he does nor name Wolf; yet it was Wolf who had
tions reproducing the manuscript tradition in the most ancient form we can demanded editions based on a solid diplomatic foundation (sec above,
attain, "without taking into the least consideration meaning or grammati- p. 7If.), even more than Bekker; and Lachmann, who studied the Homeric
cal rules." 6 Hence not only is conjectural criticism postponed until later; so question and created an analogous "Nibelungen question,'' knew Wolf's Pro-
too is even interpretation itself. Until that time, most critical editions had /cgome11a ad Homermn better than anyone else. 12
been simultaneously exegetical, even if some scholars like Heyne accorded But once the interpolated manuscripts have been excluded, how is the
the first rank to exegesis, others like Hermann to textual criticism; and all text to be constituted in those passages in which the uninterpolated ones dif-
the Romantic theorists of Classical philology, from Friedrich Schlegel to fer from one another? Lachmann answers that the original reading can be
Ast, to Schleiermacher, to Boeckh, had insisted, or would do so a little later,
on the "critical and hermeneutic circle."" Lachmann, instead, produces edi- 7. He alludes to Steinhart wirhouc naming him ("a certain illustrious man" I: cf.
tions that are purely critical: in the preface to his Propertius (1816: iv) he G. Muller 1958: :z.53nt.
8. Sec esp. Lachrnann 1816: x vii-xviii.
:z.. Wilamowin I894: 4:z.. 9. Lachmann l 81 6 : xii: here and in wh;ic follows codices i11tcrpolati and codices lta-
3. 0Iher examples could be cired. For example, Friedrich Asr, one of rhe leaders of the lici arc synonyms for Lachmann.
new historical approach ro Classical philology, took the Aldine edition of 15 r 3 as the ba 10. Lachmann 1816: xvii (where .lside from Scaliger the only scholar praised is Livi-
sis for his own edition of Plato (Ast r 819: iii) and set himself the goal of adhering to it as neius, i.e., Jean Lievens, 1546-99); Lachmann l8l.9: iii.
closely as possible! For Elmsley as a partial exception in England, see above, p. 57. u. Lachmann 18:z.9: iv: "In any case, if I seem to have made progress in this art be-
4. Lachm:rnn I8r8 - Lachmann t876: :z..r- 17. yond Immanuel Bekker, I myself will be pleased and I hope that he will not protest." For
5. Lachm:rnn lR:z.6 Lachmann t876: :z..1o:z.-45; Lachmann 1836 = Lachmann Hermann, cf. rhc review cired above (n. 4).
1876: :z..145-60. 1 l.. Certain similarities in expression arc perhaps not accidental. For example, F. A.

6. Lachmann 1876: :r..:z.. Cf. Lachmann 11176: :z..145-46: "In my edition of the Ro Wolf 198 5 ( 1795): 44: ~A true recension f.. -1 takes off bandages and lays bare the sores"
man clegists I had the modest aim of presenting the aurhentic transmission complcrely, ex (quoted above, pp. 71f.); Lachmann l Hi9: iii: "That plague breaks out which [ .. ) leaves
d uding as for as possible all later vagaries.
H bandages and disgusting scars in place of wounds."

found in these cases only "by reason and the effort of a skilled mind," 13 so
I The First Phase of Lachmann's Activity as a Textual Critic

his editions on an extremely limited number of manuscripts, selected some-


to determine the original reading he does not yet have recourse to any kind times rather arbitrarily: 18 for Catullus he used only two manuscripts, and in
of mechanical criterion (of the sort Bengel had already enunciated). Nor can this case it is quite obvious that the choice among variants could only be
he have such recourse, since he is not yet tracing a genealogy of the manu- based on internal criteria, since mechanical criteria require at least three wit-
scripts in these editions of Propertius, Catullus, and Tibullus. He does in- nesses. Yet in that very same edition of Catullus he thought he could recon-
deed refer a very few times to the fact that some manuscript "almost com- struct the pages of the lost ancestor, whose numbers he indicated in its mar-
pletely agrees" with some other one, or that some manuscripts are "derived gins-a failed attempt, despite Moriz Haupt's attempt to defend it. 1 ~ As we
from a single source," 14 but these references are so sporadic that they could shall see, in the case of Lucretius Lachmann will have far more success with
have furnished no basis for mechanical choice among the variants. From an analogous attempt.
th is point of view, he is still very far behind Heyne's and Schweighauser's ge- At the same time as he was publishing his editions of Propertius, Cacul-
nealogical attempts, imperfect as these were. His d istinction between sincere lus, and Tibullus, Lachmann was also extremely active as a textual critic in
manuscripts and interpolated ones refers only to their value, not to their ori- the field of medieval German poetry. To the flowering of Germanic studies
gin: unlike Politian and other Classical philologists who followed him,J Lach- during those years in response to the stimulus of Romanticism (it is not nec-
mann eliminates interpolated manuscripts not because they are copies of ex- essary to recall the names of Wackenroder, Tieck, Uhland, Arnim and Bren-
tant manuscripts but simply because they are untrustworthy. u And in part tano, and the brothers Grimm), he contributed a rigorously philological
because of the difficulty of procuring collations of manuscripts preserved in method that he had learned from the Classical philologists; in this way he
distant libraries, 16 in part out of a haughty disdain for anything that looked made a significant contribution to the birth of a scientific approach out of
to him like a useless aggregation of erudition, 17 he always ended up basing what had originally been a "return to the Middle Ages" espoused by pop-
ulist and reactionary literary figures. 20 In the period 1816-2.9, which we
discussed earlier with regard to his Classical philology, Lachmann published
1;. Lachmann t 816i xvii "Since in the case of good manuscripts that present differ
editions of the Nibc/1mgcn ( 1826), Hartmann von Aue's Iwein (182.7), and
ent readings and in other cases as well the correct and original reading can be restored only
by means of re:ison and the effort of a skilled mind." The phrase: " in other cases" refers, I the poems of Walther von der Vogelweide (182.7), alongside a large num-
think, to cases in which all the manuscripts present a corrupt read ing and the editor must ber of articles and reviews concerning Germanic philology. In the same field
heal it by conjecture. his editions of Wolfram von Eschenbach (183 3 ), Hartmann von Aue's Gre-
14. Lachmann l 816:xii; l8l.9:vi. gorius (18 3 8 ), Ulrich von Lichtenstein (I 84 1 ), and various articles and re-
l 5. Pasquali l95l.a {1934): 5, :i.5 justly observes that Lachmann never practiced c/imi editions will follow later.21
11atio codic11m dcscriptomm. But his formulation suggests that this operation was not per-
formed unril the gener:ition lh:ir followed Lachmann; yet we have seen (pp. 47-49) that it
w:is already pr:icttccd by Poli1ian :ind Vettori, and we shall see that it had continued to be lcssly will be able with little trouble to pile up monstro us errors and ~chobrs' innumerable
practiced by Bo1vm and Schweighauser, and then during Lachmann's lifetime by Sauppe, inrcrprelations."
with a parcicularly rigorous mc1hod (p. 100). Another correction: according to l'asquali 18. For Propert1us he used lhc Ncapolitanus (now in Wolfenbiittel: Gud. lat. 1l.4),
l95l.a ( r934): 4n1, Lachmann refused to trust the recentiores only in the c:ise of "texts rightly, but he was mistaken to ' onsider it inferior to the Groninganus {Groningen: Bibi.
which during the Middle Ages were read hardly or not at all." But in fact the testimony of Un iversitalis t 59). For T1bullus, the only really impo rtant mantMcript among the ones he
Haupt thnt he cites {Belger 1879: I l.J) refers to rccc11tiorcs in general: "He thought that used (disregarding the fragmc11t11m Cuiacia1111m ) is the Eboracensis, lost but collated by
it was almost never at all useful to distinguish those manuscripts quire devoid of author N. Hcmsius (Berlin, Die:t. B. Sanr. 55 d, pp. 15-l.-4). For Caiullus his choice was even less
ity into groups and classes with lhc same meticulousness [sc. as ancient manuscripts)." felic itous (cf. l'asquali 195l.a ( 1934): 5nl.).
t6. l'asquah 1951a {1934): 4-5 emphasizes chis difficulty. To be sure, schol:irs such as 19. Haupt 1836: t -l. = Haupt 1875-76: 1.1-1. Cf. Ellis 1878: xxxv-xxxvi; Wi-
Hcinsius and Bekker had m:inagcd to inspect an enormous number of manuscripts; but lamowin 19 8l. (19l.1 ): l J O, 141.
they had had the opportunity of traveling throughout Europe, whereas Lachm;inn never :i.o. On the relations between Lachmann and German Romanticism, see Leo 1893: 6-
left Germany except during the war against Napoleon of l 81 s. Nonetheless the basic rea 7, r 7- 18 "' L=o 1960: l..411!-19, 430-;1. I have no t had access to Sparnaay 1948
son remains the second one: sec the following note. !brought to m y attention by D. lrmer). On Lurz-Hensel t97j, cf. below, p. I Hnl 1.
17. See, e.g., Lachmann 1816: vii i: "Since so many manuscripts from this worthless l.I. For more: detailed bibliographical refcrc:ncc:s, see Hern 1851: 100-119 and xxiv-
group have been collated that we are already thoroughly glutted with their re:idings to xxxii in the appendix. The Germanistic articles and reviews arc republished in Lachmann
the pomt of nausea." Lachmann l 8l.9: vi: ..Anyway, whoever wants to waste hi$ time use 187 6: t. I have not been able to sec Lachmann's edition of lwei11 myself.
Ho CllArTER THREE The First Phase of Lichmann's Activity as a Textual Critic

Obviously, it is not up to me to judge this part of Lachmann's activity, tic and orthographic modernizations, it can still preserve readings that arc
given my ignorance of Germanic studies. Nonetheless, I might perhaps ven- more genuine than an older manuscript;25 that the only manuscripts that
ture to say something regarding the pure and simple technique of recensio can really be neglected are the apographs of ones still cxtant 2 6-all asscr
as he applied it in this field. It seems to me that the Lachmann recognizable tions that if transposed into the field of Classical philology sound almost
in these works is indeed fundamentally the same one as the Classical philol- anti-Lachmannian .27
ogist who wrote the prefaces to the Latin poets,22 but nevertheless it is no- Lachmann also attempted to apply a mechanical criterion of choice among
ticeable that as a Germanist lachmann devotes greater care to seeking out variants to the criticism of Germanic texts before he tried to do so with Clas
the manuscript material (as the mere list of the manuscripts he collated al- sical ones. Unfortunately, the rules he formulated for this purpose, in his re
ready makes clear) and greater effort to investigating the kinship relations view of F. H. von der Hagen's second edition of the Nibeltmgcn, arc a real
among the manuscripts or at least to distinguish ing among different redac- brainteaser and probably contain not only several slips but also some gen-
tions of the same work. I think that this slight difference in method can be uine mistakes: on this see below, Appendix A. Lachmann himself ended up
explained partly by the g~eater accessibility of the manuscripts (which are making no use of them in his own edition of the Nibc/1mge11. Still, even this
almost all found in Germany or Austria), partly by the fact that here Lach- first inchoate attempt has some importance for studying the genesis of " Lach-
mann's polemical target was different: in the case of Classical texts he felt mann's method."
above all the necessity of eliminating the worthless excess variants that had
been accumulated in the editions of the seventeenth and eighteenth cen- :t5. Liebmann l876: 1.163 (= Lachmann l81ob: x).
turies, and this ended up leading him co excessive simplification; in the case 2.6. Lachmann 1!176: 1.89: "The only manuscripts that can easily be m:glected arc the
copies of exemplars that arc srill extant."
of medieval texts, on the other hand, he had to refute the opposite prejudice,
27. O n L:ichma nn as an editor of medieval texts, cf. Stackmann l 964, esp. 2.55 and
that the edition had to be based on a single manuscript (as a rule, the oldest
n. 4 5. I am pleased that Stackmann, with h is speci fic experience as a Gcrmanist, has sub-
one),2-' and this induced him co collate other manuscripts and to enlarge stantially confirmed the analy~ i~ I offered with some hcs1tacion in the first editmn of this
the documentary basis for his edition. So we sec Lachmann the Gcrmanist study (and, of course, has cnrid 11.:d it with new observations).
asserting that to restore a text to its original form, at least four or five man-
uscripts are necessary; H that even if a recent manuscript concains linguis-

ll.. For example, in his prefoce ro Wolfram von Eschenbach (L:ichmann r879 l l!!33 l:

xviii) he refuses with typical impatience to study the genealogy of the manuscripts in
greater depth ("But why should one extend the investigation to the smallest details?") and
declares that he has constantly followed one class of manuscripts, however much this
"mighc impair the truth as a whole."
23. Among Medievalists this pre judice lasted for a long time: m Italy it was defended
in the second half of the nineteenth century by Ernesro Monad and acr;:icked by Rajna and
B;:irbi (cf. Pasquali '942.: u4 = Pasqu;:ili 196!!: 2..1 51-5!!>. In our century Bcdier sought
to revive it (Bc:dier t928); against his attempt see Pasqua Ii 193 2.a: r 30- 3 l and t 942: 232.-
33 (= Pasquali 1968: 2.163-64); O;iin 1975 (1949): l41; cf. also below, Appendix C,
n. 3, and Av;ille 1972.: 28-2.9 (who bases his argument in part on the methodological ob
serv;:itions of Contini 1942: 12.9-3 2). As these scholars have observed and as Barbi al
ready understood, the only case in which BCdier's cmerion is somewhat justified occurs
when every manuscript (or group of manuscripts) represents an independent rcdacuon, m
a certain sense an autonomous work, ;:is often happens in texts of popular tradition (or in
literary texts revised by the author): if practical considerations make us abandon as im
possible the idea of publishing all the different red;:ictions separately, it is indeed better to
select only one of them rather than to compile a contaminated text that does not corn:-
spond to ;my redaction that has ever really existed.
2.4. Lachmann 1876: 1.89, and cf. 16l, 2.85-!!6 (= L:ichmann 1817).
Addendum to Chapter Three

tween rccensio and cmendatio. Furthermore, when he enumerates the three

degrees of "difficulty" in restoring the genuine text, he creates a certain con-
Addendum to Chapter 3 fusion between different points of view. The "ease" of having a single
manuscript available must apparently be understood here in the sense of
"convenience": in chis case one need not bury oneself in those laborious in-
vestigations of collating various manuscripts and stemmatology that Her-
mann continued to find a heavy burden, and sometimes a useless one (and
it must be conceded that he was not entirely wrong). But although it is not
necessarily a misfortune for a philologist to have only a single manuscript
upon which to establish his text (on this point, cf. S. Mariotti 1971), that
does not mean that the presence of more than one manuscript makes it
"more difficult" to reach the authentic text, as long as one of the manu-
In the last years of his life, did Gottfried Hermann feel the need to refer to scripts preserves good readings that are lacking in the others or helps reject
manuscripts instead of taking the vulgate as the sole basis for a conjectural lectiones si11g11/ares by agreeing with other manuscripts. To be sure, it is
criticism? Did he recognize at least in part the justice of Lachmann's criti more "difficult" to work on more than one manuscript, not only because of
cisms of him in his review of the Ajax many years before (cf. above, chap. 3, the greater toil involved but also because of the often arduous problems of
n. 4)? The answer seems certainly yes. Antonio La Penna has brought to my choice among variants; but it can also be "easier" for the purpose of arriv-
attention this statement by H ermann from the beginning of his essay "De ing at the correct reading. To take a banal example at random, if only a
hymnis Dionysii et Mesomedis" (Hermann 184:i.: I = Hermann 18:i.7-77: single representative of the Palatine redaction of Plautus (and not even one
8.343): "There can be no doubt that the reliability of the written documen- of the better ones) had survived, and we possessed neither the ocher Palatine
tation must be examined first of all when an erroneous transmission is manuscripts nor, worse still, the Ambrosian palimpsest, then the work of
emended. Usually this is easy when we have only one exemplar of a text; it editors of Plautus would in a certain sense be "easier," but it would be far
is more difficult when there are several that differ from one another; it is more difficult to make a good edition. And finally, it is true that emendatio
most difficult of all when we suspect that the true form of the text has not is the most d ifficult task that the textual critic confronts, at least in very
been transmitted but must finally be tracked down by conjecture." many cases; but emendatio is not a "third degree" of difficulty with regard
Sauppe's testimony, which we quoted above and which dates from 1841, to a first degree constituted by the existence of a single manuscript. The
only one year before this essay of Hermann's, therefore corresponds at least comparison is not one between homogeneous entities: a single manuscript
in part to the truth, even if with some distortion. Indeed, as we shall see in may be swarming with errors and stand all the more in need of conjectural
chapter 5, by this time not only Lachmann but also Orelli, Madvig, C. G. criticism.
Zumpt, and Ritschl had been repeatedly declaring for many years that it was Perhaps these are pedantic and overlong criticisms of a passage written
necessary co investigate the documentary foundation of the texts, the fides by a great Classical philologist whose greatness lies elsewhere, namely, in
script11rac. Hermann could not possibly have remained entirely deaf to these that kind of conjectural criticism (or in that intelligent defense of the trans-
voices. mitted reading) which he always considered the loftiest kind of philological
On the other hand, by this rime he was too committed to his own version labor, as this very passage demonstrates with all its incoheren~e.~
of philological practice (which indeed had produced many splendid resoles)
and perhaps was also too old to "renew" himself completely. He never ap
plied the new methods in a way corresponding to that statement of prin-
ciple, not even in that very same article, except for a few sporadic references
to manuscript readings; what is more, the sratement itself is not a model of
clarity and coherence.
Above all, Hermann seems to be still committed to the distinction be-
tween emc11datio ope codic11m and ope ingenii, rather than to the one be-
L3chmann as an Editor of the New Tcsramcnr

4 Alongside Bentley's example, it was Bengel who influenced Lachmann.4

The two families, Eastern and Western, which Lachmann distinguished
Lachmann as an Editor of within the tradition of the New Testament, correspond in substance to
Bengel's two nationes, Asian and African;-' and above all, the mechanical
the New Testament method of choosing the variants derives from Bengel. Lachmann writes in
his Rechenschaft of t830: "Every reading shared by both families, whether
it is the only reading attested or both families vary in the same way, thereby
proves itself to have been widespread [verbreitetl and is worth accepting
into the rext; a reading of the one family and a different one of the other
family have equal authority for me; a reading attested only by one gart of
one of the two families is to be eliminated (even if perhaps it is the only gen-
uine one)." 6 This formulation is more detailed than Bengel's, which we
From Germanic studies let us return to Greek and Latin texts. We can dis- quoted earlier (pp. 65f.); but Lachmann's debt to Bengel is dear, especially
tingu ish two parallel lines in Lachmann's activity as a textual critic during if we compare this passage of the Rcchmschaft to the extremely confused
the fifteen years from 1830 to 1845. On the one hand he edited a series of rules Lachmann enunciated in his review of von der Hagen (see pp. 81,
texts transmitted in only one manuscript or editio princeps, for which it was 139ff.), when in all probability he was still unacquainted with Bengel's work.
naturally not recensio that posed problems but only emendatio: Genesius Of the three cases that Lachmann distinguishes in his Rechenschaft, he
(1834); Terentianus Maurus (1836); Gaius (1841), an edition begun by divides the first one into four in the preface to his editio maior of 1842, so
J. F. L. Goeschen and completed by Lachmann; and Babrius ( 184 s).I On that the rules become six. 7 The third one introduces a geographic consider-
the other hand he edited the New Testament: the editio minor appeared in ation and is particularly interesting: the agreement of some m;inuscripts of
18JI, the first volume of the cditio maior in 1842.: o ne family with some of the other has greater value if the manuscripts come
The last important edit,ion of the New Testament had been Griesbach's from places very distant from one another. In fact, distance is a guarantee
(see above, p. 70 ). Lachmann severely criticized Griesbach's persistent ac- against " horizontal transmission," against contamination.8 Already Bentley
quiescence in ~ the textus reccpt11s (Lachmann 18 76: 2..151), and he finally
achieved what Bentley had planned: an edition founded solely upon the an- 4 . Lachmann states in the preface ro his editio maior, " these men alone (sc. llcndey and
Bengel] understood what it is that I call editing (recc11serc]" (lachmann r !142.-50: 1.xxxi).
cient manuscripts and Saint Jerome's Vulgata. Not even Lachmann was
5. Already before Lachmann, Semler (in the preface to Semler r765) had designated
spared unpleasant accusations by narrow-minded theologians,l but times
the two families as " the Eastern" and "the Western" ones; he called them reccnszo11es, a
had changed and' the authority of the recept11s was no longer able to re term that suggests he conceived them as "ancient editions," "redactions," r:nher than as
assert itself. mere gene3(ogical groupings.
6. L:ichmann 1876: 2..2.57; his qualification "even if perhaps u is the only genuine
r. Shortly 3her his edition of llabrius, in the same year, r845, L:ichmann published a one" precedes the distinction between lefOlt vraie (true reading! and /cfOll a11t'1e11tiq11e
little edition of Avianus in llcrlin. In this case there were many manuscripts, bur L:ichmann (authentic reading! established much later by Havel 19 11 :42.J - 2.7. But, as L:ichmann
pushed his desire for simplification to an extreme and hence d id nor even indicate them, lumsclf declares {pp. 2.5!1, 2.69 ), in general he followed the "E:isrern" family in the cd1tio
inste3d using in the critical apparatus vague terms like "very many, " "few," "two very 3n- minor, even where the other one offered readings that he himself recognized to be supe-
cienr ones," "3 very 3ncient o ne." The only value to this edition consists in some good co n- rior. He remedied this incoherence in his editio maior, bur this too was based on a manu-
jectures. The first scho lar to undertake a true rece11sio of Avianus was Frohner in r 862.; script material that was too incomplete to allow him ro really reconstruct the original
cf. Frohncr l 861: xii. readings of the rwo families (which as it turns o ur arc more rhan two and all contami
2.. Lachmann r 8 3 r, without a preface or critical apparatus (buc Lachmann cxpl:uned nated): cf. Tischcndorf-Gcbhardr l 897: :.. 7S 8-6 r ; Gregory 1900-1909: :i..966- 8:.; Merz
rhe critcri3 he had followed in Lachmann 1830 = L:ichmann 1876: 2..2.50-72., which I ger 196!1 ( 1964): chaps. 2., 6, 8 (chap. 6 must be used wirh caution).
cite); L:ichmann 184:.-50. 7. L:ichmann 1842.-50: l .viii; Quentin 192.6: 3<1-35 gives a pa raphrase of the six rules.
3. Lachm3nn 1876 : :i..t 5t; Lachmann 1842-50: 1.xxx-xxxiii and passmi; Hcrrz 8. Lachmann t !142.-50: 1.viii: "Furthermore, the confirm:irion in the agreement of wit
1851: r6o, 165-67. ncsses brought together from different regions is greater rhan the danger arising from som e
86 CHA J>T llR FOUR Lachmann as an Editor of the New Testament R7

had made an observation of this sort, but he connected it with the reliabil- d ition of Gregory of Nyssa, he had encountered a case in which a genuine
ity of the New Testament tradition as a whole, not with the choice among reading was preserved by two manuscripts that came from "marginal"
variants: "'Tis a good providence and a great blessing, th;it so many manu- zones. 11 It is understandable that when he read this passage from Lach-
scripts of the New Testament are still amongst us; some procured from mann's preface to the New Testament he was pleased by the coincidence and
Egypt, others from Asia, others found in the Western churches. For the very thought he could detect in that preface a less mechanic;il, less Lachmannian,
distances of places as well as numbers of the books demonstrate, that there more Pasqualian Lachmann than the better-known Lachmann of the Lu-
could be no collusion, no altering nor interpolating one copy by another, cretius commentary.
nor all by any of them." 9 And we have seen (chap. 2, n. 20) that Bengel in- But it seems to me that Lachmann's preface to the New Testament, reread
duded 1 a geographic consideration among his rules as well. with greater detachment, is substantially in line with the whole develo pment
Pasquali insisted on the importance of this geographical criterion, em- of his textual criticism, except for the few indications of a more CO]llplcx
phasizing its analogy with the criterion of lateral areas used by "neolin- conception of manuscript tradition that can be found in Lachmann's work
guilitS." 111 Shortly before, while Pasquali was studying the manuscript tra- as a Germanisr, as I indicated in the preceding chapter. Ir should be noted
that the criterion of lateral areas can be understood in two different ways in
both linguistics and textual criticism, one "mechanical" (wirhout giving this
manuscripts from the same countries which differ from them either through carelessness term any pejorative meaning) and one sociocultural.~ In the former case,
or almost by intention." A few pages earlier Lachmann had expressed himself more sim- what is involved is an elementary calculation of probability: if two wimesses
ply: "Above all we shall rake account of the most ancient (sc. witnesses I, and among these at a great distance from one another agree with one another with little or
of such ones as derive from the most widely separated places" (p. vi). Cf. also: "Where no' possibility of communication, and if the agreement is such that it can
manuscripts from distant regions agree with one another, this is likely ro have been prop-
hardly be attributed to chance (that is, to polygenesis of innovations),': then
agated from very ancient sources into the various places: on the other hand, we must sus-
ir must be concluded that they preserve a genuine tradition. This is how
pect that unique readings of individual exemplars were born at home and are not derived
from a common source" (p. vii). But in this last passage what is opposed to agreement
Lachmann (and already Bentley and Bengel, as we have seen) understood
among "distant" manuscripts is not agreement among "near" manuscripts, ns being less the geographical crirerion; so too Giacomo Leopardi and Wilhelm von
decisive, but rather the readings of individual manuscripts: unless we understand c;iccm Humboldt, who applied it to certain agreements between Latin and Sanskrit
plar in the sense of "sub;irchetype," there is a certain conta mination between the "geo against Greek, in unpublished writings that Lachmann could not have
graphical criterion" ;ind elimirratio lectio11um si11gulari11111. known. 12 On the other hand, while the neolinguists' and Pasquali's formu
9. Bentley t8 J 6-38: 3.350 (from his Remarks 11po11 a Late Discourse of Free lation does not entirely neglect this "mechanical" aspect, it emphasizes in
Thi11ki11g; see ;ilso above, p. 63f.): l<Jchmann had this passage in mind and expresses him stead a sociocultural fact: the "province" is more backward, more conser-
self in a similar way on p. ix of his preface. Less convincing is Lachmann's anempt (p. vii)
vative than the "center": "victorious innov01tions," as Pasqua Ii writes (19 5 :z.a
to find an anticip;ition of the geographical criterion in S. Jerome, Ep. Ad Damasum, in
(1934): 7), "usually radiate outward from a center toward the periphery,
Migne 1846: 559.
t o. Pasquali t 951a ( t934): xvii-xviii, 7-8, 159-60, :1.2.4n 3; cf. also Pasquali 195 ta:
and they do not always manage to reach it." Here the word "center" is not
117 - Pasquali 1968: 1.440 on a similar idea already expressed in 1891 by Michele B;irbi. to be understood in a purely spatial sense, and what guarantees the greater
On the value Pasqua Ii attributed to this criterion, cf. also Pasquali 19 5:1.a ( 1934): 34 5: "in degree of archaism of facts attested in the lateral areas is not the lack of com-
deed, this is one of the ideas that prompted me to write this book." Pasquali himself cites munication between such areas bur the lesser dynamism and cultural pres-
Bartoli 1915: 6-:.3. Sec also the conclusion of Bartoli 1943: 76. Nc:olinguistics (also c;illed tige of their inhabitants- so much so that, according to neolinguistics and
area or spati:il linguistics) was 3 school originating from the teaching of Hugo Schuchardt
and Jules GilliC:ron which aimed above all to reconstruct the relative chronology of Im
guistic facts on the hasis of their geographical distribution. The "nc:olinguists" were op-
posed ro the "neogrammarians" and were: influenced by the idealism of Bencdc:no Croce It. Pasquali 1959 (1 915): xliv; cf. Pasquali 195u (1934): 158-61.
and Karl Vossler; this influence appears to be stronger in Giulio Bertoni, much less so in 11. Leopardi, Zibaldo11e, pp. :1.351- H of the autograph ( Flora 1937: 1417- 19);
Manco Bartoli, in whom it had ro struggle against a prejudicial distrust of any ph ilosophy d . Timp:inaro 1997 (t955): r61 - 63. More llccringly, Humboldt 1903- :.0: 6.t.153. The
of language (or, as he used ro say ironically, "glottosophy"I. This brief indiation is in- p;issage from Leopardi dares from 1811, the one from Humboldt from 182.7- 19, but both
tended for any non-Italian readers who may not h;ive heard of neolinguistics. were not published until long afrer Lachmann's death.
88 CHArTER FO ll

Pasquali, even facts attested in only ,one marginal area are probably more
T Lachmann as an Editor of the New Testament

methoCJ was an attempt to really achieve rece11sio sit1e i11terprctatio11e, but it

ancient. u very quickly revealed its sterility. IN
In any case, the fact that Lachmann is rigidly Lachmannian as an editor
of the New Testament too is demonstrated by the very beginning of his pref- 18. PasquaIi 193 ;z.a: 13 1 writes justly: "Already Lachmann did not trust i11dicim11
ace (p. v}, in which he not only distinguishes between recensio and emen- enough; Qucnrin exaggerates this distrust in a somewhat too C:itholic, monkish way."
datio more clearly than in his earlier writings but also repeats quite drasti- D:iin 1949: 16:i-64 demolishes Qucnrin's ~ccdonc " theory with affectiom11e irony in a
paragraph titled "La grande illusion" (The grand 1llusion) (the title reproduces that of
cally the difference between recensio and interpretation: "We both can and
Renoir's famous film :ind is already ironic); Dam's criticism of Quentin in the correspond-
must edit (rccenserc) without interpreting"-the very principle against
ing paragraph of l:iter cdinons of his work is more perfunctory :ind less effective. More re-
which Pasquali always protested. Lachmann had already objected against cently, Balduino 1979: 132-34 has raised calm but radical objections to Quentin. The
Griesbach in his Rechenschaft of 18 30 that internal criteria for the choice Quenrini:m method has been revived hy theoreticians of the automation of textual criti
among variants "by their nature almost all cancel each other out," 1 ~ and cism (froger 196!!; Zarri 1969, 1979, and other works of Zarri's which I do not cite for
this disdain for internal criticism is also why he unjustly disparaged Wett- lack of space). I have had the benefit of exchanging ideas with my friend Zarri, J nd I am
stein.11 Without a doubt, there are cases in which internal criteria cancel convinced that new methods can yield great advantages, and already have done so, for
each other out: sometimes a reading is diffecilior in one respect, facilior in unraveling to a cemun extent manuscript traditions that arc very complex and richer in
another; sometimes the criterion of the lectio diffecilior runs afoul of the 11s11s variants than in real corrupuons. More than ever I find mistaken and reactionary any hos-
tility against the methods of automation which is hased on rheroricJl claims for the unique-
scribc11di, especially if it is applied with scholastic rigor. 16 But recensere sine
ness of the "human sp1rn." But one fact remains: Quentin's method is powerless in the face
interpretatione was never anything more th:m empty boasting, even on
of the ob1ccuon that only coincidence in error can indicate the kinship between rwo miln
Lachmann's part, not only because he had at the very least to understand the usu1pts; comctdence m the correct reading proves 11otbi11g, since it is il fact of conscrva
readings of the manuscripts in order to be able to classify them, but also be- t1on th:it c:m also occur m manuscripts unrelated 10 one :mother. The recourse to Quentin's
cause after the eliminatio lectiommi singulari11m there still remained a large method on the p:irt of theorcticiilns of automation is a "sad nL-ccssity," since a computer
mass of variants of equal documentary authority from among which he too is not Cilpilble of dtstmguishing a correct reading from a corruption: for that, we would
had to choose on the basis of internal criteria. 17 More recently, Quentin's need an "artificial ph1lolog1st," something we do nor yet possess.

13. Pasqua Ii 19 pa (19 34): 160 n. 1. Barroli 19 2. 5: "The earlier phase tends ro he pre
served m the more isolated area" (p. 4); ~ 1r of rwo linguiscic phases one is found in lateral
an:as and the other in an area in the middle, the phase in the laternl :irea 1s normally the
earlier one, as lo11g as the area i11 the middle is not the more isolated area" (p. 7, my ital
ics). The clement we h:ive called "sociocultural" appears in the writings of orher srudenrs
of geolinguistics (e.g., Terracini, Bcnoldi, Devoto) even more than in B:irtoli, who usually
preferred to limit himself to statistical ohserv:itions. On the criterion of the marginal ar
eas in rextu:il criticism, cf. Corti 1961 and Sanroli 1961: 116 and n. 9.
14. Lachmann 1876: 1.;z,52.
15. Lachmann names Wettstein only rarely, unlike Benrley and Bengel, and not favor
ably; e.g., in the preface to Lachm:inn 184;z,-50: xxiii.
16. This kind of contrast not only corresponds to the one bcrween the grammatical
theories of analogy ( = 11s11s scribe11di) and anomaly (= lectio di{ficilior), even if wh:it is
invol,ed 1s neither a mechanical correspondence nor an irreconcilable opposition; the two
contrasts arc also connected historically, sraning from the philology of Alexandria, as I
shall try to explain elsewhere. For a partially similar repercussion of these two grammati
cal theories upon textual criticism, cf. Kenney 1974: 1 q.
17. Ir will be recalled that according to Lachmann's division, which is still generally
followed, recensio also includes the choice among variants of equal external authority, i.e.,
what Paul Maas calls se/c?c:tio.
Contributions of Lachm:1nn's Contcmpor:iries

5 substantially correct genealogical outline of the tradition of Cicero's Verrine

orations in his Epistola ad Orelli1m1 [Epistle to Orelli]. l
Contributions of This was the first exchange in a fruitful dialogue between the Classical
philologist of Zurich and his younger colleague from Denmark, which con-
Lachmann's Contemporaries
tinued in the following years and can be followed in the inaugural disserta-
tions and editions of individual works of Cicero's which were published by
the one scholar or the other:' This was a genuine dialogue in which, even if
from the very beginning Madvig displayed a more vigorous and original
philological personality, Orelli too made contributions that were far from
negligible and demonstrated that he knew how to study certain problems in
depth wh ich he had barely touched on in his Cicero edition.
And soon the dialogue was enriched by the addition of a third interlocu
In the meantime, during the 1830s, other philologists had made contribu- tor, Carl Gottlob Zumpt from Berlin. A student of Wolf's, Zumpt had
tions of enormous importance to the method of textual criticism, especially learned from him to disdain eclecticism and to demand a text based con-
regarding the genealogy of the manuscript tradition. stantly on the best manuscripts. He put these precepts into practice. first in
"The families and as it were lines of descent, both of manuscripts and of an edition of Quintus Curtius Rufus, then in an edition of Cicero's Verrine
editions, must be established": this was the plan Johann Caspar Orelli pro- orations that appeared in 1831.5 This latter edition bears the traces of toil-
claimed in the preface to the first volume of his edition of Cicero, which ap- some revision: Zumpt had to begin his work on the basis of old printed edi-
peared in 1826. 1 He himself only succeeded in fulfilling that plan to a very tions because of the difficulty of obtaining reliable collations; only later was
small extent: a whole lifetime would not have sufficed for a truly critical edi- he able to go back to the manuscripts, and he d id not feel up to eliminating
tion of all the works of Cicero, especially at that time, when manuscript col- completely the old and by now quite superfluous hodgepodge from the pref-
lations were so difficult to obtain; and the interests of Orelli, an attractive ace and critical apparatus.~ All in all, his genealogical reconstruction makes
figure as scholar and religious and political reformer and an intelligent fol- little progress beyond Madvig's much more concise one. which Zumpt read
lower of Pestalozzi's ideas," were too various to allow him to dedicate him- only when his own work was almost finished.' But the stemma codicmn that
self entirely to this project. For that reason his edition, which was completed Zumpt drew up as a conclusion and summary of his investigation (1831:
in only six years, from 182.6 to 1831, ended up being in large part a hurried 1.xxxviii) was a very important technical innovation:
revisionb of the preceding editions. 2 Once again we find the same contrast
between editorial theory and practice that we have already noted, for ex-
3. Madvig 1818: 7-10. Cf. Klotz r913: viii.
ample, in Ernesti. But Orelli had the merit of at least reviving the program 4. For Orelli see esp. OrelliBeier 1830, preceded by an Epistola critica ad lo. Nie.
matic demand for the investigation of the genealogical relations between the Madvigi1m1; Orclli 1831, 1835, 1837. For Madvig see the prefaces and dissertations col
manuscripts in a period in which it had fallen into neglect after having been lectcd in Madvig 1834 and 1841. On Madvig sec also below, p. 97. As is well known,
asserted with such energy at the end of the eighteenth century (see above, Orelli undertook a new edition of Cicero in collaboration with Baiter in 1845; aher his
pp. 75f.). And just as his brief reference to the manuscript tradition of Lu death it w:is continued by Baiter and Karl Halm.
cretius provided a starting point for Johan Nicolai Madvig and ocher schol- 5. C. G. Zumpt 1816: esp. xvi; later, in r849, Zumpt published :in editio maior of the
ars before the publication of Lachmann's Lucretius (see the following chap- same author, cf. K. Miiller 1954: 798. C . G. Zumpt 1831. On his relations with Wolf sec
A. W. Zumpt 185 l : 17, 31-34. Zumpt's preface to the Vcrri11cs reveals Wolf's influence in
ter, pp. 102ff.), so too his edition of Cicero stimulated Madvig to give a first,
its very terminology, for example, in the distinction between recc11sio and rccog11itio:
"Thus, employing a recent, but useful, distinction, it will be truer to say that he published
:1 revision (recognovisscl of Grutcr's text than a critical edition (rccenmisse) of Cicero"

t. Orelli 1816-38: 1.xiii. (C. G. Zumpt 1831: 1.xxv-xxvi);cf. above, p. 71f.

1. It was only for the manuscript traditions of Cicero's letters that he was able to do 6. He himself explains this in his preface, C. G. Zumpt r 8 3 1: 1.xxxiv.
an :iccurate and panly origin:il job: see the preface co Orelli rB:.6 - 38. 7. C. G. Zumpt 1831: 1.xxxv-xxxvi.
CllAP T ER FIVE Contributions of Lachmann's Contemporaries 93

Cod. antiquus dcperditus Cod. Vatican. In the field of Latin texts, in 1809 and 1813, even before Schlyter, Johann
rescriptus August GoerenzJ had summarized in the form of a tab11la (table) or tabeJJa
Nann. Fabric. Par. 7774 As. Regius !small table] his subdivision of the manuscripts of Cicero's De legibus and
(Merell.) (coll. Havn.) De fi11ibus into two classes. 9 But Goerenz's classification still referred to the
libr. II et III. libr. IV et V value of the manuscripts (on the one side the codices potiores [superior man-
Lag. 42 pr. m. uscriptsJ, on the other the deteriores [inferior ones]), not to their relations
in illis libr.
Par. A Erfurt. Par. B Lag. 29 of derivation; and his tab11lae were simple lists of the manuscripts of both
categories, not genealogical trees. Zumpt, on the other hand, already in-
I dicates the manuscripts' derivation, even if somewhat less precisely than
Par. D (Steph.)
Cuiac. Schlyter; and so far as we know,' he is also the first to incroduce the term
Guelf. 2 Leid. stemma (that is, precisely, "genealogical tree"), which will end up prevail-
(Memmiani Lamb.) ing over other synonyms.'
Guelf. 1 vulgares
The young Ritschl is also connected with Wolf, even if he did not study
directly with him. Typically Wolfian is his interest in the history of the text
Is this the first stemma codicum that was ever actually drawn up, and not (understood in a very broad sense, as the history of Alexandrian and Byz-
only planned like Bengel's tabula genealogica? No. In the first version of my antine philological culture), which he showed starting with his edition of
study (Timpanaro 1959: 213 ), I attributed the innovation to Ritschl (see be Thomas Magister's 'EKAO)'~ ovoci -rwv Kat i>11c1Twv J\TTLKWV [Selection of
low, p. 93f.), though I added, "I know well how cautious one must be in Attic nouns and verbs]. 11 If I am not mistaken, the very term history of the
such assertions of priority." My caution was all too justified! When this text comes to Ritschl from Wolf. 12 But within this broad perspective he also
study was first published as a book (Timpanaro 1963a: 46 and n. 1), I was
able to backdate the first stemma from Ritschl to Zumpt, thanks to a kind 9. Goerenz 1809-13: 1.ix -x, xxxi-xxxii, 3.vii, xxix. Ori Goercnz, sec Kiimmel 1879.
reference by Konrad Muller. But, a little later, a Swedish scholar, Costa 10. He presents his stcmm:i with the words, "Thus is productd more or less this
Holm, demonstrated that Carl Johan Schlyter had drawn up a stemma of a srcmma of the manuscripts" (C. G. Zumpt 183 t: 1.xxxviii m). The term stcmma re-
type more "modern" than Zumpt's in the first volume of an edition of an- appe:irs later in Schneidewin's preface to M:irtial (1841: cxxxi; sec below, pp. 101f.I, m
cient Swedish legal texts published in 1827 (and thus four years before Bernays's dissertation on Lucretius ( 1847; see below, ch:ip. 6, n. 8), and in Rirschl's edi-
tion of Plautus (Ritschl 1849-54: 1.xxxviil. E:irlicr, Rirschl had used other expressions:
Zumpt's edition of Cicero), in a field very remote not only from Classical
in his edition of Thomils M:igistcr he writes, "T he connection of both relations can be dis
philology but also from the study of Germanic literary texts as Lachmann
pl:iyed in :1 single di:igr:im in this form" (183;z.: xxx), :ind in his articles on Dionysius of
had practiced it, and furtht;rmore that already at an extremely early date he Halicarnassus of 1838 :ind 1847 (see below, n. 13) he spc:iks of artifici11m ldcviceJ. But,
was quite aware of certain causes of disturbance.in a purely "vertical" trans- as Konrad Muller points our to me, rhe notes Ritschl took in Italy in l 837 alrc:idy contain
mission (noc so much contamination as rather the conjectural activity of in- the Germ:in expression corresponding to stemmo: ";i form:il genealogical rrec ('einen
dividual copyists)." Given the subject matter of Schlyter's publication, it is formlichen genealogischcn Stammbaum') for the descent :ind relation of all the fathers,
not surprising that it escaped the notice of Classical philologists: it is almost brothers, cousins, :ind nephews in the great family of Pl:iurus manuscripts" (cf. Ribbeck
certain that Zumpt drew up his stemma without knowing that he had such 1879: 1.:1.01 ). M:idvig speaks of tabula in his 1833 dissert:ition (see below, p. 97), as Ben-
an acute predecessor in Schlyter.c gel already hild ii century earlier (cf. above, p. 65); Schlytcr (cf. n. 8) used the expression
schema cognatio11is CodiC11111 man11sc(riptorum) (scheme of rhc kinship rel:itions :imong
the milnuscripts), cf. Holm 197;z.: 53.
II. CollinsSchlyter 18:.7. Cf. Holm 197;z.: 74-80, SJ (reproduction of Schlyters 11. Ritschl 183:.. We must also recall Ritschl's studies, only a little later, on Orus :ind
stemmal, 77-79 (comparison with other, slightly l:iter stcmm:1s). Perh:ips Holm ovcresti Orion, and on the Alexandri:in libr:iries, which demonstrate his interest in the history of
mates somewhat Schlytcr's contribution rcg:irding the reconstruction of the tcxr of the an ancient philology.
cestor (Holoi 197:?.: 60-641: as for as I can rel! (bur it musr be borne in mind th:ir I o nly 12.. Ritschl writes, "in order to sec cle:irly rhe history of the texr, if I m:iy use a recent

have access to rhe passages rhnt Holm rr:inslares or summ:irizes), Schlyter oscillated be- term" (183;z.: xxix); cf. also below, n. 14: "the history of the rext (as we now say) ... "
tween rhe criteria of rhe "best manuscript" and of the majority of the witnesses; rhus in And Wolf had spoken of the Geschit:hte des Textes (history of the tcxtJ, for cx:imple, in
this rcg:ird he still lagged behind Mndvig (see below, pp. 97f.) and Lachm:inn. his preface to Pl:iro's Symposi11m (f. A. Wolf 1869: 1.143).
94 CHAPTER FIVE Contributions of Lachmann's Contemporaries 95

sets himself the task of reconstructing the precise genealogical relations be- ruptions and lacunas; for Plautus he was obliged from the very beginning
tween the manuscripts: on page xxx of the prolegomena to his edition of to make use above all of external indications, but he confirmed them later
Thomas Magister we find a stemma of the manuscripts and first printed edi by an internal inquiry. 14 ln both cases he summarized his results in a stemma
tions, in which lost manuscripts are indicated with Greek letters, according codicum.
to what will become the customary usage: But despite Ritschl's outstanding ability to reconstruct the history of a
text, he did not feel a corresponding need to use the stemma in order to de-
termine the archetype's readings: even where it would have been possible, he
never had recourse to mechanical rules for the eliminatio lectionum singtt
larium as they had been indicated by Bengel and, even better, by Lachmann
Rb A La G B c Ra Ox in the Rechenschaft to his editio minor of the New Testament. In his disser-
tatio n on Dionysius of Halicarnassus (Ritschl 1838: :z.6) he traced o ut this
Clearly this is not a genealogy a primo fonte ded11eta (derived from the stemma:
earliest source): Ritschl does not explain the relation among the four ances-
tors cJ>X'l>O, which, as his words and a brief examination of the critical ap-
paratus suggest, is not a pure and simple direct derivation of all four from
an archetype or Byzantine edition. Indeed, as we have seen, not even the
stemma of the manuscripts of the Verrine orations which Zumpt traced one
year earlier started out from a single ancestor for the whole tradition-from
this point of view, once again, Schlyter's stemma is more similar to the ones Chisianus Parisinus Urbinas Venetus
that became customary later.~ Ritschl makes another step forward beyond
Zumpc by specifying the position of each and every manuscript within the This is a typical case of a "mechanical recension." ' But once Ritschl had
stemma: Zumpt had been satisfied with indicating some groups of related reached the point where he could have announced the criteria for recon
manuscripts. Bue the greater specification means greater complexity: Rirschl structing 0 , instead he merely suggested " that we should side with the
was facing a very contaminated tradition-as was only natural, given that Chisianus as long as possible, but on the ocher hand should not fear to take
this was a Byzantine anthology widely used in medieval schools-and was refuge in the Urbinas whenever reason itself demands this or sometimes only
obliged to indicate a double derivation for many witnesses. With all those counsels it." Thus he did not even hint that the reading of 0 could be re
intersecting lines, his stemma already resembles the ones that are found more constructed with certainty every time one of the two apographs of <I> agreed
and more often in recent critical editions and that aim co give some idea of with the two apographs of \JI, or one of the apographs of cJ> and one of "1
the manuscript tradition in all its d isarray, without convenient but arbitrary agreed with one another against singular readings of each of the other two.
simplifications. But when the d isarray is excessive, it is better to give up on Perhaps he did not know the rules Lachmann had formulated; or perhaps
the stemma! ~ (and this is more likely) those rules seemed to him to have no practical use
Some years later Ritschl investigated the manuscript traditions of Diony- fulness in the case of Dionysius. For he was convinced that the Parisinus
sius of Halicarnassus and Plautus by the same method. 13 For Dionysius, he (Paris, Bibliorheque Nationale: Coisl. gr. 150) and the Venetus (Venice, Bi-
based his genealogical reconstruction only on internal data, on shared cor- blioteca Marciana: Marc. gr. :z.7 :z.) provided nothing useful for the constitu

13. On Dionysius of Halicamassus, sec esp. Ritschl 18 38 (incomplete reprint in Ritschl 14. Ritschl himself distinguishes chc two kmds of history of tradition, for example:
186 6 - 68: 1.47 1- 515)and Ritschl 1847 ( = Ritschl 1866 - 68: 1,51 6 -40). On Plautus, af wWhile in other cases it is possible first to invt.-stigatc the history of the manuscripts and
tcr some works completed in 1834-35, esp. Ritschl 1835: 1 n-2.1 6, 486-570 (reprinted then to move on to judging their value (this happens, e.g., in the manuscripts of Plautus),
with additions in Ritschl 1866-68: 1.1-165), see the extensive prolegomena to his cdi we have gone in the opposite direction ourselves, first weighing the value of the individual
ti.on, Ritschl 1849-54. Of course we arc selling aside here Ritschl's studies on Plautine manuscripts and then moving on 10 outlining the history of the text (as we now say) even
prosody and metrics, contained above all in Ritschl 1845. without external evidence" (18311: ::.5). Cf. Ritschl 1!149-54: 1.xxxvi.
CHAl'TEK FIVE Contributions of Lichm:mn's Conremporaries

tion of the text. 11 Later, returning to this question, he concluded that they Now let us return for a moment to studies on Cicero, which, as we have
were dcscripti and hence should be eliminated: thus the choice was really seen, in the years around 1830 were one of the fields in which the ge-
between the Chigianus (Vatican, Chig. gr. 58) and the Urbinas (Vatican, nealogical method was tested concretely for the first time. In 1833 Madvig
Urbin. gr. ro5) alone, so that it became impossible to apply any sort of me- published the first part of his dissertation De emcndandis Ciceronis oratio-
chanical criterion. 16 11ib11s pro P. Scstio ct iu P. Vatini11m [On emending Cicero's speeches Pro Ses-
Thus Ritschl and Lachmann present a striking contrast: the former was tio and In VatiniumJ, 19 in which he drew up a stemma codicum, the fourth
passionately attached not only to ars critica in the sense of Gottfried Her- one in time after those of Schlyter (Collins-Schlyter 1827), Zumpt (1831),
m:inn but also to the history of the tradition as such and tended to trans and Rirschl (Ritschl 1832). Two important innovations are found here.
form it into cultural history, like Wolf before him; the latter was completely First of all, Madvig derives his three families of manuscripts from a single
intent upon the goal of freeing the text from late interpolations and restor- lost medieval ancestor, already disfigured by lacunas and mechanical er-
ing it to the oldest attainable form bur was indifferenr to the kind of ge- rors,20 and indicates that ancestor with the expression codex archetypus [ar-
nealogical inquiry that was an essential step toward reconstructing the ar- chetype manuscript}, which thus takes on for the first time the precise and
chetype. There was also a difference in their explanatory styles: Ritschl narrow technical mea ning chat we attribute to it even today. 21 Second, Mad-
loved didactic clarity, to the poim of lapsing sometimes into a certain ver- vig explicitly sets himself the aim of using the stemma to reconstruct the
bosity, i -r while Lachmann preferred an oracular style made up of sentences readings of the archetype: although the third family has less authority in it-
imp:irted from above and imelligible only to initiates. For the same reason, self since it is represented only by recent manuscripts, it serves in case of dis-
as we shall see, Lachmann did not adopt the use of stcmmata codic11111 even agreement between the first two families to tip the balance in favor of the
in his very last editions, although these had rapidly became popular and one or rhe other.22 Thus Ritschl's genealogical requirement is joined here
were even inrroduced into his own school by Karl Nipperdey.K To use a
stemma to make the history of the text more understandable would have 19. Madvig 1833 ( Madvig 1834: 4 11- 76). In the second eduion, Madvig 1887:
seemed to him a narrow-minded pedanrry. ,333 - 407, he modified this text and suppressed some parts of the original version, because
m rhe mc:inrimc the objections of Karl Holm had convinced him 1h:i1 the tradnion shoulc.1
1S Ri1schl 1838: lS Ritschl speaks sometimes of Vcnctus, sometimes of Vencti, for be divided into only two families, not three.
he was nor enrircl) sure whether the incomplere and secondhand informarion available to 10. Madvig 1833: 7 ( M:idvig 11!34: ,416), M:idviggivcsexamplesof such lacun.is and
him referred to one manuscripr or ro two ( 1838: 19- 10). errors at 1833: 801 ( Madv1g 1834: 41601). In an edition of some of Cicero's speeches
16. See his stemma, quite diffcrcnr from the preceding one, in Rirschl 1866: 1.539, he published in 1830, he did not yet trace out any stcmm:is nor use rhe term archetype bur
490 (for rhe explanation of his symbols). In a specimen edition he published in 11146 he had alre:idy written th:ir all the manuscripts of the Pro Roscio Amcrmo are "derived from
had already written, uThus it is clear rhat for emending Dionysius ir is better to weigh one not very good m:inuscripr and share the same errors, intcrpolarions, :ind lacun:is"
carefully rhe meaning and style of rhe individual words :ind scnrcnces than to investig:ite (11!34: 118). A! we have seen, more and more scholars were beginning to rccogniic that
the value of the manuscripts" (Ritschl 11!66: 1.491). Karl Jacoby adhered to this criterion the derivation of sevcr:il wirncs~s from a single .incesror can only be demonstrated by co
in his Tcuhner edition of Dionysius's A11tiq11itates. As far as I know, no one since Ritschl incidence in lacun:is and serious corruptions, not m correct readings, and not even in in
has ever reexamined the entire m:inuscripr tradition ro sec whether there might indeed be novations rhat can derive from contamination or polygenesis; rhis conviction is already
manuscripts not copied from the Chigianus or the Urhinas. It would be highly advisable clear in Madvig, even if ir will only be explicirly codified farer.
ro do so, especially now, when historians and scholars inrcrcstcd in the most ancienr Ro i 1. Several years larer, the expression reappears in OreIii (1 8 3 7: l.) in the Greek form

man institutions have become more willing to ignore Dionysius's wordy rhetoric and 10 <irxhu11m; he h:id :ii ready used ir oe<"asionally in :i more gener:il sense. Then it was picked
arrribute a high value to his work. up by various other Cl:issical philologisrs and was fin:illy sancrioned by Lachm:inn (see be
17. He himself dccl:ires in his prolegomena to l'l:iutus, "I confess that my inborn na low, p. 103).
ture leads me ro employ a somewhat fuller style rhan experienced readers mighr seem 10 :1.:1.. Madvig 1833: 11 (= Madvig 1834: 434): "While the individuals lsc. inferior man-

need, since I bear in mind ahovc all the needs of those who wish. 10 learn this discipline" uscripts( have ahsolurcly no :iurhority, it has already been demonstrared by examples that,
(Ritschl 1l!.f!;1-54: i.lxix). On Ritschl as a philologist and te:icher, besides Rihheck's cl:is especially when they are compared with the best manuscripts from other families, re:id
sic work (Ribbeck 1 ll79), see the lively evocation in Schmid 1968: 131-37, 141-41. ings can be extracted from them that were found in much older :ind better manuscript:>
I!!. Nippcrdey 11!47-56: r.41! (I owe this indication too to Konrad Muller). Nipper and in the ocigin of rhis lineage, and that when these :ire found rhey can he used to restori;
dey studied first with Morii Haupt, rhen with Lachmann. His prolegomena to Caesar con the authentic text :ind to adjudicate the disagreemenr of rhose most ancient manuscript~
tain no genuine med1odological novelties, but his arguments arc very rigorous and clear. of the lirsr and second family which now should rightly prevail."
C HAl'TER FIVE Contributions of Lachmann's Contemporaries 99

with Lachmann's mechanical method. Madvig is also fully aware that me- Sauppe devoted particular attention to the e/imi'1atio codicum descrip-
dieval manuscripts already present not only mechanical corruptions but also tomm. We have already seen (chap. 1) that excellent examples of this proce
intentional alterations to the text, though to a lesser degree than Humanist dure were already produced by Politian and other scholars who followed his
ones; he considers this fact to be "strange" but nonetheless undeniablek- teaching in the sixteenth century.m Later, Jean Boivin (1663-172.6)-a first
an awareness that Lachmann always lacked.2 J class paleographer and Classical philologist, who unfortunately published
A few years later, Madvig will give a fuller application of these principles little in this strictly technical field and was satisfied for the most part to leave
of textual criticism in his famous critical edition of Cicero's De finibrts with the fruits of his own researches unpublished or to supply them to other
commentary.l 4 Here he mainrains the distinction of manuscripts into melio- scholars, and for this reason is still neglected by the historians of philol-
res [better ones] and deteriores [worse ones}, which Bremi and Goerenz had ogy-demonstra ted convincingly in an autograph annotation at the begin-
already made,25 but he understands the two families genealogically and not ning of Parisinus gr. 2.036 (tenth century) that the text of Pseudo-Longinus
just axiologically; he demonstrates that the manuscripts of De finibus are all nept iK!iouS' contained in that manuscript was the source for all the other ex-
derived from a single archerype and shows how the deteriores too have a tant manuscripts (except for two, which contain the so-called fragmentt1m
function in reconstructing that archerype.26 Tollia1111m): "for where the other manuscripts have a lacuna, whole leaves
A little later (I 841 ) Hermann Sauppe published his Epistola critica ad arc missing in this manuscript, exactly two or four, or even more." Boivin
Godofredrtm Hennannum [Critical epistle to Gottfried Hermann], in which also noticed that the apographs had been copied from the Parisinus (for-
he explains some of the principles of recensio and emendatio with exem merly Laurentianus) at a time when it Wbs already in bad shape but less so
plary darity.27 In th is case too the first impulse will probably have come than later: so that they have some usefulness for establishing the text of cer-
from Orelli and his circle: Sauppe taught at Zurich from 1832 until 1845, tain passages.2~ Almost a century later, Schweighiiuser, a scholar we have al-
in close conract with Orem and his student and collaborator Johann Georg ready mentioned for his edition of Epictetus, demonstrated that the Venetus
Baiter.28 of Athenaeus (Marc. gr. 447, temporarily transported to Paris in the Napo-
leonic era) is the source of all the other surviving manuscripts: where the Ve-
:1.3. Madvig 1833: 10 I Madvig 1834: 419): "But what I have: said must still be
netus presents material damage, the others have omissions or obviously con-
explained, namely, that the: Parisinus, even if it is the oldest manuscript and :is a whole jectural stopgaps." 30
the least intc:rpol:ited one I...1, nonetheless in a few passages I...I has clearly been inter These eliminatio11es follow one another at long intervals of time and do
polatc:d by the substitution of a r;ish conjecture for a corrupted reading in the ;irchetype- not seem to form a continuous scholarly trad ition: it is unattested and seems
certainly a very strange: thing to have: happened in that period." In the specific case of this unlikely that Boivin knew the results Politian and his followers had attained,
manuscript, the existence of "rash conjectures" is anything but proven. But the important nor that Schweighiiuser knew about Boivin (even if on this point more cau 6

thing is the general criterion. For Lachmann's different attitude, see below, p. 11 t.
tion is perhaps necessary); and Sauppe, a conscientious and scrupulous
:1.4. Madvig 1869 {1839); I leave out of account the later editions. See in particular, on
scholar, would have mentioned his predecessors in his turn if he had known
the concept of archetype:, Madvig 1869 (1839): xx-xxi.
of them. It is not impossible that other cases of elimi11atio11es unknown to
:1.5. He cites them himsclfot Madvig 1869 (1839): xxii. Johann Heinrich Bremi, a stu
dent of Wolf's, h;id published an edition of De fi"ibm in 1798 (Bremi 1798). For Goerc:nz
me will be indicated in the future and that the series will become more
sec above, p. 93.
:1.6. Madvig 1 !169 (1839): xii. But Madvig still had roo little faith in the deteriorcs: cf. :1.9. Nicolas Boileau publicized Boivin's discovery, and a little l:irer Z:1charias Pearce
Schichc: 1915: x - xi. supplied some fu rther details: cf. Hc:mmerdingcr 1977: 518. Boivin's annot;ition, of which
:1.7. Sauppe 1841 = Sauppe1896: 80-177. Besides the section on climinatio dcscrip I have only quoted one: phrase:, is cited fully by some editors of the nEpi lNlious, e.g., Rost3
tor11111, about which I shall speak shortly, sc:e especially the passage on the classification of gni 194 7: xxxix. The article: on Boivin in Nouvelle biographic gencralc 1 8 ss: S.4 79-80
manuscripts (8:1.: Sauppc: is quite aware that this must be based on "shared errors," bur he is inadequate. On his coniecrurcs on the Ccsti of Julius Afric:inus, of which he st:irted a
recognizes much more clearly th;in Ritschl, Madvig, and Lachmann that contamination translation that he did not complete:, cf. Vieillefond 1970: 8 5, 87; on his discovery of a p;i-
almost always makes this very difficult or impossible); the excellent criteria for using the limpsest, one of the firsr such discoveries ever, cf. Timpanaro 1980: :1.49-50.
indirect tr:idition ( 111, even if Sauppe does not evaluate some of h is particul;ir examples 30. Cf. Schweighauser t 801-7: t.lxxxviii-ci. Kaibcl 1887-90: 1.vii-viii does not
correctly); and the discussion about different kinds of corruptions :ind the rules foremen seem to me to assign Schweighiiuser enough importance: he claims th;it Dindorf, in a later
datio (1:1.1-77). anicle (Dindorf 1870), was the first to demonstrace that all the other manuscripts are de
:1.!!. Cf. Ziebarth 1910: 148-49. rived from the Venctus. But in foct Schweighauser's arguments arc already conclusive.
JOO CHA l' TER FIVE Contributions of Lachmann's Contcmpor:irics IO I

crowded, but for now I have the impression that the technique of eUminatio ning of that tendency toward as it were voluntary elimi11atio11es, which, as
was "rediscovered" independently on several occasions-after all, in cer- we shall see, spread in the second half of the nineteenth century and has still
tain cases one arrives at this technique with relative ease. In any case, even left conspicuo us traces in Maas's Tcxt11al Criticism.q..U
if Sauppe cannot claim the priority that some have awarded him, he can In the edition of Martial that F. W. Schneidewin published in r842, one
claim to have fully formulated this technique, which had been entirely ne- year after Sauppe's Epistola, he dedicates the whole third chapter of the Pro-
glected in times dose to his by Bekker, with consequent harm to his edition tegomena (pp. c-cx) to examining " what the kinship relations among the
of the Attic orators, and had always been neglected by Lachmann, as has al- individual manuscripts are, what the families share, finally what the author-
ready been indicated. He also gave an excellent application r by demon-
ity and value of each of the individual families is." Nor does Schneidewin
strating that Pa lat. Heidelberg 88 is the ancestor of all the other manuscripts stop at the bipartition of the manuscripts into mcliores and deteriores, by
of the orations of Lysias (except for the Funeral Oration, which has a sepa- now customary: he also distinguishes three families among the meliores on
rate tradition), since a lacuna that in the Palatinus is due to the loss of eleven the basis of shared corruptions (p. cv) . His claim to be the first to give a solid
pages recurs in the other manuscripts without any material damage being foundation to the textual criticism of Martial (p. c) is justified: the genealogy
detectable in them.31 To be sure, this very same Epistola ad Hemta1tnmn he established is still valid in its fundamental structure, just as his preference
and other writings of Sauppe's also contain rash climinationcs based not on for the Thuaneus (Paris lat. 8071 : p. cviii) remains by and large val id .~ 6
indisputable evidence bur on mere presuppositions; ll Sauppe got carried I hope that my brief exposition makes it sufficiently dear that the schol-
away with his first success and extended the procedure of climinatio beyond ars I have mentioned did not limit themselves to applying a method that
its legitimate limits. But it should not be forgotten that, in the case of the Lachmann had already established but that they must be considered among
manuscript tradition of Florus, it was he who reacted against the overesti- the founders of that method, alongside of Lachmann and sometimes before
mation of the value of the Bambergensis (Bamberg Class. 3 1 IE.iii. 22 )) him. This is especially true of Ritschl, Madvig, and Sauppe; and we shall see
taken alone. n A little later the Dutch scholar Carel Gabriel Cobet went on that with rega rd to Lucretius it is also true of Jacob Bernays.
to show far less caution in his famous Oratio, which laid o ut the program
of his future work: though he was a fine expert on the Greek language, he ligently discussed, also concerning earlier cases of eliminatio (Boivin, Schweighiiuscr),
was hyper-analogise as a critic, and regarding manuscript tradition was con- which I had not noticed; but all in all this essay is too apologetic with regard to Cobct.
vinced that the dependence of the rece11tiorcs on one or two late ancient or Kramer I 844- 51: d iv- Iv had warned ag;iinst too hasty climi11ationl!&hut was not lis
medieval manuscripts {which he called archetypi, with a usage closer to that tened to enough.
of certain Humanists than to Madvig's and then Lachmann's) was not just 35 Cf. esp. Maas 1958 ( r917 ): p. 4, sec. 8 (:1 ) (to which we shall return later, p. 155)
one of the possible cases but was the most frequent (indeed was almosc con- and the unsatisfactory d iscussion of recmtiores, 11011 dctcriorcs in the "Retrospect t)l5 6"
(pp. p-53), ending with :1 quotation of Cobet's motto "Comburendi, non conferend1"
stant) and, what is worse, the most "desirable" one.H This was the begin-
(They should be incinerated, not invcstigated)-onc piL-ce of evidence among many th:it
Maas understood nothing of P;isquali's work.'
31. Sauppe 1896: 83-84. Cf. the preface to Baiter-Sauppe 1839 - .0. E:irlier, Bekker 36. Cf. Citroni 197 5: xxxix n. 1, with further details concerning Schm:idewm's :il-
had considered codex C (Florence, Biblioteca Mcdicc:a Laurenriana: Laur. 57,4), one of the rcady " modern" terminology.
apographs of the Palatinus (Heidel berg, Universiratsbibliothek: Codex Pal:uinus gr. 811), Cf. n. in chap. ::..
to be superior to all the other manuscripts of Lysias.
31. Cf. W1lamowirz 1894: 41.
33 Sauppe 1870 = Sauppc 1896: 608-18. Cf. the preface to Malcovati 193 8: vi.
Only with Malcovati's edition has a fair evaluation of the Bambergcnsis been achieved; it
is certainly the best manuscript (Otto Jahn emphasized it), but it docs not deserve to be
con~tantly preferred.

H Cohee 1847: 16-17, 17n, and 101-37. Sece~p. 17: "'And we hope wuh some con-
fidence that someday either one mnnuscript or else very few will supply a complete basis
for the recension and emendation of every single author." Similar ideas recur in various
later writings of Caber's. For criticism of this approach er. Wilamowitz 1981 (I 9:z.1 ): 90-
91; Kenney 1974: 1I8-10. Hemmcrdinger 1977 conu ins much interesting material inrel-
Studies on the Text of Lucretius 103

6 light in 1850; ~ and in that very same year, at the initiative of Rits(;hl, Bonn
University established a competition for a study on the text of Lucretius and
Studies on the Text of Lucretius the criteria for a new edition. The winner was a student of Ritschl's, Jacob
Bernays; his dissertation was published in Rhei11isches Museum in 1847.6
And in the meantime, in 1846, another young scholar, Hugo Purmann, had
dealt with the same problems; he had studied with Ambrosch and Schnei-
der in Breslau but was influenced above all by Madvig's writings on textual
criticism. 7
In the preface to his Lucretius, Lachmann refers to the works of Pur-
mann and Bernays with haughty condescension: "Once Johann Nicolaus
Madvig, an extremeliy erudite man, had indic~ted the proper method, two
very well trained youths, Hugo Purmann of Silesia and Jacob Bernays of
Lachmann's last two works of textual criticism were his editions of the Agri- Hamburg, [. .. ] worked with great energy and some success on evaluating
meusores and Lucretius. The first volume of the Agrimcnsores appeared in the evidence for improving the text of this poem( ... ]. But although they in
1848: Lachmann edited the text but did not provide it with a preface. The vested their labor in this subject in a way that was no doubt quite praise-
second volume d id not appear until 1852., when Lachmann was already worthy, nonetheless, if I may speak the truth, they did nor seem to contrib-
dead; the complex history of the tradition was explained in it very fully and ute anything to it as a whole for me, since some of the things they said were
accurately by his student Friedrich Blume, an excellent investigator of man- ones of which I was already fully cognizant, while other things they either
uscripts (it should be enough to recall his Iter Italicum).1 left out or else mixed up with errors and thereby contaminated, since they
Of all the most important Greek and Latin authors, Lucretius was per were young men insufficiently experienced in recognizing Lucretius's pecu-
haps ~ the most suitable one for applying the canons of the new ars critica: liar talent" (lachmann r85ob: 4). And he absolves himself from any obli-
only a small number of medieval manuscripts, whose genealogical relations gation to cite them in the rest of his work, justifying this refusal by his ha-
are easy to reconstruct; a mass of Humanist manuscripts that can certainly bitual disdain for bibliographical minutiae: "Therefore my agreement with
be negrected w ithout detriment co the recensio, even if they do not derive them in certain matters will make it easy to tell when l think they are right:
from medieval manuscripts known to us (but today such a derivation seems but l prefer not to indicate everything in detail so as not to burden my read
highly probable).1" 1 For Lucretius too, as for Cicero, the way was indicated ers" {185ob: 4). With the same haughty tone he attributed to himself the
by Orelli and Madvig. In 18:z.7 Orelli referred briefly co the fact that all the merit of having been the first to adopt the term archetype in a technical sense,
manuscripts of Lucretius are derived from a single ancestor;.l in 183 i. Mad- although in fact, as we have seen, this was due to Madvig, and Madvig had
vig confirmed this hypothesis and noticed the particularly close relation be already been followed by Purmann and Bernays in this usage.0 K
tween the schedae Cottorpianae and one of the two Vossiani, the one that
Lachmann would later call the Quadratus.4 But it was above all starting in
1845 that Classical philologists focused their attention and rivalry on Lu S Lachmann 185001, 185ob. See Hcrrz. 1851: 139-46 on Lachmann's first idi::i in the
crecius. In that year Lachmann began to work on his edition, which saw the summer of 184 s and on the various phases of the development of his work.
6. Bernays 1847. On Bernays's education and earliest works, Uscner 1902.: 395 is still
interesting. On his personality in general, see Gompen 1905: 106-15. The most lucid and
1. Blume-Lachmann-Rudorff 1848-57.. A. Rudorff was Professor of Law at Berlin. at the same rime most passionate appreciation of Bernays is Momigliano 1975: 1.12.7- 58.
:z.. On the Humanist manuscripts, see below, pp. 108-9, 1 u-u. 7. Purmann 1846. Purmann recalls his teachers ih his autobiography on p. 70; the in
3 Orelli 18:z.7: 86n. Orclli suggested doubtfully that an exception needed to l]e made fl uence of Madvig (not only of his study of Lucretius, Madvig 1832., but also of his pref
fo r the frag111c11t11111 G11dia1111111 (i.e., the Gottorpianae, now in Copenhagen, collated in the a~ to De finibus , Madvig 1869 [1839J) is evident, for example, on pp. 6-7, 15, 19.
seventeenth century hy Marq uard Gude: GI. Kgl. S. 2.11 :z. 0 ) . But there was in fact no rea B. On Madvig, see above, p. 97. Purmann (1846) uses the expression codex arch!!
son to make this exception, as Madvig demonstrated: cf. the follow ing note. typ11s on p. 7 ("All the manuscripts we know of until now derived from one and the
4. Madvig 1831 =- Madvig1 887: 148-62.. same archetype manuscript") and them many other times; Bern;iys 1847: 570 n. (in the
104 CllAl'Tl:R SIX Studies on the Text of Lucretius IOj

In fact Purmann's work and Bernays's di ffer in originality and complete

ness. Purmann proposed several shrewd conjectures butJ did nothing more
for the classification of the manuscripts than add some further confirma
tions (not all of them sure ones) to what Madvig had already noticed, namely,
Exemplar gencris t . Exemplar gcneris 2.
that all the manuscripts are derived from an archetype and that the Quadru-
tus and the Schedae are particularly closely related.9 Purmann had no clear
idea of the importance of the other Vossianus (Lachmann's Oblongus) for
Ludg. 2. Memm. Gott. Vindob.
t. Ludg. Poggianus
various reasons, including the lack of a reliable collation, and he even in-
clined to classify it among the deteriores, since he was misled by its real af I
codd. interpol.
finities with the ltalici. The only thing that was really new, even if it was
still expressed imprecisely, was an o bservation regarding the script of the ar-
chetype or pre-archetype, which Lac hmann neglected.' 1 One can d isagree (and, as is well known / scholars still disagree) about where
Bernays too started out from Madvig, but he studied the problem in much to assign the Poggianus, the ancestor of the ltalici; and one can point out
greater depth than Purmann did and succeeded in designing a stemma that that Bernays did not demonstrate with conclusive arguments that the Qua
is basically correct: 1:!. dratus and Schedae are both derived from the exemplar generis r [exem-
plar of the first famity], even if in fact it is entirely certain; u but everything
else is unexceptionable. And unlike Purmann, Bernays fully recognized that
the Lugdunensis r (the Oblongus) was certainly no less valuable than the
stemma we reproduce on p. 105). Bue the fact th:n Lachmann stated ar the bcgmning of
his preface, "that exemplar, the ARCHETYrE of all the others (that is how I am accustomed Quadratus.
to call n)," made people believe in general that he had been the first person to use the word Bernays's article has exerted very little influence. Until the first version of
in this sense (see, e.g., Pasquali 195:z.a (19341: 3). Cf. also below, p. 113. this study of mine, everyone attributed~ co Lachmann the merit of having
9. Purmann 11i46: 7-i3. I use: L:ichmann's design:ition Schedae to indicate the total- been the first to reconstruct the genealogy of the manuscripts of Lucretius. 14
ity made: up of the Gouorpiani, now at Copenhagen (GI. Kgl. S. :z. 1 1 l. " ), and Vmdobo- Even Usener, who appreciated Bernays's extraordinary brilliance more than
nensis MS 107, which were particularly studied by Purmann 1846: 15 16 after Siebelis
1844: 781i had drawn attention to them. But it was only in 1857 thar it was noticed that,
although fols. 9 -14 of the Vindobonensis belong ro the same manuscript from which the 13. He relied only on the groups of verses of books 1, :!., and 5, which arc lacking in
Gouorpiani leaves were detached, the following fols. 15- 1fl derive from another manu their correct location in the Quadratu5 and Schedae and :ire ad ded at the end: acco rding
script (Goebel-Goebel 1fl57); cf. Diels 19:1.3 - :z.4: xix, who designates the former as V and to him, this was due to an error of the copyist of the sobarchetype from which the Quadra
the l:m er as U. In any case, I neglect this disnncrion here, both because it has no influi:ncc tus and Schedae were derived, exemplar gc:11cris 1 (Bern:iys 1847: 5) 4, 57 1- 7:z.). In real
on the stcmma w dic11111 (G + V and U both belong to the ~ame family as Q) and because 1ty, as Lachmann went on to demonstrate (see below, n. 191, tile arc/Jetypc had already suf
it was still unknown to the philologists I am discussing (Madvig, Purmann, Bernays, Lach fe red the displacement of four le3ves after the Oblongus had been copied fro m it; hence
m:inn). the Quadratus and Schcdac co uld have derived from the :irchetypc after it had been dam
10. Purmann 1li46: 16: " Given Haverk:imp's lack of consistency in indicating variant aged in this way. All the same, the existence of the exemplar ge11cris 1 is demonstrated with
readings, it is not yet clear to me to which family of manuscripts Lugd. 1 (i.e., the Ob- absolute c-crtainry by the many corruptions shared by the Quadratus and Schedae aga1nSt
longus) is to be assigned, to the genuine and more sound ones or to the more recent and the Oblongus (1t should be enough to cite r.6:z., 104, 117; :z..:z. 17, etc.). On this pomt many
inferior ones. Nonetheless, as matters now stand, I would prefer to as.sign it to the infc rcc-ent editors (Bailey, Leonard and Smith, Martin, Buchner) express themselves in" very
rior ones. ~ confused m:inner; not even Dids 19 :1.3- l.4: xvi is q uite satisfactory. This is a method
1 t. See Appendix B, n. 9. ologically interesting case, since it shows th:it lacunas that can derive from material dam
1:z.. Bernays 11i47: 57on. Lugd. 1 and:!. are the manuscripts that Lachmann will call :igc (loss of leaves, holes, etc.) suffered later by the ancestor have no Y:llue as conjunctive
Oblongus and Quadratus; Mcmm. is Lambmus's codex Mc111111ia11us (which seems to be errors (to use Paul Maas's terminology). As far as I know, this point is not to be found m
identical to the Quadratus: Bernays 184 7: 54 6, 550 already inclined toward this identifi- manuals of textual criticism (cf. also below, pp. 17374).
cation I; Poggianus 1s the: ancestor of the ltahci. On the Gotcorpiani leaves and the Vindo- 14. The only editor later than Lachmann who refers with due consideration to Ber
boncnscs sec :ibove, n. 9. n ays's article is Munro r !186: 1.:z.o .
~ -

106 C HAPTER SIX Studies on the Tcxc of Lucretius 107

anyone else in the nineteenth century, did not think it appropriate to include with the Quadratus against the Schedae gives us with certainty the reading
his dissertation on Lucretius in the collection of Bernays's articles he edited of the archetype. It is true that the lack of accurate collations of the Schedae
"since it has been rendered obsolete by Lachmann." 15 Lachmann and Ber- meant that he was only able to make a very limited use of this criterion; even
nays themselves contributed to this unjust oblivion: the former by the con- today, indeed, eliminatio lection11m singularimn is not very useful in practi-
descending reference we have cited, which gave the impression that Bernays cal terms for Lucretius, both because the Schedae are lacking for a large part
had only glimpsed in a confused manner what Lachmann had recognized as of the text and hence recensio is limited to the Oblongus and the Quadra
dear as daylight; the latter by his excessive modesty, which led him to ad- tus, and because almost all the /ectiones singulares are obvious errors chat
mit fully Lachmann's superiority. 16 in any case would not have misled any editors. 18 All the same, there is no
It would certainly be unjustified to accuse Lachmann of plagiarism. A let- doubt that from the point of view of methodology Lachmann was ahead of
ter of his to Moriz Haupt makes clear that he had decided that he would Bernays in this regard.
read Bernays's article in its entirety only after he had finished the first draft Lachmann's ability to calculate the number of lines of every page of the
of his preface to Lucretius to his own satisfaction, 17 and there is no reason archetype-and consequently the number of pages too-was based on the
to doubt that he did just this. But what matters here is not a dispute about length of certain passages that were rransposed or damaged. ~ le was above
chronological priority but rather a comparison between the one scholar's re- all this reconstruction that impressed his contemporaries: "And where is
sults and the other's; and the comparison is not entirely favorable to Lach-
mann. Lachmann's mistake does not consist in his having made use of Ber-
nays's work, but rather in his not having made sufficient use of it.h 18. For a long time I have been hoping rhac some scholar would determine the real
contribution of elimmatio lccti01111m si11g11/ari111n to consrituring the text of Lucretius.
As a matter of fact, Lachmann went beyond Bernays in two regards: eli-
Now this has been done by Alberti 1979: 60-61, and the result is just what might have
mi11atio lectionum si11g11/ari11m and reconstructing the archetype's external been expected: " the lectio11es si11g11/arcs ;:ire made up for the most part of quite banal er
form. Bernays did not think of elimittatio lectiottum singu/ariwn, just as his rors which c;in be eliminated without recourse to mechanical criteria." On the other hand,
teacher Ritschl had not thought of it in the case of Dionysius of Halicar- in about twenty cases a lcctio si11g11/aris supplies what is certainly the right re:iding, and
nassus. Although Bernays understood dearly that the Schedae were not cop- polygenesis of errors must have taken place (according to Alberti 1979: 61; but in some
ied from any of the surviving manuscripts, in practice he reduced recensio passages I think one must hypothesize contamination or, even more, :i correct conjecture
to the two Vossiani alone: "Therefore the two Lugdunenses are the basis on the part of a copyist). Of course, this does not imply a devaluation of the method of
upon which alone the textual criticism of Lucretius rests today" (Bernays eliminatio lectiom1111 si11gulari11111, which has proven its practical utility for many other
texts. And in the text of Lucretius itself there is one passage, not mentioned by Alberti, in
1847: 570). Lachmann, on the contrary, saw that the mechanical criterion
which the method functions usefu lly: 3.1, where the agreement of the Oblongus with the
could be applied to Lucretius, just as to the New Testament, so that the
Schedae serves to confirm the correccness of the initial interjection 0, missing in the Qua
agreement of the Oblongus with the Schedae against the Quadratus and
drams (the concept of lcctio sing11/aris also includes what might be called wlero readings,"
i.e., omissions): cf. Timpanaro 1978: 1}S- 93. In this case Lachmann accepted the b;id
conjecture E and thereby failed co make fruitful use of his own wmethod."
1S Usener in Bernays 188 5: v. In foct, Usener added, "and ir is easily accessible for 19. What is involved are nor only the passages transposed in the Quadratus and Sche-
the specialist." But as we have seen, among specialists too no one has acknowledged Ber- dae (these alone would have no probative value, since the displ;icement could have oc-
nays's just merits. curred in the subarcherype from which the Quadr:1tus and Schedae dcnve: see above,
16. This excessive modesty inspires Bernays's preface to his Teubner edition of Lucre n. 13) but also the transposition of 4.32.3 - 47 before 199-p1 and the mutilation of
ti us, Bernays 18 S2.. 1.1068-7 5 to which the lacuna after 1. 1094 corresponds. This damage is found in all
17. Cf. Lachmann 1892.: 180; Wolfgang Schmid has drawn my attention to this pas the manuscripts, hence ir goes back to the archetype; and since it is explained on the hy
sage. All the same it is not free from ambiguity: Lachmann himself says that he has given pothesis that each of the pages of the ;irchctype had rwentysix lines, it follows with great
a first glance ar Bernays's anicle. So the observations of Kenney 1974: ro7 and nn. 3-4 probability that the transpositions of the Quadrarus and Schedae too, which presuppose
are justified: he also recalls that Lachmann's anirude ro Bernays muse have been influenced the same number of lines per page, go back to 1he archetype (from which the Oblongus
by the very strong hostility between Lachmann and Rirschl, Bernays's reacher. Nonethe had already been copied), not to the subarchecype, as Bernays had supposed; see the note
less, I believe that the essential point, in scholarly terms, nor moral ones, is rhe one I make cited just above. On thi~ point too some of the recent editors (Ernout, Manin) arc any-
immediately below in the rexc. thing but clear.
ro8 C HArTER SIX Studies on the Text of Lucretius 109

this manuscript described with such precision? It was destroyed or lost; and Lachmann does not opt for one of these solutions (as Bernays had done) or
yet there is not a single point in the description that is not demonstrated admit he was uncertain, but instead slips from the first to the second in the
with almost mathematical certainty.., io Nowadays this kind of certainty has course of his explanation without noticing it. He begins by saying that three
been quite shaken: doubts have arisen regarding the exact number of the copies were derived from the archetype, the Oblongus, the ancestor of the
pages and the script of the archetype; above all, scholars have come to real- ltalici, and the ancestor of the Quadratus .ind the Schedae.25 Jn this case we
ize that they cannot use the reconstruction of the archetype for practical would have one of those tripartite stemmas that notoriously constitute a
purposes (that is, to justify transpositions of whole passages) as hastily as very rare piece of good luck in the criticism of ancient and medieval texts:
Lachmann supposed.2 1 All the same, the reconstruction remains valid in its
essentials and is a fine proof of Lachmann's acumen. 0
But Lachmann's explanation regarding the actual genealogy of the man-
uscripts is much more confused and contradictory than Bernays's. The most
important contradiction has not yet been noticed, as far as I know: it re-
gards the place in the stemma to assign the ancestor of the Italici. As we have
already suggested, this question was controversial for a long time and prob- Oblongus ltalici Quadratus Schedae
ably will never be finally resolved:; scholars have disagreed-and they will
continue to disagree-about whetheri the ancestor of the ltalici constitutes In this case the mechanical method would be quite fruitful: the agreement
a third branch of the tr.idition alongside the Oblongus and the shared an- of two branches would give us the reading of the archetype with certainty; 1 "'
cestor of the Quadrarus and the Schedae,22 or whether it was derived from we would only remain in doubt if each of the three branches presented a dif-
a subarchetype from which the Oblongus also descends,23 or whether, as ferent reading, but in practice this happens only very rarely.
now seems almost certain,k it was even derived from the Oblongus.24 But But just one page later, although Lachmann still speaks of three inde-
pendent apographs, he characterizes the ltalki in the following way: "they
:z.o. Hert2185t: 14:z.. Cf. Haupt 1911 (18H): 534 are extremely similar to our Oblongus in every regard and yet they are not
:u. See Merrill 1913: u7-19, 134-35. Doubts about rhe number of pages and lines derived from the Oblongus; for sometimes they disagree with it and go to-
were already expressed by Chatelain 1908: vii; cf. Ernout r948 (t9:z.o): xv-xviii, but this gether with the Quadratus, and indeed this happens in readings that no one
is barbed hy nationalistic animosity. A more balanced and precise judgment can be found could have arrived at by conjecture" ( t 8 5ob: 5). While the derivation of the
in Goold r958. For the script of the archetype, see below, Appendix B. Italici from the Oblongus is decisively excluded, no reason is given why both
z.:z.. This is the view of M:irtin 1969 (r934l; Smith in LconardSmith 194:z.: t14; Bai
the Oblongus and the ancestor of the ltalici could not depend on a shared
ley 1947: 1.4l.-43. According to Buchner 1956: :r.01 ( Buchner 1964. 1.r:z.1-13), the
subarchctype (as we have seen, Bernays's thesis was just this). The good
ancestor of the ltalici w:is in fact independent of the archetype to which both the Oblongus
and the sh:ired source of the Quadratus ::ind the Schedac go back. But his hypothesis has (and, according to Lachmann, not conjectural) readings that the ltalici share
been contested with good arguments by J:>izz::ini 1959: ll:z. - 117 (cf. 54-78) and, even more with the Quadr:1tus might indeed go back directly to the archetype, but they
effectively, by Schmid 1967: 475 . A simil::ir hypothesis had already been suggested by
Chiari 19:z.4, republished with :m addendum in Chiari 1961: 3.1- 17. Chiari 1961: :z.3 - :z.4
is right to defend his article ag:iinst one unfounded objection raised by l'asqu:ili 19 S:z.a revived with new arguments by K. Muller 1973 and, even more analytically, by Cini r976.
(1934): 11 :z.n4; but I think there is some validity ro l'asquali's other objection, that Chiari And indeed it seems to be the most prob:ible hypothesis. All the same, it docs not seem to
too easily considered to be tradition what in the Humanistic manuscripts wean derive ei be possible to perform an elimillatio descripton1111 "to a T" like those of Politian, Boivin,
ther from contamination or from conjecture," especially from conjecture: (ibid.). Schweighiiuser, a nd Sauppe, already mentioned by us. And so it is not surprising that the
:z.3. As h:is been seen, this was Bernays's opinion; i~ was taken up :igain hy Birt 1913: hypothesis of the independence of the ltalici keeps coming up again: cf. now Flores 1978:
:z.:z.. It 1s strange that it has enjoyed so little success. Whoever denies that the lt:ihci :ire de :z.1- 37 (attractively polemical ::ind intelligent but, I think, rather lacking in arguments).
rived from the Oblongus must consider their descent from a sh:ired subarchctype to be the :z.5. Lachmann r H5ob: 3: "From that [sc. archetype ... I m:iny have been derived fro m
most obvious solution, given the undeniable affinity between them. We shall sec that L:ich three copies, as far as we c:in tell." 4: "But from those three apographs," and he goes on
m:inn inclined toward this kind of solution, but incoherently. to speak about eac h one.
:z.4. This hypothesis, already maintained by Dicls m his edition (D1els r9:z.3-:z.4: xxi- :z.6. Except, of course, for the c::ises of a disturbed tradition noted hy Alberti (cf.
xxiii) and thereafter, more succinctly, by R. Heinze, H. Mewaldt, :ind U. 1'1zzani, has been ~ bove, n. 18).
110 CllAl'TE!t SIX Studies on the Text of Lucretius Ill

might also have been found in a subarchetype, which the copyist of the Pog- ther of them is in error), then judgment must be used for constituting the
gianus copied faithfully while the copyist of the Oblongus committed e"o- text, nor can it be said with as much certainty which reading is older, as
res singulares. For the moment, Lachmann does not consider this possibil- there can be disagreement about which one is true or closer to the truth"
ity; but his reference to the extraordinary similarity (in correct readings or (Lachmann 185ob: t o). Here too Lachmann is presupposing a bipartite tra
only in errors?) between the ltalici and the Oblongus seems to indicate that, dition: he says that one must have recourse to illdicium every time a reading
like Bernays, he was already inclining unconsciously toward the hypothesis attested by the Oblongus and the ltalici is opposed by another reading at
of a bipartite tradition.1.27 tested by the Quadratus and the Schedae. If so, then the Oblongus and the
For the moment this is a vague and rather uncertain suggestion; but it be- ltalici would be copies of one and the same subarchetype.
comes an explicit declaration on page 9, where Lachmann writes concern- For the practical purposes of constituting the texf, the problem of the
ing the norms for reconstructing the readings of the archetype:m "It annoys ltalici is not very important, even if they were independent of the Oblongus,
me to have to speak too much about matters of no usefulness whatsoever: as now seems entirely excluded: on this point I think we must agree in any

the whole tradition of the original reading must be derived from the Voss- easer with Luigi Castiglioni. 2~ But the contradiction into which Lachmann
ian manuscripts, except that sometimes the ltalici cancel out the Oblongus's fell confirms that he was not very interested in the history of the text as such.
testimony, and once in a while, as I said,2 H the Schedae diminish the Quadra- He would probably have noticed the contradiction if he had drawn up a
tus's authority" ( 18 5o b: 9 ). Here the o nly function assigned to the Italici is stemma, following the examples of Ritschl and Bernays; but we already
to eliminate the Oblongus's e"ores singulares, just as the Schedae serve in know that he avoided such didactic expedients, and thus he remained a vic-
parallel to eliminate the Quadratus's e"ores si11gulares. Without saying so tim of the very obscurity he sought.31>
and, apparently, withoutn noticing it, Lachmann has slipped from the hy- But there is another oddity in the passage cited last. Every time the two
pothesis of a tripartite tradition to that of a bipartite one, one family repre- families disagreed with one another, Lachmann supposed that there was
sented by the Oblongus and the ltalici, the other by the Quadratus and the a dou ble reading in the archetype. Thus he did not admit that the copyists
Schedae-just as Bernays had maintained. If Lachmann had adhered firmly of the two subarchetypes could have, I will not say voluntarily altered the
to the tripartite scheme he would have had to say that the ltalici "sometimes text, but even fallen into psychological errors of a certain magnitude, for ex
annul the Oblongus's authority, sometimes the Quadratus's and the Sche- ample, substituting one word for another similar one. As Pasquali has justly
dae'," and not the Oblongus's alone. observed, this was a consequence of Lachmann's absolutely mechanical con-
Once again we read a little later (p. 10), "I say this. Wherever it is clear ceptio n of the transmission of texts in the Middle Ages: "He had so low an
from trustworthy testimonies compared with one another that there were o pinion of the copyists' activity of thought and half-conscious transforma-
two readings in the archetype (and this is the case whenever the Oblongus tion that he immediately concluded that the archetype had a double reading
and the Quadratus disagree with one another and it is clear either from the wherever the copies disagreed and the one reading could not be eliminated
Schedae or from the ltalici, either manuscripts or printed editions, that nei- by this mechanical procedure." 31 Later it was demonstrated that there really

27. Oinfor:i 1964: 612 has drawn :itrenrion to this p:iss:igc. Bur perh:ips it 1s :in ex 29. Oistiglioni 1937: sse, 561.
:iggerntion to s:iy th:it "Lachmann lacked in part the very concept of sub:irchetypc ;md the 30. One might perhaps suppose that, before reading Bccn:iys's article carefully, Lach
ability to identify it :ind m:ike use of It in constituting the text" (even if the words "in p:irt" mann h:id become convinced that the tradition of Lucretius was tripartite, and that read
do well to lessen the force of the claim). When he wrote the fi nal version of his prcfac!!, ing Bernays then made him incline toward the other hypothesis. But it remJ ins strange that
Lachm:inn had certainly read Bernays's article with its stemma, in which the "concept of he d id nor revise his whole exposition to make it coherent. Prob:ibly the preface to Lu-
sub:irchetypc" was quite clear. What is instead the case (:ind 1he observations of ours that cretius was written hastily.
follow will demonstrate it) is that Lachmann slipped from o ne stemm:i without sub:irche 31 . Pasqu;ili 1929: 417. In the corresponding passage of P:isqu:ili 1952a ( 1934): 111,
types to another one with suh:irchctypes without noticing it. he seems to change his mind. In someone who knew the German language as perfectly :is
28. The wo rds "as I said " refer to L:ichmann 185o b: 8: "None of these Schedae, nei Pasquali d id, the stylistic harshness of the phrase "der Denk und halbbewusst umgcstal
ther the Haunienscs nor the Vindo bonenses, have any authority of their own excepr when tenden Tiitigkeit dcr Schreiber" Ithe copyists' activity of thought and ha Ifconsc ious trans
they sometimes agree with the Oblongus." formation I is strange, even if the meaning is clear (conscious and semiconscious changes

Iii C H Al'T l: k SI X Studies on the Tex! of Lucretius 113

were double readings in many archetypes, and perhaps in Lucretius's too, piece way and in the most didactic form " (Pasquali 1951.a {1934): 5). The
but on the basis of completely d ifferent arguments.J:! very contradictions we have noced in the co urse of our exposition make such
And another peculiarity needs to be pointed out. We saw (cha p. 3) that a "didactic" quality very doubtful!' And it cannot even be said that " Lach-
Lachmann's distrust for the docti ltalia11i, those insidious interpolators, was mann based his method upon the premise that always and in every case che
deeply rooted and that he tended to neglect a priori the manuscripts they trad ition of every author goes back to a single exemplar already disfigured
transcribed. In the case of Lucretius, unlike that of other authors, such a dis- with erro rs and lacunas, what he called an archetype" (1952a (1934} : 15).
trust would have been legitimate, and indeed we have seen that it ended up Lachmann never formulaced such a general theory, nor, as far as I know, did
prevailing. But in Lachmann it is far less strong than we might have ex- any of his contemporaries. I suppose it was the large number of te~ts in
pected. To be sure, in the course of his explanation he does "declass" the which all the manuscripts agree in errors and lacunas' that gradually con-
Poggianus from being a d irect copy of the archetype ro being a copy of a vinced Classical philologists that there had been an archetype in all cases.
subarchetype; but he never supposes that it might be a dcscriptus (indeed, Of course, the importance of Lachmann's commentary on Lucretius is
as has been seen, he excludes this possibility from che very beginning, and not at all limited to the preface-on the contrary, the commentary is supe
on this point he never belies himself), and he assigns it a function in recon rior to the preface if we do not wish to assume as a criterion of judgment
structing the archetype, even if only a subsidiary one. And even after that only the "method" but co consider instead its results.' In a large number of
initial passage we have already quoted, he asserts that the Humanist manu- cases, Lachmann knew how to perform extremely well not only the rccc11
scripts do indeed have "many(... Jconjectural emendations" but also other sio of the text of Lucrecius but also its eme11datio, 14 even if the i11tcrpolator
passages "which are less emended, that is, which better preserve the reading philosoplms [philosophical interpolacor) he hypothesized never existed, H
of the ancient archetype" (185ob: 6). Must we suppose- but the question and even if Munro was not mistaken in observing that other conjecturers
will have to be studied further -that Lachmann might perhaps have been were su perior to him in curiosa fclicitas !success chat seems intuiti ve but in
influenced by Madvig's less negative attitude toward the rece11tiores of Cic- fact is painstaking!. But above all, Lachmann brought together in his com
ero or jahn's toward those of Persius (see below, pp. u5f.)? In constituting mentary an extraordinary quantity of new observatio ns on the grammar
the cext, Lachmann went on ro take account almost exclusively of che Ob- and o rchography of the Latin language 1n and made an enormous conrribu-
longus and Quadratus, as was only right; but one can say that he presents
himself in his preface, paradoxically, as an almost too ardent believer in rc-
ce11tiores no11 detcriores f'\ 33. Blass 1892: 281 still considered the dcriv:ition of all the manuscripts from an ar
chctype as only one of the possible cases.
In any case, Lachm;inn's preface co Lucretius contains the results he ar
H Excellent conjectures :ire, for example, a/in :it 1.665, diri at !.421, parcat at 2.660
rived at in the course of studying that specific manuscript tradicion and not
(680), comcq11e . . rcdc1111t at 5.679. Lachmann also collected the indirect tr:idmon and
a general methodo logical exposition, such as is found in contrast in Mad used 1t \'cry well, except in :i few cases.
vig's and Sauppe's writings, which have already been recalled, and to a cer- 35. G. Muller 1958 and 1959 have tried to revive this hypothesis. K. Miillcr 1975 ad
ta in extent also in Lachmann's o wn preface to the New Tescament. On the heres ro it :is well: hence his many deletions, which arc :ilw:iys ingenious but, in my judg-
first page of his boo k Pasquali has characterized perfectl y the tone of de ment, too numerous. The problem would deserve a more detailed discussion; but my own
tached superiority that one senses in Lachmann's preface to Lucretius; but view i ~ that, even if in m:iny cases the hypothesis of an original text left incomplete by its
it is not with the same degree of justice, I believe, that he then went on to author, wirh passages not yet definitively put inro their proper IOC:Jtion, "doublets~ not
say that Lachmann explained the method in this preface "in the most com- yet climiniltcd, etc., h:is demonstrated itself to be little more than a convenient expedient
for shirking the responsibility of constituting :1 reliable text, in the case of the De rcr11111
11at11ra what we know about its incompleteness :ind its posthumous publication mah-s
introduced by copyists into the text). But :i typogr:iphical error or the omission of some that hypothesis still quite legitim:ite-withouc considering that it is characteristic of Lu
words is quite unlikely. cretius's style to repeat his verses, to retrace his steps, in a certain way. This docs not jus-
32. Pasquali 1919: 498. For Lucretius cf. Diels t 913-24: xxiv-xxvii (but :ilmost :ill ufy a hyperconservativc mode of crincism, bur it ought to counsel c:iution.
the ~ases Diels cites :ire :inything hut certain), and more recently, but quite unsuccessfully, 36. The most cclebr:itcd of these obscrv:itions is "Lachmann's law~ (185o h: 54, on
Buchner 1964: I. ull-32 (refured by Schmid 19 67! 476-78). But ~:irchetypes wirh vari- Lucretius 1.805), according to which verbs of the type fiicio-fiict11s arc opposed to rhose
ants" is another quesrion rh:it ought to he recx:imined from top to bottom. Cf. below, Ap- of the cype ago-ikt11s, :is is made clc:ir by their compounds (cffcctlls but cxaC111s): cf.
pendix C, n. 5 1. Leum:inn 1977: 114. Starting from Mommscn's dr:istic negative judgment, the question

tion to our knowledge of the linguistic and prosodic peculiarities of the 7

archaic poets: especially for Ennius and Ludlius his achievement is in a
certain sense analogous, even if of smaller dimensions, than Ritschl's for What Really Belongs to Lachmann
Plautus in the same period.37 ln this regard Lachmann's commentary on Lu
cretius is still alive and valid.

whether Lachm:rnn had an adequate knowledge and understanding of the thought of Epi-
curus and Lucretius has been much discussed: cf. Kenney 1974: 108 and n. 6. Probably he
did not, a nd certainly his commentary does not indicate that he did. But we must consider
that, with extremely rare exceptions regarding commentaries to individual books of the
De rerum natma, until now it has been almost impossible to find a commentator or edi-
tor of Lucreuus endowed in equal measure with :ibili1ies fo the fields of philology and lin-
guistics on the one hand :ind in philosophical interpretation on the other. Let us briefly recapitulate the results we have attained.
J7. He also undertook an edition of Lucilius; interru pted by his death, it was com Ordinarily, when reference is made to .. Lachmann's method," what is
plcted and published by Vahlen 1876.
meant is a complex of criteria for recensio. As we have seen, a number of
different Classical philologists contributed to its formulation.
1. The rejection of the vt1/gate and the requirement that the ma,mscripts
not merely be consulted from time to time b11t be used as the foundation of
the edition. We have seen that this point was already clear to Bentley, and
even more so to Ernesti and Wolf. Lachmann's merits consisted in his return
to insisting on it after Gottfried Hermann had caused it to be almost for-
gotten and also in his application of it to the criticism of the New Testament,
where nonscientific reasons had delayed its application, even if in that same
field it had been theorized lucidly in the eighteenth century.
2. The distmst for manttscripts of the Htm1anist period. The predecessor
for Lachmann's distrust (much reduced, as we have seen, in the preface to
Lucretius) h was Scaliger in particular (and earlier Politian and Vettori).
3. The reconstmctio11 of the history of the text a11d partic11larly ofthe ge
nealogical relations that link the extant manuscripts. This is usually consid-
ered to be an essential characteristic of" Lachmann's method,~ yet I believe
that I have demonstrated that Lachmann's original contribution in precisely
this regard was very limited and unccrtain.J The true founders of the ge
nealogical classification of manuscripts a re Schlytcr {in a field different from
Classical philology),< Zumpt, Madvig, and above all Ritschl and Bernays,
who concretely fulfilled the requirement to which Bengel had given expres
sion. In particular, for the derivation of all the manuscripts of a work from
a single archecype we must go back even further, to Erasmus and Scaliger;
and; for the technical term archetype in the specific sense of a lost medieval
o r late ancient' ancestor, to Madvig. In the procedure of eliminatio codirnm
descriptorum, the scholars who distinguished themselves were, after Poli-
tian, above all Boivin, Schweighauser,11 and Sauppe. The first history of a
text in antiquity was attempted by Wolf in his Prolegomena ad Homemm. 11
116 CHAPTER SEVEN Whal Really Belongs to Lachmann 117

4. The formulation of criteria pennitting a mechanical determination producing this result: Lachmann's oracular tone, which ended up creating
(without reco11rse to iudicium) of which reading goes back to the archetype. around him an atmosphere of veneration and "untouchability," as later
This is Lachmann's most genuine' contribution, even if his debt ro Bengel around Wilamowitz; the tendency-sincere and affectionate but too exclu-
and, nearer in time, to Madvig must be acknowledged. sive-to glorify the master on the part of certain students and friends, above
It might be objected that we have committed a sophism to Lachmann's nil Moriz Haupt.! We have also recalled rhe impression produced upon his
detriment by analyzing the method in this way into its various components contemporaries by a fact whose importance is all in all only secondary-the
and indicating the "precursors" for each one, and that his genius consisted calculation of the number of pages and lines of the archetype of Lucretius.
precisely in uniting these isolated procedures into a synthesis and making But there were also more serious and fundamental reasons for this develop-
them serve the purpose of a rigorous recensio. This is not a negligible ob- ment. Whereasm editorial textual criticism was only one of many activities
jection: many other discoveries and theories in various sciences have arisen for a Ritschl or a Madvig-Madvig was justly celebrated above allJor his
from the combination of principles and techniques that were already known conjectural emendations in the field of Latin prose, to say nothing of his in-
individually, and this "combination" has often been a creative act of high novative historical and linguistic writings; Ritschl studied all the aspects of
value. There is no doubt that tracing out genealogies of manuscripts, even if Plautine philology, and in particular prosody and metrics-this was the ac-
in a masterly way, but then not making use of them for the recensio, as Ritschl tivity to which Lachmann primarily dedicated himself starting from his ear-
and Bernays had done, remains sterile for the purposes of the critical edi- liest years and ro which therefore his name was linked. And above all, as we
rion, even if it is fruitful for the history of the text conceived as an episode have shown, Lachmann was the one who aimed in the clearest and most im-
of cultural history. mediate way at the practical goal of reconstructing the archetype without
And yet we must insist upon rwo points, not out of a narrow-minded wasting time on problems of the history of the tradition-a limit, as we have
taste for diminishing Lachmann's work but in order better to attribute to said, but also a strength. He was a great simplifier, with all the virtues and
each scholar his own role: vices this brings with it. The very one-sidedness with which he separated
rccensio from interpretation, though mistaken in itself, had a pedagogical
I. The first to make use of the stemma codicum in order to reconstruct the function : it made a powerful contribution toward recalling scholars' atten-
archetype was not Lachmann but Madvig; and if in the case of Bengel we tion to the requirement to give critical editions a solid documentary basis (a
must speak of a mere project that was not yet fulfilled, one cannot say the requ irement that other Classical philologists had expressed in a way that
same of Madvig (cf. above, pp. 97-98). was more complex and balanced, bur for that very reason less harshly ef-
2.. Perhaps more out of impatience than because of a true lack of aptitude,
Lachmann himself remained strangely unable to reconstruct those gene-
:.55-56 and Birt 1913 : 16 praise Sauppe 1841; W1lamowitt 1894: 4 :. cites Lachmann and
alogies of manuscripts which were the indispensable prerequisite for eli-
Madvig next 10 one: another. See also Vitelli 196::.: 11. I wou ld say that it was Ritschl
minatio lectionum singularittm: he needed to make use of genealogies
whose merits in this field scholars foiled to recogniie more: than anyone: else's; even Rib-
traced out by others, and his preface to Lucretius shows us how difficult bcck, in his fine biography of Ritschl (Ribbeck 1879), docs not give sufficient prominence
this was for him, so much so that he does not notice chat he contradicts to his importance for the methodology or textual criticism.
himself within a few pages.t 1. Sec esp. Haupt 1836 and 1911 (1854); passages from the latter were quoted by Bcl-
gcr 1879 before its full public:ition. Other somewhat too exclusive glorifiers of Lachmann
But at the same time as we reduce the dimensions of Lachmann's figure were Hern, Vahlen 1893, and Leo 1893 ... Leo 1960: 1.415-31 (but Leo admitted that
and acknowledge the claims of other Classical philologists who cooperated Haupt's words about Lachmann were never "characterizing but always purely admiring
and venerating ones'": Leo 1893: 4 - Leo 1960: :..416). How much effort it coo German
with him or preceded him in certain fundamental points, we must explain
Classical philolog)' to free itself from the "cult" of Lachmann is indicated by Wil:imowitz
the reasons for which his fame as a textual critic' has ended up obscuring
1981 (19:.1 1: 13:.: "The impact of his personality and speech o n his hearers was so over-
that of all the others. 1 No doubt various inferior reasons contributed toward whelming that they were sometimes paralysed and sometimes tempted into grotesque at
tempts co copy him . We have to free ourselves from his leading-strings, bur only by first
1. But this process happened gradually, and there was no l;ick of schol3rs who con surrendering willingly to him, as he himself prescribed." And Wilamowitz had not been
finued to recall the merits of some o f Lachmann's contemporaries, even if only inc1dcn a srudcnt of Lachmann's, and belonged to a generation that was already remote: in time
tally: e.g., lloeckh 18!!6: 105 recalls Sauppe together with Lachmann; Blass 1891 (1886): from him.

fective): 0 prim11m recensere [Do the recensia first}! Although Lachmann's 8

natural talent as a Classical philologist was less acme and profound than
that of some of his contemporaries (Gottfried Hermann, Ritschl, Boeckh, Textual Criticism and Linguistics, and
Karl Otfried Muller) and although he tended more toward a certain dog Their Crises at the End of the Nineteenth
matism than they did," he still deserves a place of considerable prominence
in the history of nineteenth-century Classical scholarship because of his and in the Twentieth Century
salutary insistence on the problem of recensio. And we will be able to con
tinue to speak of "Lachmann's method," even if we will have to use this ex-
pression as an abbreviation and, as it were, a symbol, rather than as a his-
torically accurate expression.

Already on one occasion rpp. 86-88) we have pointFd to a parallel between

the methods of textual criticism and those of historical-comparative lin-
guistics. Now we must linger a bit more on this subject, not for a mere os-
tentation of interdisciplioarity but because the comparison really can help
explain the difficulties that the application of Lachmann's method soon en-
countered and certain hostilities in principle to the method itself!
There is an undeniable affinity between the method with which the Clas-
sical philologist classifies manuscripts genealogically and reconstructs the
reading of the archetype, and the method with which the linguist classifies
languages and as far as possible reconstructs a lost mother language, for
example, Inda-European. In both cases inherited elements must be distin-
guished from innovations, and the unitary anterior phase from which these
have branched o ut must be hypothesized on the basis of the various inno
vations. The fact that innovations are shared by certain manuscripts of the
same text, or by certain languages of the same family, demonstr:ires that
these are connected by a particularly close kinship, that they belong to a
subgroup: a textual corruption too is an inttovation compared to the previ
ously transmitted text, just like a linguistic innovation. On the other hand,
shared "conservations" have no classificatory value: what was already found
in the original text or language can be preserved even in descendants that
are quite different from one another."
Naturally, like all analogies, this one too is valid only within certain lim
its: in linguistics there is nothing corresponding to the distinction between
archetype and origi11al; and even when a corruption has spread through the
whole manuscript tradition, it is still always felt to be something that dis
turbs the comext (and hence it is felt as a real error, or as a banalization that
constitutes a deterioration too,C not as a simple neutral "innovation" ), while
a linguistic innovation, once it has achieved success, ceases by that very fact
co be felt as an error. But the parallel between the research methods of the
two disciplines remains valid: indeed, it becomes even more evident if we
110 C HAl' T ER 1GllT
Textual Criticism :ind Linguistics l:Z.f

think of the linguistics of someone like Schleicher, who really thought he

or Madvig to comparativ e linguistics. It must also be noted that in the pe
could reconstruct the Inda-Europ ean language with the same certainty with
riod 1830-42- when, as we have seen, the fundament al principles of the
which Lachmann had reconstruc ted the archetype of Lucretius (to the point new ars critica were being establishe d-Inda-Eu ropean linguistics had not
that he presumed to rewrite the celebrated fable of the sheep and horses in
yet taken on the predomina ntly reconstruccive character that it possessed
Inda-Europ ean), and who regarded phonetic changes as a " decline" from an
for Schleicher starting in J 8 50. For Friedrich Schlegel, for Rask, for Bopp,
original state of perfection and therefore as being completely analogous to
the essential purpose was still to demonstrate the kinship among the Indo-
textual corruption s. 1 And just as the philologists, beginning with Schlyter, European languages and to go back from there to the problem of the origin
Zumpt, and Ritscht,J represented the genealogies of manuscript traditions
of the grammatic al forms; they certainly were not yet thinking of precise ge-
graphically by means of stemmata codiwm, so too somewhat later Schlei-
nealogical trees, of reconstruct ions of "asterisked " forms or even of texts in
cher introduced the use of genealogical trees into comparativ e linguistics. 2
Inda-Europ ean. So in the 1830s and 1840s the analogy between textual crir
If one wished, one could also compare the trust of Lachmann and his con- icism and linguistics was not yet as clear as it later became.
tempor:uie s in the oldest manuscript s, and their exaggerate d depreciatio n of
At first the inverse hypothesis , that the model of textual criticism had in-
the recentiores, with the analogous prejudice on the part of the linguists of
fluenced Schleicher, seemed improbabl e to me as well. I thought I saw a dif-
that period (or better, of a somewhat earlier one: Jones, F. Schlegel)< that
fic ulty in Schleicher's rigidly naturalistic mentality: he had already estaby
Sanskrit, the language of most ancient attestation, always preserved the
fished a clear distinction between philology, a "historical discipline," and
most ancient phase as well.
linguistics, a "natural science," toward the end of the I 8 5os;1 and had srud-
These analogies authorize us to ask whether there might have been a di- ied botany starting in his youth. But Henry H. Hoenigswa lds has had the
rect relation of influence of the one discipline upon the other. At first sight,
merit of recalling a fact that was nor quite unknown but had been too eas-
it might seem likelier that comparativ e linguistics, which arose between the
ily forgotten: the young Schleicher was seriously interested not only in bot-
end of the eighteenth century and the beginning of the nineteenth , supplied
any but also in Classical philology, studied with Ritschl at Bonn, and even
Lachmann , or better still Madvig and Ritschl, with a model for the method
started his career as a textual philologist. So Schleicher probably derived rhc
of textual criticism. 1 Such a hypothesis might seem to be supported by the
idea of a genealogical tree of the Inda-Europ ean languages" and of recon-
fact that the founders of that method, unlike other Classical philologists
structing their extinct mother language from Ritschl himself (one of the first
such as Gottfried Hermann, were interested in and sympatheti c with the
new linguistic science. 1 4. Schleicher 1874: 119-:.0: Ml'hilology is a historical discipline I. 1. Linguistics on
However, this hypothesis does not withstand closer examinarion.i: In the other hand is not :i hismric:il discipline but rather a natural historical one I .. ). The
fact, as we have seen, editorial textual criticism ca me about by developing o bject of glomc science is a n:itural organism." Cf. Schleicher 1K63: 711 1. Maher 1966 has
i11xta propria pri11cipia fclose to its own beginnings), and Lachmann and his well demonstrated th:it Schlcichcr's namr:ilism is pre Darwm1an, and that even the work
contempor aries did nothing but systematize and apply coherently methods of his that I have 1usr cited does nut represent a real change m his thought, notwithstand
that had already been formulated by the textual critics of the eighteenth cen- ing its mlc :ind its vague profession of D:irwinism. We should not be misled by the :idjcc
tive "natural historical" in the pa ssage just quoted : Schleicher 1s referring to l11sloria 11a
tury (or, in some cases, already by the Humanists and Scaligcr). Moreover,
l/lralis in its old sense, for which i0Tnp(11 mc:ins "description" without any diachronic
there is no reference in the writings on textual criticism of Lachmann , Ritschl,
implications. In any case, as is well known, Schleicher s:iw pure \Verd(!ll lbecoming) and
even decadence in the diachronic evolution of l:inguages (cf. above, p. 1:.oni ), while he
t. This theory of the Vcrfall 1decline I of languages in the historical period is found
in reserved the d ignity of Gcschicbte lhistoryl for conscious human history alone, m con
all of Schlcichcr's writings; sec esp. 1874: 35, 47-50, 64.
formity with his early formation in Hcgclianism, which lasted into his later materialism,
i. Schleicher 1874: :z.8, 8:z.. Sec below, pp. 1:.1-:.:z..
as into Jaco b Molcschotr's.
3. Lachmann wa~ a close friend and co!lahorator of Jacob Grimm's {l achmann's com-
5. Hocnigswald 1963: esp. 8.
mentary on Lucretius 3.198 also quotes Grimm's De11lschc Gram111a1ik: l:ichmann 18 5ob:
6. Morpurgo Davies 1975: 636n53 hns observed that there is already ;i genealogical
149-50). For Ritschl, sec the following p:iss:ige from a text of 1833: "scientific exposi-
table of languages in Klaproth 18:1.3. But as for as I know, Klaproth's work no longer en
tions of Greek literature and Latin grammar, of which the latter is quite impossible if one
joyed great prestige or diffusion at the time of Schleicher. Hence the derivation of Schleicher
does not :issimilatc research into general comparative linguistics, which scholars still view
from Riischl remains probable, even if M:ihcr 1966 and Morpurgo Davies 197 5 :ire right
all too timidly" (Grimm 1864-71: 5.171. Madvig was interested in comparative linguis-
to note thar :in image as obvious as that of a genealogical tree could have :inscn indc
tics, :ind even more in gener:il lingu1stics.
pcndently m various di:ichronic d i1ciplincs.
1 2.l. CHllrTER EIGHT Textual Criticism and Linguistics

to study the genealogy of a manuscript tradition in depth and to trace out permissible as to suggest conjectures to the text of Plautus without taking ac
stemmata codicum, as we saw in chap. 5)-even if reconstructing the arche- count of the Ambrosian palimpsest and the vetus codex, 10 or to the text of
type, the counterpart of this second linguistic operation, was practiced by Sophocles while neglecting the Laurentianus A. (Curtius 1858-61: 1.u)
Madvig and Lachmann, not by Ritschl, as we have seen and as Hoenigswald
does not sufficiently emphasize." As these last words indicate, Curtius intended that his comparison between
But beyond this direct connection (which, I repeat, is probable but still the two disciplines would help convert Classical philologists to linguisi:ics-
only hypothetical), one might think that the comparativist atmosphere wide- many of them were still persisting in the footsteps of Lobeck in constructing
spread in all of European culture at that time co uld have favored the rise etymo logies of Greek words based only on Greek. u
both of comparative linguistics and of that form of verglefr.hende Textkritik Since then, the evolution of linguistics and that of textual criticism have
(comparative textual criticism} that is "Lachmann's method." Even before continued to follow parallel lines. If the passage from Curtius tha~ I have
linguistics, comparative anatomy had stimulated a taste for the comparative just quoted still displays an unshaken faith in the genealogical method, in
method. But the relations between comparative anatomy and linguistics too the last decades of the nineteenth century such a faith began to falter among
cannot be reduced to a pure and simple influence of the one upon the other, both textual critics and linguists. In textual criticism cases of the perfectly
as I hope to demonstrate in studies to be published shortly.7 successful application of the genealogical method were not lacking (and
What is certain is that the similarity of their research methods ended up they have not been lacking even in times nearer our own): it should suffice
seeming clearer and dearer to linguists on the one hand and textual critics to recall Leo's preface to Venantius Fortunatus, of which Eduard Fraenkel
on the other) Georg Curtius, a linguist whok always claimed to be a philol- has rightly emphasized the paradigmatic value. 12 ln Romance philology too,
ogist too and who insisted more than once on the necessity of a rapproche the method has had distinguished applications; without tarrying in a field
ment between philology and linguistics, 1 developed the comparison fully in with which I am insufficiently familiar, I shall only recall two eminent
the introduction to his Gmndzilge der griechischen Etymologie [Funda names, Gaston Paris and Pio Rajna.1
menrals of Greek etymology]: AU the same, little by little scholars came to real ize that the method
achieved full success only in a relatively limited number of cases. All the
The individual languages of the Inda-European trunk resemble just as many manuscript traditions that were "too simple" (those represented by only one
old copies of a lost original manuscript. None reproduces the original text or two witnesses) remained outside its range, as did all those that were "too
exactly, but all are important for us inasmuch as they are old witnesses to complicated " {those in which the copyists not only transcribed but also col
a state of affairs of which we have no direct knowledge I .. J. If we indicate lated or conjectured so much that the kinship relations among the manu
with A the stage earlier than the differentiation of the lndo European lan scripts were obscured). As we have seen, even the founders of the method
guage5, the Greek language (C) and the Latin one (D) cannot be derived from had some trouble before they found in Lucretius an author to whom the
it directly, but both go back to an apograph (B) which is lost for us, Greco~ method was fully adapted; and even the tradition of Lucretius is absolutely
Italian,9 which itself descended directly from A. In the same way there is a clear only if we set aside the problem of the ltalici and leave out of con
particularly close affinity between Sanskrit, which deserves the first place sideration certain cases of contamination among the Oblongus, Quadratus,
among all the copies of A for its legibility and correctness, and Persian, as and Schcdae as well! m
wdl as between the readings of the Germanic languages on the one hand and In the face of these difficulties, some scholars followed a tendency that,
those of the BaltoSlavic languages on the other[ ... ]. To try to deal with et~
ymological questions by limiting oneself to a single language is just as im
ro. The so-called vet11s codex Camerarii is Palat. Vat. 16r5, usually indicarcd as D.
u . Lobeck 1853: vii.; "I have decided IO discuss rhe language not of Ogygia but of
7. Cf. fo r now Timp;inaro 197;1. and 1973 on the Schlegd brothers and Franz Bopp, Greece, which is more than enough work in irself." Cf. Curtius I 858-6:z.: 1.10- 1 J. Cur
and Timpanaro 1979: esp. 474-86, on Giacomo Lignana. But ~ee already Maher 1966. tius returns to the analogy between textual criticism and linguistics on pp. :r.6 and 98. So
8. E.g., Currius 186:1.. too Lcfmann 189 r: 5:z.; and still Kretschmer 19:z.7: 5, even if his ideas on linguistic kinship
9. Curtius still believed in a particularly close kinship between Greek and Ita lic, as did and affinity were by now very different from 1hose of the genealogical comparativism of
most of his contemporaries, for that matter: cf. Meillet-Vendryes 1963: Introduction; De the nineteenth cenrury (cf. below, p. u6 and n. :z.o).
VOID 1958: l.ll.9. u . Fraenkel in Leo 1960: t.xxii-xxiii.
CHArTl!k EIGH T Textual Critkism and Linguistics

as we have seen, had its first representative in lachmann himself: they pre- creasing distance benveen the history of tradition and textual criticism. The
ferred to cut the knot rather than untie it. They tried to eliminate as many history of tradition became more and more the history of ancient and me-
manuscripts as they could, as suspected in general of being interpolated or dieval culture; in Wilamowitz, in Traube, in Eduard Schwartz, for example,
descripti. l l Once the manuscript tradition had been reduced to one or two it acquired a richness and complexity unknown to the scholars of the pre
manuscripts, every genealogical difficulty conveniently vanished; and so, for ceding generation, but at the same time it became less and less capable of fur-
example, Wilhelm Dindorf could prepare editions of very many (too many)" nishing a secure criterion for constituting the text, since in highly disturbed
Greek authors with little effort, and Eyssenhardt could publish a Macrobius traditions no reading could be rejected a priori and in a particular passage
based arbitrarily on only two manuscripts; even a critic as prudent as Vahlen even the most suspect manuscript might preserve the correct reading.
mistakenly eliminated one of the best manuscripts (the Heinsianus: Leiden On the problem of the rece11tiores too, scholars retreated little by little
B.P.l. u8) in his edition of Cicero's De legibus and refused to change his from Lachmann's total condemnation (which is already weakened anyway
mind explicitly even when faced by Jordan's and C. F. W. Muller's stringent in his preface to Lucretius, as we have seen) to the more balanced attitude
objections; 14 even Leo, whom we have already mentioned as an intelligent of Semler and Griesbach in the eighteenth century (cf. p. 69f.) and Madvig
Lachmannian for his edition of Venantius Fortunatus and whose edition of as late as the nineteenth century (p. 97). In fact Otto Jahn, 16 the most intel-
Plautus and Plautinische Forsch1mgen would much later contribute to over ligent and least servile of Lachmann's students, had already shown his im-
coming Lachmannism, published an edition of the tragedies of Seneca in patience with his teacher's too rigid criteria in the period of full lachmann-
187 8 that remains important for the metrical studies contained in its first ism. Jahn's editio maior of Persius, published in Leipzig in 1843, is dedicated
volume but is fundamentally mistaken in its prejudice that the only inde to Lachmann, "the incomparable teacher, the incorruptible friend": but in
pendent witness is the codex Etrusc11s (Florence, Laur. 37. 13). 15 0
its opening dedicatory epistle Jahn already expresses his fear that his teacher
Other critics realized that this was not the right way: instead, it had to be will not approve its apparatus criticus, so full of variant readings derived
recognized that many traditions were extremely complex, since contamina- from Humanist manuscripts. 17 And toward the end of its extensive prole-
tion and the innovations introduced by copyists and ancient and medieval gomena (which constitute a real history of the text of Persius, more in Wolf's
"editors" r had played a large role in them, starting from the most ancient and Ritschl's style than in Lachmann's),Jahn makes a declaration that sounds
stages we can reach; therefore the mechanical method of choice among vari obviously critical if not of Lachmann himself then certainly of too scholas~
ants adopted by lachmann was not applicable to these (or if it was, then tic a way of understanding his method: "In general, the nature of the man
only with many precautions and reservations).q This meant a positive re- uscripts of Persius is such that one cannot constitute different classes and
assessment of the internal criteria (/ectio di(ficilior, usus scribe11di) which families to which the individual manuscripts could be assigned. Everyone
lachmann had despised, a return to principles already maintained by Clas- understands that the manuscripts differ in age and authority (which is usu-
sical philologists before Lachmann or those contemporary with him (recall ally linked with age in this case) and that in different cases different ones
how Sauppe had insisted on contamination); and, at the same time, an in come close to other ones in the quality of their readings; but since no man
uscript is so uncorrupted that you could use it as a basis for the recension, it
13. Cf. Pasquali 1951a (1934 ): 15-40, and above, p. 100., for Cobct's unhealthy in- cannot even be observed that in particular passages certain manuscripts agree
fluence in this direction. See also below, p. 15 5 and n. u. in genuine or corrupt readings .... Therefore I could not select certain man-
1-4 See the preface to Vahlcn 1883, and now 1hc preface to Ziegler 1950: 16. uscripts according to whose rule I could establish the author's words, but I
1S As is wcU known, the positive revalu;uion of the manuscripts of the soclllled fam- always had to consider all of them" (Jahn t843 : cxciii-cxciv)- even if he
ily A (cetcris parib11s less reliable than the Etruscus, but superior in very many passages)
is due above all to Carlsson 192.6 and his later contributions; cf. Pasqua Ii t 9 sl.3 ( t 9 34 I:
116-i9 and now many rl'Cenr studies (by Philp, Giardina, Tarrant, Zw1erlem, and vari- 16. I have learned much about this brilliant and anticonformist scholar, whose activ-
ous others). Fraenkcl in U:o 1960! t.xviii-xxi does indeed mention the defects of U:o's ities extended to many fields (ph1lology, archaeology, history of ancient and modern mu-
edition, but out of love for his teacher tends rather to minimize them. Besides the question sic), from conversations with Eduard Fraenkcl. On Jahns attachment to Lachmann's mem
of the manuscripts, I do not think one can say that Leo contributed ua l:nge number of ex- ory, see the testimony of Gompcrz 1905: i9.
cellent emendations": Leo was a great interpreter and historian of Latin literature and cul- 17. Jahn 18.13: cxciv: "Although I do not know whether I will succeed in convincing
ture, a distinguished metrician, but an infclicirous conjecturer-of his conjectures on you chat I have done well to indicate the readings of so many recent manuscripts too." Jn
Seneca, in partir.:ul:ir, almost none arc still remembered. his cditio mirror (Jahn 18 5 r ), Jahn provided a much more concise apparatus.
l:Z.6 CHAPTER EIGH T Textual Criticism and Linguistics

immediately goes on to indicate that in those "all of them" he does not in the NeoLatin ones, can be compared with highly contaminated manu-
dude all the most recent ones. In rhe specific case of Persius, Jahn was ex scripts (for which it is therefore impossible to trace out a stemma), and that
aggeraring somewhat: a genealogy of the manuscripts can be traced out, these authors conceive the original Inda-European language as an "arche+
even if it is nor cmirely rigorous and goes back nor ro a medieval archetype type with variants."
but to at least two ancient editions, and ir was Jahn himself who traced out In this period, lingu istics certainly directly influenced textual criticism
its first outline; nor should the importance of the recentiores be exagger with regard to explaining the origin of corruptions. Schuchardt's great
ated. l* Bur all the same rhe passage we have quoted has a considerable meth work, Der Vokalismus des Vrtlgiirlateins [The vocalism of Vulgar Larin]
odological value: it is a first rebellion (even an excessive one, let us repeat} showed Classical philologists that corruptions d ue to phonetic vulgarisms,
against orthodox Lachmannism by a disciple and admirer of Lachmann.1 As or to psychological phenomena such as occur in the evolution of languages
we shall see, these ideas on comamination and on the impossibility of fol (assimilation and dissimilation, metatheses, etc.), are just as numerous as
lowing mechanical criteria were taken up again and developed at the end of purely graphic corruptions, or even more so. 21 This was not an absolute
the nineteenth and in the twentieth century. novelty,22 bur in the period of Lachmannism almost all medieval corruptions
The crisis of comparative linguistics occurred at the same time as the cri- had tended to be explained as graphic errors.23
sis of .. Lachmann's method ... The concept of an absolutely unitary mother Then Louis Havet, an eminent French scholar, demonstrated with an ex
language, from which two daughter languages branched out and then went traordinary r ichness of examples the complex character of many corrup-
on in turn to produce by successive differentiations the various historically tio ns, which originate in graphic errors or, more often, in psychological
attested Inda-European languages, began to seem unsatisfactory. Already in ones, upon which clumsy attempts at correction are then supertmposed; he
1872 Johannes Schmidt, a student of Schleicher's, had opposed the " wave too brought to textual criticism the experience not of a pure Classical philol-
theory" to the theory of the genealogical tree. 19 His ideas were developed ogist but of a linguistic philologist and expert on metrics. 24 And some of his
further by his student Paul Kretschmer, who emphasized more and more the
importance of "horizontal transmission" of linguistic facts as compared 1 1. Schuchardt wrote, with a bit of exaggeration, " The number of genuine graphic
with "vertical transmission," the only one that the theory of the genealogi mistakes is very limited; most of what are called laps11s calami [slips of the pen] are laps11s
cal tree had considered. 20 Hugo Schuchardt had arrived at analogo us results li11g11ae !slips of the tongue)" (Schuchardt 1866 - 68: i.17).
starting out from the study of the Romance languages. More and more- :z.:z.. Le Clerc had attributed considerable importance to psychological corruptions and
and with undeniable exaggerarions u- linguistic kinship started to become chose due to the vulgar pronunciation: see above, p. 6:z.f. Cf. also Hermann 18:z.7-77:
something that was not inherited but acquired by means of contacts. And 6.:z.3: " Substitutions of words which occur ro the copyist at che wrong moment because
more and more the "intermediate unities" between Inda-European and the they are in constant use cannot have a diplomatic explanation because their motivation is
psychological. "
historically attested languages were dissolved: after ItaloGreek was dis
:z.3 . This tend ency is noticeable, for example, in Rihbcck 1866; it reached grotesque
solved, it was ltalo-Celtic's turn (later, with Devoto, came halo-Latin's too}.
extremes in Hagen 1879. On the other hand, Bruhn 1887 demonstrated the frequency in
The original IndoEuropean language itself was conceived more and more che text of the Greek tragedians and Homer of psychological corruptions, including more
as having already been rich in dialectal differentiations that could not be lo- complex ones than those investigated by Schuchardt: hanalizations, su bstitutions of words
cated geographically. with a basically similar sound and length, ere. Bruhn, repeating observations of Le Clerc's
Here too the analogy with textual criticism is clear. In both disciplines Cd . above, p. 6 :z.f.) in all prob a biliry unconsciously, called :mention to the fa ct that che
the claim for the importance of "horizontal transmission" was made in the copyist reads relatively long passages of the o riginal and hence when he copies them d own
same period. Whoever wanted ro have some fun writing a showpiece like is exposed ro errors of memory and of " self-d ictation," especially at the end of the sen-
Curtius's on the analogies of the Schleicherian-Lachmannian period could tence (or at the end of che verse in poetic texts).
:z.4. On Havet's studies as a young man, cf. Chatelain 19:z.5: 1:1.; and Frere 19~6 : i.1 ,
note that for Kretschmer the Inda-European languages, or for Schuchardt
who observes correctly, "He enriched the doma in of ph ilolo gy by making his concerns as
a linguist penetrace into it." Among Ha vet's methodological works preceding his great
18. Cf. Sdvoletto 1961: v-vi, xiv-xv. Clausen 1956 is more inclined ro rrust the Ma1111el de critique verbale [Manual of verbal criticism) (Haver 1911 ), one of the most in
recellliorcs. teresting is Haver 1884; see on p. 1104 his d istinction between " servile mistakes" and "crir
19. J. Schmidt 187:z.: esp. 15, :z.7-:z.8. ical mistakes" and his assertion thar rhis second kind of error is more frequent. Cf. also
i.o. Kretschmer 1896: see esp. chap. 4, 93-t:z.4. below, pp. r:z.9- 30.
118 CHArTER E ICllT Tcxrual riricism :ind Linguistics 129

best students were linguistic philologists, like Jules Marouzcau and Alfred tory in the fullest sense, which constitutes his glory, led him in this case to
Ernout. underestimate the strictly graphic aspect of textual transmission, which
All the same, in the second half of the nineteenth century, just as in the rarely is the sole cause of errors, but very frequently acts as one cause among
preceding period, the analogies between linguistics and textual criticism are others.w Among Traube's conjectures to the Anthologia Sa/masia'1a there
the result nor only of undeniable direct influences but also of a shared cul- are some splendid ones, but others are faulty precisely by reason of their pa-
tural atmosphere. Just as around the 18 50s and 186os linguists and philol- leographical improbability; 2 ~ and the Anthologia Salmasiana is one of those
ogists (and philosophers, and scientists) had breathed a common compara- texts in which mechanical corruptions (graphic ones, or else ones due to the
rivisr and evolutionary air, so too at the end of the nineteenth century people vulgar pronunciation, but not ro the copyist's whims) are prevalent. We
began to breathe an air of reaction against positivism. As is well known, this must also remind ourselves that "mechanical corruption" and corruption
was a reaction that combined some justified elements (an impatience with due to confusion of graphic signs are not identical: many psychological cor
hasty schemes and generalizations, the need for greater faithfu lness ro the ruptions are just as unconscious and involuntary (and hence "mechanical ")
complexity and variety of historical facts) with other more dubious ones (a as graphic corruptions are, or sometimes even more so. This is an ambigu-
return to a spiritualist metaphysics far older than the old positivism, a so- ity inm whose trap many have fallen, sometimes even Pasquali (e.g., i952a
phistic rejection of empirical classifications in the name of rhc uniqueness of [1934): xviii, II4-15, and 481-86 (Appendix 2, "Conjectures and Diplo
the individual phenomenon, irrationalist and antihistorical tendencies now matic Probability"]).Y
named "historicism"). So too, there can be no doubt about the extraordinary methodological
It is nor up to me to recount how these latter aspects became stronger and value of Eduard Schwanzs "Prolegomena" to his edition of Eusebius's His-
stronger in linguistics, especially in Italy, during the course of the nineteenth toria ecclesiastica [Church history); yet all the same Schwartz went too far
century, and how scholars fell headlong from a conception of language that in his distrust for genealogical classifications based on shared corruptions
was still as historical and as rich in problems as Schuchardt's into "aesthet- and in his view that horizontal transmission (not only of correct readings or
ics as general linguistics." 21 But textual criticism too, a much more special- of conscious innovations but also of rea l errors) was just as frequent as ver-
ized and less ideologized discipline, witnessed not only justified criticisms of tical transmission, or even more so.19 It is true that prose texts have in gen-
the excessive schematism of Lachmannism but also exaggerations in the op- eral a much less mechanical transmission than poetic ones; it is also rrue,
posite direction, caused by a desire to deny to every manuscript tradition and Madvig had already noted it very exactly, 111 that mechanical corruptions
any "mechanical" character whatsoever and to demonstrate that the history are less frequent in Greek texts than in Latin ones, because the Byzantine
of textual transmission is essentially a "spiritual,. histo ry. For example, Middle Ages had no Dark Ages comparable with those in the medieval West.
when the great Ludwig Traube wrote, "A conjecture does not become bet- But even so, a text like Eusebius's Historia ecclesiastica is quite an exceptional
ter because it can be explained paleographically, and certainly it does not case, even among Greek prose texts, since it was linked with theological dis-
become correct because in the best of cases it is paleographically pos- putes and changes in political affiliation and hence was exposed all the more
sible," 26 without any doubt he was right regarding the seco nd proposition, to conscious reworkings;' so it is risky to assign it a paradigmatic value.
but not at all regarding the first one; or, at least, that expression contained The French school was better at avoiding dangers of this sort. As we have
a certain ambiguity.v.~? Traube's lofty concept of paleography as cultural his- already suggesred, Havet's Ma1111el de critique verbale (Ha vet r 91 r ) does in
deed represent a reaction against the Lachmannians' oversimplification re
::.5. Many o hscrvarions can he found in Nencioni 19_.6. On Croccan and Vosslcrian garding the genesis of corruptions and the affiliation of manuscripts; 11 and
linguisrics and a ccrrain kind of Srrucruralism as rwo differcnr fo rms of anrimarcrialisric
reaction (the one subjectivist, inru itionist, and acstheticizing, the other marhcmarizing and 18. Traube 1909- 10: 3.5 1-59 Among his bcsrconjecturcs may be cited those to A11th.
~Platonic"), sec my observations in Timpanaro 1975: 137- 39. But it cannor be denied Lat. 198.38 and 304.4; amo ng the lcasr successfu l ones, the o ne to 83.88 :ind his complcrc
rhat nonetheless Structuralism has bc:en much more fruitful and richer in scientific dis rewriting of poem 377.
covcrics than inruirionisr linguistics has been. 19. Schwartz 1909: cxlvi- cxlv1i. Even so, it cannot be denied rhat real corruptions can
:i.6. Tr:iubc 1909-20: 3.113. he transmitted horizonrally, even if this happens only rarely: cf. below, Appendix C, n. H
:1.7 . Naturally, the poss1hility of jusrifymg a conjecture palcographically docs nor mak<" JO. Madvig 11171: 1.13 and n. t.
ir better from the poinr of view of meaning and style. Bur ceter1s {1aribus, palcographic:il 31. ~My whole book, in any c:isc, is a conrinuous protest :igainst mendacious ovcr-
prob:ibihty is a strong argumcnr m favor of a conjecture. simplific.:ition," declares Haver m hss prefocc, 1911 : xii. And sec ar r911 : 41 8 - 14 his ex-
t30 CHArTER EIGHT Textual Criticism and Linguistics 131

yet Ha vet has no wish at all to dissolve textual criticism into a multitude of Schwam, "the greatest textual critic of the century, the first one to over-
individual problems that cannot be compared with one another. On the come Lachmann in his method" (Pasquali 1952a (1934]: 471).
contrary, he even aspires to turn it into a rigorous science, a "pathology and Other approaches that were akin to his own despite the difference in
therapy of errors": the study of the genealogy of manuscripts is replaced by formation and cultural background remained little known to him, especially
the study of the genesis of corruptions .'~ As Ernout writes, "Louis Haver because of a prejudiced aversion, which went back to his youth and which
was an enemy of empirical conjecture, of what he called amateurs' criticism. lasted for a long time, against Classical studies detached from a general
Between two corrections of different value for the same error, I am sure that conception of Altertumsrvissenschaft [the science of antiquity], such as was
he would have chosen the inferior one-or at least the one that most schol- practiced in France and England from the beginning of the nineteenth cen-
ars would consider such-if this allowed him to determine more exactly the tury on.34 In Pasquali's last years he did indeed manage to approach the
process by which the restored word, had it been the original reading, could French school, some characteristics of which we traced out a little earlier
have given rise to the faulty reading." Jl Here, certainly, he went coo far: al- in summary fashion, with interest and almost with an astonished sympathy:
though explaining easily the genesis of a corruption has great importance in his lengthy review of Alphonse Dain's Les ma11uscrits [Manuscripts] 15
(more, I think, than Maas and Pasquali attribute to it), no genetic consider- he observed many points of agreement (especially regarding the genesis of
ation can ever induce one to prefer the "inferior" correction.~h And in fact corruptions and the history of tradition conceived as cultural history rather
most of the too many conjectures Havet published in the Revue de philolo- than as an abstract stemmatics) alongside some differences in intellectual
gie, many of which he collected in his Manuel de critique verbale, are un- character (Dain's interest in text constitution, in rece11sio and emendatio,
acceptable precisely because they try only to explain how the corruption was weaker than his predominant inrerest in codicology and the vicissitudes
was produced, not above all to find the right word for that particular con of the manuscript tradition: the title of Dain's book had already indicated
text. Havet had many abilities, but he was not a great interpreter, and this this, and it was confirmed later by his mediocre edition of Sophocles}.
harmed his activity as a textual critic too. But Marouzeau and especially Pasquali's death nipped in the bud what might have become a very fertile ex-
Dain, his students, knew how to profit from his extraordinary experience change of ideas and experiences between Dain and himself.
and, together, partly to overcome his limits. Pasquali's relations with English textual philology followed a rather dif-
Among these new approaches to textual criticism, Giorgio Pasquali's ferent course. There really had been a period of depression in English Clas-
book occupies a special position. As is well known, Paul Maas provided the sical studies, more or less from x825 (Dobree's death and, a little earlier,
occasion for inspiring this work with his Textual Criticism; but the contrast Elmsley's) until the last years of the nineteenth century (when the star of
between the mathematical mentality of Maas (who was interested above all Housman rose on the horizon),36 a depression partly, but only partly, com-
in the rigorousness of his formulations, without their always being rigorous pensated for by the many good commentaries that appeared in England dur-
in fact "'), and the lively sense of the uniqueness of each manuscript tradition ing that long interval. It cannot be said that Pasquali did not recognize
that animates Pasquali's exposition, leaps to the eyes of every reader.Jl The Housman's brilliance, but his image of Housman as a "Humanist," not a
true inspirers of Pasquali's book were Wilamowitz, Traube, and above all
34. On French philology see the conclusion of Pasquali 1964 (t9~0): 89-90, with its
bizarre final juxtaposition of Kanr and Treitschke, a genius on the one hand and a medi-
cellent discussion of the "pitfalls of genealogical classification" and his observation that ocre and n:arrow minded Realpolitiker[pragmatic politician} on the other, as the two great
groups of manuscripts can be formed by "convergence," by processes of contamination, est representauves of modern German culture. An important reason for Pasqu:ali's low
even from originally distinct branches of a cradition. Here too we must note the analogy opinion of French Classical philology were the alltoonumerous bad, and sometimes ter-
with the linguistic concept of "affinity" between languages, as o pposed to genealogical rible, editions published in the first years of the collection Bt:lles Lettres: cf. Pasquali
"kinship." 1952.b (1933): 2.42.- 43, 2.49; :and esp. Pasquali 1935: 2.06-7 P:asquali t968: 1.191,
3 2.. Ernout t 9 2.6: 2.4. 196, 379. On English Classical philology, :aside from Housman and Lindsay, to whom we
3.3 Cf. S. Marioni 1952.: 2.13-14 !this article is also important for its new observa shall refer shortly, cf. P:asquali 1964 (t92.o): 81.
tions and examples); Canfora 1968. I intend to write something myself abouc some other 35. Pasquali 195 tb (in Italian, translated by Timpnnaro, in Pasquali 1952.:a [1934):
cases of lack of rigor in Maas for all his ostentation of rigor. Sec in the mcanume below, 469-!10).
pp. 155f., t62.- 70. See now Canfora t982. and Timpanaro 1985a.JU 36. See Brink t978: 1 t96-1z.13.
131 CHAPTER EIGHT Textu:il Criticism and Lingumics

"scientist," remained too reductive, a bit because of insufficient know! in essence unmethodical, indeed antimethodical. Certainly, Housman had
edge,J? a bit because Housman's own aggressive character, his contempt for words of contempt for Textgeschichte; but his emeudatio was always guided
all routine and mediocrity, inevitably provoked in those who did not know by rigorous methodical criteria, and the material on various types of cor-
him very well either irritation or a fanatical enthusiasm, but only rarely a ruptions and their genesis which he collected in the prefaces and notes to his
reasoned admiration.JRCertainly, one notes in Pasquali a growing admira- editions and in many articles confirms what he himself always repeated, that
tion for Housman, from the reference in Filologia e storia {"a scholar of the "intuitive" element which one cannot do without in conjectural activity
acute but uncontrolled natural talent"; 1964 ( 192.0]: 81) to the exclamation must receive the confirmation of experience and reasoning. His syntactical,
witnessed by Otto Skutsch, "There is only one man who knows how to stylistic, prosodical, and metrical observations, which he always considered
make emendations, and that is Housman!" 19 to his judgment on Housman's to be a necessary support for his conjectures (or for his defenses of trans-
edition of Lucan ("for knowledge of the poet's highly individual language mitted readings: these too exist, and for the most part they are excellent), go
and style, for judgment, for sureness in emendation, it is a masterpiece, not- in the very same directi!->n.~2
withstanding the author's well-known eccentricity"; 1952.a (1934) : 4 32.nt); Wallace M. Lindsay, an English contemporary of Housman's, is cited in
and in its spontaneity the assessment Skutsch reporrs is even quite exagger- Pasquali's Storia de/la tradizione more often than Housman is-very often
ated. But although Pasquali always defended the legitimacy and value of with agreement but with the conviction that Lindsay's "histories of the text"
emendatio against short-sighted Italian critics, he was more interested in re and his editions remain below the best results obtained by German Classi-
censio; and his lack of familiarity with Housman's edition of Manilius (in cal philology.4 1 And it is certain that Lindsay was much better as an expert
which the genealogy of the manuscripts is traced out with a sure hand), an on Latin linguistics (his old Latin Language, in certain points, is still supe-
evidently infelicitous hypothesis in Housman's edition of Ju venal, 40 Hous- rior to Leumann's La11t- imd Formculehre (Phonetics and morphology)), on
man's fundamentally correct (but still somewhat too summary) liquidatio n grammatical works and Latin glossaries, above all on paleography, than as
of the genealogy of the manuscripts of Lucan as being completely contami- an editor of texts (although his edition of Martial remains exemplary, 44 and
nated 0-all this convinced Pasquali that Housman, brilliant as he was, was his editions of Festus, Nonius, and Isidore of Seville, for all their defects, are
unl ikely to be replaced in the foreseeable future). But the greater accusation
37. It seems cenain that Pasquali never had direct acqu:iint:ince with Housman's edi- of neglect and forgetfulness of Lindsay must be lodged not against Pasquali
tion of Manilius: Pasquali 1964 (1 910): 81; 1951a ( 1934): 391n3. This was observed si- and not even aga inst the Germans, but against the English themselves.4 s It
multaneously by Kenney l 97 4: r:z.9nr and Mo migliano 197 4: 370. is not a question of opposing Lindsay to Housman: there is no doubt that
3 8. It must be acknowledged that, even if Housman has exerted an extraordinary pos-
itive influence on the rigo r :ind sense of style of che English textual critics of rhe following
gener:iuons (iris in large p:in his merit, besidC5 rhar of Gcrm:in refugees in England, if all fo re approves the in fact truly exemplary review of Fr:icnkd 19:.6 = Fraenkel 1964:
in all the Classical philology of this n:nion today enjoys an md1spurable superiority over :z..167-308. We cannot lmger here on lhe later relations between Fraenkel :ind Housm:in,
all others), he has nonetheless :ilso stimulated an arrog:ince and an exaggera1ed snobbism which arc foscm:iting not only from a human perspective but also because of the difficult
in some English Classical p hilologists to which their value :is scholars does not alw:iys cor- symbiosis between English philology :ind the German philology tr:insplanted into England
respo nd (:ind even when the v:i lue 1s rhc:re, it would be berrer if rhe arrogance were not!). as a resuh of the N:izi persecutions; there is a brief testimony in Housman 1972: 3.1177.
Cf. my observations in Tim pan:iro 1964: 790-91; Mo migliano 1974: 36!1 on the i111ita- 41. Kenney 1974: 117- :1.9 provides the best acco unt of Housman's profoundly me-
t10 Ho11smani; :ind now Salemmc 1981: 190. thodical char:ictcr, despite his :ittacks against wmcthod" :is a form of routine; there arc
39 Skutsch 1960: 6 -7. I myself can ndd another oral testimony: the admiration with Dlhcr excellent observ:itio ns in Brink 1978: no6- 13.
which durmg a seminar Pasqu:ili spoke of Housman's cclebrared punctuarion and inter- 4 3. It is especially in the case of Plautus that P:isquali is consmmly concerned to place
pretation of Catullus 64.32.4, as simple :is it is brilliant. Lind53y below Leo and other German Classtc:il plulologists: e.g., Pasqu:ili 1951:1 (1934):
40. Housman 1931 : xxxix. Pasqu:ili's o bjection (19 513I 19341: 43on:z.) is entirely cor- 33 rn1, 337n1, 338n"4, etc.
rect, as far as I understand it; but it is expressed with :i polemic animosity all the more ex- 44. I :im referring both to his edition (l inds:iy 1903) and to his A11cic11t ditio11s of
cessive as he agreed with Ho usman abour the authenridry of that passage of Juvenal and Martial (Linds:iy 190:.); P:isquali 195i.a ( 1934): 416-i.6 too gives a clearly positive judg
disagreed o nly about rhe expl:inarion for its absence in almost all the manuscripts. ment of both works, and he even tended to agree too much with Lindsay on the problem
41 . Housman 19:1.7 (19161: vii. Pasqu:ili :icknowledges thar Housman "judges the tra of authorial variants, even though he was alrc.idy then more cautious than Lindsay him-
dition in its totality more correctly than :iny of his predecessors"; but he insis1s that even self. See now Citroni 1975: xli- xliv.
a contaminated tr:idnion should be disentangled as for or as little :is is possible and there- 45 . Unless I am mist:ikcn, neither Kenney nor Brmk names linds:iy even once.
134 CllArTER EIGll T Texrnal Criticism and Linguisti s 135

Lindsay did not possess Housman's genius. It is a question of reconsidering tion of the kinship relations among the lndo-European languages on the ba
a scholar who was certainly one of the greatest Latinists of the period be- sis of these data.
tween the end of the nineteenth century and the first decades of the twenti- All this might make one conclude that Pasquali's book merely system-
eth, and who did not follow in others' footsteps but almost always worked atizes Schwartz's ideas, supplies them with many examples, and lends them
in fields that had not been sufficiently explored hitherto.ff further emphasis in certain points. Indeed, many felc that it did just this.
Let us return to Pasquali.~ His defense of the rece,,tiores against preju- Above all in Italy, in an acmosphere saturated with Idealism, and hence with
diced and hasty condemnations, hh and his insistence on the importance of the indiscriminate polemics against any classification and "mechanicism" to
contamination in rich traditions and on the nonmechanical character of very wh ich I pointed earlier, the work of Pasqua Ii seemed to bean invitation tone-
many corruptions are in keeping with the Wilamowitzian and Schwarczian glect altogether the greater or lesser authority of the manuscripts, to a ban

inspiration of his work." Pasquali was oriented in chis direction not only by don any effort at genealogical classification, to put rccentiores and vetustio-
the teachers we have mentioned but also by his own direct experience as a res on an equal footing, and to constitute the text solely on the basis of
textual critic (the leners of Gregory of Nyssa, which he edited, have a very internal criteria. 47 The hypothesis of authorial variants, which is legitimate
rich and very contaminated and interpolated tradition) and his growing in- only as a last resort, 4~ was invoked to explain obvious banalizations or even
terest in medieval and modern philology and particularly in halian rexes, es graphic corruptions. Pasquali himself, in his preface to the reprint of 195:z.,
pecially following the fruitful exchanges of ideas and experiences between wro te, "I fear that in this regard my work has done even more harm than
himself and Vinorio Rossi, Giuseppe Vandelli, and above all Michele Barbi.ii good, and I feel the duty to warn beginners, even o lder beginners,01Jna8EiS
This last aspect is especially evident in the lase chapter of his Storia de/la (late learners] in that kind of philology, to be cautious." 49
tradizionc, dedicated to authorial variants: here the rich documentation of- But in reality, even in the first edition, Pasquali had been very far from
fered by the texts of Petrarch, Boccaccio, and Manzoni provides a starting proclaiming a pure and simple return to subjective i11dici11m. In this regard,
point for going back to analogous phenomena in antiquity whose attesta- indeed, a difference in mentality and orientation chat had existed from the
tion is for more infrequent and uncertain. very beginning between Pasquali and one of the teachers he loved most,
Pasquali was one of the Classical philologists most interested in linguis- G irolamo Vitelli, became clearer. It was to Vitelli, next to Schwartz, that
tics, not old-style lndo-European linguistics- and, what is more, not Struc- Pasquali had dedicated his Storia de/la tradizione. But for Vitelli, just as for
turalism eitherk-but the history of the Greek and Latin languages, and, in
his last yea rs, that of Italian. That is why the analogy of method between lin-
guistics and textual criticism, to which we pointed earlier, is far more ex- 47. In foct, this happened not only in the years immedi:itcly fo llo wing the appearance
plicit in his book than in the writings of his predecessors. 4 6 The very ten- of Pasquali's book, and not only in Italy. Sec Dawe 1964, :1cu1e and li:arned, but too " an
gry" and destructive; the author learned of Pasquali's book after he had already begun his
dency we just spoke of, to apply methodical procedures elaborated in the
research and reacted to it with enthusiasm ( 1 57n ), but he ended up agreeing more with
study of modern texts to the criticism of ancient texts, has its co uncerparr in
Schwartz than with Pasquali ( 160).
the linguistics of the late nineteenth and twentieth centuries, which started 48. Already in the first edition ( 1934: .p9- :to), Pasquali showed himself more pru
out from the study of Neo-Lacin origins or even living languages (and not, dent than other scholars, although he sometimes exnggcr:ited in hypothesizing authorial
like Bopp and Schleicher, from Sanskrit), and modified the earlier concep vari:lnts, and he warned, ~Authorial variants' are the last resort of textual criticism, :ind
it is not legitimate to have recourse to them so long as the divergences can be explained in
any other way."
46. See, e.g., Pasquali 195 :ta ( 1934): xvii, 160 (o n which cf. above, pp. 86f.), and 472; 49. Pasquali 1952a (1934): xxii; and already P:isquali 194:t: 137 = Pasquali 1968:
l95:tb 119331: 133 , 135, 136 - 39 = Pasquali 1968: 1.104, 106, 107-9 (acute observa 1.166, and t 947: 16 1. This c:iution ended up becoming excessive, also because of the po-
tions on the parallelism between linguistics and paleography, though here and there they lemic, often acute, but nor devoid of quibbles, that Gunther Jachmann directed against the
arc a bit flawed by Idealistic influences in linguistics); Pasquali 19 3 :tb. The fact that the hypothesis of authorial variants. Sccvola M;triotti, a scholar who at first had effectively
analogy between the two disciplines is still capable of being further developed and con and justly attacked some unmethodical hypotheses of authorial variants (S. Mariotti 1947:
tinucs to allow fruitful exchanges of experiences is demonstrated by many of the comm uni J03; 1950), later maintained that in certain cases the probability of autho ria l variants
c:uions at the Congress of Italian Philology collected in Studi e problemi di u itica testualc must be indicated, even if absolute certainty is unattainable, as is almost always the case
1961; see esp. Folena 196t. Out of this volume lacer arose a periodical on 1he subject of in ancient texts (S. Muiotri 19 54; 195:i.: 2181. Cf. the methodologically rigorous analysii
textual criticism. in Nardo 1967: J :t1-81.
136 CHArTER EIGHT Tcxrual C riticism and Linguistics 137

Gottfried Hermann, ars critica was identical with the perfect knowledge of third century AD (195z.a I 1934): 339); and point I 1 of the "Decalogue of
the style: although Vitelli was an extremely expert paleographer and inves- u Articles" in his preface (1951a [1934]: xix-xx) proves that he consid-
tigator of manuscripts, nevertheless he did not admit that external consid- ered this hypothesis valid for other authors too. In any case, the polemic

erations based on the authority of the witnesses had a significant weight in against Lachmannism in Pasquali's book is never divorced from a recogni-
the choice of readings, and he felt distrust and even dislike for research into tion nor only of its historical function but also of the value it still maintains
the history of tradition and the genealogy of manuscripts. 11 So one can un- in the present when we have to deal with mechanical recensions.
derstand that he did not fully appreciate the value of a book on methodol- Certainly, the criteria of lectio diffecilior and us11s scribendi acquire pri
ogy, even if the methodology was as devoid of precepts as Pasquali's was.""" mary importance in nonmechanica! rr recensions.51 But Pasquali's discus-
Commemorating his deceased teacher a few years lacer, Pasquali referred sion of these two criteria (1951a J1934): xu-14) aims precisely to free
with restrained bitterness to that lack of understanding: "In these last years, them from the dominio n of pure subjective taste (or from that of an abstract
in which I knew him better, I sometimes even suspected that he condemned rationalism, no less subjective even if it deludes itself that it is "universal")
systematic disquisitions on the relations among the various manuscripts of and to demonstrate that knowledge of the history of the tradition is neces-
an author as useless and found them distastefu l (... ). Certainly, even a few sary if they arc to be applied well. "Easy and difficult arc not absolute terms,
months before his death, he claimed the right to consriruce the text of a verse and what is difficult, that is unaccustomed, for us could have been easy for
of Aeschylus according to his own taste without submitting to canons that people of other periods. Judgment regarding the ease or difficulty of a read-
he found mechanical; I have not succeeded in convincing myself that he was ing will be all the more secure, the better the judge knows the customs of
right, either in that particular case or in general." 50 language and thought of the periods that transmitted it and that might have
Even Pasquali's great admiration for Schwartz did not prevent him from coined it. The best critic of a Greek text with a Byzantine tradition will be
noticing the exaggerations in certain methodological pronouncements by the one who is not only a perfect Hellenist but also a perfect Byzantinist.
the editor of Eusebius to which we have already referred. Reread pages I 36 - The best ~ditor of a Larin author transmitted in medieval or postmedieval
41 of Pasquali's book, and you will see chat his agreement with Schwartz is manuscripts will be the one who knows the Middle Ages and Humanism
accompanied by reservations: " I consider exaggerated only the first of these just as well as he knows his author and his author's language and times and
words, which attack the concept of archetype" (1951a (19341: 136); " It the language of his times. A critic of this sore is an ideal that no one can in-
would be mistaken co derive a presumption against the existence of an ar- carnate perfectly in himself, bur toward which everyone has the duty to try
chetype from the number and quality of the variants" (I 37); "Not even here to come as near as possible." ~ 2
is everything right[ ... ). If it is true that errors can be transmitted by colla- But Pasquali did not even abandon external criteria altogether for non-
tion just as much as genuine read ings can [... ], nonetheless it is certain that mechan ical <t<I recensions. He never completely resigned himself to Paul
the transmission of the text, the 'tradition,' occurs on principle in a vertical
'd irection,' as is only natural" ( 140). Pasqua Ii also rejected Leo's opinion
s1. I avoid Pasquali's terms closed rece11sio11 and of1c11 rccc11sio11, despite their popu-
that the whole manuscript tradition of Plautus goes back d irectly to an edi- l:1nty in Italy and abroad, because Alberti 1979: 1- 18 h:is demonstrated that Pasqu:1li al
tion by Valerius Probus, and considered it indispensable to return to the hy- ready uses both terms, but espcci:1lly rhe second o ne, in 100 many different senses. Thii
pothesis of an archetype- even if not a medieval archetype, but one of the r:1ki.-.: nothing away from the fo cr rhat in Pasquali's time these two expressions were sig
nific:int and effective in the polemic ag:1inst orthodox Lachmannism.
J 2. See :1lso the discussion with K. Ziegler about clausulas in the So11111i11111 Scip ionis
50. Pasquali 1936: 9- 10 Pasquali r94.z.: 300 = Pasquali 196H: :z..:z.07. Bur sec al- (Pasqua Ii 19p.a I 19341: 117- 18): Pasquah mamrains rightly rhat a text cannot be consu-
ready Pasquali r964 (1 920): 77. I can add an oral testimony. Pasqu;ili told me once (;ind ru1ed on the sole basis of 1111111cms without taking the documcnrary auihomy or the vari-
he certainly will have told others as well) that Vi1elli, after having read or skimmed rhe ous readings into account. Later, Ronconi ;ind Castiglioni rook up the problem or the di-
book-whrch was dedicarcd 10 him, with words full of admiration and affec1ion which vergences in the collocation of words in 1hc So11111i11111 once ag:1in and gave two differenr
concluded, "I n him I revere rhc greatest expert on Greek poerry among all living mcn"- explanations for it (cf. now Ronconi 1961: 40 , 61 ). My own view is thar Ronconi was
5;i1d to him, with forthright frankness bur also wirh a total lack of understanding, "You righr, at least in most c:iscs, and I think he should have defended his hypothesis more de
would have done bcner to wrirc a book about an ancient :1uthor instead of a book on cisively against C:isriglioni's. In any c:isc Pasquali's requirement for the constitution of the
methodology." text rem a ins valid .

Maas's aphorism, "No specific has yet been discovered against contami- APPENDIX A
nation" (Maas 1958 (1927]: 49).n A good part of the fifth chapter of his
book is dedicated to the search for new "objective criteria," ones more so- .Lachmann's First Attempt at a
phisticated than Lachmann's and capable of "resolving disagreements in the Mechanical Recensio in r81 7
case of an open recension" (19 52a [1934]: 160). Pasquali indicates one such
criterion in the norm of lateral areas, which we have already had occasion
to mention/' and another in Ulrich Knoche's attempt to classify contami
nated manuscripts of poetic texts genealogically on the basis of lacunas that
impair the meter (1952a [1934): 180-83). If there is a defect in this part of
Pasquali's book, rich in erudition as it is, it consists in his excessive faith that
these criteria could be fruitful "not only for the history of the text, but for
the text itself" (1952a (1934]: 177n1). The impression remains that when
the history of a text is very complicated, it is not very useful for textual crit- In July 1817 Lachmann published a long review of Friedrich Heinrich von
icism, but otherwise has a value in itself, that it belongs to the history of cul- der Hagen's edition of Der Nibelu11ge11 Lied (Hagen 1816) and Georg Fried-
ture, to the Fortlebe11 [survival] of the Classics. And the practical exigency rich Benecke's edition of Bonerius's Der Edel Stein (Benecke 1816). 1 Lach-
remains that certain critical editions not be postponed forever for the sake mann distinguished rwo redactions in the manuscript tradition of the Nibe-
of studying the history of the tradition in all its smallest details, that schol /1111ge11 Lied: a shorter and more genuine one contained in the manuscript
ars no t bury themselves so deeply in the study of medieval and Humanist he called B; and another, longer and heavily interpolated one, represented
culture that they forget to return to textual criticism. Nonetheless, although by the manuscripts GEM. 2 According to Lachmann, both redactions have
in more recent works the separation between history of the tradition and reached us disfigured by corruptions and secondary interpolations; but
textual criticism has now taken place and been codified,55 in Pasquali's book while the first one cannot be reconstructed in its original form until another
the two disciplines are still conjoined. Pasquali's inrerest in the vicissitudes manuscript, a brother of 8, is discovered, the second one can be recon-
of Classical texts in the medieval and Humanist periods, lively as it was, structed by comparing GEM.
never makes him forget his job as textual critic and interpreter. It is in this According co Lachmann, such a comparison reveals that the ancestor of
combination of a broad perspective on cultural history" with an acute GEM was still fairly free of interpolations in the text wrirten by the first
philological intelligence directed to the individual passage of an ancient au- hand but that a second hand inflicted many changes and arbitrary additions
thor that the unmistakable character of Pasquali's work resides. upon the original text. Each copyist of GEM reproduced now the reading of
the first hand, now that of the second hand, and also inrerpolated on his
53. Cf. the preface to M:ias 1958 (192.7): viii-ix. own. So in order to make Lachmann's thought easier to understand we
54. See :ibovc, pp. 86-87. could trace out the following stemma:
SS See, e.g., Hunger ct al. 1961 - 64, which even imroduces a further distinction be
tween Tcxtgcscllichte (history of the text I and Textubcrlic(cru11g ltransm~sion of rhe text), 1 . Lachmann 1817. I warmly thank the director of the university library at Jena for
and supports it with a cert:i in doctrina ire ostentatiousness. Against this di5tinction cf. also sending me :i photograph of this review. The review is republished in Lachmann 1876:
S. Mariotti 1966: :?.36. In the last few years there has been a salut:iry reaction: more and 1.81 - 11 4.
more philologists first tre:it the manuscript tradition of :i text in :i mt>nograph, making i . B (now designated conventionally as A: Munich, Bayerischc Sta:itsb1bliothck, cod.
:imple use of the assistance of codicology and cultural history, and then go on to do a Cnt germ. 34 ( and M (- D: Munich, Bayerische S1aa1sb1bliothck, cod. germ. 3 1J arc m Mu-
ic:il edition of it with the same success. Other scholars, on the other h:ind, in tcxtu:il crit nich, E ( C: Karlsruhe, cod. Donaueschigen 63) m Donaueschingcn (now m Karlsruhe},
icism as in many other disciplines, have preferred the e:isy path of terminological exhib1 G ( B: St. Gallen, Stifstsbibliothck MS 8 571 at Sankt Gallen. Lachmann 1876: 1.84n in-
tio nism to which no genuine conceptual progress corresponds; one of the worst examples dicates the correspondences between Lachmann's symbols and those th:it von der Hagen
of this tendency-it is useless to fool our~ lvcs: it will certainly find admirers and follow had used and which Lachmann himself went on to ado pt in his own cdiuon of the Nibc
ers-is the interdisciplinary seminar published as De/ tcsto 1979. /1mgen Lied. Nowadays, in contrast with Lachmann's view, scholars generally believe that
the shorter redaction 1s the more reccnl one.
140 APPENDIX A L:ichmann's First Attempt at a Mechanical Rcccnsio l ..p

there was a reading x, that it was reproduced faithfully by B and cjl, and that
4>2 added a variant y in the margin. Let us also suppose that the copyists of
G and E reproduced the reading of the first hand (x) and the copyist of M
B cf> (cf>' ) th<lt of the second hand (y). In that case the agreement of three manuscripts
'-v--' (BCE} against one (M) would give us not cp2 but cf>.

The same thing m:Jy be said in case the readingx was reproduced by GM
and y by E, or x by EM and y by G. Obviously the agreement of BGM (or
of BEM) would give us q, and not cp2
In reality, in contrast with Lachmann's rule, the agreement of three man
where w would indicate the oldest redaction, cf> the first hand of the an uscripts against one gives us with certainty the reading of cf> (:Jnd of w), if one
cestor of GEM, and q, 2 the interpolations by the second hand in the same of the three that agree is B. Only if none of the three that agree is B-that
manuscript. is, only if GEM agree against B-is it possible that the three preserve the
Lachmann adds that the editor's task is co identify the interpolations in reading of $ 2 For one can suppose that <I> has the same reading :JS B (and as
the ancestor of GEM: cha~ is, as we would put it, to reconstruct the readings w), and that GEM have all three reproduced the variant of .p1 . Such a hy
of q, 23 To achieve this purpose he provides the following rules, in which the pothesis is not improbable, especially if we admit the possibility that the
signs > and < signify, respectively, "better than" and "worse than," just as scribe of cf> 2 did not limit himself to adding the variant in the margin or
in mathematics they signify "greater than" and "less than": :Jbove the line, but also deleted the reading of the first hand. In :Jny case this
is only one of the possible hypotheses: the agreement of GEM against B
r. Three manuscripts out of our four outvote one every time. could also be explained on the supposition chat cf> (not <!> 2) had innovated
2.. When any two agree, then BG < EM {i.e., where B agrees with G, the with respect to w and that its innovation had been reproduced by its three
unanimous reading of E and Mis to be preferred), GE > BM, GM > BE. apogmphs, or else that B had innovated 4 and that the reading of w had been
3. Where there arc three readings, then BG > E- M (the reading shared reproduced by cf> :Jnd hence by GEM. So the probability that the agreemenr
by Band G is preferable [to) the two others in E and M), GE > B- M, of GEM against B reprcscnrs <1> 2 is not very high.
GM > B-E; on the other hand, EM = B- G (the agreement of E and M The second rule, on the other hand, is correct. When there arc two read
leads to no secure decision against the two readings of B and G), BM "" ings, each one attested by two manuscripts, it is certain that the reading at
G- E, BE = G-M. tested by B and by one apograph of <I> W<lS <llready found in <I> :Jnd w, while
4. When all four disagree, the original reading is just as uncertain. the other two apographs of <I> reproduce the variant of the second hand (<1> 2).
{Lachmann 1817: 117-18"" Lachmann 1876: r.8 6- 87) So too the fourth rule is correct (and obvious).
But the third one presents real absurdities in the form in which it is
We have reproduced the passage from Lachmann 1817. The reprinted printed in Lachmann 1817. The agreement of BG against isolated readings
version in Lachmann 1876 reads "then BG < E-M" in the third rule. of E and M certainly gives us cf> (and w), not .pi, nnd hence it cannot be "pref-
Aside from this variant, the text of Lachmann 1876 is identical to chat of erable" in trying to reconstruct cjl 2 The ngreement of GE against isolnted
Lachmann 1817. The eminent Germani.st Karl Miillenhoff, who edited this readings of B and M, or of GM against isolated readings of B and E, can
first volume of his teacher's Kleincrc Schriften (Lachmann I 876), did not represent <1>1. but can :Jlso represent <fl. No less erroneous are the formulas
add any notes to these rules, as though they were perfectly clear. But in fact BM = G - E and BE - G- M in the second part of the same rule: this time too
they are a real brain-twister. the agreement of B with an apograph of cf> gives us with certainty cf> and not
The first rule already seems to be unacceptable. Let us suppose that in w
4. Although Lachmann considered B ro be the bcsr manuscript, he srill maintained
3. Lachmann 1817: u8 - L:ic hmann 187 6: 1.87: "The editor's task is ro discover tha t it was not free fro m corruptions and interpolations: "since it must be expected that
these changes, which now this copyisr, now th:i,t one overlooked, and every one increased nehher the older recension in B nor the more recent one m G will have been transmitted
wirh new changes. " Anto nio L:i Penna has help!:d me undersrnnd rhe rask Lachm:mn set to us without errors and arbitrary changes, partly negligent and partly intentional, on the
himself. pa rt of the copyist5" IL:ichma nn 1817: r 17 "" Lachmann 1876: 1. 86).
APPENO J X A Lachmann's First Anempt at a Mechanical Rl!cc11sia 143

cj> 2: hence for reconstructingcj> 2 such an agreement is "worse" and not "equiv- B this is impossible. But in reality, as the parenthetic explanation ("the
alent" compared to the isolated readings of the other two apographs of cj>. agreement .. .") also makes clear, the rule must be understood to mean that
Such serious inconsistencies (to which must be added the syntactical ab- the shared reading of two apographs of <I> and the different reading of the
surdity of the parenthetic expression "the reading ... is preferable [to) the third apograph of B have an equal probability of reproducing cj> 27
two others") suggest that the genuine text of the rule has been distorted by Thus there is perfect agreement between the second and third rules: both
typographical errors. Fritz Bornmann has made a very acute and persuasive rules mean in substance that a reading attested by one or two apographs of
attempt at correcting it in an article dedicated to this problem,i..s in which q, can represent 4>l as long as such a reading is not found in B too. As Born-
he has proposed that the third rule should read as follows: mann notes, the two rules are parallel formally: in both cases Lachmann be-
gins with a negative formulation ("such a reading is to be refused in favor
3. Where there are three readings, then BG < E-M (against the reading of such another one") and then goes on to expound the second and third
shared by Band G, the two others in E and Mare preferable), G-E > BM, cases in a positive form ("such a reading is to be preferred to such another
G-M > BE; on the other hand, EM = B-G (the agreement of E and M leads one").
to no secure decision against the two readings of B and G), B-M = GE. All the same a striking contrast remains between these two rules and the
B-E GM. first one, which we discussed above (pp. 14of.). In this case we cannot hr
pothesize a typographical error: the incoherence was in Lachmann's mind.
In the first formula, Bornmann evidently accepts Miillenhoff's correction Bornmann acutely detects the reason for this incoherence in the "superpo~
in Lachmann 1876, already cited. But while Mullenhoff had limited himself sition of two criteria, one purely recensional, aiming to identify mechani-
to this single change, which is not enough to restore sense to the whole pas- cally the interpolated readings of cj>, and the other, which I would call edi
sage, Bornmann gives us for the first time a coherent text (except for the first torial, of reconstructing a text as close as possible to the original (which,
rule, of which I have already spoken and to which I shall return)' and ob however, is supposed to be capable of being reconstructed only in the class
tains this by means of relatively minor corrections: indeed, his restitution of of manuscripts derived from 4> ). " An examination of the concrete examples
the text requires changing not the letters but only the mathematical signs, that follow the formulation of the four rules in Lachmann's article seems to
which are often subject to displacement and confusion by typesetrers, espe- confirm Bornmann's hypothesis.K In fact there is something paradoxical
cially in difficult conrexts. 6 about this whole complicated attempt to reconstruct cj> 2 (a series of interpo
If we accept Bornmann's text-and I believe we must- then Lachmann's lat ions!) when Lachmann could have even reconstructed the first source
th ird rule can be paraphrased as follows: the agreement of one of the three of the whole tradition, w, with much less effort and much greater profit. For
apographs of <I> with B gives us the certainty that that is the reading of cj>, not the agreement of B with one of the apographs of <I> gives us with certainty
of cj> 2, while the isolated readings of the other two apographs can both re- the reading of <I> and of w. It is perfectly natural that although Lachmann had
produce cj> 2; on the other hand, the agreement of two apographs of cj> against set himself the difficult and all in all not very useful goal of reconstructing
isolated readings of the third apograph and of B does not give us any guar Q2 , he then tended more or less unconsciously to reconstruct w. But it must
antee of reproducing the reading of cj>2, which might have been preserved by be added that the first rule is erroneous not only from the point of view of
the third apograph. Strictly speaking, the formulation of the second part of reconstructing <fi 2 bur also, and just as much, from the point of view of re
the rule (starting with "on the other hand" ) is not very felicitous, because it constructing cl> and w: the agreement of GEM ("three manuscripts out of our
might suggest that the readings of B and of the third apograph might both four") against B can represent w, but it can just as well reproduce an inno-
reproduce q, 2, like the shared reading of the other two apographs-while for vation of 4> or even of <j>2 At this point Lachmann seems still to have been a
victim of the false criterion of the majority, from which he will free himself
only in his edition of the New Testament (see above, p. 8 5).
5. Bornmann c96:1.. In a first vcn ion o f this :!;!udy of mine, I had accepted a different
anempc at correction, proposed to me by Eugenio Grassi and Antonio La Penna; although
that :mempt too was very astute, it does not now ~ccm to me easy to defend it :igainst the 7. Cf. Bornmann c96:1.: 49n:1.: "The meaning [of EM 8-G) is: the coincidence o f
objections of Bornmann 196:1.: 48-49. the two readings in E :ind M does not lead to a secure decisio n in comparison with G when
6. For a detailed reconstruction of the origin of the errors, sec the observations of G differs fro m B, that is, it docs not permit us to reject G."
Bornmann c9 6:i.: 50. 8. Cf. Bo rnmann 19 6:1.: 50- 53, and 4 7n5.

Three years later, in a review of a new edition of the Nibelu11ge11 Lied by APPENDIX 8
the same von der Hagen, Lachmann reconfirmed the value of his rules,
though he admitted the possibility that they could be rendered more pre- Determining the Script of Lost Manuscripts
cise., Had he therefore not noticed that the rules were almost useless, at
least in the form in which he had published them in Lachmann 1817?
Finally Lachmann published a critical edition of the Nibe/1mgen Lied him-
self in 1826. 10 But this time he did not try to reconstruct. the longer and (ac-
cording to him) interpolated redaction and adhered fundamentally to the
manuscript that he had called Band for which he now adopted von der Ha-
gen's designation, A. 11

9. Lachmann 182.oa: Ergiinzungsbliiner 70-76 = Lachm:mn 1876: 1.:1.16.

1o. Lachmann 18 SI ( 1816). We have seen how Scaliger already tried to determine the script of the arche-
11. For an anempt to demonsmue the correctness of Lachmann's rules without hav type of Catullus on the basis of some characteristic errors that appear in his
mg recourse to corrections of this sort, cf. Lurz-Hensel 1971 and 1975: i.:z.8-39. Lurz manuscripts, and how Lachmann did the same thing for Lucretius (above,
Henscl's interprccacion, which Bornmann and I do not consider to be at all persuasive, will pp. 51- 52, 107-8).
soon be discussed by Bornmann in Zcitsc/Jrift fiir Dc11/sc'1c Pl1ilologie. In any case this cm Attcmprs of this sort have a good chance of hitting the mark if they arc
inent scholar's book, praiseworthy as it is fo r its rich and derailed information, is ruined
performed with a rigorous method, and they are useful both for the history
by its nebulousness and captiousness. Cf. now also Cecchini 1982., an intelligent article,
of trad ition in itself and for textual criticism: for example, if one succeeds
which however docs not resolve the question once :1nd for all in my view. But Fritz Bo rn
mann will deal with all this in the article whose appearance I announce in this note.
4 in demonstrating that the archetype was in capital script, conjectu res that
presuppose a confusion between minuscule letters become less pro bable,
and so forth.
But if we wish to avoid getting lost in unfounded hypotheses, we must
rely above all on a large number of readings that arc certainly erroneous and
cannot be explained otherwise than by the similarity between certain letters
in a certain script. Errors that can be attributed to other causes too (that is,
to confusions between graphic signs in other scripts, or to nongraphic rea~
$Ons)' have no probative value. For example, confusions between c and g,
or between e and i, might indeed be due to erroneous reading of a model
written in capitals; but, at least in many cases," they might also be phonetic
vulgarisms (the confusions c-g might also provide an indication of a model
in uncials, and an even better one)," and hence it will certainly be better to
neglect them. Matters become even worse if we rely on readings that might
be correct, or that could be corrected just as well (or better) J in a different
way too.
Once a large number of really probative errors has been collected, we
must bear in mi nd the following points: (1) if each of the apographs of a lost
manuscript a presents errors of its own due to misunderstanding a given
script, then that was the script of manuscript a; (i.) if the apographs of a
present shared errors due to misunderstanding of a given script, then that
was the script not of a but of the model from which a was directly o r indi-
Al'l'ENDIX B Determining the Script of Lost Manuscripts t.f?

reedy copied-for in that case the errors shared by the apographs were al some passages in his Storia de/la tradizione make clear, 1 even if a few other
ready (ottnd in a, and hence they were due to misunderstanding the script passages show some uncertainty.; And yet especially the second criterion has
of a preceding manuscript. often been neglected. For example, W. V. Clausen quoted a large number of
Neither of these criteria is immune to objections. Against the former it corruptions common to the two manuscripts of Persius A (Monrepessulanus
can be noted that an error already found in a might have been reproduced 212) and B (Vat. tabularii basilicae H 36) in his editioni of that poet in or-

by a copyist A and corrected conjecturaUy by a copyist 8: in this case we der to demonstrate that they were both derived from a model in minuscules
would be attributing to a the script that was instead that of a 's model. (Clausen 1956: vii-viii). But no: those corruptions demonstrate that what
Against the latter it can be noted that various copyists might have misun was in minuscule was the manuscript from which the model of AB was cop
derstood the same graphic sign of the model a independently of one another ied. Certainly, this makes it likely that the model of AB too was in minus-
(for example, in pre-Caroline minuscule they might all have mistaken for a cule, even without going in search of minuscule corruptions characteristic of
u an "open a," since this letter was particularly subject to being misread in A alone and of B alone-but only as a secondary inference. 2
that script): in this case we would be attributing erroneously to the model of For the same reason, even if the examples of corruptions of scriptura
a the script which was instead that of a. l.Angobardica (in the sense of pre-Caroline minuscule) that Scaliger col-
If the examples on which we base our conclusions are few, these objec- lected in the whole tradition of Catullus he knew of had been numerous
tions doubtless have a considerable weight; but in the face of a large num enough and had all been certain enough,k.3 they would have served to dem
ber of examples, it becomes quite improbable that various copyists might in-
dependently have made a mistake every time at the same point, or that no w
1. Pasquali 195:La ( I9H): 149 and esp. 194-95. This question has also been wcll
one copyist, now another, might have made a felicitous conjecture, espe-
treated by Fabre 1947 (1936): lvi- lvii; Andrieu 1954: lxivn:z.; Pfeiffer 1949-53: :z..lxxxiv;
cially if there arc not only two apographs of a but three or more. K. Muller 1954: 185-97 . Sec also the beginning of the following note. We shall discuu
Giovan Battista Alberti has called my attention to another possible ob- Lachmann and Duvau shonly.
jection against the former criterion: mistakes due to misunderstanding a :z.. Later, Clausen himself made this point quite well: Clausen 1963: :z.55. On the
given script which are found in the individual apographs of a m ight indicate ancestor of the Palatine manuscripts of Plautus, see also 0. Scyffenh 's correct objection
the script not of a but of lost apographs intermediate between a and each of (Scyffenh 1896: 1551) against Lindsay, quoted and confirmed by Questa 1963: u8. In his
the extant manuscripts. This objection is valid above all when our data are edition of Caesar's Bel/um civile, Klotz ( 1950: v) erroneously cites examples of errors due
to misunderstanding of abbreviations in order to demonstrate that uthe man who made
few or contradictory, as in the case of Lucretius (see below). Bur if the
1he archetype of the manuscripts used many abbreviations, some of them rare ones," even
graphical errors in the individual apographs A, B, etc. are numerous and
1hough in the meantime the problem had been discussed in 1he right terms by Fabre (I 94?
they all derive from misunderstanding minuscule letters, then it is highly I 19361: lvi-lvii): since those errors are found in all our manuscripts, the abbreviations that
probable that they reflect the script of a and not of hypothetical intermedi- gave rise to them must have been found in a prcarchetype, not in the archetype. An anal
ate links, since it would seem strange if no trace of misunderstanding of rhe ogous mistake is 10 be found in Klotz's preface to S1atius's Tbebaid, a preface that, in other
script of a had survived. Furthermore, if the graphical errors of the individ- regards, remains fundamental even today: he ciics many eJCamples of errors of minuscules
ual manuscripts are all derived from misunderstanding majuscules, it is even (insular and perhaps, more generally, preCaroline ones?) in i11dividua/ manuscrip1s of the
more improbable that the common model a was nor in majuscules: other w class and concludes on this basis that between them and their ancestor there were inter
wise we would have to suppose a passage from an ancestor in minuscules to mediate eJCcmplars in insular script (Klotz t908: lvii-lix). In theory this is possible (sec
Albeni's observation cited above, p. 146); but precisely the fo rge number of examples-
various descendants in majuscules-somethi ng not impossible, to be sure
whfch, as Klotz himself says, represent only a selection-provides a sufficient demonstra-
(especially now that we know that as a rule the first ancient copies of our
tion that already bl was wrinen in insular script or in pre-Caroline minuscule: the winter
Larin classics were nor written in capitals: see below, p. 152 and n. 16), but mcdiarics~ ore neither a necessary nor a demonstrable hypothesis, one suggested, I
difficult to postulate when all the manuscripts descend from a medieval ar suspect, by the preconception that the archetype necessarily had to be in capitals. The mi
cherype, and virtually impossible for Greek texts.' nuscule errors shared by all the descendants of w which Klotz cites at 1908: Ix serve to
The two criteria enunciated above are not to be found formulated ex- demonstrate that not w but one of its ancestors was written in minuscules as well.
plicitly in the manuals of textual criticism or paleography I have seen, bur 3. Grafton 197 5: 17 rn 58 cites from Scaliger's Castigatio11cs five examples of the con
they do not represent a real novelty.ii Giorgio Pasquali knew them well,11 as fusion a-u, three of i-/, seven of c-t, one of c-g. But the confusions i-1 and c-g can oc-
Al'l'ENDIX II Determining the Script of Lost Manuscripts 149

onstrate that a manuscript earlier than the archetype was in pre-Caroline certain examples of corruptions due to misunderstanding of minuscule
minuscule, but not the archetype itself. 4 script; 8 but in fact, despite the claims of those who cite him without hav-
Considerable confusion reigns on this subject in many editions of Lucre ing read him with a minimum of attention, he derived from them the infer
tius. To read the prefaces, one would think that Lachmann relied on cor ence not that the archetype was in minuscules, but rather that what was in
ruptions shared by the whole tradition in order tO maintain that the arche minuscules was the pre-archetype: "If all our manuscripts present shared er-
type was written in rustic capitals, and that Louis Duvau later demonstrated rors deriving from a resemblance between certain letters that exists in mi-
on the same basis that the archetype was in insular minuscule, copied in its nuscule script, and exists only in this script, then it follows: ( 1) that since
turn from an archetype in capitals.5 If this were correct, then both Lach these errors are shared by all our manuscripts, they were found in their ar-
mann and Duvau would have fallen victim to the same error that we have chetype; (:z.) that the origin of these errors, that is, the minuscule script, was
just now noted in Scaliger and in some recent Classical scholars. Bur in fact found in the manuscript to which this archetype goes back directly or indi
Lachmann limits himself in his preface ro stating, "Many indications prove rectly" (Duvau I 888: 34). His judgment regarding corruptions due to the
thar rhe script of that manuscript [sc. the archetype] was in rather thin cap- misunderstanding of abbreviations was just as precise: "Since they are found
ital letters, not in uncials" ( 18 5oa: 1.3 ); in the commentary, unless I am mis in all our ma nuscripts, they were made at the latest by the copyist of their
taken, he notes explicitly only one of these indications, the reading homo- archerype: hence their cause, that is, the use of abbreviations, was found in
fomerian of the first hand of the Oblongus at 1.830, where the Quadratus a manuscript earlier than this archetype itself" (1888: 33).
and the Schedae have the correct reading homoeomeria11. 6 He would cer- According to him, the archetype too was in minuscule script.~ He did not
tainly have done well to cire other examples to confirm his thesis, for that give a positive demonstration of this claim; the fact that the pre-archetype
single case is not enough, as we shall soon see more clearly. But it cannot be was in minuscule script must have seemed sufficient to him to demonstrate
said that in itself it was chosen poorly: Lachmann deduced the script of the that the same thing was true a fortiori for the archetype. Against the Lach-
archetype not from a corruption shared by rhe whole tradition but by a cor- mannian hypothesis of an archetype in capitals he observed: "I seek in vain
ruption in a single apograph, in conformity with the former of the two cri- for the reasons Lachmann thought he was obliged to believe that the origi-
teria we enunciated earlier. Nor can there be any doubt that the confusion nal of the Oblongus was in capital script. The fact that no doubt led him to
between e and f presupposes a model in capital letters. conceive this idea, the frequent confusion of letters that resemble one an-
In an article of exemplary clarity and rigor,7 Louis Duvau collected many other only in capital script (for example, i, c, /, t), seems to me to prove ex-
actly the opposite. For these errors are not committed separately by each of
the copyists of the Oblongus, the Quadratus, and the Schedae: the same
cur with other scripts too, and among the other confusions only very few indeed are pro words are altered in the same way in all our manuscripts( ... J. The conclu-
bative: most are connected to completely fanciful conjectures on the part of Scaliger and sion is obvious: these errors existed in the arch~type, either because the
collapse together with those conjectures.
copyist of this manuscript introduced them himself when he copied an orig-
4. There is also the possibility that some minuscule errors m Catullus's text, as in other
inal in capitals or-and this is the hypothesis that must be accepted( ... ]-
authors', go back to a much earlier ph:asc, th3t of "ancient minuscule": cf. Brunholzl t 97 r:
2.t-:1.2., and below, pp. 1 p-53 and nn. t6, 18. because they already existed in the manuscript that he was copying" {Du
5. See, e.g., Diels 192.3-2.4: vii, xxiii; Ernout 1948 (19 ;:;0 )~ xvi- xvii; Leonard-Smith vau 1888: 34 ). Hence this was an archetype in minuscules, copied from a
1942.: 107, 115; B:ailey 19:1.2. (190 1): 1.) 7-38; Buchner 19561 2.0 1 = Buchner1964: 12.1. manuscript in minuscules too, that in irs turn was derived directly or indi
Brunholzl too {196z: 98) rightly doubts the insular origin o r the archetype and prearche rectly from a manuscript in capitals.
type (cf. Reynolds-Wilson 1991 11968): z61) but incorrectly anributes to Duvau the opin- Now, there is no doubt that Duvau was perfectly correct in explaining
ion that the script 0 1he archetype can be ascertained on the basis of the corruptions
shared by the whole tradition. Martin 1969 (1934): viii repom Duvau's real opinion cor
reedy, even if a bit too briefly, and without distinguishing between the Quadratus and 1he 8. Duvau t88ll: n-36; some examples are repeated in Ernout 1948 (1 92.0): xvii.
subarchetype common to the Quadratus :and the Schcdae. 9. Duvau 11188: 33-34: ~1 believe that not only 1his direct original of the Oblongus
6 . L:achmann 185ob: 2.. 57 on 1.830: "The first hand 0 the Oblongus does not have (i.e., the archetype I was in minuscule scripr, but so too w:as rhe lost original of this manu-
homoiomeriam, as Havcrcamp reports, but homofomcrian, from which we can tell what script." Purmann (1846: 14) had 1hought of nn archetype or prcarcherype in scriptura
kind 0 script the archetype was written in." La11gobard1ca (i.e., Bcnevcnran? or rather here too pre-Caroline minuscule, as in the pas-
7. Duv:au J 888. sages from Scaliger we cited above, at chap. 1, n. i.o?}, but without :idducing any proof~.
150 Arl'ENDIX 8 Determining the Script of Lost Manuscripts 151

the capital errors shared by the whole tradition as errors inherited from a with less confidence on 2.685 (noscat for noscas in Q and the Schedae) and
very ancient phase. But he was mistaken in supposing that Lachmann relied 2.839 (remotu for rcmota in 0): the first case may be a substitution of the
on these errors for his hypothesis concerning the archetype: we have already more common third person for the second, while remot11 may be the result
seen that Lachmann relied on a corruption in the Oblongus alone. Evidently of attraction from the preceding sonitu. 12 Jn the parts of the poem where
Duvau did not notice this passage in the commentary to i.8 30, and so he the Schedae are lacking, the Quadratus presents some evident errors of pre-
attributed co Lachmann the very same error of method that the editon of Caroline minuscule; 5.374 aequori frmdis for aequoris u11dis, 482. g1trgites
Lucretius would later attribute just as unfairly to himself. ossas for gurgite fossas, t 147 l11sa for iura {the Oblongus has /ura); but since
In fact, although there can be no doubt concerning the existence of one the subarchetype was certainly in pre-Caroline minuscule, 1.1 these errors
pre-archetype in capitals and another one in minuscules, as we have seen may be peculiar to the Quadratus, so that we cannot make any inferences
just now, the indications regarding the script of the archetype are not un- from them concerning the script of the archetype.
ambiguous. And, as far as I have been able to detemrine, this is a sit uation Thus the indications we have are few " 1 ~ and, what is worse, contradic-
that occurs quite 1 frequently: it is much m easier to determine the script of tory. Anyone who wishes to keep believing in the hypothesis of an archetype
manuscripts earli<!r than the archetype than the script of the archetype itself. in minuscules will have to suppose that the capital corruptions of the Oblon-
In support of an archetype in capitals, there are other indications besides gus alone, or of the Quadratus + Schedac alone, were already found in the
the homofomerian of the Oblongus at 1.830 noted by Lachmann (to which :irchetype and were corrected conjecturally by the copyist of one of the two
we shall return shortly}: only 0 has iam for tam at :z..1088 and 5.901, ve- apographs. This is certainly possible for banal errors like eacies for facics;
niorum for vcntontm at 5.12.30; Q and the Schedae (but not 0) have ant- and even the sole error that, taken in isolation, might seem difficult to cor-
facta for anteacta at 1.:z.33 and tinguntque for finguntque at 3.9o;n only Q rect conjecturally, that is, the homofomerian at r.83 0 upon which Lach-
has eacies for fades at 4.733, forum for eontm at 5.1337 10-all" corrup- mann insisted, could have been healed by the copyist of the Quadrarus (or
tions that do not seem to be explicable otherwise than by a model in capitals. of the subarchetype fro m which the Quadratus and the Schedae descend) by
Less certain is reficit for reiicit in 0 at 1.34: this could be a semiconscious comparison with 1.834, where the whole tradition has the correct reading
banalization, since reficit seems to yield good sense, and the erroneous pros- homoeom crian (as has been co rrectly pointed out to me by E. J. Kenney 1 ~).
ody re- could have been influenced by apparently analogous cases like re- On the o ther hand, anyone who wanted to exclude the hypothesis of con-
ligio (aside from the fact that unmetrical readings are anything but rare in
the manuscripts of Lucretius).r u . Ar :z..833 (where the Oblongus hos disced11nt, the Quad rntus and the Schedae have
disced11a111, and the correct reading is discedam , :is Marullo saw), it is possible th:it the ar
Important indications in favor of an archetype in pre-Caroline minuscule
chetypc already had a doub le reading, disccdwzt wi1h a n a added above the line: cf. Lach-
are 1.282, where Q and the Schedae have mtget for auget (and hence so too
mann 11i5oa: :z..12.3.
did the subarchetype from which they descend), 4 13 (meos aiauis for meo 13. Sec, e.g., 1.4 :1.4 re(ercntcs G, sc {erellles Q; :z..::.06 volat1tcs Q, vo/1mtas G; ::..891
s11a11is), 506 (p11mmq11e for puramque)<i-all confusions between 11 and scdus Q, (cdm V; 6.1106 britta11is Q, brit1mis U; 6.117 s mcrsa11s 0 (correc tly), messam
" open a." 11 So too, parcis for partis in 0 at 5. 354 is quite probably an er- U, incrra11s Q; 6.1 :z.68 vidcrcs Q, videscs U.
ror of minuscules, certainly not one of capirals- if anything, then of uncials, 14. So me others con no doubt be found; bur most of the corruptions peculiar to one
even if the confusion is not between words of generally similar phonic and of the two branches of the tradition consist in vulgarisms or elementary psychological er-
graphic appearance but instead between individual letters.' I would rely rors (artraction of endings, ere.).
15. As Ke nney reminds me, "The Middle Ages knew the word homoios from theo-
1o. In these last two passages the testimony of the Schedae is lacking; but nonetheless logical d isputes." I add that Dr. E. Hulshoff Pol of the university library of Leiden, whom
these cannot possibly be errors committed by the copyist of the Quad ratus in 1he course I asked to confirm the reading homofo111cria11 of the first hand o f rhe Oblongus, kindly
of transcribing the a rchetype, since the subarchetype was certainly in pre-Caroline minus warns me that that reading is not at all certain: "The c of homoeomcria11 in line 830 is
cule {see below, n. 13). So these must be errors committed by the copyist of the suban:he- wrincn with a darker ink over an erasure. But I do not venture to say thar what the firsr
type m the course of transcribing the a rchetype. hand wrote was an f. Even under the infrared lamp no trace remains. The only thing that
t 1. David A. West has drawn my attention to some of these errors of minuscule. At can be affirmed is that the erasure has more or less this form (the form of a rightangled
1.:z.8:z. the second hand of the Oblongus ha5 urget, preferred by Woltjer and Diels. Most triangle with the right angle :ir the top lcftl extending to the left and not to rhe right un
editors prefer auget, rightly, I believe; in any case there is no doubt that this was the read- dcr the e. - All 1hc same, Lachm:inn's reading remains highly probable given the form of
ing of the archetype. the erasure, and it has been accepted by all subsequent editors.
ISl. Arl'ENDIX n

jectural interventions in both branches of the tradition would have to take

T Determining the Script of Lost Manuscripts

what I wrote already in r960 (Timpanaro 1960: 6:z. and n. 1). Since then,
recourse to a more complicated stemma: a pre-archetype in ancient minus there have been great steps forward in this field. On the one hand, scholars
cutes (which would have caused the minuscule corruptions shared by the have studied in greater depth the paleographical and cultural aspects of the
whole tradition), followed by another pre3rchetype in capitals (which would "common script" in antiquity and of its various later (or sometimes contem
have caused the capital corruptions shared by the whole tradition) and by porary) i:;anonizations. 17 On the other hand, especially Franz Brunholzl's ex-
the archetype likewise in capitals ("Lachmann's archetype," which would amination of the minuscule corruptions already present in the most ancient
have caused the capital corruptions peculiar to each of the two branches);u surviving manuscripts in capitals (or in the manuscripts considered to be di
an intermediate apograph in pre-Caroline minuscule (which would have rcct descendants of such manuscripts which have been lost), performed in
caused the minuscule corruptions peculiar to each of the two branches) be- conjunction with the examination of the most ancient Latin papyri, has led
tween the archetype and the Oblongus, and also between the archetype and to the conclusion that the most ancient literary manusi:;ripts of the Republi-

the model of the Quadratus and the 5':h~da e: can and Imperial ages did not resemble the Palatinus or Mediceus manu
scripts of Virgil (Vatican, Pal. lat. 163 1; Florence, Laur. 39. 1 + Vatican lat.
w J :z.:z.5, fol. 76) but were volumina of semicursive script, minuscule or rich in
I minuscule aspects.' 18

x Even if these recent studies seem capable of confirming the stemma I have
proposed now,;,.> all in all the first hypothesis (an archecype in minuscules)
I continues to seem prefcrablehb to me, now that Kenney has overcome what
"' had seemed to me the only serious obstacle, as I indicated above. Of course

~ T
this does not in the least exclude the possibility that the corruptions due
to ancient minuscule might in part go back to copies much older than the
archetype." 1,
O blongus
into hy accident (orhers can no doubt be found): Cicero, de rep. l..39, where the fi rst hand
of the Vatican palimpsest has certamine for ce11t11riac. The phrase wKursiver-Kapitalis "
Quadratus Schedae
!cursive capirall, used by Brunholzl prec isely with rega rd to certain corruptions of the
manuscriprs of Lucreu us ( 197 1: 16 -17, 21 ), docs not seem appropriate. Brunho lzl ( 1971:
The phase in "ancient minuscule" (111)'"' would not constitute any diffi- 16n1) cites the first edition of this book of mme and even the earlier version (Timpa naro
culcy at all. One need only think of the studies of Jean Mallon and Robert 1959, 1960), but he docs not seem IO have read them; and alrhough he cites interesting
Marichal and their followers, who have demonstrated the early diffusion of examples of co rruptions in Virgil {197 1: l.J - :?.4), he docs not refer to llibbcck, who at an
this kind of script, and also of the fact that the Virgil manuscripts in capi early time made a precious collection of material, even if it 1s in need of classification, and
rals of the fourth and fifth centuries present many corruptions that can be who ha<l certain ly understood that many errors of the o ldesr extant manuscripts of Virgil
16 presupposed a script earlier than boo k haod capitals.
explained only in terms of misunderstanding of minuscule letters. This is
17. I do not intend to burden m y expositmn with lo ng bibliographical hsts, which in
16. Out of Ribbeck's long and nor entirely certain list (1866: l.35-38) I select some any case would be inadequate. I only wish to recall rhe studies of E. Casama ssima because
particularly significant examples, which I have checked with the apparatus of M. Gcy- of the novelty of their o wn com ributions, and m particular Casamassima-Staraz 1977.
monat's edition of Virgil (Geymonar 1971): the first hand of the Mediceus has mm1tc for 18. Brunholzl t 971 (conclustons at 29- 3 1). Sec also the following note. Cf. Cavallo
t 97 5 for the general aspects (script, volumc11 and codex, Christian influences, the rise of
morte (Georg. 3. p8), oriori for Orion (A cm. 4.51. ), am am fo r A1111am (4.634 ), ta for Ill
(11.3 84 ); Fulvio Orsini's Schcdac Vaticanae have {111ct11r for {111cllls {ACll. 3.665); the Pa a new type of luxury manuscripts in Late Antiquity, including pagan ones); sec also the
latinus has fec1111d11t for (ecimdat (Georg. 4.l.91), 11e11tas for vc11111s (Acn. 5.777), cyprism other essays collected in this volume and the fina l bibliography.
for cypn1Pn (u > i + s, Acn. 1.6 u),J miseretd11111s (1. 145 ), 11itu for 11is11 (5437), tar1111s 19. On this point, with regard to Lucretius, Brunhiilzl 1971 : l. 1 is right to com:ct him
fo r Sam11s (7 .738), 11/ta for alta ( 10.197). Ribbcck attributed these errors to cursive ma self wi1h respect to llrunholzl 196 2. In my opinion, Brunholzl's study, which he presents
juscules (he was thinking o f ~Pompeian" script, as was o nly natural at the time), but there in any case only as a first version of a fuller wor k, has not only many virtues bur also three
is no doubt that these arc errors of minuscule. Cf. also the following example, which I ran defects: ( 1I for each text, the exemplification i5 too skimpy; (:.) this skimpiness is wors
ArrENDIX B Determining the Script of Lost M:inuscrip1s

Naturally this whole discussion can be neglected by an cdiror of Lucre viving manuscript and thar in consequence the correct readings that it con-
rius, since the only thing that matters for the practical purposes of consti rains may represent tradition and not conjecture.
tuting the text is to know that there are errors both of capitals and of mi Now, one good proof of nonderivation can be provided by certain
nuscules in the tradition of Lucretius. It can be neglected; but it should not graphic substitutions. If, for example, a manuscript that is suspected of be-
be explained in a confused and erroneous manner, as almost all the editors ing derived from another surviving manuscript in Caroline script presents a
have done hitherto.2 certain number of rnrruptions typical of pre-Caroline minuscule, then we
can conclude with certainty that that suspicion was groundless and that this
manuscript must not be eliminated.
Pasquali warns quite correctly against hasty eliminations based on
Finally, there is one field of textual criticism in which the examination "graphic signs that lend themselves to being confused with one another"
of graphic corruptions can render a useful service: eliminatio codicum and observes, "When someone maintains thar A is derived from B because
descriptorwn. A has misread a case ending in B which was expressed by an abbreviation
As is well known, this operation is particularly difficult when a manu 'by suspension' that can be resolved in more ways than one, or by a letter
script presents not only a large number of corruptions in common with an drawn so poorly that it can be confused with another one, then it is legiti
other older manuscript, but also a number of certainly or probably correct mate to observe that any manuscript of the same scriptorium, or even of
readings in passages in which the older manuscript is corrupt. This raises the more or less the same period as B, used or could have used the same abbre
following problem: can rhese readings be the fruit of conjecture, or are they viation, which could easily be confused wirh anorher one every time it was
such that they could not have been excogitated conjecturally in any way not written impeccably" (1952a [1934): 35). But although these corrup
whatsoever? In the latter case, the more recent manuscript is a brother of the tions due to misunderstanding of graphic signs count little or nothing for the
older one, not its son, and therefore must not be eliminated; in the former purposes of the elimination of a manuscript, they can be decisive for the pur
case (which unfortunately occurs quite frequenrly), the more recent manu poses of its nonelimination. Whereas five or six certain confusions between
script can be a copy of the older one, bur there is no certainty that this is s and r do not at all serve to demonsrrate that a manuscript A was copied
how matters really stand. from an extanrJJ manuscript B in insular script, inasmuch as any other in
Hence iris a great piece of luck to be able co find positive proofs that one sular model would have lent itself to analogous misunderstandings, they
manuscript is derived from another, consisting in material damage to the serve perfectly well to demonstrate instead that A was not derived from a
older manuscript (displacement of leaves or fascicles, "windows," etc.) to manuscript C in Caroline script, even if by chance the two manuscripts pre-
21 sent many coincidences in corruption.
which transpositions or lacunas in the more recent manuscriptcorrcspond.
But it is just as useful to find proofs of no11derivatio11, which let us assert Hence Maas's formulation, "If a witness, J, exhibits all the errors of an
with certainty that rhe more recent manuscript is not a copy of another sur- other surviving witness, F, and in addition at least one error of its own ('pe
culiar error'), then J must be assumed to derive from F" (1958 f1927]: p. 4,
sec. 8 [a]), is not only open to the more general objections that can be raised
ened by rhe facr thnt the nuthor is too sure that he cnn explain ns being due m gr3phic con-
fusions errors rh:it could be psychologic3l; (3) when hue 3ncienr m3nuscriprs m ca pimls against such a rule;ll what is more, one must in any case specify: "and in
:ire nor preserved bur :ire reconstructed from C:iroline copies, one muse never forger the
:iltcrn:itive hypothesis of errors of prcC:iroline minuscule: cf. Timpnn:iro 1970: 188. In l.l.. Wh:it is more, this rule is presented by Mn:is himself :is :i deduction from :i pre

the c:ise of Lucretius this hypothesis becomes a certa inty when wh3t is involved arc the er- supposition that need nor correspond to re:iliry: cf. M:i:is 1958 ( 19 27): pp. 8- 9, sec. 11;
rors of the common model of the Qu:idrams and the Schedne, cf. above, n. 13. and P:isquali 1952.3 (1934): 3on3 (bur Pasqu:ili d ispl:iys :i cert:iin rcluct:ince to discuss d i
10. In his edition of Lucretius, K. Muller provides :i far better stcmma than those of reedy with M:i3s nnd minimizes rhe difference between M:i:is's position :ind his own :in31
enrlier editions, even if ir too is not free of Mdoubtful p:iss:iges" (upon which I sh311 not in- ysis of the problem of clm1i11atio dcscriptomm). In 3ny C3se, :i m:inuscript th3t h3d the
sist, since my doubts c:in e:isily be inferred from what I h3ve s3id unul now): K. Muller ch:ir:ictemtics of], even if it h3d not been copied from F, would h3vc to be elimi001red 3ny-
197 5: 197-300. l 3m disregarding, of course, the problem of the M lmlic i,~ to which I have way because it would be complerely useless for constituting the text. The interesting (3nd
:ilre3dy referred (:ibove, pp. 108-9, 111-12.). difficult) case is not this one, which M:ins considers Mtypic:il," bur :inothcr one, in which
11. See :ibove, pp. 47(., 99- 101; and the m:igisreri31 chap. 3 of P:isquali 195101 ] presents good readings th3t :ire nor found in F :ind th:ir can (but need not) be rhe fruit of
(1934): 15-40. conjecture; nnd in this case M:i:is :ilwnys inclines 1oward elimination, not in his theorc:ti

addition at least one error of its own that does not presuppose a script (or A PPE N D IX C
even a context} different from that of F."
In fact, Pasquali already indicated a specification of this sort when he Bipartite Stemmas and Disturbances
wrote, "When, after collaring a more recent manuscript in its entirety with of the Manuscript Tradition
an older one, one has found no probative indications of dependence, but
neither has one discovered better readings or individual divergences, and
not even corruptions that ca11not derive from the older manuscript but ra ise
the suspicion of a tradition that may be extremely disfigured but is still dif-
ferent- in that case, and in that case alone, one can be satisfied with the
'presumption' that the more recent MS is a copy of the older one" (1951a
[1934]: 35, our emphasis}. All the same, as is dear from the words "but
raise the suspicion of a tradition that may be extremely disfigured but is still
different," Pasquali was not thinking so much of graphic errors that can be In the last fifty years (though often with long intervals, as is only natural),"
explained by one determinate script and not by another, but rather of larger the scholarly discussion regarding the extraordinary frequency of bipartite
corruptions such as presuppose at their origin a- different reading from that stemmas has given rise to the most varied positions and to results of con-
of the presumed model. And th is too is a possible case; but I believe that the siderable methodological interest. Until about twenty years ago,J Classical
examination of purely graphic corruptions as well will be able to rescue philologists often displayed no knowledge of the contributions of their col-
from elimination manuscripts that have been unjustly suspected hitherto, if leagues in Romance languages (sometimes the opposite occurred too, but less
care is taken to collect a considerable number of certain examples. often); more recently the field of Greek and Latin studies too has witnessed
a reawakening of interest in this problem, which is less marginal and less
strictly technical than might be supposed. The first version of this appendix,
cal p ronouncements, but in his scanty exemplification: cf. 195 8 (19:z.7): pp. i 6 -:1.7 , sec.17;
in the 1963 edition of the present volume, may perhaps have made some con-
cf. Maas 1960: 3:z.. Bur I intend to return elsewhere to the problem of the dcscripti.
tribution to this state of affairs. The second version, which I present here,
does not differ substantially from the first one, bur without aspiring to what
would necessarily be a confusing bibliographic completeness it does take ac-
count of those works that have appeared since then, some of which are very
important; and it aims to draw attention more decisively to what, in my
opinion, has always been the central point of the whole controversy.
In his brilliant and paradoxical article on the manuscript tradition of
the Lai de l'Ombre, which we have already had occasion to cite more than
once,2 Joseph Bedier observed that in the overwhelming majority of cases
the manuscript stemmas traced out by the editors of medieval texts had only
two branches: even if the extant manuscripts were very numerous, they were
almost always made to derive not directly from the archetype, but through
two and only two subarchetypes; and even when preparatory works on the

1 . Besides the studies I shall cite in the course of the following pages, see the copi

ous bibliographical indications in Frank 19 s5: 463n 1 (brought to my attention by Alfredo

Srussi). Cf. also Balduino 1979: z.3 7-41.
2.. Sec above, pp. 44, 8oni3 . Bedicr had already presented the same ideas in his pref
ace to Renart 1913: xxv-xli, hut Jess fully and with less polemical verve.
Ar r ENOIX c Bipartite Stemmas and Disturbances of the Manuscript Tradition

manuscripts of a given text had hypothesized the existence of three or more tended it far enough." And so, if a first phase of his research has led him to
families, the scholar who finally set about to do the critical edition had re- establish the existence of three families of manuscripts, that "anxiety" in-
duced the number of families to two. duces him to look for readings that unite two families against the third one,
So, Bedier observed, it would appear that by a very singular chance al- to convince himself that such readings are erroneous and hence to make
most all archetypes had possessed a direct progeny consisting of only two these conjunctive errors go back to a shared subarchetype, from which the
apographs, or, at least, that it was from only two apographs that all the ex- ancestors of the two families would have descended (instead of d irectly from
tant copies had been derived. "The flora of philology knows only trees of a the archetype). "It is not with impunity," Bedier concludes with ironic em-
single kind: the trunk always d ivides into two dominant branches, and only phasis, "that he has accustomed himself to oppose the good reading to the
into two.. .. A bifid tree is not at all strange, but a clump of bifid trees, a bad one or ones, the rays of light ro the darkness, Ormazd to Ahriman: once
grove, a forest? Silva portentosa [a wondrous forest]." Bedier observed an the dichotomic force has been aroused, it continues to act to the very end"
analogous frequency of bipartite stemmas in the manuscript traditio ns of (Bedier 192.8: 176).
Classical texts as well, though he admitted that his research had been less From rhis denunciation of Lachmannism, Bedier derived an exhortation
extensive and profound in this field than for French medieval texts (19:.8 : to abandon any attempt at recensio and to adhere instead to a single man-
171-Tt). uscript. The illogicality of this exhortation has already been demonstrated
Bedier identified the cause of this strange phenomenon not in the objec- too clearly for it still to be necessary for us to linger on this subject.J What
tive conditions by which medieval manuscripts were produced but in the interests. us here, and what constitutes the only really interesting and acute
philologists' unconscious desire to maintain their freedom of choice when part of Bedier's article, is the question of the frequency of bipartite sremmas.
choosing variants. In fact, if the manuscript tradition has three or more On this point, two very different answers were given to Bedier, by Gior
branches, the reading of the archetype can almost always be established by gio Pasquali on the one hand, and by Paul Maas and various Romance phi
a mechanical procedure, as long as the eventuality of contamination is ex- lologistsi on the other. Pasquali, at least at first, denied for Latin and Greek
cluded (see above, p. 109); but if ic has only two branches, the mechanical texts the datum from which Bedier had started out, namely, the extraordi-
method 1 can only serve to eliminate chose innovations that have been pro- nary rarity of stemmas with more than two 1 branches: " I would like to ask
duced in descendants of individual subarchetypes," whereas the decision Bedier to extend his inqu iry to Classical texts; there he would find umpteen
must be entrusted to internal criteria whenever the one subarchetype's read- three-, four-, five-branched stem mas." 4
ing is opposed to the other's. Was this answer correct? To a certain extent it was, but to an entirely in-
Thus, according to Bedier, Lachmann's method, which had been elabo- adequate extent. There is no doubt that Bcdier exaggerated the extreme rar-
rated precisely so as to expel subjective judgment from textual criticism, had ity of stemmas with mo re th an two branches when he performed his exam-
been applied by philologists in such a way as to preserve the widest pos- ination of the editions of medieval French texts: Arrigo Castellani redid that
sible field of application for subjective judgment. If Bedier had noticed that examination with great accuracy and precision and reached less radical con-
Lachmann himself had unconsciously slipped from a tripartite classification
of the manuscripts to a bipartite one in his preface to Lucretius (see above, 3. Sec the works cited above, chap. 3, n. 23. And setting aside, as always, the case in
pp. 108-1 r) and that Madvig had consciously changed his mind in an anal- which each manuscript represents an independent Mrcdaction," it should be noted that it
ogous way with regard to a gro up of Cicero's orations {see chap. 5, n. 19),h is not 3 1 all true that the Mlesser evil" is to follow a single manuscript when no stemma can
he would have seen in these facts the best possible confirmation of his thesis! be reconstructed. In these cases the lesser evil is to choo se the variants ;iccording to inter
But following a suggestion of Mario Roques, Bedier also indicated an- nal criteria, without 3b3ndoning the attempt to provide a complete evaluation of the greater
or lesser tendency of each manuscript's copyist to reproduce the model fa ithfully even
other cause, this one subjective too, for the philologists' tendency to bipar-
where it is corrupt or on the other hand to "parch it up," ro " prettify," to fals ify. It is sense
titism, besides their unconfessed desire to preserve freedom of choice: their
less to reject such a procedu re as Mcclcctic." Every time more th;m one copyist transcribes
habit of always seeking new connections between groups of manuscripts, a model, "eclecticism" is o bjcttively created, inasmuch as they make different mistakes in
and thus of ascending to more '1 nd more encompassing; groupings, until different parts of the text, with rare exceptions. To this random and ir rational eclecticism
they have reduced the fundamental rcgroupings to only two. According to we must oppose our choice,. which is based on rational argument and therefore is not
Bed ier, the Lachmannian textual critic feels " the persistent anxiety that, eclectic in the pejorative sense.
however far he has extended the criticism of variants, he has still not ex- 4. Pa~qu:1 li 1932a: 130-p .
Ar r ENOIX c Bipartite Stemmas and Disturbances of the Manuscript Tradition

manuscripts of a given text had hypothesized the existence of three or more tended it far enough." And so, if a first phase of his research has led him to
families, the scholar who finally set about to do the critical edition had re- establish the existence of three families of manuscripts, that "anxiety" in-
duced the number of families to two. duces him to look for readings that unite two families against the third one,
So, Bedier observed, it would appear that by a very singular chance al- to convince himself that such readings are erroneous and hence to make
most all archetypes had possessed a direct progeny consisting of only two these conjunctive errors go back to a shared subarchetype, from which the
apographs, or, at least, that it was from only two apographs that all the ex- ancestors of the two families would have descended (instead of d irectly from
tant copies had been derived. "The flora of philology knows only trees of a the archetype). "It is not with impunity," Bedier concludes with ironic em-
single kind: the trunk always d ivides into two dominant branches, and only phasis, "that he has accustomed himself to oppose the good reading to the
into two.. .. A bifid tree is not at all strange, but a clump of bifid trees, a bad one or ones, the rays of light ro the darkness, Ormazd to Ahriman: once
grove, a forest? Silva portentosa [a wondrous forest]." Bedier observed an the dichotomic force has been aroused, it continues to act to the very end"
analogous frequency of bipartite stemmas in the manuscript traditio ns of (Bedier 192.8: 176).
Classical texts as well, though he admitted that his research had been less From rhis denunciation of Lachmannism, Bedier derived an exhortation
extensive and profound in this field than for French medieval texts (19:.8 : to abandon any attempt at recensio and to adhere instead to a single man-
171-Tt). uscript. The illogicality of this exhortation has already been demonstrated
Bedier identified the cause of this strange phenomenon not in the objec- too clearly for it still to be necessary for us to linger on this subject.J What
tive conditions by which medieval manuscripts were produced but in the interests. us here, and what constitutes the only really interesting and acute
philologists' unconscious desire to maintain their freedom of choice when part of Bedier's article, is the question of the frequency of bipartite sremmas.
choosing variants. In fact, if the manuscript tradition has three or more On this point, two very different answers were given to Bedier, by Gior
branches, the reading of the archetype can almost always be established by gio Pasquali on the one hand, and by Paul Maas and various Romance phi
a mechanical procedure, as long as the eventuality of contamination is ex- lologistsi on the other. Pasquali, at least at first, denied for Latin and Greek
cluded (see above, p. 109); but if ic has only two branches, the mechanical texts the datum from which Bedier had started out, namely, the extraordi-
method 1 can only serve to eliminate chose innovations that have been pro- nary rarity of stemmas with more than two 1 branches: " I would like to ask
duced in descendants of individual subarchetypes," whereas the decision Bedier to extend his inqu iry to Classical texts; there he would find umpteen
must be entrusted to internal criteria whenever the one subarchetype's read- three-, four-, five-branched stem mas." 4
ing is opposed to the other's. Was this answer correct? To a certain extent it was, but to an entirely in-
Thus, according to Bedier, Lachmann's method, which had been elabo- adequate extent. There is no doubt that Bcdier exaggerated the extreme rar-
rated precisely so as to expel subjective judgment from textual criticism, had ity of stemmas with mo re th an two branches when he performed his exam-
been applied by philologists in such a way as to preserve the widest pos- ination of the editions of medieval French texts: Arrigo Castellani redid that
sible field of application for subjective judgment. If Bedier had noticed that examination with great accuracy and precision and reached less radical con-
Lachmann himself had unconsciously slipped from a tripartite classification
of the manuscripts to a bipartite one in his preface to Lucretius (see above, 3. Sec the works cited above, chap. 3, n. 23. And setting aside, as always, the case in
pp. 108-1 r) and that Madvig had consciously changed his mind in an anal- which each manuscript represents an independent Mrcdaction," it should be noted that it
ogous way with regard to a gro up of Cicero's orations {see chap. 5, n. 19),h is not 3 1 all true that the Mlesser evil" is to follow a single manuscript when no stemma can
he would have seen in these facts the best possible confirmation of his thesis! be reconstructed. In these cases the lesser evil is to choo se the variants ;iccording to inter
But following a suggestion of Mario Roques, Bedier also indicated an- nal criteria, without 3b3ndoning the attempt to provide a complete evaluation of the greater
or lesser tendency of each manuscript's copyist to reproduce the model fa ithfully even
other cause, this one subjective too, for the philologists' tendency to bipar-
where it is corrupt or on the other hand to "parch it up," ro " prettify," to fals ify. It is sense
titism, besides their unconfessed desire to preserve freedom of choice: their
less to reject such a procedu re as Mcclcctic." Every time more th;m one copyist transcribes
habit of always seeking new connections between groups of manuscripts, a model, "eclecticism" is o bjcttively created, inasmuch as they make different mistakes in
and thus of ascending to more '1 nd more encompassing; groupings, until different parts of the text, with rare exceptions. To this random and ir rational eclecticism
they have reduced the fundamental rcgroupings to only two. According to we must oppose our choice,. which is based on rational argument and therefore is not
Bed ier, the Lachmannian textual critic feels " the persistent anxiety that, eclectic in the pejorative sense.
however far he has extended the criticism of variants, he has still not ex- 4. Pa~qu:1 li 1932a: 130-p .
160 ArrENDIX c Bip;inite Stemm;as and Disturbances of the Manuscript Tradition

clusions. And yet he had to admit that Bedier's accusation remained sub- nation may have been too strict: I still incline to believe that the stemma of
stantially valid. As he wrires, "Bifid trees arc about 7 5-76 percent; they are Calpurnius's eclogues is tripartite, because I do not believe that a subarche-
S:z.-83 percent if the uncertain trees are not included in the total. Even if type can be postulated on the basis of a single sha red corruption, aside from
their predominance is not as overwhelming as in Bedicr's st:uistics, it still re- coincidences in hardly significant innovations: if there really had been such
mains quite remarkable. Four bifid stemmas for every multifid one: that is a a subarchetype, it would have left greater and less ambiguous traces.' But
rario which does indeed seem 'surprising."' As for Greek and Latin texts, in Alberti's arguments are incontrovertible in the vast majority of the eighty
1932 Pasquali was still able to collect a respectable number of stemmas with or so Greek and Latin manuscript traditions that he examines. The path
more than two branches on the basis of the most authoritative editions; he which we saw that Pasquali temporarily chose, to deny the very existence of
himself cited some of them two years later in his Storia de/la tradiziorie (e.g., the fact observed by Bedier, must be abandoned.
1952a [1934]: 149, 195, 270, 303 ). And yet it is not by chance that his Maas tried another path. He fully admitted the overwhelming prevalence
brash phrase about the "umpteen" multipartite stemmas is no longer to be of stemmas with two branches, but sought its explanation" above all in con
found in that book; it is not by chance that Pasquali later returned to at siderations of a statistical nature: "First of all we must remind ourselves that
tacking Bcdier rightly for his demand chat editions should be based on a of the twenty-two types of stemma possible where three witnesses exist, only
single manusc ript but never again breathed a word regarding the question one has three branches [... ]. Furthermore it is in the very nature of the me-
of stemmas with more than two branches. Those very same passages of his dieval tradition that in the case of little-read texts three copies were only
Storia dclla tradizione that we have cited just now, and others which could rarely taken from the same archetype; more ra rely still have all these cop-
also be cited, demonstrate that even when Pasqua li accepts a multipartite ies, or descendants from each of them, survived; on the other hand, where
stemma as a starting point for his discussion about the tradition of an au- texts were much read there is a tendency for contamination to creep in, and
thor, he almost always ends up observing later that the distinction into where contamination exists the science of stcmmatics in the strict sense
"branches" or families is far less clear than is customarily believed, or even breaks down. In the lacer subbranches it would certainly have been easier to
making such branches (it matters little whether they arc two or more in num- presuppose the existence, and survival, of three copies from the same ar-
ber) go back to stages earlier than the archetype, to ancient editions com- chetype; but in these cases the editors were often able, without doing any
pared with which the medieval archetype, even if it is still hypothesized, is
conceived as a "collecting basin" for different traditions, or else as a true
Lachmannian archetype, but one whose descendants have received by col- 6. Cf. Alberti 1979: 67-68, who discusses the trip:irtitc stemma proposed by Casta-
gna 1976: 1 !11- 243. Apart from three other passages, which Alberti too :idmits are not
lation the contribution of other streams of ancient tradition, now lost (on
decisive, the only conjunctive error between two of the three branches traced out hy C;ista-
this possibility see below, p. r83 and n. 51).
gna is altcro for orida :it C;ilpumius Siculus 2.48. The error belongs to the category of con-
Thus Pasquali already admitted implicitly that there were only a few true fusions between words of similar graphical or phonic " tot:il appearance," on which cf.
tripartite stemmas, and later studies have reduced their number rather than T1mpanaro 1976: 64-71, 97, with further bibliographical references; it may already have
increasing it. In Classica l philology the same phenomenon has occurred that been present in the archetype, and the correct re:idmg may be the fruit either of conjecture
Bedier observed in his own field of studies (and chat anyway had already be- or of coll:uion with :inothcr manuscript unknown to us (extra-stemmatic contamination:
gun with Lachmann and Madvig, as has been seen): manuscript traditions cf. below, p. 179). Alberti declares that he is skeptical with regard to the former possibil-
that had first been coruidered tripartite were later reduced to only two ini- ity but he docs not consider the latter one. But I rep~t that the price th:it must be p;a1d in
tial subarchetypes. This has happened, for example, for Plato's fourth tc- order to obtain a bipartite stemma is a $ubarchetype whose copyist committed only one
serious error. On the tradition of Calpumius Siculus sec :ilso below, p. 175. One problem
tralogy and for Macrobius's Saturna/ia. 5 And just now Giovan Battista Al-
I would like to reexamine sometime is th:it of Cicero's Catilmoria11s (Alberti 1979: 61-67).
berti has performed a strict examination that leaves very few stemmas with
On the tradition of 1he lesser l..:itin bucolic pocu , sec now, after C:m;ign;i 1976, Reeve 1978.
more than two branches intact (Alberti 1979). In certain cases this exami- I have not yet had the time to reexamine this problem with the attention it deserves.m
7. As we shall see, such arguments only rarely induce Alberti to construct bipartite
S For Pl:ito sec the introductmn to Carlini 1963; other Platonic tctralogies or indi- stcmmas (cf. below, p. 182); in subst:incc his conclusions coincide with my own regarding
vidu:il dialogues cannot he traced b:ick ro mediev:il :irchetypes, so char the problem that the role pl;ayed by contamination. Alberti is alw;iys or almost alw;iys right in his opposi-
interests us here docs not even arise. For M:icrobius, Willis 1957: 156-57 against a tri- tion to hypothesized sJemm;as with more than two branches which permit a mech:mic:il
partite stemma proposed earlier by A. La Penn:i. recension.
Al'l'ENDIX C B1panite Stcmmas and Disturbances of the Manuscript Tradition

harm, to avoid adducing more than two of these copies in order to recon- other twenty-one is tripartite. Here is an outline of the various species of
struct a hyparchetype of no stemmatic importance." 8 stemmas (the Greek letters indicate lost manuscripts):
Objections against thisr statistical argument have been raised correctly,
if a bit too briefly, by Jean Irigoin, Arrigo Castellani, and Istvan Frank;~ and A A A
more recently, after the first edition of the present study, Alexander Klein-
logcl has expressed even more radical objections, as we shall see shortly.q I
A c
But since Maas (an exceptional philologist, but entirely impervious to other
people's objections, entirely unable to understand from others anything that
he did not understand from himself)' always continued to believe in the va-
c B C
lidity of that absurd argument," and later as well, and 1 even recently has re- I II III
ceived the approval of scholars who are excellent but too. inclined to iurare (6 combinat.) (3 combinat.} (3 combinat.)
in verba magistri [to swear upon their teacher's words],u,lo it will not be use-
less to recum to this question in a rather detailed way, even at the cost of a a a
causing the reader some fatiguc.v
According to Maas (1937: z87-89 = 1958 {1927] 44-47}, with three A A
witnesses there are twenty-two possible stemmatic types. More precisely, A p A c A B C
there are six combinations in which from a first manuscript a second one de-
rives, and from this latter a third one; three combinations in which from one
A c
of the three manuscripts derive the other two; three combinations in which v VI
two manuscripts derive from the third by means of a lost intermediary; three (3 combinat.) (6 combinat.) (1 combinat.)
combinations in which a lost archetype has produced on the one hand one
of the surviving manuscripts and on the other a lost manuscript from which
But to combine all these possibilities under the common label of .. types
in tum the other two surviving ones derive; six combinations in which a lost
of stemma possible where three witnesses exist" means to lump together
archetype has given rise to two surviving manuscripts, from one of which
things that ought to remain quite distinct for the purposes of the calculation
the third surviving one is then derived; and finally one combination in which
of probabilities. For in the stemmas of the first and second species there was
each of the three surviving manuscripts is derived independently from a lost
an original total of three manuscripts, all three preserved; in the stemmas of
archetype. This last combination is the only tripartite one, while none of the
the third, fifth, and sixth species there was an original total of four manu
scripts, of which one has been lost and three are preserved; in the fourth spe-
8. Maas 1937: 289-91 =Maas 1958 (1927): 48-49. cies finally the lost manuscripts are two, that is, the original total was five
9. lrigoin 19 54; Castellani 1980 ( 19 57) (an excellent treatment: Castellani gives a par manuscripts. The only element common to the twenty-two types listed by
t1cularly derailed refutation of j. Fourqucr's arguments, cf. below, p. 165 and n. 13, but his Maas is the fact that the surviving manuscripts are always three. But the
refutation is also valid against Maas, to whom he replies briefly at p. 17on5); Frank 1955:
number of genealogical combinations of three surviving manuscripts out of
46 5 (correct, but a bit too general).
a11 1mdefined m1mber of originally existing ma1111scripts is not twenty-two
10. Roncaglia 1952: 281-82.; Erbse 1959: 97; Hering 1967. Hcring's anicle is an at-
tempt to develop Maas's argument funher by correcting it in some secondary points but
but infinite. Maas's list does not include, for example, stemmas like these:
leaving its substance intact; it is wonh reading for some acute observations on individual
points and for its author's full knowledge of manuscript traditions, but Hering docs not in- a a
validate the basic objections th;u had already been raised against Maas by myself and oth
ers; indeed I would go so for as to say that he has not even understood them, even though
they arc not difficult. Even more rcccndy, in a valuable anthology of writings on philolog- 'Y
A B c
ical method, Bruno Basile selected from Maas precisely that wretched passage on stem-
matic rypes and on the rariry of stcmmas with more than rwo branches (Basile 1975: 59-
/ 1
B c
64)! Cf. Bdloni 1976: 507.
APPENDIX C Bipartite Stemmas and Disturbances of the Manuscript Tradition

and like all the others that can be traced out by multiplying the number of ducing them to the o'11y tripartite stemma uaced out by Maas and attribut-
lost manuscripts at will. ing only an eq11al probability co this stemma, which includes within it those
It will be objected that from the point of view of recensio these infinite others as well.
stemmas can be reduced in every case to one of the twenty-two listed by So if we ask ourselves: "Given three surviving manuscripts out of an un-
Maas. The stemma that the textual critic ends up tracing out on the basis of defined number of originally existing manuscripts, what is the probability
the indications furnished by the shared corruptions is in fact a greatly sim- that those manuscripts belong to a bipartite stemma or to a tripartite one?"
plified one, as has long since been made clear: 11 the method of shared cor- the answer must be that the problem is insoluble. The probability that three
ruptions allows one to establish that two manuscripts AB both descend surviving manuscripts belong to two or to three different branches of the
from a lose manuscript a, but except in rare cases it does not allow one to manuscript tradition is obviously quite different, depending on whether the
identify possible intermediate copies between a and A and between a and B. original total in that tradition was four manuscripts or fifty, and depending
If it were possible to trace out the genealogical tree of all the manuscripts of on whether three different branches of tradition originally developed out of
a given text that really existed (what Fourquet and Castellani call "the real the archetype or ten.
tree" ), then this would almost always turn out to be much richer than the If, on the other hand, we define not only the number of manuscripts that
stemma that we end up reconstructing on the basis of shared corruptions. 12 survive but also the number of the ones that originally existed, then a cal-
For the purposes of recensio, this causes no problems: our simplified stem- culus of the probabilities is indeed possible, but it does not at all indicate
mas function just as well for reconstructing the reading of the archetype as the kind of preponderance of bipartitism over multipartitism imagined by
they would if we were able to trace out the "real stemmas." But when the Maas and, later, Jean Fourquet. Castellani has demonstrated that if the " real
point is to calculate the pro bability chat three manuscripts belong either ro stemma" is rich enough, the chances are virtually equal that three surviving
a bipartite stemma or to a trip3rtite one, then it is not legitimate to neglect manuscripts will belong to a bipartite stemma or to a tripartite one. 13 Cas
the real stemmas as Maas does and simply to assign an eq1tal probability to tellani's polemic is directed against Fourquet's arguments, but, as he himself
each simplified stemma. indicates in a note, the same objections are also valid against Maas.~
Let us consider, for example, these four stemmas: Let us add that ir cannot even be said, strictly speaking, that when Maas
listed those "twenty-two types of stemma possible with three witnesses" of
his he was restricting himself to types that are significant for the purposes of
recensio. In fact, most of these Maasian stemmas present manuscripts that
should be eliminated as descripti: in as many as eighteen cases out of twenty-
two, one or even two of the three manuscripts are copies of a surviving man-
uscript, and hence either recensio docs not take place (in the case of a unique
c manuscript) or else it is reduced to only two manuscripts, and then it is ev-
A B C A B A B C A B c ident that even the possibility of a tripartite stemma cannot be conceived. 11
For the purposes of recensio, the first two have the right to an autonomous
existence, but the third and fourth do not, since even if the existence of an
intermediate member <I> between w and A or between w and B could be dem- t ) Castellani 1980 ( 19 57): 164 70. Castellani 1980 [ 1957]: 17on4 cites another ref

onstrated, this would in any case be irrelevant for reconstructing the arche- utarion of Fourquet by Whitehead-Pickford 1951: I w:is able to sec this thanks to Castel-
l:ini himself, who kindly lent me an offprint. See now also Whitehead-Pickford 1973; but
type's readings. But when we calculate the probabilities, it falsifies every-
this article has too m:iny g;ips to provide a good summ:iry of the whole discussion, nor
thing if we take account of the first and second type, attributing an equal
docs ii seem to me to contain new conrnbutions of any great importance.
probability to each one, and then eliminate the third and fourth ones by re- r4. Castellani 1980 (1957l 17rn5: "Answer to Mr. Maas: the number of b1fid srem
maric rypcs that three manuscriprs c::in form has no importance. What matters is rhe place
11. Cf. Seel r936: 19, and more recemly, Fourquet r946: 4-5. occupied by these three manuscripts in the real tree."
One way of approaching (but only approaching, though srill remaining quite dis
1 :z.. 1 S This objecrion was already raised by lrigoin 1954: :z.u, but he st1ll gave roo much
tant) the complexiry of the "real rrcc" is to investigate the scripts of lost manuscripts of credir to Maas, inasmuch as he maintained thar only the twelve cases in which a surviving
which we would otherwise have no trace, as we have indicated in the preceding Appendix B. manuscript is the source of the other two were ro be elimimited from hts hst. But so too
166 Arl'ENOIX C Bipartite Stemmas and Disturbances of the M:inuscript Tradition

But when Bedicr noticed to his astonishment the great prevalence of bipar- demonstrates that Maas presupposes tacitly and unconsciously, and with
tite stemmas over multipartite ones, he was referring to stemmas traced out out the least experimental basis, that the different stemmatic types arc to
after climinatio codicum descriptorum had already been performed. o that be considered as "equiprobable events" (Kleinlogel 1968: 66-68).17 Quite
while on the one hand Maas excludes from his list the infinite "real stcm- correctly he objects: "In Maas's typology we neither have statistics which es-
mas" to which three surviving manuscripts can belong, on the other hand tablish that each type is found with the same frequency nor have we any rea-
he inflates the list illegitimately by including within it aU the cases in which son to assign them the same probability a priori. The structural differences
one or two of the three manuscripts ought to be eliminated. that provide the only criterion for classifying the types imply nothing about
In reality, the fundamental defect of all these probabilistic arguments is their probability or frequency" (1968: 67). I myself am not capable of fol-
that they start "from the tail" instead of "from the head." People tend to lowing Klcinlogel's argument in all its steps and details (he also availed him-
forget that the real historical process is that a certain number of copies are self of the advice of a mathematician, H. G. Kellcrer; cf. Kleinlogel 1968:
derived from a model, and then a certain number of subcopics arc derived 6301), but the conclusion is dear (1968: 74-75) and agrees with what oth-
from them, and so on. The inverse process, grouping together a certain num- ers and I had afready observed: it is not possible to solve the problem of the
ber of copies so as to form different stcmmatic figures, is purely abstract. great prevalence of bipartite stcmmas by merely deductive means, by an ab-
Hence we should be trying not to see into h_o w many bipartite or tripartite stract calculation of probabilities unsupported by empirical data.w
combinations a given number of manuscripts can be grouped together but to The second of Maas's arguments that we reported above seems to lead us
establish whether it is more probable that only two copies were initially de- onto a genuinely historical and empirical terrain, and hence a much more
rived from an archetype, or three, or more. Once we put the problem in this concrete one:x "it is in the very nature of the medieval tradition that in the
way, we immediately see that it cannot be resolved by means of a mere math- case of little-read texts three copies were only rarely taken from the same ar-
ematical calculation. Whether a manuscript is copied only once, twice, or chetype; more rarely still have all these copies, or descendants from each of
ten times depends on a complex set of cultural and economic conditions: the them, survived; on the other hand where texts were much read there is a ten
number of persons who wish to read that text, the number of copyists who dency for contamination to creep in, and where contamination exists the
arc available for copying it, the cost of the writing materials, and so forth. science of stemmatics in the strict sense breaks down." And yet this argu-
And in the same way, the greater or lesser probability that all the copies of ment too turns out to be fallacious if it is examined with a minimum of at-
that text have been preserved or that they have been destroyed to a greater tention. Maas's reference to "little-read texts" can certainly explain well the
or lesser extent is determined by quite variable historical conditions. 16 lack of stcmmas with ten or twenty branches, for example, but it docs not
The defectiveness of Maas's reasoning is already clear from what has al- at all suffice to explain the enormous divergence in frequency between
ready been said, but Kleinlogel's arguments make it even more evident. He stcmmas with two branches and stemmas with three. We would obviously
expect to find a gradually decreasing frequency of stemmas corresponding
inversely with an equally gradual increase in the number of their branches:
where only rwo manuscripts remain, the problem of a tripartite sremma docs not even
arise: hence six orher combinations musr also be eliminated.
for example, given fifteen different manuscript traditions, five stemmas with
16. According to Greg 1930-31: 401-3, Andrieu 1943: 461, and Ullman 1956: 580, two branches, four with three, three with four, two with five, one with six-
the preponderance of bipartite stemmas is precisely a consequence of the "decimation" un obviously I have artificially regularized this example for the sake of exposi
dergone by most manuscript traditions. But Castellani 1980 (1957): 174 demonstrates tory convenience: in reality the decrease would turn out to be more "capri
that it is only in special cases 1hat the decimation could have increased the number of bi cious." But what we cannot understand is such an abrupt "jump" between
partite stcmmas (somewhat contradictorily, he later assigns greater importance to the dee the number of bipartite stcmmas and the number of tripartite ones. It seems
imanon; 1980 1l9 57 I: l 81 ). Indeed, why must the decimation every time have been pre that Maas and the many scholars who have found this argument of his con-
cisely so destructive as to let the descendants of exactly rwo subarchetypes survive and not
vincing have forgotten that the difference between two and three is only one!
of thn'C? This is already unlikely from the point of view of :m abstract calculation of prob
To listen to them, one would think that the category of "poor traditions" is
abiliry; it becomes even more unlikely (and here Castellani perhaps did not go into the
question deeply enough) if we consider that decimation too is a series of historical events
that depend on accidental causes (fires, etc.) and degrees of wcultural depression" that are 17. Even before Kleinlogel's article, this was pointed our to me privately by my friend~
highly variable and cannot be calculated in the absence of detailed documentation. Giuseppe Torresin and Giampiero Zarri.
168 ArPENDIX C Bipartite Stemmas and Disturbances of the Manuscript Tr;idition

constituted only by traditions with a single branch (for which stemmatic the model copied, then have the model and the first copy copied simultane-
problems do not exist) and by those with two branches; the domain of "rich ously by two different copyists, and so on. By means of this procedure, which
traditions" would suddenly begin starting with three branches. This is not Castellani calls "production maximum," three copies can be obtained in
at all "in the very nature of the medieval tradition," and it is not even in the two units of time, seven copies in three units of time, fifteen copies in four
nature of common sense: just because a tripartite tradition is only ricbcr by units of time. And whereas with the method of successive copies the pro-
a very little bit than a bipartite one, it ought to have occurred fairly often duction of seven copies results in a stemma with seven branches, the produc-
even in the case of little-read texts. 18 Such a "jump., between bipartite stem- tion of fifteen copies results in a stemma with fifteen branches, and so on, in
mas and tripartite ones could only be explained if some regulation or invet- the case of "production maximum" the production of seven copies results
erate habit in the Middle Ages were known that ensured that every ancient in a stemma with three branches, the production of fifteen copies results in
manuscript was copied not more than two times and was then destroyed. a stemma of four branches, and so on:
But obviously we do not have the slightest indication of such a regu lation,
and it is entirely improbable. Among other considerations (but I realize that w
I am discussing absurdities at too great a length), how are we to imagine
such a regulation remaining in effect just as much in the Latin and Romance A
and Germanic West as in the Byzantine East, in different cultural environ-
ments, in very different periods?
As we shall see later, the only thing that is really correct in that passage II B

from Maas is his appeal to contamination. Bur Maas uses this too in a dis-
torted way, exonerating the bipartite traditions from contamination (so that Ill
"stemmatic rigor" would rule in them) and postulating a contamination, be-
ginning with tripartite traditions, so intense as to cancel out any genealogi-
cal relation. Since all traditions (including those with two branches: cf. IV H I J K L M N 0
above, chap. 6, n. 18; and below, pp. 182f.) are more or less contaminated,
and since richer traditions as a rule are more contaminated, one wonders Thus th is type of diffusion leads to a considerable reduction in the initial
why there should not have been numerous intermediate cases of traditions ramification of the stemma, compared with the method of successive copies
not so contaminated as to be irreducible to any stemma but on the other from a single model. But only to a considerable reduction: nm necessarily to
hand not so meager as to be limired ro only two branches. bipartite stemmas, unless other factors (nonproductivity of certain copies,
Setting aside Maas's pseudoexplanations, we can still wonder whether rapid withdrawal or rapid destruction of the initial model, etc.) intervene
the conditions of the transmission of medieval texts were such as co explain that may in fact have operated in many cases but whose o peration has no
this stubborn preponderance of bipartite stemmas. Various hypotheses have character of necessity at all and is in any case independent of the mechanism
been put forward in this regard. For example, Castellani has acutely ob of "production maximum" as such. 1" It also_ remains to be seen how far
served chatr the widespread diffusion of a text and a stemma with many such a mechanism, which concentrates the production of many copies into
branches need not be connected with one another (1980 (1957): 175-82.}. a relatively brief time, corresponds to what we know about the transmission
Anyone who wants to obtain a certain number of copies from a manuscript of Classical texts in the Middle Ages. Ir seems likely that copies of a given
in the shortest time possible should not have the model copied successively Classical text were not produced intensely but were staggered in time, at
as many times as he wants there to be copies of it but rather should first have least in many cases: a manuscript of a work of Cicero, preserved in a mon-
astery, will have been copied once or twice, and the copies will have been
18. As for the improbabihry, asserted by Maas, that "all these copies( ... j survived,"
sec what I h;ive ulrcady observed inn. 16 about "decimation." Once :igain we rcrurn to 19. Perhaps Castellani 1980 (1957): 177-81 insists a bittoo much on these additional
the improb:ible "leap" in frequency between tripartite stemmas and bipartite ones. And it factors. And in any case these do serve to expl3in 3 further reduction in l hc number of
is ridiculous to describe just three copies as "a// these copies." branches but nae their almost const3nt reduction ro only two.
Al'l'ENDIX C Bipartite Stemmas and Disturbances of the Manuscript Tradition 17 1

senr elsewhere; then a third copy will have been made of that same manu- manist age.14 None of this can be imagined in the conditions of slow textual
script years later, and so on. The model of "production maximum" is much transmission of the High Middle Ages,25 when a limited number of copies
better suited to the "publication" of a new work by a medieval author (in seems in general to have been derived from one manuscript, often at inter-
fact Castellani, a Romance philologist, was thinking of cases of this type) vals of time. Avalle's .. embryonic pecia,. seems an artificial expedient, an un-
than to the medieval transmission of an ancient text. 20 But even with these attested qt1id medium [something in the middle] between two essentially
reservations, such a scheme remains valid to a certain extent for Classical different types of production of copies. It must also be added that Avalle
texts too and can at least make some contribution toward explaining the himself seems to have abandoned this hypothesis, which he had originally
rarity of stemmas with many branches. The most difficult fact to explain still proposed with caution.Z+l6
remains the rarity of sremmas with three or four branches-I hope the We shall now propose two other hypotheses that fit better with what is
reader will forgive my insistence, which may seem a bir obsessive but is, I known about the transmission of Classical texts, even if they are only suited
believe, justified. for explaining certain cases of bipartitism.31 In any case, a separate cate-
Another hypothesis has been proposed by D'Arco Silvio Avalle: 21 the bi- gory is formed by those cases in which the bipartitism really consists of the
partitism of so many stemmas would depend on a procedure of "embryonic opposition between a single manuscript (or, more rarely, very few manu-
pecia," by means of which the model was divided into two parts and each scripts) of the High Middle Ages or even of Late Antiquity on the one hand,
of two copyists copied one half and then exchanged their parts; in this way and a considerable number of more recent manuscripts on the other. Cases
two transcriptions could be performed in the same time in which a single of this sort are often cited in Pasquali's work and in the first volume of the
copyist would have performed a single transcription; then the two branches Geschichte der Textiiberlieferzmg [History of text transmission; Hunger er al.
of the stemma originated from these two copies once the model had been 1961-64: vol. 1): it should suffice to recall Aeschylus, Sophocles, Isocrates,
lost or destroyed. I shall not repeat here at length the objections that I have and Theophrastus among Greek authors; Plautus, Terence, Seneca's trage-
expressed more fully elsewherel 2 but shall confine myself to a single point. dies, and Statius's Thebaid and Acl1illeid among Latin onesP It is perfectly
It seems hard to believe (alas, I must return to the leitmotif of all my ob- natural chat the more recent manuscripts should be united by a certain
jections!) that exactly two copies were derived from the model with such number of banalizations and conjecrural "improvements," or else by real
consistency and that each time the model was not transcribed again before corruptions, from which the oldest manuscript is immune: for they go back
being lost. In the real system of pecia, which became widespread in the thir- to an edition .. revised" in the Carolingian period or even larerhb that intro-
teenth and fourteenth centuries for the use of the students of the great medi-
eval universities, each copyist transcribed a fascicle, not a half manuscript; 23
i4. Rizzo 1973: 196 and n. 1, cf. 43. Rizzo too emphasizes the connection between
and this would lead to highly ramified sremmas. An analogous procedure
pccia :and these procedures of the Humanisr age.
of transcription divided up among various copyists was in use in the Hu-
is . Some of the essays collected in Cavallo 1977 provide a clear idea regarding these
conditions: see especially G. Cavallo's inrroduction and the essays by A. Petrucci and
i o. Cf. also Balduino 1979: i38 - 41 on the usefulness of Casrellani's hyporhesis with B. Bischoff. It goes without saying that I ought to cite many other recent and innovative
regard ro the edition of medieval works; perhaps Brambilla Ageno 1975a: 158 is too works, many of them written by the same scholars I have mentioned now; but I cannot
quickly negative. On the foct that it is much less applicable to rhe transmission of ancient leave my subject too for beh ind. In any case, see the full bibliography in Cavallo 1977 and
1exts cf. also Kleinlogcl 1,968: 72 and n. r. In any case it should be borne in mind that for bear in mind that this research is still in full and rapid development.
Castellani the ~production maximum ~ i:;. not the most important cause of biparmism; he i6. He no longer mentions it in Avalle 197i; on p. 9i he speaks of pecia but without
subordinates it to other cause11 that more specifically concern brief medieva I 1exts and a re connecting it to the problem of bipanite stemmas.
rherefore not applicable 10 ancient ones texcept for decimation, on which see 3bove, n. 16; i 7. Cf. Pasquali r95ia (1934 ): :z.5-30, u6, 175-80, i 94, 331- 48, 354 -74; Hunger
Castellani 1980 11957): 181). et al. 1961-64: l.i64-65 , i74-7 5, 37 5- 78, 404-6, 409 (this latter work must be used
l.l . Avalle l ,6u 515-96 =Hunger et al. 1961 -64: i.i97. with caution). See in particular Pasquali's formulatio n at 195ia (1 9 34) 12.6: "Like so
i i . Timpanaro 1965: 394 - 95. many other texts, Seneca's tragedies are transmitted in an ancient manuscript that stands
i3 . Destrez 1935 is still fundamental. Fink-Errera I,96i is more up-to-date and dis apan and in a large number of more recent manuscripts linked together by close kinship
agrees with Oestrez on various points but, at least according to my impression, does so with relations." For lsocrates the existence of an archetype can probably not even be demon-
a certain oslentation of novelty to which real progress sometimes does not correspond. strated, as Dicier lrmcr points out to me, referring to Seek 1965: 106-7.
APPENDIX C Bipartite Stemmas and Disturbances of the Manuscript Tradition 173

duced those changes and from then on constituted a vulgate, bur also, un Dain has drawn attention to another possible reason for stemmatic bi
surprisingly, preserved correct readings that were corrupted in the oldest partition.31 Like any other manuscript, archetypes can undergo more or less
manuscript.<<.28 In this case the bipartition is really caused by a chronologi- serious alterations in the course of time: addition of interlinear or marginal
cal discrepancy-so long as we understand this expression to designate not variants, abrasion of earlier readings, substitution of new ones. If first an
a mere interval of years or centuries but a changed cultural environment, the apograph A has been copied from an archetype, then two other copies Band
arrival of those "medieval Renaissances" that occurred several times in C are copied from that same archetype after it has been greatly altered, the
western and central Europe, and even more in the Byzantine Empire. To this tradition really has three branches but che philologist will be led to attrib-
must be added the fact that the group of more recent manuscripts owes its ute the innovations shared by B and C to an imaginary subarchetype and
homogeneity, at least in many cases, not only to medieval or even Human- hence to trace out a bipartite stemma. jean Irigoin has demonstrated that
ist innovations but also to their shared derivation from an ancient edition this possibility really happened in the descendants of the Vaticanus graecrts
that was different from the one of which a single witness remains.dJ On the 1 of Plato.
Erbse (1959: 97-98) has objected against Irigoin that none of
other hand, given that medieval manuscripts earlier than the ninth century this has any practical importance, since it matters little for the purposes of
are relatively rare and manuscripts going back to Late Antiquity are even recensio whether the agreement BC represents a subarchetype or a "second
rarer, it is not at all strange that for many texts only a single codex vetustior stage" of the archetype. But Dai n's observation, developed further by lrigoin,
[more ancient manuscript] has been preserved, which then ends up consti- was aiming precisely at explaining the surprising frequency of biparcite
tuting by itself a "first family" as opposed to the "second one" represented stemmas; and for this purpose it is much more valid than Maas's mathe-
by the recentiores. 2~ This was more or less how Marouzeau already ex- matical considerations, in which Erbse still unreasonably believes (see above,
plained the prevalence of bipartite stemmas in the Latin tradition; analo- p. 162n10). Moreover, the " alterations" to which an archetype is subject
gous considerations are applicable to Greek texts as well, but they cannot consist not only in corrections but also sometimes in mechanical damage,
be applied to all Latin texts: Marouzeau apparently limited his hypothesis for example, the loss of leaves or the transposition of fascicles during re
too much in one direction and extended it too far in the other.cc.Jll binding.Jl If copy A has been derived from the archetype before the damage
and copies BC after the damage, and if the philologist considers the lacuna
:z.8. An analogous but more complex case is constituted by the tradition of Hippocra- or transposition to be a conjunctive error and therefore traces out a bipar-
tes: see D. lrmer's interesting note lo Timpanaro 1971: 13 m i 86a. tite stemma instead of a tripartite one, this erroneous genealogical recon-
:z.9. In cases of this sort scholars have almost always proposed the hypothesis that the struction can also have adverse consequences for the recc11sio: he will at
reccntiorei derived from the most ancient manuscript; this hypothes is is sometimes cor- tribute the value of "one against one" to the agreement of BC against A
rect but more often erroneous (sec above, pp. 100, 12.4; and chaps. 3 :ind 4 of Pasquali
195:.:i (1934]).
30. I say "apparently" because Marouzeau's thesis, which I have reformulated quite JI Dain 193:z.: 79-80; cf. Dain 197S {1 949): I t :z.. Yet in 1964, the year m which the
freely here, is known 10 us only from a brief reference in Dain 197 5 (1949): 11 3: " He ex second edition of his Ma1111scrits appeared, Dain 1964: 1:r.1 refu5ed to accord any signifi-
plains this bifidiry by the foci that one part of the tradition came from the copy that had cance 10 the problem of bipartite or tripartite stemmas; but the example he adopts reveals
been transliterated around 800 while the other p:irt was derived from preCarolingian that he w:is now thinking of traditions without a medieval archetype (however conceived),
copies antedating the transliter:ition." At the time of the first edition of this work of mine, in which the various manuscripts or groups of manuscripts represent the conunuation of
Dain and Marouzeau were both still alive; but when I asked them, neither one could re- different ancient editions. This is certainly the case in many traditions, especially Greek
member where this explanation had been formufoted for the first time and more fully. "I ones (cf. p. 182), but not in all. So in my view one cannor speak of a "false problem." On
am ninety-two years old, and I do not remember!" Marouzcau wrote to me with some sad- the contrary, as we shall see, the discussion of the rarity of multipartite stemmas makes an
ness. Pasquali too speaks of a "Caroline vulgatc" for Sratius (Pasquali r95l.a I 1934(: T75). important contribution 10 the criticism of Lachmannism.
The formation of such a vulgate need not be connected in every case with the tra nslitera- Jl.. Irigoin 1954: :z.13-14. A "fluid archetype" is also hypo thesized by Reynolds
tion, conceived too simplistically as an operarion performed once and for all. Such a con t96s: 56.
ception has been shown to be fallacious for many Greek texts and is even less certain for 33. We have run into cases of this sort with regard to ancestors that are preserved and
Latin ones, for which indeed there was srrictly speaking no transliteration in the "Byzan therefore allow the dtscripti to be eliminated (pp. 47f., 99; and many other examples
cine" sense. Those vulg:ites were formed gradually by processes of cont:imination :ind con co uld be cited). But entirely analogous damage might have occurred and did in fact occur
vergence caused by shared cultural environments and hence by the intensification of con to archetypes in the Madvii;i11n-Lachmannian sense, that is, to lost manuscripts, beginning
tacts and exchanges (see also above, chap. 8, n. 31 ). wirh Lucretius (sec immediately below).
174 APPENDIX C Bipartite Stc:mmas and Disturbances of the Manuscript Tradition 175

when in reality it is worth "two against one." .l~ We have already drawn at- scholars not to waste their time classifying Humanist manuscripts.36 His ex-
tention to this danger above (chap. 6, n. 13), and we have explained that hortation was largely followed, not only because such manuscripts were dis-
major lacunas or transpositions that can be attributed to the loss or dis trusted but also precisely because they are more contaminated and therefore
placement of leaves or fascicles are not enough by themselves to define a harder to classify than medieval ones. Sometimes the old editors added a
subgroup. third group of mixti alongside the two families of meliores and deteriores;l7
With this hypothesis of the "mobile archetype" we have entered into the but tripartite classifications of this sort were destined to be short-lived, since
field of those manuscript traditions that have a bipartite appeara11ce, al- later editors ended up either adding the mixti to the family of the deteriores
though in reality they were tripartite or had even more than three branches. o r else noticing that almost all manuscripts are "mixed," that is, contami-
But this hypothesis, like Marouzeau's discussed just now, only explains a nated {we shall return to this point shortly) ..i:
limited number of cases. It is not reasonable to suppose that between the Anotherhh cause of erroneous bipartite classifications is partially con-
first transcription and subsequent ones almost all archetypes underwent nected with the preceding one but is more strictly derived from u logical mis-
corrections so numerous and so extensive as to produce an "apparent sub- take: i the tendency to identify one class of manuscripts a on the basis of
archetype." Moreover, it is easy to identify the case of later material dam- shared characteristics and then to call Ii everything that in reality is merely
age to the archetype, and most manuscript traditions are free of it. There " non-a ." There is a danger of falling into a similar error not only in tex-
must be other reasons for this "deception," which produces errors of clas- tual criticism but wherever classifications need to be made: Aristotle already
sification and makes bipartite stemmas seem much more numerous than fought against it in zoology.38 If a certain number of shared corruptions de-
would seem probable. We shall go on now to examine them, beginning with fines a family of manuscripts, the lack of those corruptions does not define
the most banal and avoidable ones and finally arriving at the ones that are another family: so after having identified a fa mily a it wi ll be necessary to
harder to avoid and all in all more frequent.ff see whether the other manuscripts are connected by shared innovations in
In my opinion, the old custom of classifying manuscripts not genealogi- their turn, or whether instead they constitute diffe rent groups, or whether,
cally but axiologically by dividing them into the two categories of mcliores as is also possible, they are so contaminated that their derivation from o ne
and deteriores may have contributed toward increasing the number of bi- or more subarchetypes cannot be detected. An error of th is sort was com-
partite stemmas beyond due measure. We have already seen how the young mitted by Heinrich Schenk! when he divided the V class of the manuscripts
Lachmann still followed this custom, and how some of the first genealogi- of Calpurnius's eclogues into two subclasses v and w. Cesare Giarratano (a
cal classifications (e.g., Madvig's of the manuscripts of De fi11ib11s) origi- Classical philologist who may not have had the gift of genius bur was rig-
nated in earlier axiological classifications and inherited their bipartite struc- orous and scrupulous as few others) observed that w was a real family but
ture. 35 It is likely that scholars often merely transferred into the genealogical v was a heap of manuscripts not united by particular affinities. Schenk) re-
domain the old bipartition based on a judgment of value, and therefore de- plied that he had indeed used the sign v in the sense of " V minus w," but
rived from the archetype two apographs, a "good" one (the work of a stu- even then he refused to recognize the illegitimacy of a stemma traced out ac~
pid and faithful copyist) and a " bad" one (the work of a deceitful interpo- cording to such criteria.3 !> The mistake Mario Casella committed in his at
lator), from which the two races of the meliores and the detcriores would tempt to classify the manuscripts of Dante's Commedia was analogous: the
have originated. The deteriores in particular will often have been considered family a which he believed he could identify " is characterized in exclusively
too hastily to represent a single class from the genealogical viewpoint as
well; it will be remembered that Moriz Haupt, like Lachmann, exhorted
36. Above, chap. 3, n. 15 .
37. So, e.g., Goerenz 1809- 13: vi; N ipperdey 1!147-56: 1.46- 47 (see above, p. 96
34 I do not und erstand the reply of Kleinlogcl 1968: 79n6 o n this point. He objects and n. 18); and Orelli r816-.)8: 4.6.
against me that the rwo more recent apographs have the value of two against one "only 38. Topics 6.6. r43b11 - 13; 0 11 tl1e Parts of Animals r.1-4; cf. Gomperz r9u- 15:
(... ]when the change of condition consisted in a purely mechanical corruption." But this
is the very case I was considering above; and the fact that this is not a purely theoretical 39 See Schenk! 19r 3: :t65, and the correct reply in Gi.irratano r943 : vi-vii. Cf. Cas.
case is demonstrated by the example of Lucretius, discussed in chap. 6, n. l 3. tagna 19 76 on the tradirio n of Calpurnius Siculus 1- 4. 1:t and on some defects of Giarra-
35 Sec above, pp. 77-78, 91-93, 9!1. tano's edition, which do not, however, invalidate his objection to Schenk!.
APl'ENDIX C Bipanite Stcmmas and Disturbances of the Manuscript Tradition 117

negative terms, as not p." u.4o Here Bedier would indeed have been right to apyo, a felicitous emendation of an error of the archetype performed by a
point to Schenk! and Casella as two victims~k of the dichotomic aberration! copyist of the a branch can be transmitted by contamination top, so that yo
Only, such an aberration is not without remedy.11 seem to be connected by a conjunctive error and therefore to derive from a
But the most serious and insidious causes of "apparent biparticism" are subarchetype that in reality never existed.
contamination ("horizontal transmission), the copyists' conjectural activ- Castellani has already drawn attention to these two phenomena,41 but
ity, and, even if to a lesser extent, polygenesis of innovations. For the sake has only assigned them a secondary imponance, wrongly in my view. This
of brevity I use the term disturbances to indicate all three of these phenom- underestimation derives from three causes: ( 1) Castellani believes that the
ena, although I distinguish them whenever it is necessary to do so for my greatest significance should be assigned to other "factors of dichotomy," at
argument. least for medieval lyrical texts (but as we have already seen, p. 170 and
I intend to insist here not on the general difficulties that these phenom- n. 20, such a significance cannot be defended for the transmission of ancient
ena cause the textual critic-I would merely be repeating what is to be found texts during the Middle Ages); (2) he asserts that "from the point of view of
in every good manual of textual criticism, and Pasquali's book can be said textual constitution, a scheme in which two contaminated families form
to be a full critical examination of "disturbed" traditions, each with its own only one functions just as well as the genetically correct scheme in which the
particular problems-but rather on the mechanism by which they lead schol- two families go back to the original" (but this is not true in all cases: com
ars to interpose fictitious subarchetypes between the ancestor and its descen- pare fig. I above, which represents the genetically correct scheme, with
dants and hence to confer a bipartite appearance upon manuscript tradi- fig. 2, which represents the scheme altered by contamination, and it will be
tions that in reality had three or more branches. seen that in the former case a reading shared by py would have the value of
Let there be given a tripartite tradition: from an archetype w let there "two against one" and hence would give us the reading of the archetype but
be independently derived three copies that produced three streams of tra- in the latter case is demoted to "one against one" and hence must be en
dition apy by successive "vertical" transcriptions. If at a certain time a pro- trusted to the editor's i11dici11m); (3 ) Castellani considers it relatively easy
cess of horizontal transmission intervenes, because of which a certain num ("one need only pay attention"; Castellani 1980 [ r957): 182) to identify the
ber of errors of p are transmitted toy or vice versa, or else if a copyist of the corruptions that can be healed conjecturally, so that for these one must deny
a branch corrects a good number of errors of the archetype by felicitous the possibility that the same errors, uncorrected in other manuscripts, might
conjectures, the shared errors of py (which in the former alternative had falsely appear to be conjunctive errors; """ and although he docs not fail
originally been errors of p alone or of y alone and in the latter one were orig- to recognize the deceptiveness of contamination, he recalls the well-known
inally errors shared by the whole tradition) will be attributed to a sub principle that some errors (lacunas, gross errors) cannot be attributed to
archetype, and the tripartite tradition (fig. I) will assume a deceptive bi par horizontLll cr:msmission. We shall return to this point shortly.
tite appearance (fig. 2): More recently, Romance philologists and Germanists among others have
excogitated more subtle "remedies against contamination," for the most

part of a statistical character.~ 2 I have already expressed my position with
regard to such methods: I acknowledge their acuity and, more importantly,
their usefulness for disentangling to some extent, even if only approxi-

p y a p 'Y
mately, the genealogy of those manuscript traditions in which genuine cor-
ruptions or mechanical lacunas arc lacking or arc too infrequent; but I have
fig. 1 fig. 2. also discovered chat such methods are obliged to ignore completely or al-

The two phenomena can also be associated with one another, and this
41. Cas1cll:mi 19!10 (1957): 1111-!l:i.; these phenomena had already been mentioned
will certainly have happened in many cases: given four branches of tradition m passing by 8t!d1er and Fourquet, both cited hy Castellani.
4:1.. C f. above all Segre 1961; Avalle 1961: 159-78 (Appendix 1, "On Some Remedies
40. Quaglio 1965: :i.48; cf. Folcna 1965: 75. Recently W:mink 1979: 84 has once again againsr Contamination~ ; more briefly Avalle 197:z.: 81 ); Okken 1970 (I owe ro D. lrmcr
warned againsr this classificatory error. Cf. also Kenney 1974: 134. my knowledge of this work, which reexamines a 1cxt already edired by Lachmann).
APrENDIX C Bipartite Stemmas and Disturbances of the Manuscript Tradition 179

most completely the distinction between coincidence in conservation (which nas that evidently impair the meaning, without, however, constituting a
in itself proves nothing) and coincidence in innovation;""H and I have al- "copyist's trap"-jumps du meme all meme, similar beginnings or endings
ready indicated the reasons why I think (and others agree) that clinging to of words, letters or abbreviations that can easily be confused in a given
Dom Quentin's method is a bad way of escaping the difficulties of Lach- script, words that can be confused with one another because of phonetic
mann's method (see above, chap. 4, n. 18). vulgarism, etc.-can almost certainly be attributed to vertical transmission
Another, much surer criterion was already enunciated a long time ago, and not to contamination or polygenesis; so too, the copyists' emendatory
was repeated with particular rigor by Ulrich Knoche and Pasquali,44 has activity can be excluded when an error found in the rest of the tradition is
been reaffirmed, as we saw just now, by Castellani and in general by all greater than the conjectural capabilities of the copyists of a given period.
those who have recently written treatises on textual criticism, and is recog- Nonetheless these principles themselvesN are not always easy to apply.
nized as fundamental by Avalle himself: corruptions and, even more, lacu- With regard to the second point, it is often difficult to set precise limits to
the conjectural acumen of medieval copyists and ..editors," especially if they
lived in one of those periods of partial cultural "renaissance" that, as has
43. Timpanaro r965: 398 and n. 14.
been well known for a long time, interrupted the long depression of the
44. See above, p. r 37; and Pasquali l95;z.a ( 19341: xvii: wAlso errors rhar would seem
ro us to be obvious ones ofren penetrate into manuscripts by collation. It is only the lacu-
Middle Ages.45 And all the same this is still the lesser danger: there are cor-
nas that, at least as a rule, are transmitted directly." This principle would seem to be fo r- rect readings at which no medieval copyist philologist (in certain cases not
mulated almost too restrictively. And yet some cases, naturally very rare ones, can appar even the best modern philologist) could arrive conjecturally. A more serious
ently be documented m which even omissions that destroy the meaning have been danger consists in the possibility that a copyist, for example, of the a branch
transmitted by contamm:uion: cf. Willis 197 ;z.: u, who mentions rhc case of a copyist of (see fig. l on p. 176) might have healed errors or filled lacunas not by con-
Macrobius's Sat1m1alia ( 1.6.14)1 who "has performed rhe remarkable feat of inrerpolating jecture and not even by checking one of the other witnesses that have sur-
an omission," deriving rhe omission itself from anorher family of manuscriprs (rhc case is vived to our day, but by collating a manuscript of a completely different
highly likely, even if not quire cerrain, since matri111isq11e preceded by patrimis and fol
branch or tradition which was later lost. In his book Pasquali cites many
lowed by prommtia11tib11s runs a certain risk of being omitted independently by two dif
cases in which one must have recourse to this hypothesis 46 even if he seems
ferent copyists). It would be worth developing further an interesting pomt made by La
Penna 1964: 369: wThe need for a critical apparatus, for a recensio musr have existed in to have forgotten them when he discusses Knoche's criteria (Pasquali 19 52.a
nucc in certain medieval compilers of variants, even before rhe Humanistic ones: some [1934 1: r81-82.). At Timpanaro 1965: 397 I suggested designating th is
times they note for themselves not only varianrs which yield a plausible meaning, as is their phenomenon by the term extra-stemmatic contamination (thut is, con-
custom, but also meaningless variants: to the variants thar serve to correct the text, or that tamination deriving from manuscripts that do not form part of the tradition
produce another acceptable text, and which could be called 'correction variants,' :ire that has survived more or less completely) and to distinguish it from intra
sometimes ;idded others that serve to show how rhe 1ext is transmitted elsewhere, and stemmatic contamination (relations of collation between surviving man
which could be called 'apparatus variants.'" Well, ler us suppose that an "apparatus vari uscripts which form part of the stemma we can reconstruct).4 7 Now, even
ant," noied down by a learned copyist in a manuscript that is no longer extant, has been
extrastemmatic comamination will have the effect of increasing the number
mistaken for a "correction v;iriant" by a later, ignorant copyist who has introduced 11 into
rhe text: the result will be an error of the so" that we are accusto med to attribute with cer-
t;iinry to vertical transmission, and that instead has been transmitted horizontally (there 1s
a case of this kind, even if it is somewhat more complicated, in T1mpanaro 1978: 408 - 45. For Greek texts, see, after Maas 1935 and 1936, also Erhse 1959: 99- 100. For
10). Marginal and interlinear w:ippararus variants" are also found in ancient manuscripts: Latin ones see, e.g., Munari 1970: xxii-xxiii; but Kenney's studies have demonstrarcd that
see the case, cited by Pasqu:ili 195;z.a (1934): ;z.56,of P. Oxyrhynchus 10 17 (a passage from more readings of the rec:entiores represent genuine tradition and should be preferred to the
Plato's P'1aedr11s) furnished with variants wthar cannot be conjectures, because loo often vctustiores than had been thought.
they make the text obviously worse." Papyrologists will be able to cite other examples; it 46. E.g., Pasqu:ili 195;z.:1(1934):48, 49, 304-5, 318-:i6 (but on the tradition of Thu-
is likely rh:it some variants of this sorr have in sinuated themselves into the mcdiev;il tradi cydides sec the progress later achieved by B.moletti, Alberti, and Klemlogcl, and briefly
tion, or into one part of it.'"' On the possibility of the horizontal tr;insmission of evident summarized in Alberti 1979: 10-u), 381-85, and elsewhere. Cf. also Andrieu 1943:
errors, sec also Pasquali 1951a (1934): 87. An extremely interesting example is furnished 468; Di Benedetto t96S (see the index under wcomaminazione" ).
by the readings of the second hand in 1he Ambrosian Vatkan palimpsest of Fronto (Mi 47. In German, the term Fremdles1111g [foreign readingl, coined by Frankel 1964: 78
lan, Ambros. E.147 sup.+ Vatican lat. 5750 ): cf. Zetiel 1980 (pp. 49- 57 on Fronto: a and n. 2., can be useful for indicating " a ~admg rhat derives from a line of tradition oth
precious collection of material, not always well interpreted).~ crwise unknown to us.n
180 AP P ENDIX C Bipanitc Stcmmas and Disturbances of the Manuscript Tradition 181

of apparent bipartite scemmas: " since the damage that goes ba<:k to the ar- fewer ones with B, we can neither affirm nor deny that AC are derived from
chetype has been corrected in a, such damage will be attributed co a sub- a subarchetype.11
archecype, and the stemma will once again take on the false appearance of These are the cases in which the philologist who wishes to reduce the dis-
figure l. on page 176; and what is worse, if the error remaining in ~'Y is of turbances to a minimum and to present in a stemma as systematic as pos
the sort that <:annot be healed conjecturally, it will all the more have the ap- sible an image of t'1e manuscript tradition will "verticalize" the insignificant
pearance of an unquestionably genuine conjun<:tive error. errors too (or the largest group of them) and will postulate a new subar-
And yet the difficulties we have reminded the reader of, serious as they chetype, which can but need not ever have existed.
are, do not always prevent the construction of a plausible stemma based on This is the criterion of the "most economical hypothesis" theorized by
rigorous criteria, that is, solely on "significant errors," excluding all coinci Avalle (with a certain ostentation of analogies with the physical and math-
dences not only in correct readings but also in banalizations, in "errors with ematical sciences) but already broadly followed in practice by Classical and
the semblance of truth" that can 'easily be transmitted by collation, and in Romance philologists.~ Recently lrigoin coo has reaffirmed the importance
errors that can be polygenctic, and taking account also of the possibility of of reconstructing the skeleton of the vertical transmission as far as possible,
conjectural a<:tivtty on the part of the copyists. The possibility of "extra- in an important article that was inspired by the legitimate desire to react
stemmatic contamination" will always remain, alongside less frequent risks against Dawe's exasperatedly and exaggeratedly antistemmatic tendency.~9
like those indicated in note 44; but if this is limited to a few cases it will not In itself, the criterion of the most economi<:al hypothesis is probabilistic
completely alter the fundamental outlines of the stemma; whereas if it is sys- and entirely reasonable. Since, as Pasquali reminded Schwartz (see above,
tematic it will be recognized as another source of tradition that can be in- p. 136), vertical transmission is a constant fact, while disturbances are ex
dicated with figures like this (which have already been used by a number of tremely frequent phenomena that exist in all or almost all manuscript tra-
editors): ditions but are nonetheless always desultory and accidental, every single
coincidence in innovations (even in insignificant innovations) has a greater
chance of deriving from a lost shared model than of being due to contami-
nation or polygenesis or a copyist's corrective activity. But if we prefer the
hypothesis of vercical transmission every time, the total picture of manu-

But it is here that the most insidious dangers begin. Let us suppose that 48. CL Avalle 196 1: 171.1 194 (and alre:idy 195 7~ 64) ~ 1971: S:z.-8 6. In this bst work

a scemma with three branches has been produced on the basis of " significant Avalle (pt=rhaps mswering implicitly my objections in Timpanaro 1965: 397} vigorously
assens that the principle of the " most economical hypothesis ~ should not be followed to
errors" alone. What is to be done now with the coincidences in insignificant
the detriment of stemmatic research performed with strict criteria, and in particular should
errors? We cannot either amibute them with certainty co vertical transmis- not lead to an undue increase in the number of subu cbetypcs. An excellent warning.: and
sion or exclude the possibility that they might be the "residue" of errors al- yet I have the impression that in this way the principle itself is invalidated, and survives
ready present in the archetype and corrected conjecturally in one branch of more as an "epistemological coquetry" than ;is an instrument that can be used for textual
the stemma-this is precisely what their insignificance consists in. But we criticism.
cannot even excl11de the possibility that such coincidences derive from ver- 49. lrigoin 1977; "Without pt=rmitting himself to be led astray by superficial and con-
tical transmission, chat is, chat they go back to a subarchecype. To put chis tradictory links, he [i.e., the textual critic) must try to discover underneath them the con
point better: we can exclude this possibility only if it happens to contradict stanu of the 'vertica l' tradition with all the means ;JVailable to his scholarship and to his
native talent. It is at this price, and at this price alo ne, that he will be able to determine the
a stemma chat has already been constructed on the basis of significant errors
rea lity, and perhaps also the extent, of a horizontal transmission" (113). There is nothing
alone. If, for example, a manuscript A shares lacunas and ocher gross errors
really erroneous in this formulation (but in fact 1here is: there is also the case of "total pre
with B but only a few banalizations or "prettifications" or unimportant al- traditional co ntamination");n but anyone who sets out to do stemmatic work in this spirit
terations with C, we will certainly attribute these latter to disturbances. But will end up vctticalizing everything that can be " erticalized and admitting d isturbances
if A does not visibly share significant errors with B and shares only insignifi- o nly as a last resort. And by multiplying rhe subarchetypes he will naturally obtain a bi-
cant errors with C, or else if it shares more insignificant errors with C and partite stemma.
APPENDIX C Bipanitc Stcmmas and Disturbances of the M;inuscript Tradition

script craditions will certainly be distorted, because we have thereby reduced genealogical relations; in still others, summarizing or developing further the
to zero an event that will have happened in a minoriry of cases but will none- conclusions of earlier scholars, he questions the very hypothesis thot our
theless almost certainly have happened, and in a rather large minority of witnesses are derived from a medieval archetype, a hypothesis that is a pri-
cases. For example, instead of 60 percent of coincidences due to vertical ori extremely improbable for texts that continued to be very widespread
transmission and 40 percent due to disturbances (percentages that obvi- throughout the Middle Ages; and even in those cases in which such an ar-
ously are hypothetical and unfortunately cannot be tested, but all the same cherype must have existed (as the presence of serious damage shared by all
seem plausible), we shall obtain 100 percent of coincidences of the former the extant tradition demonstrates), its descendents must have been enriched
sort-with an evident error. The main reason (not the only one, as we have very early with contributions deriving from extrastemmatic contamination,
seen) for the improbable prevalence of bipartite stemmas consists precisely so that we are very far indeed from a Lachmannian model."
in this paradox: if in every case of coincidence in innovation one prefers the Alberti adds (1979: 94) that an examination of the stemmas that were
more economical hypothesis, the majoriry becomes a totaliry, while the mi- hitherto considered bipartite "would lead one to doubt many stemmas [... )
nority is cancelled out. Naturally the paradox only persists because in tex-
tual criticism, unlike other situations in which a statistical prediction can be
tested empirically post eventum, no similar verification is possible: no one SI As is well known, the problem of the existence or nonexistence of a medieval (or
can tell us whether certain presumed lost subarchetypes really existed or l;ite ancient) ;irchetype is rendered difficult by the contradictory situation we find in m;iny
not.uu,SDAnd if we studied this question in greater depth, we would discover surviving texts: on the one h;ind, the existence of gross errors and above all lacunas, and
that this serious limit to the application of statistical methods is found in sometimes :ilso disloc:itions of whole passages, in all the: ext;int manuscripts, which make:
it hard to exclude the existence of a single lost ancestor that exhibited that d:image; on the
many other diachronic disciplines as well, in which it is not possible to per-
other hand the fact th;it the ;incicnt tradition (papyri, quotations in other authors, etc.) of
form repearable experiments and test their outcome.
ten agrees in inferior readings now with one pan of the medieval tradition, now with an
These considerations of ours are substantially corroborated by the book other. Pasqua Ii 19 5:z.a ( 1934) is full of examples of this sort, which have increased even
of Alberti's I have already mentioned, even if, as I have indicated, Alberti dis- funher in the last decades. Against the hypothesis of the " archetype with variants" (to
sents more than once from my earlier statements regarding stemmas that at which P;isquali still granted too much), cf. esp. Di Bencdcno 1965 : 145 - 46, t49 - 5:z.. Ir
the time I presumed to be multipartite. To a hasty reader (especially to one has been possible to demonstrate in many cases that the errors sh;ired by our whole tra
who does not read closely the book's brief but fundamental last chapter), dition were of a sort judged "tolerable" by ancient and Byzantine philologists (for Euripi
Alberti's treatment might seem to confirm the thesis of "extreme bipartit- des, but with observations that can be applied ro other manuscript tr:iditions as well, sec
ism." Certainly, as we have seen (pp. r6of.), he demonstrates the fragiliry Di Benedetto 1965: 163-93; for Hesiod's Thcogotty, Arrighetti 1961: :i.66-79 has cast
doubt on the existence of an archetype with good arguments). A loose conception of the
or improbabiliry of most of the few multipartite stemmas that scholars of
archetype, not as a unique exemplar but as a "kind of text," as a protomcdiev;il vulgatc
Greek and Latin manuscript trad itions have hitherto hypothesized. But, as
already contaminated by horizontally transmitted errors, is proposed by West 1973: 41-
he himself takes care to emphasize (Alberti 1979: 9 3 ), only very rarely does 42. and Waszink 1979: 79. In my opinion, rhe concept of archetype formulated by Dain
he replace them with bipartite stemmas: instead he speaks for the most part and taken up again by lrigoin (also in lrigoin 1977) has not been very useful and has also
of "probable contamination, which has rendered the relations among the caused terminological confusion. I ;im sorry ro see how little the: difficulties regarding the
various branches uncertain but without eliminating them" (1979: 93); in concept of archetype pointed out by Eugenio Grassi have been taken into consideration,
other cases he concludes that contamination has completely obscured the except for Di Benedetto (cf. above, p. 38). In the face of so complicated a problem, which
in many aspects concerns Romance texts as well, it seems strange how h:ist1ly a schol;ir ;is
learned and competent as Franca Brambilla Ageno ( 197 5b) peremptorily asserts that there
50 . It now seems to me so certain that errors of classification result in cxasi;erating the always was an archetype, relying on arguments of statistical probability that unfonunatcly
number of bipartite stemmas, at least in the overwhelming majority of cases, th:it I con recall the ones Maas used when he tried to demonstrate the numerical prevalence of stem
sider it useless to discuss once again the opposite thesis of Fourquct 1946 :ind then c:spe mas with two branches. Prof. Brambilla Agcno kindly informs me by letter that her ideas
cially Fourquet 1948, who thought that the philologist ran the risk of not noticing "hid regarding the archetype only concern traditions of vernacular texts with specific char;ic-
den subarchetypes ~ and hence of mistaking bipanitc tr.iditions to be multipartite. Anyone teristics: despite the title of her article and some formulations it contains, they arc not in
who still feels the need for a refutation of this thesis can read Timpanaro 1963a: 1;0 and tended to set forth a theory covering all cases. She also informs me that she intends to re
especially my observations in Tiimpattaro 1965: 395-96. turn to this subject with more detailed arguments."
ArrENDIX c Bip;mire Stemmas and Disturbnnces of the Manuscript Tradition

not in the sense of making them become tripartite or multipartite, but rather or more branches, at the end of the passage we cited on page 161, Maas
in the sense of introducing elements of doubt, due above all to contami- wrote, "In the later subbranches it would certainly have been easier to pre-
nation." suppose the existence, and survival, of three copies from the same arche-
So must we abandon altogether the attempt to trace out stemmas (cf. type; Sl but in these cases the editors were ofren able, without doing any
Munari)?ww No, except when the tradition is totally disturbed. Instead, harm, to avoid adducing more than two of these copies in order to recon
above all, we must continue Alberti's work by examining many stemmas struct a hyparchetype of no stemmatic importance."
that are asserted to be certainly bipartite. I do not excludexx the possibility Two objections come to mind upon reading these words. Above all, it is
that such an examination might sometimes correct earlier errors of classifi- curious that Maas envisioned the case of "later subbranches" only in a man-
cation and lead to the construction of multipartite stemmas still sufficiently uscript tradition that presencs more ancient witnesses and ones sufficient for
solid despite some inevitable traces of disturbance; 52 in this regard I would reconstructing the archetype. It did not occur to him that many trad_itions
be a bit more confident than Alberti. All the same I too believe that results of Classical texts (those going back to a manuscript discovered in the Hu
of this sort will be quite rare. I think instead that in more cases those "ele- manistic age, copied or even reproduced in print by scribes or philologists
ments of doubt" to which Alberti refers will induce us to propose two or of that age, and then lost) consist entirely of "later subbranches," which
three stemmas as equally probable or one as just a little more probable than scholars must of necessity not "avoid adducing," at least if they do not wish
the orhersl 3 and in the prolegomena of the editions or in separate works to to give up reading those texts; it did not occur to him to investigate to find
state clearly which subarchetypes can be postulated with almost absolute out whether such traditions too would turn out to be prevalently bipartite
certainty, which ones with a good probability, which ones only with con- or nor.56 Second (and this was observed by La Penna 1964: 374), even in-
siderable uncertainty. Sometimes it will also be necessary to trace out stem- vestigations of recent manuscripts not useful for reconstructing the arche-
mas of only one part of the tradition but co give up on them for another part type are interesting for the question of bipartite stemmas, in order to deter-
that is too disturbed.-H In this way the user of a critical edition will know for mine whether the greater production of copies that began with the srart of
the practical purposes of constituting the text to what point the agreemenc the Humanist age in fact produced an abundance of stemmas with more
of certain manuscripts invests a certain reading with "authority" (an au- than two branches...
thority that is always relative and in need of confirmation).n Now, Alberti draws attention to the fact that, among the few manuscript
Finally, a last problem. In his notorious treatment of stemmas with two traditions that remain multipartite even after his strict examination, almost
all "are represented by rather recent manuscripts"; 5" so too, though he has
not exhaustively studied the recent ramifications of traditions that also pre-
52.. Somcrimcs a "1hird branch" can be obrained by examining a series of corrections sent older manuscripts, he does mention the "lower part" of the stemma of
derived from a losr manuscript: cf. Nardo 1966. In ocher cases (as, for exnmple, for Thu
Aristotle's Metaphysics, in which Bernardinello's research indicates a filia-
cydides; sec above, n. 46} an analogous technique, which for Greek rexts can also take ad
tion of as many as nine branches, and that of Cicero's De legib11s, in which
vantage of papyri, makes ir possible to rcconsrrucr lines of extrastemmatic tradirion, even
if only parrially.
the extremely extensive stemma traced out by P. L. Schmidt presents "ram
53. This cuscom is already followed by some scholars (e.g., La Penna r 9 57: cxlv1), but ifications with eight, nine, eleven branches.'' -18
in my opinion ir should become more widespread:, nor so as to derive from ir rhe purely
skeptical and desrrucrive consequences Bedier was aiming at when he hnd fun rracing out
the ten possible sremmas of the Lai de /'Ombrc (many of which, in fact, were unfounded: S) "Easier," as the reader will recall, b~ause of the ab5cncc of the medieval ~pov
Casrcllani 19!10 (19571: 182.-92. ~ , but in order to distinguish better the various possihih erty " of the ma nuscripr tradirion (cf. above, pp. 1 67- 6 8 ).
rics and 1he various degrees of probabiliry. Ed iiors still ha~e recourse too often ro the rwo 56. Ir is superfluou$ ro add that, in the case of rrndirions containing medieval manu-
extreme solurions: either ro rrace out with excessive confidence a single sremma or co ab scripts, Ma as~ contempr for "a hyparcherype of no stemmatic importance," although
srain from any discussion of the genealogy of the manuscripts. slighrly moderated by his warning rhar edirors can neglecl them "often" (and not "al-
H This kind of solution, in my opinion, can he recommended, for example, for the ways"), is excessive and reveals once again his prejudicial distrust for the rcce11tiorcs (see
E/1hemcris of Dictys Cretensis (L. Septimius), where family y can be reduced ro a fairly above, chap. 5, n. 35).
rigorous sremma, while family Eis too highly disrurbed: cf. Timpanaro 197!1: 397-42.2., 57. Alberti 1979: 94-95 (cf. 57, 9:1., and n. 104).
and now Eisenhur 1973, who partially rakes account of my observarions bur docs not 58. Bcrnardinello 1970: u5; P. L. Schmidt 1974: sremma ar the end of the volume.
abandon rhe acrempr to trace out ::1 complete srcmm11. Sec also Irmer 1972.: sremmas on pp. 116- 17, 12.0.
186 ArrENDIX c Biparcite Stcmmas and Disturbances of the Manuscript Tradition

To me it seems necessary to make a distinction here. Filiations as rich as tious subarcherypes" or made it impossible to disentangle the manuscript
those with eight or eleven branches do indicate that the production of cop- traditions with absolute rigor. .lbb
ies in the Humanist age increased so conspicuously that not even distur-
bances have succeeded in reducing them to those holy two copies, even 1 still believe subst;mtially that my observations in this Appendix Care correct. Bue
though disturbances doubtless became more frequent in chat age, given that it is my duty to indicate that the discussion will continue, and that in particular Michael
Renaissance copyist-editors conjectured more often and more felicitously Weinman will go on in the near future to use statistical arguments (much more sophisti-
and collated more than medieval amanuenses had done.59 So here Maas's cated ones than Maas's, as is only obvious) in order to argue thnt there is nmhing strange
in the preponderance of bipartite stemmas. On this point and on others objections will
reference to "much-read texts" becomes valid-for the very first time. But
also be made og;iinst me by M. D. Reeve, with whom I have had a fruitful epistolary ex
none of the multipartite manuscript traditions, "all recent," that Alberti
chance, and to whom I om indebted for corrections and suggestions, as I have already in-
scrutinized presents more than three or four branches: 60 and in these cases, dicated at the end of the preface (but, at least for now, we are not in agreement on every
as Alberti himself recognizes (1979: 95 and n. 18), Maas's opposition be- thini;). With regard to what I interpret as Maas's failure to understand J>asquali, I had
tween "poor" and "rich" traditions, once more, can have no validity, for the written that Maas "understood nothing of J>asquali's work and perhaps did not even have
reason upon which we have already insisted too much (esp. p. 167). At least the desire and the patience to read it" (chop. 5, n. 3 5). Prof. Reeve informs me that he pos
provisionally I would suggest a different explanation: these are traditions in sesses a copy of Pasquali's book densely annotated by Maas: hence that hypothesis of mine
which there has been a relatively brief interval of time, a relatively small was entirely unfounded. I do not know the character of Maas's annotations: corrections
number of intermediate links, between the archetype and the extant copies, of mistakes? additions of further examples besides the ones J>asquali cited? objections o n
principle? In any case, I hope that Reeve will publish them soon. Naturally the mosc in
so that although ceteris paribus disturbances acted with greater intensity in
teresting ones would be the objections on principle. The fact char Maas did nor make them
the Humanist age than in the Middle Ages, they ended up being a bit less in-
public remains strange. For the present, on the basis of his " Retrospect l 9 56" I remain
tense than in traditions in which the archetype produced descendants for convinced that "Maas understood nothing of Pasquali's work," even though I recognize
centuries and the first copies were already often subject ro exrra-stemmatic that I expressed myself with excessive severity (in any case it should be obvious char with
contamination. But let us not forget chat ro these few examples of tripartite the word 1111derstood I meant, and mean, "knew how to evaluate che criteria of the work
or quadripartite Humanist traditions other traditions can be opposed, bi- and PosquaH's concept of the history of tradition," certainly nae 1mdcrstood in the literal
partite ones (think only of Catullus) and above all highly contaminated ones sensel. I know well that Maas was a greac philologist. His best work, in my view, is to be
(think only of the integri of the De oratore, or of Tibullus). Fundamentally, found in the notes on individual passages of ancienr authors now collected in Maas i973
if we disregard a few exceptions, not even the Humanist age represents a and in the contributions he made, with equal acumen and modesty, to many edicions of
che Bibliotheca Oxonie11sis, of which he was an invaluable proofreader. His conmburion
sudden break with regard to the mulripartition of stemmas. Here too dis-
to Greek metrics is also of i;rcar importance, even if che rigor of his formulations is too of-
turbances, and in the first instance contamination, have either created "ficti-
ten an end in itself and is sometimes more apparent than real. For me, the Maas of the Tex
t11al Criticism is the weakest one. But, as is. obvious, the discussion on this point too is any
59. Naturally, without wishing ac all to case doubt on the seriousness of Bemardi- thing but concluded....
ncllo's work and, even more, of Schmidc's, it would not be a bad idea to sec whether the
"nine" and "eleven branches" really are free of coincidences in innovation among smaller
groups of manuscripcs, due ro contamination or other disturbances. (As Schmidt himsel f
points om ro me, he does speak of contamination in the course of his book, 1974: i.51 ,
343, 359, 386, etc.I
6e>. Four for Corycius's Dialcxcis (or ac leasr for one of them, studied by Peros:i-
Tim panaro 1956), four for C:illim:ichus's Hymns, three for Galen's On DiagnosiJ from
Dreams, rhree for most of Pausanias's Descriptio11 of Greece, three for Scxtus Empiricus ~
011tlincs of Py"bonism, and three for Lihanius's fifty-first oration (but where one of the
manuscripts goes back to the tenth century). Alberti is inclined to hypothesize a multi-
partire stemma (hue apparently one wirh few branches) for Cicero's Brut11s (Alberti 1979t
7 4-7 5 ). In ocher cases, ones that are more uncertain or are limited to brief texts, the stem
mas would always have three or four branches.

Alberti 1979: G. B. Alberti, Problcmi di critica test11alc. Florence, 1979.

Andrieu 1943: J. Andricu, Principes ct rcchcr, hcs c11 critiq11c tcxt11cllc. Paris, 1943.
Andricu 1954: J. Andricu, Cesar: Guerre d'Alexandric, ed. and 1rans. J. Andricu. Paris,
Arrighcni 196 1: G. Arrigheni, "II tcsto dclla Tcogonia di Esiodo," Atlm1ae11m 39 (19611:
Ast 1819: F. Ast, Plato11is opera, vol. 1. Leipzig, 1819.
Avalle 1957: D.S. Avalle, "Una "editio variorum' dc:llc c:inzoni di P. Vidal," in Studta
Ghislcriana, ser. 2.: vol. :z., St11di lcttcrari per ii d11eccntocirrq11a11tesimo amro dalla 11a-
scita di C. Goldo11i, 57-78. Pavia, 1957.
Avallc 1961: D.S. Avallc, La lcttcrat11ra medicvalc i11 ling11a d'oc nella s11a trad1zio11c ma-
11oscritta. Turin, 1961.
Avallc 197:z.: D. S. Avalle, Pri11cipi di critica tcst11ale. Padua, 197:r..
Bailey 194 7: Titi L11crcti Cari De rcrum 1iat11ra libri sex, 3 vols., ed. with prolegomena,
translation, and commentary by C. Bailey. Oxford, 1947.
Baitcr-Sauppc 1839-43: J. G. Baiter-H. Sauppe, Oratores Attici. Zurich, 1839- 43.
Balduino 1979: A. Balduino, Mam1alc di {ilologia italia11a. Florence, 1979.
Bartoli 19:z.5: M. Bartoli, lntrod11iio11e a/la 11eoli11g11istica. Geneva, 19:z.5.
Bartoli 1943: M. Bartoli, "Una nuova riforma della lex Lachmann," St11di /talia11i di Filo-
logia Classica, n. s., :z.o ( 1943): 59-77.
Basile 1975: B. Basile, Lcttcrat11ra e filologia. Florence:, 1975.
B<:dier 19:z.8: j. Bedic:r, "La tradition manuscrite du Lai de /'Ombre," Ro111a11ia 54 (1918):
Belger 1879: C. Belgc:r. M. Ha11pt als akadcmischer Lehrer. Berlin, 1879.
Bc:lloni 1976: G. Bc:lloni, "Rassc:gna di studi e manuali filologici," Lettere ltalia11c :z.8
(1976): 481-514.
Benecke 1816: G. F. Benecke, Der Edel Stein vo11 Bo11eri11s. Berlin, 1816.
Bengd 17:r.5: )ohamri Chrysosto111i De sacerdotio libri sex, ed. J. A. Bengel. Stuttgart, 17 :z. s.
Bengel 1734:J. A. Bengel, Nov1mr Testame11t11m GraeC11m. Tiibingen, 1734.
Bengel 1763a ( 17 34 ): j. A. Bengel, Apparallls critic11s ad Nov11111 Testamc11t1m1. Tu bingen,
1763 (the 1st ed. is included as an appendix in Bengel 1734j.
Bengel 1763b (1741): J. A. Bengel, G110111011 Novi Tcstamenti. Ulm, 1763 (1st ed. Stutt
gart, 174:z.).
Bibliography Bibliography

Bentley 1711: R. Bentley, ed., Q. Horatius Flacc11s, 2. vols. Cambridge, 17 I J. Buchner 19 56: K. Biichner, "Priiludien zu einer Lucrczausgabe," Hermes 84 ( 19 56): 198-
Bentley 1713: R. Bentley, Remarks 11pon a Late Disco11rsc of Free-Thinking. London, :r.33.
1713 (::o Bentley 1836- 38: 3.2.87-368). Buchner 1964: K. Buchner, St11dic11 z11r ronrischcn Literat11r, 2. vols. Wiesbaden, 1964.
Bentley 172.1: R. Bentley, Proposals of Printing a New Ed1tio11 of tbc Greek Testament. Canfor;i 1964: L. Canfor;i, review of S. Timp:inaro, La gmesi de/ mctodo def Lachmam1
London, 172.1 (::o Bentley 1836-38: 3.477- 86). (Florence, 1963), Belfagor 19 ( 1964): 611-15.
Bentley 1836-38: R. Bentley, Works, 3 vols., ed. A. Dyce. London, 1836- 38. Canfora 1968: L Canfora, "Critica textualis in cac:lum rc:voc;ita," Bclfagor 2.3 {1968):
Bernardincllo l 970: S. Bernardinello, Eliminatio codic11111 de/la Meta{isica di Aristotelc. 361-4.
Padua, 1970. C:infora 1981: L. Canfor:i, "Origine della 'stemmatica' di P:iul Maas," Rwista di F1/o/o.
Bernays 1847: j. Bernays, "De emend:nionc Lucretii," Rbei11isches M11se11111 5 (1847): gia e lstr11:tiot1e Classica 110 ( 198:tl: 36:1.-79.
533 87. Caprioli 1969: S. Caprioli, lndagi11i sul Bolognini: Giurispr11de11za e {ilologia 11cl Q11attro-
Bernays 18 52.: Titi L11cretii Cari De rer11m nat11ra, ed. J. Bernays. Leipzig, l 8 52.. cento italiano. Rome, 1969.
Bernays 1855: J. Bernays,].]. Scaliger. Berlin, 1855. Carlini 1963: Plato: Alcibiades I, Alcibiades II, Hipparchus, Antcrastae, ed. with an in-
Bernays 1885: J. Bernays, Gcsammclte Abha11dl1mgen, 2. vols. Berlin, 1885. troduction, translation, and notes by A. Carlini. Turin, 1963.
Bertheau 1908: C. Bertheau, "Wettstein, Johann Jakob, gest. 1754," Rcalcrrcyklopiidie fiir Carlini 1967: A. Carlini, "L'attivita filologica di F. Robortello," Atti Accad. di Udinc,
protestalltischc Tbcologic 1111d Kirche, 3rd ed. ( r908), u.198 - 103. ser. 7, 7 (1967 ): 53-84.
Birt 19r3: T. Birt, Kritik 1111d Hen11e11c11tik. Munich, r913. Carlsson 19:r.6: G. Carlsson, Die Oberliefer11ng dcr Seneca-Tragodicn. Lund, 19:r.6.
Blass 1892. (r886): F. Blass, "Hcrmcneutik und Kritik," in Ha11dbuch der Altert11111s Casamassima 1964: E. Casamassima, "Per una storia dcllc dottrine palcografiche dal-
1vissc11schaft, l.nd ed., ed. I. Miillcr, J.149 - 195. Munich, 189:1. (1st ed. Nordlingen, l'UmanC$imo a jean Mabillon," St11di Medievali, ser. 3, 5 ( 1964): 5i.5-78.
1886). Casam;issima-Staraz 1977: E. Casamassima-E. Smraz, "V:irianti e cambro grafico nclla
Blok 1949: F. F. Blok, N. Heillsi11s i11 Dienst va11 Christi11a va11 Zwcdc11. Delft, 1949 scrittur;i dei papiri latini: Note paleografichc," Scritt11ra e Civilta 1 ( 1977): 9 - 110.
Blume-LachmannRudorff 1848 - 51: F. Blume, K. Lachmann, and A. Rudorff, Die Schrif Cast;igna 1976: L. Castagna, I B11colici /atit1i minori. Florence, 1976.
tender romiscben Fcldmcsscr: vol. 1, Textc und Zeiclm1111ge11; and vol. 2., Erlii11tcn111- Castellani l 980: A. Castellani, Sagg1 di ling11istica italia11a e roma11za, 3 vols. Rome, 1980.
gc11 1111d fodiccs. Berlin, 1848 - 51. Castellani 1980 ( t9 57): A. Castellani, Bedicr avaitil raiso11I La mct/1odc de Lachma1111
Boc:ckh l 886: A. Boeckh, Encyklopiidie 1111d Methodologie der plnlologischen Wissen- dans /cs editio11s de tcxtcs d11 Moycn Age. Fribourg, 1957 ( C:utcll:ini 1980: 3.161
scbaften. Leipzig, 1886. :too).
Bornmann 196:r.: F. Bornmann, "Sui criteri di 'recensio' meccanic;i enunc1ati da Lach Castiglioni I 9 37: L. Castiglioni, review of T. Lucreti Cari, De renm111at11ra libri sex, ed.
mann nel 11117," Rivista dr Lcttcrat11re Modcmc e Comparatc l 5 ( l96:r.): 46 - 53. J. Martin (Leipzig, 1934), Gnomon 13 (1937): 558-6 5.
Bramb11la Agena 1975a: F. Brambilla Ageno, L'cdizio11e critica dci tcsti volgari. Padua, Cavallo 1975: G. Cavallo, "Libra e pubblico alla fine del mondo antico," in L1br1, ed1tori
1975 c pttbblico 11el mondo antico, 81 13 :r.. Bari, 197 s.
Brambilla Agena 1975b: F. Brambilla Ageno, "Ci fu semi;re un archctipo?" Lcttcrc lta Cav:illo 1977: G. Cavallo, Libri e lettori ncl Mcdiocvo. Bari, 1977
liane i.7 (1975): 308-9. Cecchini 1982.: E. Cecchini, "Sulle 'quattro regole' di Lachmtinn," Orphc11s, n. s., 3
Branca 1973: V. Branca, "Mcrcanti e librai fra Italia e Ungheria," in Vc11czia c U11ghcria (198:t): 133-39.
ncl Ritrascime11to, ed. V. Branca, 33 5-sz.. Florence, 1973. Chatelain 1908: E. Chatelain, L11cr. Cod. Voss. Oblo11g11s pbototypicc edit11s. Leiden, 1908.
Bremi 1798: Marci T111/ii Ctccro11is De finibus bo11om111 et malomm, ed. J. H. Bremi. Chatelain 1915: E. Chatelain, "Louis Haver," Rev11c des P.tudcs Lati11cs 3 ( 1915): :t:t-2.4.
Zurich, 1798. Chiari 1914: A. Chiari, "A proposito di una nuov;i edizione di Lucrezio," Rw1$ta di Filo
Brink 1978: C. 0 . Brmk, "Studi classici e critica testu:ile in lnghilterra," A1111ali de/la logia c lstmzionc Clas.srca 52 ( 192.-1): :r.33- 45 ( = Chiari 1961: 1- 2.7).
Sc110/a Nom1alc S11periorc di Pisa, scr. 3, 8, no. 3 {1978): 1071-12.l.8. Chiari 1961 : A. Chi:iri, fodagi11i e lett11re. Florence, 1961.
Br&ker 1885: L. 0 . Br&ker, "Die Methoden Galens in der literarischcn Kritik," Rhci- Cini 1976: G. Cm1, "la pos1z1one degli 'ltalici' nello stemma lucreziano," Atti e Mcmorit:
nisches M11se1111140 (1885): 415-38. de/la Societa Colombaria 41 ( 1976.I: 116- 76.
Bruhn 1887: E. Bruhn, "Lucubrationum Euripidearum capit:i sclecta," jahrbiichcr fiir Citroni 1975: M. Citroni, ed., Martialis: Epigrammaton libcr I. Florence, 197 5.
classische Pbilolog1e, Suppl. 15 (1887): u7-3:r.4 . Clausen 19 s6: A. Pcrsi Flacd ct D. l1mi /11Venalis Saturac, ed. W. V. Clausen. Oxford, 19 56.
Brunholzl 1961: F. Brunholzl, "Zur Oberlieferung des Lukrez," Hcm1cs 90 (196:r.): Clausen 1963: W. Clausen, "Sabinus' MS of Pcrsius," Hermes 91 ( 1963 1: i 5:t- 56.
97-10 4. Caber 184 7: C. G. Cobct, Oratio de artc interprctandi grammaticcs ct critices fimdamcn
Brunholzl 197 1: F. Brunholzl, "Zu den sogenannten codices :irchctypi der romischen Li- tis ilmixa primario pbilologi officio. Leiden, 1847.
tc:ratur," in Fcstschrift B. Bischoff zu sei11em 65. Gcb11rtstag, ed. J. Autenrieth-F. Brun Collins-Schlyter 18:t7: D. H. S. Collins and D. C. j. Schlyter, eds., Corp11.s i11ris S11eo-
holzl, 16- 31 . Stuttgart, 1971. Gothor11m a11tiq11i. Stockholm, 1 B:r.7.
Bibliography Bibliography 193

Conrini 194i.: G. Contini, Un mmo di lctteratura. Florence, r94i.. Ernesti t 761: Callimachi l1yn111os, ep1gra111mata ct fra~:menta cum 110tis varionim, ed.
Contini 1955: G. Contini, ~Ancon1 sulla canzone 'S'eo trovasse pictanza,'n Syculorum J. A. Erncsri. Leiden, 17 6 i.
Gymnasium 8 (1955): rn-38. Emesti 180 I ( l 771): Taciti opera, vol. 1, ed. J. A. Ernesci, rev. J. J. Oberlin. Leipzig, 1801
Corri 1961: M. Coni, wNote sui rapporti fra localizz::m one dei MSS e 'rccensio,"" in Studi (ut ed. Leipzig, 177:1.).
c proble1111 di critica testuale, 8 5- 91.. Bologna, 1961. Ernout 1916: A. Ernout, .. Ala mcmoirc de L. Havcr,n Re1n1e des Et11des Latines 4 (19161:
Courtney 1968: E. Courtney, "The Textual Transmission of the Appendix Vergiliana ,n z.o-::.6.
811/tetin of the lustit11te of Classical St11dies (London) 15 (1968): 133- .p. Ernour 1948 ( t92.0): Lucrcce: De la 11at11rc, 2. vo ls., ed. and trans. A. Ernout. Paris, 1948
Cunius 1858-62.: G. Curtius, Crund:;iige dn gricch1Sche11 Etymologie, 2. vols. Leipzig, ( rst ed. Paris, 1910).
1858- 62.. fabre 1947 (1936): Cesar: La gucrrc crvile, 1 vols., ed. and trans. P. Fabre. Paris, 1947 (tst
Cunius 186::.: G. Curtius, Philologie 1111d Sprachwisse11schaft. Leipzig, 186z.. ed. P:iris, 1936).
Dain 193:1.: A. Dain, WEdition des textes classiques: Theories er methodes," in Assoc. G. Fink-Errer:i 1961: G. Fmk Errer:i, "Une institution du monde medieval: la pcc1a,n Re
Bude, Congres de Nimes 1932, Actes du Congrcs, 61-88. Paris, 1931. v11e Philosophique de Louvai11 60 ( 1962.I: 1!17- :1.10, 2.16-43 ( Cavallo 1977: 131-
Dain 1949: A. Dain, Les mam1scrits. Paris, 1949. 66).
Dain 1964: A. Dain, " Apropos de 1:1 'methode' de Lachmann,'' 811/tetm de l'Associatio11 Flora 1937: T11tte le opere di G. Leopardi, ed. F. Flora: vol. 1, Zibaldone di pensieri. Mi-
Guillaume Bude, ser. 4, no. 1 (1964): ll6-.u. lan, 1937.
Dain 197 5 (1949): A. Dain, Les ma1111scrits, 3rd ed. Paris, 1975 (1st ed. Paris, 1949). Flores 1978: E. Flores, "Ecdotica e tradizione manoscritta lucreziana,'' Vichiana, n. s., 8
Dawe 1964: R. D. Dawe, The Coltatio11 and l1111estigatio11 of MSS of Aeschylus. Cam- (1978): 11 - 37 .
bridge, 19 64. Folena 1961: G. Folcna, " Filologia restualc e stona linguistica,'' in Studi c probltmi di cri
Dawe 1973: R. D. Dawe, Studies 011 the Text of Sophocles. Leiden, 1973. tica tcst11a/e, 17-34. Bologn:i, 196 i.
de Jonge 1978: H. J. de Jonge, "Jeremi:is Hoclzlin, Editor of the Text11s rccept11s printed Folen:i 1965: G. Folena, "Ln tradizione dellc opcrc di D:intc Al1ghieri,n in Attt de/ Co11
by the Elzcviers, Leiden 1633," in Miscclta11ea Ncotcstamc11tica, 1. J05- 18. Leiden, grcsso mtema:.ionale di st11di da11tcschi, r.1 -78. Florence, 1965.
1978. Fourqucr 1946: J. Fourquct, wlc paradoxc de Bc:dier," Mcla11ges 1945, 11: Emdes Litre
de Jonge 1984: H. J. de Jonge, wNovum Tesramcnrum a nohis versum: The Essence of raires (Publications de la Facuhe des Lcttrcs, Strasbourg): 1- 16.
Erasmus' Edition of the New Tesr:imcnt, n jo11mal of Theological Studies, n. s., 3 5 Fourquet 1948: J. Fourquct, wFaures communes ou innovarions communes?" Roma11ia 70
(1984): 395- 413. ( 1948): 115 - 95.
Del testo 1979: Del testo: Semi11ario i11terdisciplillarc s11//a costit11zio11c de/ testo. Naples, Fox 19 54: A. Fox,}. Mill a11d R. Bentley. Oxford , 19 54.
197 9. Fraenkel 1916: E. Fraenkcl, "Anzeige von Lucani Bellum C iv1lc ed. A. E. Housman, Ctr() h

Dcstrez 1935: J. Destrcz, La pecia da11s /cs 111an11scrits rmiversitaires. Paris, 1935. mo11:. ( t9 2.6): 497-531 { =- Fraenkcl 1964, :1..167-30!1).
Devoto 19 58: G. Devoto, Scritti minori, 3 vols. Florence, 19 58. Fraenkel 1950: E. Fraenkcl, Aeschy/115 Aga111e11111011. Oxford, 1950.
Di Benedetto 1965: V. Di Benedetto, La tradizione ma11oscritta c11ritndca. P:idua, 1965. Fraenkcl 1964: E. Fraenkel, Klefoe Beitrtigc, 2. vols. Rome, 1964.
Dicls 1913- :z.4: T. L11creti11s Cam s De Rerum Nat11ra, ed. and trans. H. Diels. Berlin, Frank t 9 55: I. Frank, "De !'art d'editer Jes textcs lyriques,'' in Rcc11eil de trava11x offert a
19:1.3-:1.4. C. Brimel, l.463-75. Memoires et documents publics p:ir la Socictc de !'Ecole des
Dindorf 1870: W. Dindorf, "Ober d ie Vcnetianische Handschrifr des Arheniius und dcrcn Ch:irtes. Paris, 19 55.
Ahschriften," Pl1ilolog11s 30 ( 1870 ): 7 3- 1 t 5. Fr:inkel 19 64: H. Frankel, Ei11/eit1111g t ur kritische11 A11sgabc dcr Argo11a11tika des Apol-
Dindorf 1876: W. Dindorf, Lcxico11 Aeschyle11111. Leipzig, 1876. lo11ios. Giiningen, 1964.
Duvau 1888: L. Duvau, " Lucretiana,n Revue de Plnlolog1e 11 (1888): 30-37. Frere t92.6: A. Frere, "A In memoire de L. Haver," Rcv11e des f.t11dcs Latincs 4 (1916):
Eiscn hur 1973 : Dictys Cretcnsis E.phemeridos bc/11 Tro1a11i libri a L11cio Se/itimio ex :z.0-16.
Greco in Lati1111111 scm1011em translati, ed. W. Eiscnhur. Leipzig, 1973. Froger 1968: J. Froger, La critique des textes ct so11 a11tomatisatio11. Paris, t968.
Ellis 186::.: A. A. Ellis, Be11tlcii critica sacra. C.1mbridge, 1861. Friihncr 1861: W. Friihner. ed., Avia11i fab11lac. Leipzig, 1862..
Ellis 1878 ( I 867): Cat11J/i Verm1C11sis Jiber, z.nd ed., ed. R. Ellis. Oxford, 187 8 (1st ed. Ox Gambcrale 1975: L. G:imberale, "No1e sulla rradizionc di Gellio: In margine :ilia piu
ford, I 867). recente cdizione dellc Noctes Atticae, ~ Rivista di Filologia e lstruzio11c Classica 103
Elmsley 1810: P. Elmslcy, review of C. J. Blomfield, ed., Acsdry/i Prometheus Vi11ct11s (1975): 35-55.
(Cambridge, 181 0), Edinburg/, Review 33 (18 to): 1 n-4:1.. Gcymon:it 197:1.: P. Vergili Maro11is 0/1cra, ed. R. S:ibb:idini :ind A. C:istiglioni, rev.
Erasmus 1538 ( 1500): Erasmus, Adagiorum Chi/1adcs. Basel, 1538 (1st ed. Paris, 1500). M. Gcymonat. Turin, 197:..
Erbse 1959: H. Erbse, review of P. Maas, Tcxtkritik (Leipzig, 1957), G110111011 31 ( 1959): Giarratano 1943: C. Giarratano, ed Calp1m1ii ct Ne111esia11i 811co/ica, :z.nd ed. Turin,
97- 103. 1943 ( 1st ed. N:iples, i9 10).
Bibliography Bibliography 195

Giarratano 1951: C. Giammmo, "La critica dcl testo," in lntrod11zione afla filologia clas Hermann 1841: G. Hermann, "De hymnis Dionysii et Mesomedis." ( Hermann 18:1.7-
sica, 73 - r3:1.. Milan, 195 r. 77: 8.343-5:1.).
GocbelGoebel 1857: A. Goebel.. Goebel, "Die Schedae Vindobonenses und der Codex Hertz 1851; M. Hertz, K. Lachmann: Eine Biographic. Berlin, 1851.
Victorianus des Lucrez," Rheirrisches M11seum u (1857): 449-56. Heyne 1817 (1755): C. G. Heyne, ed., Tilmlli cam1ina. 4th ed., rev. C. F. Wunderlich,
Goerenz 1809-1 J: J. A. Goerenz, ed., Ciceronis phifosophica omnia, 3 vols. Leipzig, vol. l. Leipzig, 18 l 7 (1st ed. Leipzig, 17 H ).
1809 - 13. Hoenigswald 1963: H. Hoenigswald, "On the History of the Comparative Method," A11
Gomperz 1905: T. Gompen, Essays und Eri1merungen. Stuttgart, 1905. thropological Linguistics 5, no. 1 (19631: 1-11.
Gomperz 19u-:1.5: T. Gomperz, Griechische Denker, 4 vols. Berlin and Leipzig, t9:1.:1.-:1.5. Holm 197:1.: G. Holm, Carl Johan Schlytcr and Textual Scholarship. Uppsala, 1971.
Goold 1958: G. P. Goold, "A Lost Manuscript of Lucretius," Acta Classica 1 (19581: Housman 1917 ( 1916): M. Annaci Lucani: De be/lo civi/i, ed. A. E. Housman. Oxford,
:1.1 - 30. 1917 ( tst ed. Oxford, 19:1.6).
Goold 1963: G. P. Goold, "Richard Bentley: A Tercentary Commemoration," Harvard Housman 1931: D. /unii luvenalis, ed. A. E. Housman. Cambridge, 1931.
Studies ill Classical PIJilology 67 (1963): :1.85 - 301. Housman t937 ( 1903): M. Manifii Astro11ornicon, book l, :1.nd ed., ed. A. E. Housman.
Grafton 1975: A. T. Grafton, "J. Scaliger's Edition of C.1tullus ( 1577 ) and the Traditions Cambridge, 1937 (tsted. Cambridge, 1903).
of Textual Criticism in che Renaissance," Jo11rnal of the Warb11rg and Carmauld lnsti Housman 197:1.: The Classical Papers of A. E. Housman, 3 vols., collected and ed.J. Diggle
tutes 38 h 975): 155 - lh. and F. R. D. Goodyear. Cambridge, 197:1..
Grafton 1977a: A. T. Grafton, "On the Scholarship of Politian and Its Context," Journal Humboldt 1903-20: W. Von Humboldt, Gesammeltc Schriftcn, 15 vols., ed. A. Leitz
of the Warburg a11d Courta11ld lnstitllfes 40 (19771: l 50- 88. mann, 8. Gebhardt, S. Kiihler, and E. Spranger. Berlin, 1903-:1.0.
Grafton 1977b: A. T. Grafron, "From Politian ro Pasquali," Journal of Roman Studies 67 Hunger et al. 1961-64: H. Hunger et al., Gcschichte der Textiiberlieferimg, l. vols. Zu
(1977): 171-76. rich, 1961-64.
Grafton 1983: A. T. Grafron, Joseph Scaliger: A Study i11 the History of Classical Schol lrigoin 1954: J. lrigoin, "Srcmmas bifides et etats de manuscrits," Revue de Pl1ilologie 80
arship: vol. l, Textual Criticism a11d Exegesis. Oxford, 1983. (1954): 211-17.
Grassi 1961: E. Grassi, "lnediri di Eugenio Grassi (a cura di V. Bartoletti, F. Bornmann, Irigoin 1977: J. lrigoin, "Quelques reflexions sur le concept d'an:hetype," Revue d"His
M. Manfredi, S. Timpanaro)," Alene e Roma 6 (1961): 119 - 65. toire des Textes 7 {19771: 235-45.
Greg 1930-31: W.W. Greg, "Recent Theories of Textual Criticism," Modem Philology lrmer 1972: D. lrmer, Z11r Gencalogie der iiingere11 De111ostheneshandschrifte11. Ham
28, no. 4 (1930-3t): 401-4. burg, 1971.
Gregory 1900- 1909: C.R. Gregory, Textkritik des Ne11e11 Testame11tes, 3 vols. Leipzig, Jahn 1843: A. Persii Flacci satllrarmn libricum sc/10/iis antiquis, ed. O.Jahn. Leipzig, 1843.
1900 (vol. l), 1901 (:1.), and 1909 (3) Jahn 1849: 0. Jahn, G. Herman11. Leipzig, 1849.
Griesbach 1796 ( 177 4 I: Novum Testamentum Graece, l. vols., :1.nd ed., ed. J. J. Griesbach. Jahn 18 51: A111i Persii Flacci saturar111n libri, ed. 0. Jahn (smaller ed.). Leipzig, t 8 51.
Halle, 1796 (1st ed. Halle, 1774). Jebb 1889: R. C. Jebb, Bentley. London, 1889.
Grimm 1864 -71:). L. Grimm, KlcineSchrifte11, 5 vols. Berlin, 1864-7 1. Jebb 1900 (1886): R. C.Jebb, Sophocles: The Plays: vol. :1., The Oedip11s Colcme11s, 3rd tit.
Hagen 181 6: F. H. van der Hagen, ed., Der Nibe/11nge11 Lied, ind ed. Bresla u, 181 6. Cambridge, 1900 (1:;t ed. Cambridge, 18 !16 ).
Hagen 1879: H. Hagen, Gradus ad criticen. Leipzig, 1879. Jocelyn t984: H . D.J<X:dyn, review of A. Grafton, Joseph Scaliger: A St11dy i11 the History
Haupt 1836: M. Haupt, Quaestiones Cat11/lia11ae. Leipzig, 1836 (= Haupt 1875-76: of Classical Scholarship: vol. l, Text11al Criticism a11d Exegesis (Oxford, 1983), Liver
1.1 -7i.). pool Classical MoP1thly 9 ( 1984): 55 - 61.
Haupt 1875 - 76: M. Haupt, Op11scl#la, 3 vols. Leipzig, 1875-76. Kaibel 1887-90: Athe11aei Na11cratit11e Deip11osophistam111 libri XV, 3 vols., ed. G. Kai
Haupt 191 l ( 1854): M. Haupt, "De Lachmanno cririco (Rede 1854)," Nme Jahrbiicher be!. Leipzig, 1887.
fiir das klassische Altertum i.7 ( 1911 I: 529- 38. Kammel t879: H. j. Kammel, ~J. A. Goerenz," Allgemei11e De11tsche Biographic 9 {1879):
Ha vet 1884: L. Haver, ~Les fautes issues de corrections dans les mss. de Noni us," Mcfan 373.
ges Gra11x, 803- 14. Paris, 1884. Kantorowicz 191r: H. Kanrorowicz, Ei11fiihru11g in die Textkritik. Leipzig, 19:1.1.
Ha vet 19 1 l: L. Haver, Ma1111el de critique 11erbale. Paris, 1911. Kenney 197 4: E. J. Kenney, The Classical Text: Aspects of Editi11g ill the Age of the Printed
Heinsius 1 661: N. Heinsius, ed., P11blii Ovidii Nasonis opera omnia. Amsterdam, 1661. Book. Berkeley, Los Angeles, and London, 1974
Hemmerdinger 1977: B. Hemmerdinger, ~rh1lologues de jadis (Bentley, Wolf, Boeckh, Kenney 1980: E. J. Kenney, ~A Rejoinder," Gionrale lt11/ia110 di Filologia 3:1. (t980Y,
Cobet)," Belfagor 32 (r977): 485 - 5:1.1. 3:1.1-13.
Hering 1967: W. Hering, "Zweispaltige Stemmara: Zur Theorie der textkritischen Me Kirner t901: G. Kirner, "Conuibuto alla critica de! testo delle Epistolae ad Fanriliares di
lhode," Philolog11s 11 l (1967): 170- 85. Cicerone," Studi Jtaliani di Fi/ologia Classica 9 (t901): 369-433.
Hermann 18:1.7-71: G. Hermann, Op11sc11/a, 8 vols. Leipzig, r8:1.7-77. Kb proth 1813tJ. Klaprorh, Asia polyglotta. P::iris, t8:1.3.
Bibliography Bibliography 197

Klcinlogc:I 1968: A. Kleinlogcl, wDas Stemmaproblem," Philolog11s 1 u (19611): 63-S:z.. Lachmann 1876: K. Lachmann, Kleinert Schriftet1, l. vols. Berlin, 1876.
Klotz 1908: P. Papi,,ius Stat11s: Tbebais, ed. A. Klon. Leipzig, 19011. uchmann 1879 ( 18J;): Wolfr4111 von Eschenbach, ed. K. Lachmann. Berlin, 1879 ( ut ed.
Klotz 19:z.3: M. Tullius Cicero: 1" Vc"em actio"is s~cu,,dae libri I-Ill, ed. A. Klorz. Lcip Berlin, r833).
zig. 19:z.3. Lachmann 1892.: K. Laclm14nn's Bricfe an Mori;; Haupt, ed. J. Vahlcn. Berlin, 1892.
Klorz 1950: C. I. Caesar: Commentari1 Belli Civilis hbri, ed. A. Klorz. Leipzig, 1950. Le Clerc 1730 ( r697): J. Le Clerc, Ars critica, 5th ed. Amsterdam, 1730 (1st ed. Amster
Koenc 1833: W. Koene, Leben 11nd Studicn F. A. Wolfs, :z. vols. Essen, 1833. dam, 1697).
Kramer 18H-s:z.: Strabonis Gcograph1ca, 3 vols., ed. with criucal commentary by Lefmann 1891 : S. Lcfmann, fr4nz Bopp, 2 vols. Berlin, 1891.
G. Kramer. Berlin, 18H-5:z.. Lehrs 1882.: K. Lehrs, De Aristarchi studiis Homericis. Leipzig, 188:1..
Kretschmer 1896: P. Kretschmer, Ei,,leit11"g in die GesclJ1chte dcr gricchischeri Sprache. Leo 1893 : F. Leo, Rede wr Saca1larfeier K. Lachman,,s. Gottingen, 1893 (= Leo 1960:
Gottingen, 1896. 2.415-31 ).
Kretschmer 19 :z.7: P. Kretschmer, "Sprachc," in Ei,,/citung in die Altm1m1swissenschaft, Leo 1960: F. Leo, A11sgewiilte Kleine Schri{tcn, :z. vols., ed. E. Fraenkel. Rome, 1960.
3rd ed., ed. A. GcrckeE. Norden, 1.1- 12.1. Leipzig and Berlin, 19:z.7. Leona rd-Smith 194 :z.: Titi Lucrctii Cari De rcrum 11at11ra, ed. W. E. Leon~ rd and S. B. Smith.
Kuster 1710: L. Kiistcr, ed., Novum Testame,,tum Graecum. Rotterdam, 1710. Madison, 1942.
La Penna 1957: A. La Penn;i, Publius 011idi11s Naso Ibis, ed. with a commentary by A. u l..cumann 1977: M. Leumann, Lat. La11t-1md Fonnc11/ehre. Munich, 1977.
Penna. Florence, 1957. Leutsch 18 SI : Corpus Parocmiographormn Graecorum, vol. l., ed. E. L. Leutsch. Gottin-
La Penna 1964: A. La Penna, review of S. Timpanaro, La gcnesi de/ mctodo del Lachmann gen, 11151 (reprint: Hildesheim, 1965).
(Florence, 1963), Critica Storica 3 ( 1964): 369-74. Lindsay 1902.: W. M. Lindsay, The Ancie11t Editiom of Martial. Oxford, 190z..
Lachmann 1816: Sext11s Aurelius Propcrtius: Carmina, ed. wi1h noies by Karl uchmann. Lindsay 1903: M. Valerii Martialis Epigra111111ata, ed. W. M. Lindsay. Oxford, 1903.
Leipzig, 1816. Lobeck 1853: C. A. Lebeck, Patl10fogiae Graca scrmoms elementa, :z. vols. Konigsberg,
uchmann 1817: K. Lachmann, review of F. H. von der Hagen, Der Nibel11ngc" Lied 1853.
(Breslau, 181 6), and of G. F. Benecke, Der Edel Steii1 von Bo11cri11s (Berlin, 1816),]c Ludwich 1885: A. Ludwich, Aristarchs Homerischc Tcxtkritik, vol. 1 . Leipzig, 1885.
11aische Allgcmci11e Litcratur Zeit1mg ( 1817): 114- 42. ( = Lachmann 1876, 1.81- u 4 ). Lutz-Hensel 1971: M. LunHensel, "lachmanns textkritische Wahrschcinlichkcitsre
Lachmann 1818: K. Lachmann, " Ober G. Hermann's Ausgabe von Sophoklcs' Ajax," geln," Zeitschrift fiir Dc11tschc Plnlofogic 90 (1971): 394 - 408.
]cnaische Allgemci11c Litcratur Zeitrmg l.03-4, no. 4 (1818): l.49-63 ( Lachmann Lutz-Hensel 1975: M . Lurz-Hensel, Prmzipicn dcr crsteri tcxtkr1tischet1Editio11cn111ittcl-
1876: 2.1-17). hochdc11tschcr Dichturrg: Briider Grimm, Benecke, Lach111an11. Berlin, 197 5.
Lachmann 182oa: K. Lachmann, review of Der Nibelmtgc" Noth 11nd dcr Nibe/11ngen Maas 1935: P. Maas, " Eustathios als Konjekturalkritiker, 1," Byuwtinische Zeitschrift 35
Lied, ed. F. H. von der Hagen (Breslau, 18'1.o),jenaische Allgcmcine Litcratm Zcit11ng (1935): 299 - 307.
18:1.0, Ergiinzungsbliiucr 70-76 (= Lachmann 1876: 1.206- 77). Maas 1936: P. Maas, "Eustathios als Koniekturalkritiker, II," Dyumtinischc Zc1tschrift 36
Lachmann 18:z.ob: K. Lachmann, A11swahl a11s den l10chdc11tschen Dichtcm. Berlin, 182.0. (1936): l.7- 31.
Lachmann I 8:1.6: K. Lachmann, Ober Vossens Tib11/I und ei11igc a11dcrc Tib11ll-Obcrsct:t1m- Maas 1937: P. Maas, "Leitfehler und stemmatische Typen," Byumtimscbc Zcitscbrift 37
gen, ]e11aischc Allgemcine Literat11r Zcit11ng 63 - 67, no. 2. (I 8:z.6): II 3-s :z. ( ... u ch- (r937): l.89-94.
mann 1876: l..10'1. - 45). Maas 1958 (1927): P. Maas, Textual Criticism, trans. 8. Flower. Oxford, 1958 (=Text
Lachmann I 8 :z.9: K. Lachmann, ed., Albii Tib11/li carmirra. Berlin, 1Ii29. kritik, 3rd ed. Leipiig, 1957; 1st ed. Leipzig, 19'1.7).
Lachmann 1830: K. Lachmann, "Rechenschah uber Lachmanns Ausgabe des Neuen Tcs- Maas 1960: P. Maas, Textkritik, 41h ed. Leipzig, 1960.
tamenis," Theologische S111dic1111nd Kr1tikt11 3, no. :z. (1830): 81 7- 45 ( Lachmann Maas 1973: P. Maas, Kfci11e Scbrifte11. Munich, 1973.
1876: l..l.50-7 2). (Mace) 1729: (D. Macef, ed., The N~v Testament i11 Greek a11d E11glish. London, 1729.
Lachmann 1831 : K. Lachmann, ed., Nov11111 Tc.sta111c11t1m1 Gracce. Berlin, 1831. Madvig 18:z.8: J. N . Madvig, Ad virum celebcrrim11m lo. Casp. Orelli11111 epistola critica
Lachmann 1836: K. Lachmann, "Ober Dissen's T1bull," Halbschc Allgcmcme Literal/Ir de oratio1111m Vcrri11ar11111 /if1ris II extremis e111e11da11dis. Copenhagen, 18:z.8.
Zc1t1111g 109- 10 , no. 2 (1836): l.50- 63 ( Lachmaon 1876: l..r45 - 60). Madvig 183:z.: J. N. Madvig, De aliquot lacimis codic11111 L11cretii. Copenhagen, 1832.
Lachmann 184'1.- 50: K. Lachmann, ed., Novum Tcsta111e11tum Graccc ct Latinc, .,_vols. ( Madvig 1887: l.48-6:z.).
Berlin, 1842. (vol. 1), r850 (2.). Madvig 1833: J. N. Madvig, De e111enda11dis Ciceronis oratio11ibus pro P. Sestio et in
Lachmann r85oa: K. Lachmann, ed., L11cretii de rcrum natur4 fibri VJ. Berlin, 1850. P. Vatinium. Copenhagen, 1833 (= Madvig r887: 333-407).
Lachmann 18 sob: K. Lachmann, lir Lucrctii de TCrllltl 11at11ra libros co111111c11tar111s. Ber- Madvig 1834: J. N. Madvig, Op11sc11la acadcmica. Copenhagen, 1 !134.
lin, 1850. Madvig 184:z.: J. N. Madvig, Op11sc11{a acadcmica altera. Copenhagen, 1841.
Lachmann 18 s1 (18:z.6): K. Lachmann, ed., Der Nibel1mge11 Not mit dcr Klage, 3rd ed. Madvig 1869 ( 1839): M. T11/lii Ciccronis De finibus bo11or11111 ct mafonmt, :z.nd correct!
Berlin, 18 51 <1st ed. Berlin, 182.6). ed ., ed. J. N. Madvig. Copenhagen, 1869 (IS! ed . Copenhagen, 1839).
Bibliography Bibliography 199

Madvig 1871 : ]. N. Madvig, Adversaria critica, 1 vors. Copenhagen, 1871. Moricc:i 1946- 47: L. Annaei Senecae Tragoediae, 1 vols., 2nd ed., ed. U. Moricca. Turin
Madvig 1887: J. N. Madv1g, Op11scula academica ab ipso iterum collecta, emendata, 1946- 47.
a11cta. Copenhagen, 1887. Morpurgo Davies 1975: A. Morpurgo Davies, wLanguage Cl:issification in the Nineteenth
Maher 1966: j . P. Maher, "More on the History of the Comparative Method: The Tradi- Century," in C11rrc11t Trends i11 Linguistics, ed. T. A. Scbcok, 13.607-716. The Haguc:-
tion of Darwinism in August Schleicher's Work," Anthropological Linguistics 8, no. 3 Paris, 1975.
(1966): 1- ll.. G. Miillcr 1958: G. Muller, " Die Problcmatik des Lucrcztextes seit Lachm:mn, I," Philo
Malcovati 1938: L. Annaci Flori q"'1e extant, ed. E. Malcovati. Rome, 1938. log11s 102 (1958): 147- 83.
Miilzer 1970: G. Miilzer,j. A. 8e11gel: Leben u11d Werk. Stuttg.:irt, 1970. G. Millier 1959: G. Muller, "Die Problematik des Lucrcztextes seit Lachmann, II," Philo
Marichal 1961: R. M:mchal, La critique des textes. Paris, 1961. logus 103 (1959); n - 86.
I. Mariotti 1947: I. Mariotti, "Un pa~o di Sallusrio folsamente :ittribulCo a C icerom:," K. Miillcr 1954: Q. Curtius Ruf11s: Geschithtc Alexanders des Grossen, ed. K. Muller and
Studi ltalia11i di F1lologia Classica l.l. j 1947): 157. H. Schonfeld. Munich, 1954.
S. M:11iotti 1947: S. Mariotti, "Varianti d'autore nella tr;idizione dirett;i dell'Eneide?" K. Miillcr 1973: K. Miiller, "De codicum Lucretii lt:ilicorum originc," Museum Hclvcti
Paideia l. ( 1947):. 303. cum 30 ( 1973): 166-78.
S. Mariotti 1950: S. Mariotti, "Ancora su varianti d';iutore," Paideia 5 (1950): :z.6-28. K. Miiller 1975: K. MUiler, ed., Titus L11creti11s Carrts: De rerum 11atura. Zurich, 197 S
S. Mariotti 1951: S. Mariotti, "Rileggendo la storia ddla tradizione," Ate11e e Roma 11 L. Miiller 1869: L, Miiller, Geschichtc der klassischen Philologie Irr dc11 Niederlande11.
( 1951): 112.- 19. Leipzig, 1869.
S. Mariotti 1954: S. Mariotti, "Probabili varianri d'autore in Ennio, Cicerone, Sinesio," Munari 1950: F. Munari, "Codici heinsiani degli Amores," S111di Italiani di filologia
La Paro/a de/ Passato 9 ( 1954): 368-75. Classica, n. s., 14 (1950): 161-65.
S. Mariotti 1966: S. Mariotti, "Qua rationc quave via huius saeculi philolog1 veterurn Munari 1957: F. Munari, "Manoscritti ovidiani di N. Heinsius," Studi ltaliani di Filolo-
opera edenda curaverint," in Acta onmium ge11ti11m ac 1iatiom1111 co1iverrt11s fat. Litte gia Clamca, n. s., 19 (1957): 98-114.
ris li11g11aeq11e fovcndis. Rome, 1966. Munari 1970: P. O vidti Nasonis Amores, ed. with an introduction, tr;inslation, and notes
S. Mariotti 1971: S. Mariotti, "'Codex unicus' e editori sfortunati," St11di Urbmati, ser. B, by F. Mun:iri. Florence, 1970.
45 (1971): 837-40. Munro 1886: T1ti L11<:retii Cari De rerum 11atura libri sex, ed. H. A. J. Munro. Cambridge.
Marshall 1977: P. Marshall, The Ma11uscript Tradition of Cornelius Nepos. London, 1886.
1977. Murgia 1980: C. E. Murgia, "A Problem in the Transmission of Quintilian's Iiistitutio
Martin 1969 ( 1934): T. L1tcretii Cari De ren1111 11atura libri sex, 6th ed., ed. j. Manin. Oratoria," Classical Philology 75 (1980): 3 u-10.
Leipzig, 1969 ( t st ed. Leipzig, 1934). Mynors 1964: Parregyrici Lati11i, ed. R. A. B. Mynors. Oxford, 1964.
Mastricht 1711: G. von Mastricht, ed., Novum Testament/1111. Amsterdam, 1711. Nardo 1966: D. N;irdo, "Le correz1oni nei due codici Medicei 49, 7 e 49, 9 dcllc 'Fami-
Mclachlan 1938-39: H. Mclachlan, "An Almost Forgotten Pioneer in New Testament liares' d i Cicerone (una terza tr:idizionediretta?)," Atti lstit11to Vet1eto di Scie11zee Let
Criticism," Hibbert ]011rnal 37, no. 4 (1938- 39): 61 7-25. terc, "Scien:u: Morali" 12.4 (1966): 337- 97.
Meillet-Vendryes 196 3: A. MeillctJ. Vendryes, Traittf de grammaire comparee des lang11es N ardo 1967: D. Nardo, "Varianti e tradizione manoscritta in Ausonio," Atti lstituto Ve-
classiq11es. P:iris, 1963. neta di Sciem;e e Lettere, "Scienze Morali" 1::.5 (1967): 311- 81.
Merrill 1913: W. A. Merrill, "The Archetype of Lucretius," University o{Califonria P11b Nardo 1970: D. Nardo, I/ Commentario/11111 pctitiomi . Padua, 1970.
licatiom i11 Classical Philology 1, no. 10 (1913): u 7- 35. Nardo 1981: D. Nardo, "Scienza e filosofia nel primo Scttecento padovano: Gli studi
Metzger 1968 (1964): B. M. Metzger, Tile Text of the New Testament: Its Tra1tsrnissior1, d:issici di G. B. Morgagni, G. Poleni, G. Pontedera, L. Targa," Q11ademi per la Storia
Com1ptio11, and Restoratio11, ind ed. Oxford, 1968 (1st ed. Oxford, 1964). dell'U11iversitiz di Padova 14 (1981): 1- 40.
Migne 1845: Patrologiae cursm completus, ed. J.P. Migne, vol. 21: S. E11scbii Hicro11ymi Nencioni 1946: G. Nencioni, ldealismo e realismo 11ella scienz.a de/ linguaggio. Florence,
opera on111ia, pt. 1. Paris, 184 5. 1946.
Migne 1846: Patrologiae c1irs11s complet11s, ed. j . P. Migne: vol. 19, S. 11sebii Hiero11ymi Nestle 1893: E. Nestle, "Bengel als Gclehrtcr: Ein Bild fiir unsere Tage," m Marginalic11
opera omrria, pt. 10. Paris, 1846. mrd Materiale11, 1.1-143. Tubingen, 1893.
Momigliano 1974: A. Momigliano, review of A. E. Housman, The Collected Papers, col N1pperdcy 1847-56: Caesaris Commentarii, ~vols., ed. C. N1pperdey. Leipzig, 1847- 56.
lected :ind ed. j. Diggle and F. R. D. Goodyear Ill (Cambridge, 1971), Athenae11m 52 No he 1913: F. Nolte, }. A. Bengel. Giitersloh, 1913.
(1974): 368-71. Nouvelle biographie gencrale 1855: "j. Boivin," in No11vclle biographic gtf11eralc, 5.479-
Momigliano 1975: A. Momigliano, Quinto co11trib11toaf1a storia degli studi classlCI e de/ 80. 1855.
motrdo antico, 2 vols. Rome, 1975. Okken 1970: L. Okken, Ein Beitrag .:.ur 11twimmg eiller ko11ta111mierten Ma1111skript
Morel 1766: j . B. Morel, Elcmmts de critique. Paris, 1766. tradition: St11die11 z;11r Oberlieferrmg vo11 Hartma1111 vo11 Aue "lwein." Utrecht, 1970.
l.00 Bibliography Bibliography 1.01

Orelli 1Hl.6-38: J. C Orelli, ed., Ciceronis Opera quac supersunt omnia, 5 vols. Zurich, Peros:iTimpanaro 19'5 6: A. Peros:i ands. T1mpanaro, Mlibamo lo Coricio?I: roliziano e
182.6 - 38. Lcop:irdi," St11di ltaliam di F1lologia Classica, n. s., i7- i!I (1956): 411-i.5.
Orclh 18i7: J.C. Orelli, review of H. Meyer, Ciceronis Orator (Leipzig, l8l.7), jahrbii Per Sebastiano Ti111pa11aro i.001: M. Feo and M. Rossi, eds., Per Se{1ast1a110 T1mpa11aro,
cher fur Philologie und Paedagogik, ser. 3, 4 (18i7): 84-91. II Po11te 57, nos. 10-11 (October-November :z.001 ).
Orclli 183 i : J. C. Orclli, De provi11ciis conmlaribus. Zurich, 18 Jl.. Peticmengin 1966: P. Pccitmcngin, MA propos des editions patristiqucs de la Conrre
Orclli 1835:J. C. Orelli, lnterrogatio i11 Vatinium. Zurich, 1835 . RCforme," Rccherch1ts Augustinien11es 4 ( 1966): 199-151.
Orelli 1837: J.C. Orclli, Aratea. Zurich, 1837. rfeiffer 1949 -53: Callimacl111s, 1. vols., ed. R. rfciffer. Oxford, 1949-53.
Orelli-Beicr 1830: J. C. Orclli and K. Beier, eds., Ciceronis Orator Brutus Topica De op Pfeiffer 1968: R. Pfeiffer, History of Classical Scholarship from the Begirm;,,g to the End
tm10 genere oratomm. Zurich, r 8 JO. of the Hellenistic Age. Oxford, 1968.
Pascal 1918: C. Pascal, "Emcndarc," Athenaeum 6 (19111): io9-16. Pfeiffer 1976: R. Pfeiffer, History of Classical Scholarship from 1300 to 1850. Oxford,
Pasquali 19i9: G. Pasquali, review of P. Maas, Tcxtkritik (Leipzig, 19i7), G110111011 s 1976.
(19i9>: 417-35, 498-52.l. Pighi 1973 : P. Ovidii Naso11is Fastorum libri, ed. G. B. Pighi. Turin, 1973
Pasquali 1931: G. Pasquali, "Teologi protcsranri predeccssori del Lachmann," St11di /ta Pizzani 1959: U. Pizzani, 11 problema def tt:sto e de/la co111posi:io11e de/ "De renrm 11a-
lia11i di Filologia Classica, n. s., 9 (1931 ): 2.43-54. tura." Rome, 1959.
Pasquali 193u: G. Pasquali, review of P. Collomp, La critique des textcs (Paris, 1931 ), Politian 1971: A. Poliziano, Misccl/dnl!orum c