Académique Documents
Professionnel Documents
Culture Documents
VOLUME 34
ON SCIENTIFIC
DISCOVERY
The Erice Lectures 1977
Edited by
ROBERT S. COHEN
Boston University
GUIDO CIMINO
Istituto della Enciclopedia /taliana, Rome
This volume was prepared for publication with the editorial assistance of
Carolyn R. Fawcett.
Support for the translations from Italian was provided by the
Istituto della Enciclopedia Italiana.
PREFACE vii
INTRODUCTION 1
The 1977 lectures of the International School for the History of Science at
Erice in Sicily were devoted to that vexing but inexorable problem, the
nature of scientific discovery. With all that has been written, by scientists
themselves, by historians and philosophers and social theorists, by psycholo-
gists and psychiatrists, by logicians and novelists, the problem remains elusive.
Happily we are able to bring the penetrating lectures from Erice that summer
to a wider audience in this volume of theoretical investigations and detailed
case studies.
The ancient and lovely town of Erice in Northwest Sicily, 750 m above the
sea, was famous throughout the Mediterranean for its temple of the goddess
of nature, Venus Erycina, said to have been built by Daedalus. As philosophers
and historians of the natural sciences, we hope that the stimulating atmo-
sphere of Erice will to some extent be transmitted by these pages.
We are especially grateful to that generous and humane physician and historian
of science, Dr. Vincenzo Cappelletti, himself a creative scientist, for his
collaboration in bringing this work to completion. We admire his intelligent
devotion to fostering creative interaction between scientists and historians
of science as Director of the School of History of Science within the great
Ettore Majorana Centre for Scientific Culture at Erice, as well as for his
imaginative leadership of the Istituto della Encic10pedia Italiana.
Our thanks, too, to Professor Mirko Grmek for his cooperative work as
Co-Director of the School and in the preparation of this volume; and especially
to Dr. Guido Cimino for his extraordinary and energetic competence at all
stages of the organization and planning of the School and of this book.
vii
INTRODUCTION
Every two years the International School for the History of Science, located
at the Ettore Majorana Center in Erice, Italy, invites historians, epistemologists
and research scientists from all over the world for meetings devoted to discus-
sion of some of the significant problems currently facing the scientific com-
munity and the world of culture. Furthermore, the nearly sixty other schools
of the Majorana Center contribute an indirect but effective indication of the
lines along which scientists and historians can work together.
The purpose of the course of the International School for the History of
Science held at Erice, 16-22 February 1977, sponsored by the Majorana
Center, the Domus Galilaeana and the Istituto della Enciclopedia Italiana,
was to produce a critical examination of the logical, psychological, cultural
and social aspects of scientific discovery. This book is the fIrst publication to
be issued as a result of one of the School's courses, and in presenting it to the
reader, we must emphasize that it is not the 'proceedings' of the course,
namely a faithful record of the lively reports, communications and discus-
sions held at the Ettore Majorana Center. Rather we preferred to ask the
speakers to prepare a new version of their papers which took into account the
subsequent debate and also the requirements of a monographic publication.
Thus, after two years the problems and positions expressed at Erice have
crystallized to produce an image of the course which, though perhaps less
lively than the event itself, is more rigorous in its presentation.
This book sets out to ascertain the contemporary state in a particular
fIeld of the historiography of science and to make an original contribution
to topics bordering on both the philosophy of science and the history of
science. The contributors and the editors are pleased to have the volume
published under the auspices of the Boston Studies in the Philosophy of
Science.
For a long time epistemologists have considered scientific theories to be
parts or examples of a body of knowledge which is free of contradictions
in its structure and based on 'foundations', on propositions which cannot be
reduced to others. A more recent, different and reciprocal way of consider-
ing science is that of an expanding body of knowledge which increases the
domain of the known at the expense of the unknown. The analysis of scientifIc
1
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 1-6.
Copyright © 1980 by D. Reidel Publishing Company.
2 INTRODUCTION
Helmholtz went on to say that a printer's case of type contains the entire
wisdom of the world. A discovery must be formulated as a rational tran-
scription, but the transcription must be organic, and it must be in agreement
with empirical data.
Thus the rational character of scientific discovery had already been under-
stood and affirmed during the last century. A new fact, a new 'property of
Nature', as Mach would say, became known to others through the proposi-
tions by which the discoverer describes, dermes and explains the factual. Even
from the discoverer's viewpoint, however, what is discovered is what he him-
INTRODUCTION 3
structure and the occasion of its offering itself, or at least showing itself, to
the observer; on the part of the subject, personal and social events, psycho-
logical and sociological factors which have been and still are the interpretative
elements for historians and epistemologists. Yet eventually, discoverer and
discovery must feed into a cognitive process, and anything can contribute to
the growth of knowledge, which is all that truly matters. Rational objectivity
is what counts when vicissitudes come to an end, vicissitudes which at times,
have the savor of adventure.
That discovery is the statement of a fact has already been said. We can
now add that this statement must be comprehensible, as expressed in an
accepted and codified language and logic. Discovery therefore becomes a test-
ing ground for scientific knowledge, no matter whether in a positive sense, in
the form of verification, or in a negative sense, in the form of refutation.
However, innovation, the essence of discovery, can never be reduced to more
validation or confutation. Frequently, therefore, existing language does not
suffice. The discoverer thus coins a new terminology, even a new axiomatic
system: he codifies a language that his discovery will justify. Just as it is the
creation of new knowledge, discovery also implies the creation of a new
language, words, definitions and descriptions which had not previously existed.
If science is expanding knowledge, renewing itself on the periphery as well
as at the heart of its own domain, and if innovation is identified with scientific
discovery, then the history of discoveries will be the essential part of the his-
toriography of science and their logical and ontological analysis will be the
core of epistemology. Cognitive growth must be recognized in order to con-
tinue. How far have we come? Only history can provide the answer. How is
our knowledge structured, on which axioms does it depend, what inferences
does it use? The answers lie within the philosophy of science. But there also
exists a pedagogy of science and it too fmds its principal subject-matter and
instruments in scientific discovery in order to advance towards new and open
knowledge. To eliminate the possibility of the above-mentioned meta-obstacles
from the spirit of research is the greatest contribution that can be made to
the progress of knowledge.
In the papers we have brought together in this volume; scientific discovery
is studied not only as an original acquisition, a new body of knowledge
analyzed in its connection with the entire body of constituted knowledge,
but also and above all as a creative act, a mental process by means of which
one goes beyond the known.
If we take scientific discovery into consideration in so complex a manner,
three different approaches arise, overlap, and to a certain extent contradict
INTRODUCTION 5
M.D.GRMEK
R. S. COHEN
G. CIMINO
NOTES
GENERAL PROBLEMS
MIRKO D. GRMEK
Our beloved, for us, historians of science, is the whole of res gestae et
scriptae of past scholars. We embellish these in our fashion and we contrive
to give them a lustre which is in harmony with our general ideas on the nature
of knowledge and on the way it is acquired. So, the dry branch of 'historical
facts' is, by a process of progressive theorization, enveloped little by little in a
thick layer of seductive 'myths'.
Let us take for granted that it is utopian to wish to produce, in accordance
with Ranke's precept (or, to keep within the more limited field of the history
of science, with Sarton's positivist recommendations), a history claiming to be
the absolutely objective mirror ofwhat really happened. Let us take for granted
that historical reconstruction carries out selections, establishes links and pro-
poses 'explanations' which necessitate a philosophical stand, indeed an involve-
ment of an ideological nature. But this recognition of a certain lack of objec-
tivity in historical research should urge us not to relativism and an attitude of
resignation, but to a careful ~nd systematic analysis of the factors which inter-
vene in 'crystallization', to the exposure of the processes which condition our
search for truth. It seems inevitable to us that we demand nowadays that the his-
torianof science submit himself to an epistemological self-examination. In short,
at the moment when he becomes fully conscious of the impossibility of total
objectivity in his discourse, the historian should feel more than ever the necessity
to elucidate, as much as possible, the modalities of his deviations therefrom.
9
M. D. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 9-42.
Copyright © 1980 by D. Reidel Publishing Company.
10 MIRKO D. GRMEK
Under this heading I wish to denounce both (a) the pseudoproblem of the
conflict between 'internal' history and history referred to as 'external', and
(b) the illusion resulting from the confusion between the personal history of
the scientist and the general history of the progress of scientific knowledge.
Without doubt, a dual aspect exists of the iter mentis in veritatem: on the
one hand the actual itinerary of ideas in the 'discoverer's mind', and on the
other, the historical development of 'scientific thought' considered in abstracto,
that is to say without taking into account the particularities of those individ-
uals who are unavoidably its mainstay.
Experience proves the possibility of a historic discourse on the transforma-
tions of scientific theories which disregards considerations of a biographical
or psychological nature on those who have in effect created that world of
ideas. Is such a discourse legitimate? Although we are convinced that thought
is exclusively the product of concrete and individual nervous systems, we
acknowledge the pragmatic value of studying an 'internal dynamics' of
scientific ideas. However, to overcome the paradox of a thought without a
head to think it, it appears necessary to call upon the help of an intermediary
notion: that of 'scientific community'.
I therefore define 'scientific thought' as a structured aggregate of known
facts and propositions of a certain type which are accepted or debated by a
group restricted in historical time and in social space. In this way, discourse
on the general development of scientific thought relates to a historical reality
and not to an idealistic schema.
Evidently, if the 'internalist' approach is conceived, justified and realized
with the help of the notion of 'scientific community', it concerns, strictly
speaking, not the 'general' development of scientific ideas, but only
numerous sectors of it. The success at the monograph level of such a histori-
cal presentation will provide proof of the reality of a community of scholars
within well-dermed spatio-temporal co-ordinates, with due regard to the
problems they study. 'Internalist' scientific historiography requires a com-
plementary inquiry: the historical examination of the community concerned,
18 MIRKO D. GRMEK
a history of science which limits itself to tracing the skyline of the world of geniuses
prevents itself from understanding the mechanism of scientific research; inversely, the
genuinely philosophical task is not to locate truth and to brand error, but to understand
intellectual processes which are now alien to us.
The discovery of the circulation of the blood has often been cited as a
typical example of the rigorous and successful application either of Baconian
induction, or of the hypothetico-deductive method. These approved interpre-
tations, dear to the particular inclination of every pOSitivist-historian, are not
altogether consonant with what the documents say about the real genesis of
William Harvey's discovery. We may not disregard the influence of the mysti-
cal symbolism of circles, Gnostic philosophy and Peripatetic metaphysics. In
physics even, Harvey drew his inspiration as much from Aristotle as from
GaWeo's new mechanics. When he gives the heart the distinguished title of
sun of the microcosm, Harvey uses something more than a poetic image.
Physiological cardiocentrism is actually historically bound up with astronomi-
cal heliocentrism. The most decisive step in Harvey's reasoning was the
transformation of the Aristotelian analogy between the irrigation of a garden
and the distribution of blood in the body, a transformation carried out with
the help of another 'model' likewise taken from his reading of Aristotle,
namely the place of the sun in the circular motion of terrestrial and atmos-
pheric waters.
I will not dwell on the role of chance, for I have recently published a study
specially devoted to this topic (see Grmek, 1976).
To illustrate how the rational reconstruction of discoveries serves as a
model for a positivist history, let us remember the introduction of vaccina-
tion against smallpox by Jenner (Razzell (1977) has shown certain of its
'mythical' aspects), the first vaccinations against rabies by Pasteur, and
Bernard's experiments on glucosuria. The events leading to these achieve-
ments are not in reality as simply and logically linked up as we are asked to
believe.
A particularly instructive example is provided by the history of the dis-
covery of insulin. It is admirably clear in the justification of the Nobel prize
award by the chairman of the selection committee, and in the autobiographi-
cal accounts of the recipients, but becomes more and more confused as one
proceeds in the critical analysis of the facts. In the first place, it does not
seem permissible to eliminate from the historical report the contribution of
certain 'forerunners' (see Murray's article on experiments by Paulesco, not
unknown to, but wilfully deformed by the winning team). What is even more
interesting, from our point of view, is that one becomes aware (see especially
FREEING SCIENTIFIC DISCOVERIES FROM MYTH 23
the excellent study of Pratt, 1954), that the official version of the way the
discovery was arrived at, does not stand up to an examination of the original
publications nor to a repetition of the experiments. The clue of Banting's
reasoning was simply an illusion: obstruction of the pancreatic ducts does not
give rise to a selective degeneration of the tissues producing trypsin. Contrary
to the working hypothesis he proposes, the extract obtained from a normal
pancreas contains a higher proportion of insulin than that yielded by a
'degenerate' one. The results of Banting and Best's experiments are at vari-
ance with the conclusions that they drew from them. We know today that the
success of the Toronto team is due to the precipitation of proteinic hormone
by 95 percent alcohol, which eliminates toxic ingredients and allows medical
use of the extract. The inventor of this process, the biochemist CoUip, was
put on the side, when the Nobel prize was awarded, to make room for the
head of the laboratory in which the discovery had been made.
Continuity and rupture co-exist and any historical venture based on an exclu-
sive mythicizing of one of these two complementary aspects would seem
doomed in advance.
If, in the personal history of a researcher, there are sudden moments of
enlightenment, comparable to Archimedes' Eureka, there is also, inevitably,
prior to them an incubation period, a slow, continuous process of intel-
lectual ripening. In the general history of the sciences, there are likewise
revolutionary reversals - historical falsifications in Popper's sense of the
expression, or paradigmatic reversals in Kuhn's terminology - but as the
latter so well emphasizes, that does not exclude the existence of a slow, not
really discontinuous accumulation of scientific knowledge in the period
separating two crises.
From this standpoint let us look at the history of the discovery of the
circulation of the blood. Baldini has examined Harvey's work in the light of
Popper's epistemology and has severely criticized the idea that his discovery
resulted from a progressive and gradual piecing-together of a mosaic of minor
discoveries, of tiny steps made by his 'forerunners' and by Harvey himself,
necessarily leading by their added momentum, to the new pattern of circu-
latory physiology. Baldini's analysis Seems to us convincing: Harvey's achieve-
ment is revolutionary and, in essence, cannot be explained by the summation
of a series of partial discoveries. It is nonetheless true that such a summation
24 MIRKO D. GRMEK
exists as well, and if it does not necessarily produce nor explain the decisive
discovery, it is nevertheless, for its production, an indispensable precondition.
The advent of Harvey's new physiology is unthinkable without the cumula-
tive contribution of Vesalius, Colombo, Cesalpino, Fabrizio d' Acquapendente
and several other researchers who breached Galen's system and allowed
Harvey to find himself confronted with a particular panorama of known facts.
The 'heroic' view of all scientific research idealizes historic reality. Without
doubt, a 'normal science' (in Kuhn's sense) exists, and we could not partici-
pate, in this connection, in the disillusioned sigh (,yes, alas!') of certain of
Popper's disciples. The existence and proper working of this 'normal science'
is a historical sine qua non precondition of 'extraordinary science'. The alter-
nation of 'puzzle-solving' periods and emergence of new scientific theories
seems to belong to the very nature of the process which produces the growth
of our scientific knowledge.
Certain epistemologists, for example Hanson and Kuhn, compare the act
of scientific discovery to the shift in the interpretation of an ambiguous
drawing (the famous rabbit-duck). According to this analogy, a scientific
theory would explain a group of facts exactly as a perceptual interpretation
organizes lines and spots into a Gestalt. We would thus pass from one inter-
pretation to another, both in the change of paradigm and in the setting-up of
a new apperception, by a total transformation determined by insight and
therefore of an essentially unforeseeable and radically discontinuous nature.
The reversal of the scientific paradigm has been compared to a kind of mystic
crisis.
Without wishing to go as far as the 'anarchical' excesses of Feyerabend,
who speaks of the 'incommensurability of theories' and the poetic determina-
tion of the choices made by scholars, I am ready to accept the existence, in
the progress of knowledge, of qualitative jumps, that is to say, changes which
are irreducible to the accumulated mass of variations of secondary importance.
But we do not wish to fall into the trap of the myth of a permanent revolu-
tion and deny the fundamental importance of slow transformations of scien-
tific theories, which come about by a progressive adaptation, not by a rever~l
of paradigms. The two aspects appear to me complementary.
To go back to the analogy between scientific discovery and the interpreta-
tion of the rabbit-duck drawing, I intend to consider it in a more dynamic
way than that in which it is usually described. Let us imagine that the draw-
ing in question is not a state but a becoming, that is to say that it is not static,
but mobile, in slow, continuous transformations. This can be realized, for
example, on the screen of an analogical computer. Let us imagine further that
FREEING SCIENTIFIC DISCOVERIES FROM MYTH 25
Mendel was expecting certain numerical data; he was testing his hypothesis
and did not infer it from observed facts.
However, certain new historical and mathematical precisions, due especially
to Van der Waerden and Welling, seem to temper Fisher's conclusion accord-
ing to which Mendel first formulated his theory and then performed his
experiments. Theorizing and experimenting went hand in hand: the essence
of Mendel's discovery was acquired during the carrying out of the experi-
ments.
There are certainly no logical rules for acceding with unassailable certainty
to a so-called inductive assertion. But must one go as far as Popper and say
that induction is an illusion in the strongest sense of this term, that is to say,
a mental process not only without logical foundation but also, as it were,
without existence? According to Popper, induction exists only in appearance,
thanks to the selection of hypotheses obtained by a completely different
process. An assertion aspiring to general validity would then be in no way
influenced, in its genesis, by observed facts. It would secondarily be brought
in harmony with these, and not be inferred from them.
For a historian of science, convinced as he is that reality ranges wider than
the excessive rigidity and extreme simplifications of logical patterns, it
appears rash to reject induction in so drastic a fashion. While admitting the
non·validity, in pure logic, of inference from the particular to the general, we
are not obliged to deny the existence of a mental inductive process which
generates hypotheses, without, however, justifying these. If there is no induc-
tion in the logical sense of the word, it may well be that there is a kind of
psychological inference, an orientation of thought via knowledge of empirical
data.
The 'discovery' of the 'cell' by R. Hooke and the progress of micrographi-
cal studies in the seventeenth and eighteenth centuries did not lead directly
to the formulation of the cell theory. This is a historical example of the im-
potence of the inductive method, such as it was taught, precisely at the time
in question, by the brilliant English Chancellor.
Various theories on fibers, globules and 'molecules', understood as ele-
mentary carriers of life, did not give birth to the authentic cell theory before
Schleiden conveyed to Schwann certain data gathered from observation. I
interpret this historic case (for a good exposition onthis, see Florkin's, Klein's
and Canguilhem's (1965) publications) as an instance of subtle dialectic
between attempts at generalizing theorization and concrete observation.
FREEING SCIENTIFIC DISCOVERIES FROM MYTH 31
Faced with the logicians' method of reasoning, we are forced to admit that
scientific theories are not verifiable in the absolute sense of the word. Histori-
cal examples of theories which have never been falsified evidently prove
nothing: we do not know their future. But a different demonstrative im-
portance would attach to the historical existence of theories which, after
being falsified, have later been recognized as 'true'. Such cases are known;
indeed they are quite numerous. So, to our surprise, historical inquiry seems
to show that, in fact, scientific theories are not falsifiable either in a definitive
way.
The logical explanation for this can be easily found: data provided by
observation can be improperly noted because of technical or psychological
errors and, most important, the reasoning process which leads to falsification
always presupposes tacit acceptance of a conceptual framework which may
be called in question after the event. Such an alteration of certain tacitly
accepted presuppositions leads to a disavowal offalsification: which explains,
on the one hand, the historical possibility of a rebirth of scientific theories
judged at a given moment as certainly false, and on the other, the impossi-
bility of experimentum crucis.
Let us quickly look at a historical example: choosing once again, to make
our analysis easier, the discovery of the circulation ofthe blood. Harvey was
at grips with serious difficulties and could not decide, for many long years,
to publish his new physiological system. Without a knowledge of the capil-
laries and pulmonary alveoli, it was impossible to explain properly the passage
of blood at tissue-level (from the arteries to the veins) and at lung-level (from
the veins to the arteries). Assuming the existence of porosities, invisible to
the naked eye, Harvey acknowledged implicitly the Galenists' right to believe
in the existence of invisible pores in the interventricular partition of the
heart: in other words, he weakened an important criticism of the ancient
system. To explain the venous return to the heart, one had to tum to non-
FREEING SCIENTIFIC DISCOVERIES FROM MYTH 33
problem. They are often 'falsified' out of hand, yet their discoverers do not
abandon them, but cling to them and end up falsifying the first-stage 'refuta-
tions'. The best hypotheses often contradict one part of the experiment:
they call for a re-examination of the 'facts' which, at that point in history,
seemed established in definitive fashion.
The strength of a new theory does not therefore lie solely in its resistance
to attempts at refutation. There must be, particularly in the initial phase of a
discovery, factors which incline one in its favor and which are irreducible to
the Single-handed logic of falsification.
Why out of two (or several) non-falsified theories, or two (or several)
theories historically falsified and saved by auxiliary hypotheses, does one
decide in favor of one rather than the other? Neopositivists put forward the
criterion of probability, especially the degree of corroboration. Popper appeals
to the principle of maximization of empiric content. Serious authors have
spoken of factors of an aesthetic nature, of 'poetical', even 'mystical', pre-
ferences. I shall draw attention to a characteristic which could be important
for the decision scholars make (without believing, however, that it is the only
one to turn the scale): the 'perspectivity' of a theory.
To abandon a scientific theory and accept another is more often a matter
of methodological opportunity than of strict refutation. Formal logic inter-
venes generally after the event. It justifies more than it decides the way of
scientific research.
Now, as a result of my last statement, we find ourselves faced with the partic-
ularly delicate problem of the relationship between the logic, psychology
and sociology of scientific discovery. I do not have time to analyze properly,
in this context, the three myths which consist in reducing the total reality of
'science in the making', or at least its essence, into just one of these three
approaches. Let us be satisfied with a quick overall view, hoping to be able to
deal with this subject more fully in the course of another seminar.
My account thus concludes with a summary and curt denunciation of
three illusions which, because of their methodological importance, should
have been placed at the head of a list of errors to be avoided, and which, by
their special epistemological nature, would have deserved the detail and
finesse of a.well-documented monograph.
The three myths in question have a common source: the dissociation of
FREEING SCIENTIFIC DISCOVERIES FROM MYTH 35
the unitary triad made up of the three aspects of the acquisition of knowl-
edge, corresponding to the constitutive triad of the concept of man:
individual
species
/ "'- society
(Concerning the latter, see Edgar Morin's publications.) The distinction be-
tween the logical, the psychological and the social lies not in reality, but in
our means of intellectual analysis. It is a necessary and fruitful process of
dismantling, but it creates difficulties and exposes one to the risk of mythiciz-
ing, in an exclusive, imperialistic fashion, each of the components of the
global system. The three approaches are concurrent, mutually antagonistic
and complementary. They should not be impervious to intercommunication.
In the analysis of scientific discovery, two key-concepts are used: induc-
tion and intuition. These are in my opinion 'amphibian' concepts which,
because of their polyvalence, have a very great operational value and, at the
same time, are the despair of certain philosophers and psychologists who
would like to lock them in the field of a single investigatory discipline. Induc-
tion is a concept invented by logicians, and yet they have never succeeded in
reducing it, in a completely satisfying manner, to formal logic. Factors of a
psychological nature intervene in inductive inference. As for intuition, while
it is a concept belonging to the field of psychology, it cannot be successfully
studied if it does not take into consideration the categories and rules of logic.
Positivists with varying allegiances and neo-Baconians still fight tooth and
nail in favor of the old rationalist myth which reduces scientific investigation
to a practical application of logic. It is an undeniable fact, which we must not
forget in this debate, that a considerable number of scientific discoveries
result simply from the application of an impeccable logical reasoning to con-
crete problems. In the genesis of such discoveries the historian can easily
grasp the 'leading idea' and is bound to admire the precision of observation
(Leeuwenhoek, Redi, Fabre, etc .... ), or the persistence in the application
of certain clear, relatively simple hypotheses (Pasteur, Ehrlich, Ludwig, etc.).
When it is a matter of 'deductions', that is to say, applications of a theory
or an investigatory technique to cases not previously contemplated, rationalis-
tic explanation is triumphant. Ascertaining the origin of a completely new
idea is a more complex matter. It seems that to explain the genesis of certain
discoveries, and precisely the most original and fruitful ones, it is necessary to
refer to irrational processes.
It goes without saying that the term i"ational does not indicate anything
36 MIRKO D. GRMEK
This myth rests, on the one hand, on a strict and rigid rationalism, that is to
say, on the considerations of the logicians we have just mentioned, and, on
the other, on a romantic, irrational, sometimes even surrealistic idealism.
These last tendencies are in fashion again today.
A psychology of scientific discovery, which invokes in a quasi-mystical
manner the 'personal genius' of the researcher, must be judged with the
greatest severity. Even in its variants bearing semblances of scientificity (the
analogy, for instance, between the birth of a new idea and biological muta-
tion, which was proposed by C. Nicolle, or the quantum leap; speculations of
certain psychoanalysts and biographical studies with a psychological veneer),
FREEING SCIENTIFIC DISCOVERIES FROM MYTH 37
Translated by
MARGARET ROUSSEL
BIBLIOGRAPHY
Agassi, J.: 1973, 'Continuity and Discontinuity in the History of Science', J. Hist. Ideas
34, 609-626.
Antiseri, D.: 1976, 'Prova di una teo ria ed educazione al riconoscimento dell 'errore',
Med. nei secoli 13, 307-364.
Anzieu, D. and coil.: 1974,Psychana/yse du genie createur (Dunod, Paris).
Bachelard, G.: 1938, La formation de ['esprit scientifique (Vrin, Paris).
Bacq, Z. M.: 1974, Les transmissions chimiques de ['influx nerveux (Gauthier-Villars,
Paris).
Baldini, M.: 1974, Epistem%gia e storia della scienza (Citta di Vita [for Univ. di Siena),
Florence).
Bellone, E.: 1976, II mondo di carta (Mondadori, Milan).
FREEING SCIENTIFIC DISCOVERIES FROM MYTH 39
Gruber, H. E., Terrell, G., and Wertheimer, M. (eds.): 1963, Contemporary Approaches
to Creative Thinking (Atherton Press, New York).
Hadamard, J.: 1954, An Essay on the Psychology of Invention in the Mathematical
Field. 2nd ed. (Dover, New York).
Hanson, N. R.: 1965, Patterns of Discovery (Cambridge Univ. Press, London).
Hanson, N. R.: 1965, 'The Idea of a Logic of Discovery', Dialogue 4,48-61.
Harding, S. G. (ed.): 1976, Can Theories be Refuted? (Reidel, Dordrecht).
Hesse, M. B.: 1974, The Structure of Scientific Inference (Macmillan, London).
Holmes, F. 1.: 1974, Claude Bernard and Animal Chemistry; the Emergence of a Scien-
tist (Harvard Univ. Press, Cambridge, Mass.).
Horton, M.: 1973, 'In Defence of Francis Bacon: a Criticism of the Critics of the Induc-
tive Method', Stud. Hist. Phil. Sci. 4,241-278.
Inhelder, B. and coli.: 1976, 'Des structures cognitives aux procedures de decouverte',
Ann. Psycho/' 44,23-60.
Ivanov, G. M.: 1969, 'K voprosu 0 ponyatii fakt v istoricheskoi nauke', Vop. istorii,
No. 2,73-87.
Jewkes, J., Sawers, D., and Stillerman, R.: 1969, The Sources of Invention. 2nd ed.
(Macmillan, Edinburgh).
a
Jung, C. G. and Kertlnyi, C.: 1953, Introduction I'essence de la mythofogie (Payot,
Paris).
Kedrov, B. M.: 1969 'Sur la psychologie de la creation scientifique', Organon, No.6,
49-68.
Koestler, A.: 1967, The Act of Creation (Laurel, New York).
Kordig, C. R.: 1971, The Justification of Scientific Change (Reidel, Dordrecht).
Kuhn, T. S.: 1962, 'Historical Structure of Scientific Discovery', Science 136,760-764.
Kuhn, T. S.:1963, The Structure of Scientific Revolutions (Univ. of Chicago Press,
Chicago).
Lakatos, I.: 1973, The Changing Logic of Scientific Discovery (London).
Lakatos, I. and Musgrave, A. (eds.): 1970, Criticism and the Growth of Knowledge
(Cambridge Uniy. Press, Cambridge).
Largeault, J.: 1976, 'Questions de logique et d'epistemologie', Rev. Synthese 97, 279-
310.
Loewi, 0.: 1965, The Workshop of Discoveries (Univ. of Kansas Press, Lawrence).
Lorenz, K.: 1973, Die Riickseite des Spiegefs (Piper, Munich).
Machamer, P. K.: 1973, 'Feyerabend and Galileo: The Interaction of Theories and the
Reinterpretation of Experience', Stud. Hist. Phil. Sci. 4, 1-46.
Medawar, P. B.: 1969, Induction and Intuition in Scientific Thought (Methuen, London).
Meyerson, E.: 1931, Du cheminement de fa pensee. 3 vols. (Alcan, Paris).
Mikulinski, S. R. and Yaroshevski, M. G. (eds.): 1971, Nauchnoe otkrytie i ego vospri-
yatie (Nauka, Moscow).
Monod, J. et al.: 1975, 'The Analysis of Scientific Method and the Logic of Discovery',
in H. A. Krebs and J. H. Shelley (eds_), The Creative Process in Science and Medicine
(Excerpta Medica, Amsterdam), pp. 3-28.
Morin, E.: 1973, Le paradigme perdu: fa nature humaine (Le Seuil, Paris).
Murray, I.: 1971, 'Paulesco and the Isolation ofInsulin', J. Hist. Med. 26,150-157.
Nagel, E.: 1961, The Structure of Science (Harcourt, Brace and World, New York).
Nicolle, C.: 1932, Biologie de l'invention (Alcan, Paris).
FREEING SCIENTIFIC DISCOVERIES FROM MYTH 41
Olby, R.: 1974, The Path to the Double Helix (Macmillan, London).
Pagel, W.: 1967, William Harvey's Biological Ideas; Selected Aspects and Historical
Background (Karger, Basel).
Pagel, W.: 1976,New Light on William Harvey (Karger, Basel).
Pfliiger, E. F. W.: 1905, Das Glykogen und seine Beziehungen zur Zuckerkrankheit
(Hager, Bonn).
Piaget, J.: 1967, Biologie et connaissance (Gallimard, Paris).
Popper, K.: 1959, The Logic of Scientific Discovery (Basic Books, New York).
Popper, K.: 1963, Conjectures and Refutations (Routledge and Kegan Paul, London).
Popper, K.: 1972, Objective Knowledge; an Evolutionary Approach (Clarendon Press,
Oxford).
Porep, R.: 1971, 'Der Prioritatsstreit urn die Entdeckung des Glykogens zwischen Claude
Bernard und Victor Hensen', Med. Mschr. 25,314-321.
Pratt, J. H.: 1954, 'A Reappraisal of Researches Leading to the Discovery of Insulin',
/. Hist. Med. 9,281-89.
Razzell, P.: 1977, Edward Jenner's Cowpox Vaccine: the History ofa Medical Myth,
(Caliban Books, Newhaven).
Rey, A.: 1907, 'L'energetique et Ie mecanisme au point de vue des conditions de la
connaissance', Rev. phi/os. 46,495-517.
Rossi, P.: 1977, Immagini della scienza (Editori Riuniti, Rome).
Rupke, N. A.: 1976, 'Bathybius haeckelii and the Psychology of Scientific Discovery',
Stud. Hist. Philos. Sci. 7,53-62.
Russo, F.: 1974, 'Typologie du progres des connaissances scientifiques', Rev. Quest. Sci.
145,345-363 and 479-502.
Schaff, A.: 1970, Geschichte und Wahrheit (Europa, Vienna, Frankfurt, ZUrich).
Schaffner, K.: 1974, 'Logic of Discovery and Justification in Regulatory Genetics',
Stud. Hist. Philos. Sci. 4, 349-385.
Scheffler, I.: 1963, The Anatomy of Inquiry (Knopf, New York).
Schilpp, A. (ed.): 1974, The Philosophy of Karl Popper (Open Court, La Salle).
Selye, H.: 1964, From Dream to Discovery (McGraw-Hill, New York).
Stendhal: 1959, De I 'amour, ed. with an introduction and notes by H. Martineau (Garnier,
Paris).
Stent, G.: 1972, 'Prematurity and Uniqueness in Scientific Discovery', Scient. A mer.
227,84-93.
Szent-Gyorgyi, A.: 1962, 'On Scientific Creativity', Persp. Bioi. Med. 5, 173-178.
Taton, R.: 1955, Causalite et accidents de la decouverte scientifique (Masson, Paris).
Thorn, R.; 1974,Modeles mathematiques de la morphogenese (Union generale d'editions,
Paris).
Waerden, B. L. van der: 1968, 'Mendel's Experiments', Centaurus 12, 275-288.
Watson, J. D.: 1968, The Double Helix (Weidenfeld and Nicolson, London).
Weiling, F.: 1971, 'Neue Ergebnisse zur statistischen Vorgeschichte der Mendelschen
Versuche',Biometrics 27, 709-719.
Wertheimer, M.: 1945, Productive Thinking (Harper, New York).
Westcott, M. R.: 1968, Toward a Contemporary Psychology of Invention (Holt, Rine-
hart and Winston, New York).
Worden, F. G. (ed.): 1975, The Neurosciences: Paths of Discovery (MIT Press, Cam-
bridge, Mass.).
42 MIRKO D. GRMEK
Wright, G. H. von: 1957, The Logical Problem of Induction. 2nd ed. (Macmillan, New
York).
Yaroshevski, M. G.: 1971, Problemy nauchnogo tvorchestva v sovremennoi psikhologii
(Nauka, Moscow).
GERARD RADNITZKY
all research were like the unique acts of creativity in art, all the decision-
making involved might be governed by 'tacit knowledge' in the sense of not
being articulable. If so, a critique of decision-making in research would be like
art criticism, since there would be no general criteria, no statute law. Faute de
mieux, the critic would have to rely on his intuition, sensitivity, Finger-
spitzengejiihl. He would have to bring his personality into play. If on the
other hand all of research were like a routinized procedure of 'problem solv-
ing', the methodology could be an algorithm. Clearly either view is a totaliza-
tion - and patently false. Research obviously contains moments of both
types and above all moments which in various degrees approximate to each of
the above extreme types. Insofar as research or an important section of it is
a rational goal-directed activity, at least some part of the decision-making can
be elucidated by a praxiological study. (,Praxiology' is being used in Kotar-
birlski's sense: roughly as the theory of effective and efficient action.) If so -
and we submit that this is a correct assumption - methodology is a viable
project. Moreover, the researcher cannot avoid it since, whether he likes it or
not, part of his time as researcher is spent in this sort of decision-making:
every researcher is his own part-time methodologist. There cannot be any
methodology-free or methodology-neutral research. (As little as there can be
observation sentences or even communicable, hence formulable, perception
reports free from theoretical ingredients.) The suspicion that methodology as
such, i.e. as a diScipline, might claim to be able to prescribe to the researcher
what he should do is, as already hinted at above, based on a misunderstand-
ing. The existence of certain misguided paternalistic methodologies which
oversell themselves does not warrant such a generalization to the discipline as
a whole. The researcher who does not recognize the interdependence of
research and methodology will be a 'methodologicien malgre lui'. The meth-
odological criteria and gambits he uses in his research activity will remain
latent, and so long as they remain latent they cannot be criticized and hence
there will be little chance of their being improved. We have belabored the
obvious because in contemporary discussion the raison d 'etre and even the
possibility of a methodology have been questioned.
0.1. What Contributions Can Methodology Make?
Most important of all is that it should contribute to the refinement of our
image of science. This image is becoming increasingly important for our
image of man. Questions such as 'Can scientific knowledge be rationally
justified?' 'How does knowledge grow?' etc. have to be attacked if we are to
refme our view about man's capacity of knowing, which is a central part of
PROGRESS AND RATIONALITY IN RESEARCH 45
decisions and so on, it may be said to develop what Popper calls a 'situational
logic'. When it attempts to assess the efficiency of ways of proceeding, of
actions, of research undertakings, etc., it qualifies as a praxiological study in
Kotarbmski's sense; and insofar as it looks at research as processing a 'research
program', i.e. a system of hypotheses (knowledge) - problems-instruments
(techniques, calculi, etc.)-plans-etc. into a more refined system of that sort,
it might be said to exemplify what nowadays is often labelled 'systems
thinking'. 2 It constitutes a sub-field of philosophy rather than anything else,
a sub-field that is secondary to some other fields as it is secondary to science
itself in the sense that first there must be empirical research before a need for
methodology can arise.
Methodology obviously does not and cannot deal with certain problem
clusters of traditional philosophy. But nonetheless philosophers often re-
proach it for not attacking certain philosophical problems. Hence some proca-
taleptic remarks about its limitations are called for. (i) Some philosophers
(e.g. H. Spinner) accuse Popperian methodology of conventionalism, asserting
that in spite of its being a non-justificationist approach it has not been able to
come to grips with the problem of the so-called 'basic' sentences. So far as
methodology is concerned Popperians hold that no type of sentence is to be
accorded an epistemologically privileged status. In the empirical testing of
hypotheses, data sentences are used, e.g. in physical research statements about
material objects and processes but not about perceptual experiences. Percep-
tual reports form part of the good reasons for the conjecture that a certain
'basic' sentence or data sentence may be accepted pro tempore. When there
appears no reasonable doubt concerning a data sentence (a matter to be
decided by the researcher) it would be pointless for the methodologist to
emphasize that it can always be questioned. For this reason the relationship
between perceptual reports and a data sentence about physical objects is not
analyzed by Popperian methodology. Such an analysis is regarded as a topic
of general epistemology, and ontological analyses of acts of perceiving, etc.
are left to philosophers. (E.g. 'epistemology as the ontology of the knowledge
situation' as it is developed by Gustav Bergmann and his followers is a field
that methodology cannot encompass.) (ii) Similarly Popper's three-world
ontology is intended as means to an end: to provide a suitable ontological
ground-plan for discussing certain methodological problems. It is not in-
tended as an 'Aufbau der Welt', as e.g. 'ontology as an argumentative struc-
ture upon a phenomenological base' by means of which to construct or recon-
struct reality including our experience. This again is a task of 'first philosophy'.
(iii) Popperian methodology presupposes that there exists a language and that
50 GERARD RADNITZKY
and the Popperian ideal of science has arisen from criticizing the explicatum
proposed by logical empiricism. We propose first to give some hints at the
ideal as explicated by logical empiricism, then to move to the Popperian
criticism of that ideal, and eventually to the alternative offered by Popper.
1.1. The guiding idea of the ideal of science of logical empiricism, of the
foundationalist approach, is to give the desideratum certainty, top priority. As
a result the most important task of methodology becomes that of formulating
and legitimating a role of acceptance in accordance with the basic conviction
that a scientific proposition is ultimately acceptable only if it is true. To
make such a role operational one needs a method of establishing in concrete
cases whether a proposition is true or not, and the method must give us an
infallible criterion of truth. This is the position of verificationism, the justi-
ficationist approach to rationality in science: a sentence is acceptable only if
true and recog~ed as such. It is then weakened in probabilistic verification-
ism: a scientific proposition is acceptable only if it has been probabilified
to a 'sufficient' degree. In this scheme experience plays a positive role (hence
the tag 'positivism'): it serves as the final establishing arbiter. Verificationism,
absolute or probabilistic, seems attractive only if one is willing to countenance
a particular class of empirical propositions whose certainty does not need to
be called in question: a sure source of knowledge about reality as an epistem-
ological fundament on which to erect the edifice of science. Ideally, systems
of propositions would be generated by deductive connections. This is a sort
of 'proof-empiricism'. Again, in view of its obvious unattainability - since
universal propositions infmitely transcend any fmite set of singular sentences
serving as a fundament - one lowers his demands and attempts to construct
propositional systems with partial information-covering from a selection of
'basic' sentences assumed to be certain. Underlying this probabilistic verifica-
tionism seems to be a principle of hope that, as more and more evidence
comes in, it may in the long run be possible at least asymptotically to approach
that ideal state in which the evidence completely 'covers' the information of
the complex sentences. Inductive logic or theory of confirmation is to pro-
vide the connections and to measure the degree of 'coverage'. Cognitive
progress would then be defmed in terms of better and better approximation
to this ideal of science. Logical empiricist philosophy of science may be
viewed as an articulation of various key aspects of this ideal of science. 4
1.2. Popper's critique may, in accordance with what has been said above, be
divided into two parts: the critique of the problem solutions offered, and the
52 GERARD RADNITZKY
critique of the problems. (i) The critique of the solutions offered amounts
roughly to pointing out that the various models (such as the various explicata
offered for the concept of 'empirical significance', the models of explanation
which promise an explication of the idea of causal explanation by stating not
only necessary but also sufficient conditions, etc.) have not been able to help
researchers approach certainty, the ideal's own centerpiece in measuring the
results of research, not even when these results are highly stylized. This must
be so because, firstly, the presupposed certain 'basis' does not exist, and,
secondly, even if for the sake of argument such a 'basis' were conceded to
exist, not only absolute verificationism but also probabilistic verificationism,
inductivism, founders on Hume's criticism, since it follows from the proba-
bility calculus that the logical probability of a universal proposition on the
basis of a finite set of evidential statements is zero. Even a partial retransmis-
sion of truth from verified conclusions to premisses is logically not possible.
In addition, probabilistic verificationism, with its attempt to develop an in-
ductive logic or similar method, will continue to lack the crucial deductive
structure - which is one of the desiderata of the logical empiricists' own ideal
of science. It must ex definitione introduce amplificatory logical moves. (ii)
Popperian criticism of the explicatum of our intuitive ideal of science pro-
posed by logical empiricism asserts that this explicatum is neither fruitful
nor sufficiently similar to the reasonable part of the explicandum to be
acceptable as the result of a successful explication attempt. The desideratum
given top priority, certainty, is unattainable in principle - utopian. Since
certainty and informative content are inversely proportional, and since cer-
tainty is given top priority, the value of 'high content of empirical informa-
tion' must be sacrificed. However explicated, the desideratum of 'depth' -
an important component of the intuitive ideal - is not only lost, but has also
become anathema. (Another reason for the tag 'positivism'.) In sum, the price
to be paid for the search for certainty is a total loss in all dimensions even of
the ideal of science as explicated by the logical empiricists themselves: not
only is certainty unattainable in principle, but in striving for it nonetheless,
other desiderata of the logical empiricist's own ideal become unattainable.
The very idea of knowledge in the sense of certain knowledge, which appears
to be a secularization of the theologican's concept of revealed knowledge, can
have no place in empirical inquiry.
From logical empiricism's explicatum of the intuitive ideal follow certain
problems for methodology. Since the explicatum is mistaken, these problems
are inappropriate. From the ideal it follows that the main task of methodo-
logy is to search for an acceptance rule, to formulate and legitimate such a
PROGRESS AND RATIONALITY IN RESEARCH 53
rule. It conflicts with that idea of scientific progress which sees novelty as
the essence of progress: insofar as we want theories to go much beyond 'back-
ground knowledge', to lead to new insights and thereby to new and deeper
problems, it is unreasonable to hold that we are ultimately after theories
whose information content will eventually asymptotically approach the state
of being completely covered by the information carried by that 'complete'
evidence. The justificationist, cumulative view of science is mistaken if only
for that reason, and if, interpreted as a descriptive picture of historical
science, it is refuted by the history of science because theories once regarded
as certain, later were falsified and superseded by new theories. In short, the
Popperian side shows that logical empiricism's commitment to foundational-
ism and inductivism has proved untenable - that the foundationalist approach
founders like all Begriindungsphilosophie. Scepticism would be a possible
reaction to this insight, but it is not the only possibility.
pair of competing theories in fact comes closer to the truth than the other.
The indicators will be fallible but must and can be objective. s (E.g. whether
or not a prognosis about, say, an eclipse is falsified is objective in the sense
that the fate of the prognosis is independent of human influence.) As is well
known, the notion of degree of corroboration is offered as a fallible indicator
that sets out the balance of attempted falsifications to date. The role of
experience here is exclusively that of a critical arbiter. No 'founding' is
sought after. To interpret degree of corroboration as designed to function in
the long run much like degree of inductive support - as some philosophers
have suggested (e.g. G. H. v. Wright) - has no grounds in Popperian method-
ology. (2) Since degree of testability is equivalent with content of empiri-
cal information, the more content a theory has, the greater is the risk of
falsification and thus also its corroboration potential. From the desideratum
of increase in degree of corroboration (as a fallible indicator of increase in
truthlikeness) follows that of content increase (potential explanatory power).
This in turn sets methodology the task of clarifying the concept of empirical
content and of providing instruments for making content comparison. (3)
A dramatic increase in content is possible only when the successor theory
contradicts the predecessor, for only then does it really introduce new
concepts and open new perspectives, thus leading to deeper explanations and
to deeper and deeper problems. This sets methodology the task of explicating
the concept of depth and of providing an indicator of an increase in depth.
One indicator that functions similarly to a sufficient condition 6 is that the
successor theory, in attempting to explain the predecessor theory, corrects
it, i.e. from the successor theory (e.g. Newton's theory) a hypothesis is
deduced (e.g. Newtonian versions of Galileo's law of free fall or of Keplerian
laws of planetary motion) which, although contradicting the original ex-
planandum, may be regarded as an improved successor to the original ex-
planandum (e.g. Galileo's law of fall), or the explanandum mathematically is
an approximation within a limited realm (in the example, when the height
of fall is negligible in relation to the earth's radius) to the improved successor
hypothesis. (The original explanandum could be derived if we make the false
assumption that the earth's radius is infinite or the height of fall zero.) The
successor theory (in our example, Newton's) introduces a new sort of con-
cepts (causal concepts) which are not used in the predecessor theory (Galileo's
and Kepler's law hypotheses do not involve any causal concepts); and it is
plausible that it is these new concepts that enable us to look at the world
in a new way, which in turn makes possible deeper explanations and gives
rise to new problems of a greater level of depth. Therefore the desideratum
58 GERARD RADNITZKY
increase in depth appears to hold a key position. In this way theory appraisal
leads to appraising metaphysics or cosmological hypotheses (as Feyerabend
prefers to call them), to appraising the comparative value, fruitfulness for
research, of competing world-picture hypotheses. Scientists were once de-
servedly termed 'natural philosophers'. The best of them did and still do face
up to the philosophical issues posed by their own work. As Joseph Agassi has
argued, 7 giving top priority in the explicatum of progress to degree of testa-
bility carries an anti-metaphysical flavor; we would add that it is a hangover
from positivism.
Since 1941 Popper has drawn attention to the phenomenon that a deeper
theory corrects the 'observationa1' law-hypothesis - independently of whether
or not the latter has been falsified when the deduction is made - in the very
process of explaining it. 8 It is important to notice the continuity in the
empirical as well as in the mathematical aspects between the hypothesis
corrected and the improved successor hypothesis deduced from the new
theory, although the successor hypothesis contradicts the hypothesis that gets
corrected. Thus, in spite of the break constituted by the new concepts in-
troduced by the successor theory - the new perspectives it opens up and the
new, deeper problems it poses - there still is an element of continuity in
these two aspects.
1.5. In the Popperian ideal of science the essence of progress is seen in moving
from problems to deeper problems. The Kantian-Popperian thesis of the
propagation of problems claims that every solved problem generates new
problems 9 (objectively, i.e. independently of the researcher's wishes, indepen-
dently of whether or not he formulates them or even recognizes them). The
deepening of problems is seen as a measure of progress.! 0 "Science should be
visualized as progressing from problems to problems - to problems of ever
increasing depth."!! When we confront this ideal of science with the picture
of historically given science, we find that the history of science illustrates it
well. Although this fact per se could not be used as a good reason for recom-
mending the ideal of science, it nonetheless shows that the ideal is not utopian
- as is the ideal of science of logical empiricism.
From the point of view of the history of the philosophy of science it is
sweeping, but correct, to say that Popper is the chief critic of the methodology
developed by logical empiricism and of the ideal of science at its root, an
ideal which owes its decisive impulse to the philosophy of the early Wittgen-
stein. The polemic between Wittgenstein and Popper - which has remained
implicit - in fact has a continuation: the most important contemporary critics
PROGRESS AND RATIONALITY IN RESEARCH 59
smaller than that of T and its amount of true information not smaller. A set-
theoretical interpretation of the comparative concept, called 'verisimilitude',
takes the statement "the amount of information conveyed by the true con-
sequences of T' is larger than that conveyed by the true consequences of T"
to mean that the true consequences of T are a proper subset of the true
consequences of T', and similarly, mutatis mutandis for the other parts of
the definition. This interpretation proves inadequate if taken as an expli-
catum of our intuitive idea of 'more accurate representation than'. Miller
(1974) showed that if one uses this set-theoretical interpretation of the
definition of verisimilitude, the required subset relations can only obtain
between axiomatizable theories if both are true. The next explicatum proposed
hinges on the idea that the information conveyed by the true consequences of
one theory minus the information conveyed by its false consequences must
be larger than the information conveyed by the true consequences of the rival
theory minus the information conveyed by its false consequences. This
approach too has proved to be unfeasible. Andersson 14 has shown that some
points of the criticism advanced by Miller and Tichy can be met, but con-
cedes that there remains a fundamental difficulty which cannot be overcome
even with fmer measures of content and verisimilitude: it turns out that all
false theories with the same measured content have the same degree of veri-
similitude if this explicatum is used. This is an absurd consequence since,
according to this explicatum, the verisimilitude of a false theory depends only
on how much it says, not on what it says. Hence this explicatum does not
meet the metacriterion of 'sufficient similarity' between explicatum and
explicandum, a necessary condition for fruitfulness. Moreover the comparison
of false theories is very important, since cognitive progress often consists in
one falsified theory's being replaced by another which, although likewise
falsified, is regarded as closer to the truth. (We need not to go to the history
of science; a primitive example can illustrate this: the hypothesis 'The planets
move in triangular orbits', although false, contains a kernel of truth (e.g.
that the planets have closed orbits); the hypothesis 'The planets move in
circles' is likewise false, but intuitively is closer to the truth than the first one.)
In the literature various positions can be discerned. Some writers go so far
as to deny that our intuitive idea of one hypothesis being closer to the truth
than its rival is fruitful. (For example, K. Hiibner, A. J. Ayer and G. S. Robin-
son hold this view.) The rest would certainly agree that it is fruitful. Some of
them have shown that the problem of explicating our intuitive idea via a con-
cept introduced by a formal defmition (Le., a defmition formulated in an
idealized language schema, IL, based on standard logic) remains unsolved
PROGRESS AND RATIONALITY IN RESEARCH 61
1.6.1. In this situation it seems appropriate to recall the real issue here: our
aim is to legitimate (in the sense of giving good reasons) a preference rule,
which is based on a rule of theory appraisal- at the moment in the informa-
tion-theoretical and epistemic dimensions. Hence the adequacy of the expli-
cata proposed must be judged in terms of their fruitfulness for this meth-
odological task. Throughout, of course, we must clearly distinguish between
two sorts of tasks, that of explicating concepts and that of developing indica-
tors.
are needed) i.e., the set of all conjunctions of appropriate statements of initial
conditions and the negation of a hypothesis deducible from T in the presence
of auxiliary hypotheses A, so that this conjunction contradicts the conjunc-
tion of Tand A.
Thus it appears advisable to stipulate that the assertion that T' is better
than T in the information-theoretical dimension is to mean that the informa-
tion conveyed by the set of potential falsifiers for T' with the auxiliary
hypotheses A necessary for deducing them is greater than the information in
the potential falsifiers deducible from T and A. It is presupposed here that
the hypotheses deduced concern scientific problems and that the theories are
competitors, i.e., that they attempt to answer the same questions.
This yields a preference rule: before testing, prefer that theory for which
there are good reasons to conjecture that its empirical content (in the pre-
sence of the necessary auxiliary hypotheses (part II, Section 2.1.1) is greater
than that of its competitor, that it is more falsifiable than its competitor.
1.6.2.2. All this has to do only with the issue of explicating concepts. The
good reasons for such a conjecture hinge upon the use of some indicator
(fallible but objective) of relative falsifiability. Such an indicator is, in prin-
ciple, provided by the information conveyed by each of the sets of potential
falsifiers thus far deduced as potential answers to our scientific problems.
Thereby the proposed explicatum has plainly guided the production of
indicators - as should be the case. In actual practice, of course, there is no
usable measure of a unit of infOImation, and comparing two theories here
seems feasible only if one theory entails the other. On the other hand, there is
no point in content comparison anyway unless the theories are competitors,
and ultimately the decisive question is whether what T' says in answer to
the problems common to T and T' is a more accurate representation than
what T says; what counts is the situation after testing. In brief, the proble-
matic of content comparison seems, at least to this writer, to have received
undue attention considering that the global aim is to legitimate a preference
rule for the situation after testing.
In conclusion we can now return to the distinction initially mentioned.
The information in a theory, the group of formulated theses constituting it,
or, more accurately, constituting a particular version of it, is so condensed
that in practice it is difficult to test it directly. The more general and the
deeper a theory is, the more highly condensed the information will be. This is
why the whole business of deducing testable consequences has to be gone
through to make empirical criticism possible at all. Since every sentence
(trivially) entails itself logically, the theses themselves are of course included
64 GERARD RADNITZKY
in the set of the theory's consequences though not in the set of its testable
consequences. However, although for our present aim only the content of the
potential falsifier (or, in certain contexts, of the potentially falsifying hypoth-
eses (Part II, Section 2.2.2» matters, the two intuitive ideas - the content
of a theory and the set of its consequences - are not identical. And it seems
inadvisable to make them identical by defmition. The hypothesis that a
particular formulation of a theory has such-and-such content is again a con-
jecture for which good reasons can be provided, good reasons to be based
upon our interpretation of the theory's theses. Making such a conjecture is a
hermeneutic task. The conjecture may be criticized, inter alia, by confronting
it with the information in the consequences including that in the 'metaphysi-
cal' component of the theory.
1.6.2.3. The above preference rule is applicable in the situation before test-
ing. The simplest case would be that in which the two competitors give incom-
patible answers to the same scientific question. In this case, the theory that
gives the correct answer or a more accurate answer than the rival does would
be preferred. This consideration goes beyond the information-theoretical
dimension. In the epistemic dimension, the assertion that T' is better than T
is to mean (Le., it is so explicated) that T' is closer to the truth than T. But
this is just another way of saying: by 'truthlikeness' we mean that the theory
is nearer to the truth whose answers to our scientific questions represent the
relevant aspects of reality more accurately than those given by the competitor
theory. The preference rule is: after testing, prefer the theory for which the
conjecture that it is closer to the truth than its competitor is supported by
good reasons. These good reasons will make reference to an indicator (fallible
but objective): whether the potential falsifiers (or the potentially falsifying
hypotheses (Part II, Section 2.2.2» thus far deduced have or have not stood
up to the empirical tests thus far carried out.
This is all that is needed to reach our aim. It is presupposed that in our
language the idea of truth and the concomitant idea of truthlikeness (mehr
zutreffende Darstellung als) function successfully. This may seem problematic.
But it cannot be stressed too strongly: without the descriptive function
(Darstellungsfunktion) there is no language in the full sense, no human
language. Karl Buhler's work (1934) is highly pertinent here; Tarski's famous
semantic definition of truth, pace Popper, is not. 15 It is, of course, a task of
philosophy to clarify and explicate the idea of truth and its derivative con-
cept of one hypothesis being closer to the truth than another. This is indeed a
perennial task of philosophia prima. It cannot be a task for methodology
(cf. Section 0.2, s.f. on the division of labor). Nor can methodology wait until
PROGRESS AND RATIONALITY IN RESEARCH 65
'fust philosophy' has produced answers that are deeper and more relevant
than those it has produced during the last two and a half thousand years.
Hence to reject these concepts because we have not been able to provide (and
perhaps never will be able to provide) an explication of them such that the
explicatum is introduced via a formal defmition couched in an idealized lan-
guage schemaIL based on standard logic would be pathetic, ineffectual,
indeed even quixotic. Moreover it would be self-stultifying: the very argu-
ment for rejecting them would use the concepts of truth and truthlikeness
and would presuppose their functioning - surely such an argument would
claim to be true, correct, if it were to be taken seriously.
1.6.2.4. Thus we can conclude this section by returning to our starting
point, the typology of positions on the issue of truthlikeness. I would join
H. Albert and G. Andersson, but would also conjecture that the attempts to
explicate the idea ' ... is closer to the truth than .. .' in terms of an IL will
carry with them a repetition of the degenerating problem shifts we have
witnessed in the logical positivists' attempts to explicate the concept of
'empirical significance' by means of the IL : a host of problems will be induced
by the very instruments introduced in order to solve the originall~xplicatory
problem. 16
2.1.1. Realism. We do not use the customary label 'critical realism' because
naive realism, the view that the world is what it appears to be, has been dis-
carded since antiquity. Realism is a necessary presupposition of methodology.
Falsification presupposes the idea of error, hence that of truth and of the idea
that one hypothesis may be a more or less accurate representation/description
of certain aspects of reality. Whether or not a hypothesis is falsified depends
on reality: the idea of an experiment is that reality 'gives an answer' which is
independent of human influenceP (To take a trivial example, whether or not
gold is heavier than iron is something that cannot be influenced by human
beings.) Ontological realism as the thesis that material entities 'exist' in the
full sense (are, inter alia because of their independence, given full ontolOgical
status) or as the thesis of the existence of the external world and of other
minds is scarcely contested. Its main support is the unattractiveness of its
denial since ontological idealism relentlessly leads to solipsism, a position,
which, even if it may be consistent, is patently absurd. Also its corollary, the
thesis that only what (physically, materially) exists can be the object of cogni-
tion for the natural sciences, appears scarcely controversial. Epistemological
realism is the thesis that at least the properties of physical entities, physical pro-
cesses and the 'reality' of a physical event are independent of any process of
cognition, in particular of observation. This sort of independence is a precon-
dition for the possibility of objective indicators of comparative truthlikeness.
One can combine epistemological idealism (the denial of epistemological
realism) with ontological realism. This position implies an instrumentalistic
view of scientific theories. It has become increasingly popular among quantum
physicists. Popperians would not distinguish 'theory realism' as a special
sort of realism since such a distinction is the result of the artificial distinction
between 'theoretical language' and 'observation languages', of the two-
language approach characteristic of positivistic-foundationalist philosophers
who wish to give 'observation sentences' or 'observation predicates' a pri-
vileged epistemological status. For Popperians a theory - this holds good for
theories of quantum mechanics no less than for Newton's gravitational
hypothesis - talks about the world and makes truth claims which are in
principle the same as those made on behalf of a data sentence deduced from
the theory in presence of suitable additional premisses. The epistemological
idealists hold that the theories of microphysics do not represent/describe
aspects of the micro-world, but are, rather, nothing but instruments for
PROGRESS AND RATIONALITY IN RESEARCH 67
instrumentalist view does not follow from quantum theory, the intellectual
motive underlying it stems from the historical situation of quantum theory.
Elementary particles have properties which do not fit in with the world-
picture, neither with that of common sense nor with that built out of the
contributions of classical theories. Hence the question 'What sort of entities
are they?' becomes disturbing. All difficulties are avoided if one holds that
the statements about the behavior of elementary particles are not descriptive,
that they are nothing but fictions to be used as instruments for predicting
what will happen in the laboratory when a certain experiment is carried out;
in the last resort they are predictions about the perceptual experiences of
the experimenters. If this gambit is adopted, the question of the ontological
status of these entities does not arise - they are but fictions and the theories
nothing but black boxes used as instruments. If so, then the results of quantum-
physical research have no repercussions on the level of world-picture. The
task to examine the mteraction between the results of quantum-physical
research and our world view is eschewed. Hence instrumentalism is a lazy
philosophy. (It may be convenient for experimental researchers if it is used
only as a short-term moratorium on metaphysical questions in the hope later
on to be better equipped to deal with them.)
For Critical Rationalism the ontological pillar, realism, is a philosophical
presupposition of methodology. The defence of the realist posit cannot be
the task of methodology because a methodologist embarking on this enter-
prise has eo ipso turned ontologist. Nor can methodology aspire to develop
ontological analyses of, say, acts of perceiving, thinking, etc. The division of
labor within philosophy requested in Section 0.2 and Section 1.5 is indeed
indispensable. But methodology gives problems to ontology: e.g. Popper's
precious insight that certain problems are literally discovered poses the
problem, for ontology, of accounting for their partial independence (of
Popper's world-3 entities), which problem is at the same time a test case for
any ontological groundplan.
o. Among other things, I wish to make good the claim that Popperian meth-
odology implies pluralism rather than pure falsificationism, hence that it is
not a form of 'logicism'. Two diagrams will be used as a means of exposition.
PROGRESS AND RATIONALITY IN RESEARCH 73
f
IDEAL OF SCIENCE
/_.---..... EXPLICATION OF
/' CRITICAL,
I SCIENTIFIC PROGRESS
( RULE SYSTEM /" RATIONALISM
meth~ology ) /'."" REAL'SM I
(
/·~nlj..................... FALLIBILISMI MORE AND MORE ACCUAATE
< conventionalism
.ppriliUI
:,;- REPRESENTATION
. / .......................... ,/"/
,
'''''''''''''>
lUll /'/
y
I
......., . / ' I
Y i
world picture hVPQlhne'S ("metil phYI'CS'"
BACKGRO"ND
KNOWLEDGE
scienullc thrones about rewarch irei X:X
'l
~
1-+-1 -+-"'- ---~
I 1
, <p
I I ··1 P il
I 1
r-
r
-----n:iEccoC-Rl--A,."','-.A-IS-A-L----++-.
I 1
1
....,...
phil_ion
I
I
IL..-~)""
I
I
-'.........
cl'it:icitm
I
B i_i+1
I
I
I
I FALSIFICATION """.. 'oFH
~----------
CORROBORATION ..'...... ------- 0 FH
x x x x
Fig. 1.
(This figure copyright © 1978, G. Radnitzky.)
76 GERARD RADNITZKY
appraisal of the output comes afterwards, i.e. intermittently during the various
stages of interim products. Only the structural characteristics of hypothesis
generation are the concern of the methodologist. In Part I (Section 0.2) we
mentioned that Popperians by no means treat the hypothesis generation as a
black box whose study is to be left to the psychologist. Rather, psychological
investigation or hypothesis formation needs the methodological study of
structural characteristics (world-3) before getting off the ground. 34
In Fig. 2 (p. 82) hypothesis generation is represented by a box marked
'HG'. This box has inputs from all levels of background knowledge. Hence its
operation must include a selector. In Figure 2 this is symbolized by a fre-
quency fIlter. The selection will be governed by the prior assumptions about
the general nature of the phenomena studied. In Figure 2 this is indicated by
the 'internal steering- factors' JSF (whose operation has been mentioned in
Section 1.0). When working within a research tradition, a selection criterion
for both input assumptions and output hypotheses will be their coherence with
the cosmological hypotheses of the tradition in question. When a dramatic
shift in perspective is in the offmg, the assumptions about the phenomena
under investigation may conflict with the cosmological assumptions of the
reigning tradition since an output theory (Ti+ 1) may conflict with the pre-
decessor theory.
A pOSSIble source of input into the 'hypothesis-formation box' are theories
about other sorts of phenomena, which may belong to the same or to another
discipline. Hence in Figure 1 this input is shown as coming from Y. (i = abbre-
viation of the diScipline studying the realm X, the X-ology in question, Y=
abbreviation for studies other than the one in which the theory under consi-
deration is developed.) Example: Bohr conjectures a possible analogy between
atoms (a realm about which no scientific knowledge existed at the time) and
the planetary system (a realm about which full-fledged scientific theories
were extant). Such a guess would belong to what is here labelled 'JSF'. It
governs input from the extant theories about planetary motion into HG to
generate hypotheses about electrons, which are then subjected to falsification
attempts. One expects that if there is an analogy there will also be negative
aspects to it. Whether there are, and, if so, how far the analogy can be carried
must be found out by empirical testing. In this way a basic conjecture of
analogy provides a structural heuristic governing selection of input from
neighboring scientific theories as well as experiments.
The overall operational characteristics of hypothesis formation (HG box)
will be governed by methodological rules; in particular by the 'preference
rule' (cf. Figure 1), because in order to make severe testing possible the out-
78 GERARD RADNITZKY
put theory should have as much empirical content as possible: preferred are
'daring conjectures'.
We said above that in hypothesis formation 'everything goes' since criticism
follows upon each interim product. (In actual research there is a continuous
interplay between hypothesis formation and criticism - which cannot be
mirrored by Figure 1.) However, inductive procedures will not be used - not
even on the level of mental processes (world-2). This of course is a question
for the psychologist to investigate, an interface between psychology of re-
search and history of science. The methodological considerations strongly
suggest that in mental processes a number of observed cases will only be
recognizable as similar if at least an implicit, perhaps often subconscious
conjecture has been made that they are entities that have something in com-
mon. That means that hypotheSis formation as a mental process will never
proceed in a way that would correspond to the algorithmic functioning of a
Baconian inductive machine. (That is why Popper puts 'observational' law
between shudder quotes; as there are no theory-independent descriptive state-
ments, so there are no inductively, 'purely observationally-experimentally'
produced law hypotheses.)
1.4. Appraisal of the Successor Theory
How is the output of the hypothesis-generation-station to be subjected to
criticism, appraised? Most commentators regard this as the core of Popperian
methodology. We hope to have made it apparent that the critical element is,
however, comprehensible only as an integral part of the prescriptive picture
of research. Obviously appraisal makes sense only in the light of a guiding
ideal of science, an explicatum of progress. This ground has been covered in
Part I. Figure 1 gives a skeleton diagram of research, Figure 2 a blow-up of the
component theory appraisal.
The output of the hypothesis-generator automatically creates an objective
problem: to examine the output critically, to appraise the successor theory.
This criticism is basically a falsification attempt. It has two logical steps. To
answer the question, 'What exactly does the theory say?', we have to deduce
testable consequences, and in particular, consequences which enable us to test
the theory severely. This stage is represented in Figure I by the left box within
the box marked 'theory appraisal'. Then we have to answer the question, 'Is
what T says true?', by ascertaining whether or not it stands up to the empiri-
cal tests made possible by the consequences deduced in step one. In Figure 1
the second stage is represented by the box marked 'HC' (as short for 'hypoth-
esis-checking/control'). The HC-box is provided on the output side with a
PROGRESS AND RATIONALITY IN RESEARCH 79
decision triangle. The plus and the minus signs indicate the answer decided
upon to the question, 'Did the attempts to falsify the potential falsifier (de-
scribing an individuated event) succeed or not?' If it has not succeeded then
we have a falsifying hypothesis (describing a reproducible effect, which has a
higher level of generality than the potential falsifier, which is a singular state-
ment). In both figures the falsifying hypothesis is symbolized as 'FH'. If we
accept pro tempore the falsifying hypothesis, then there must be at least one
false premiss in the argument for the consequences whose negation constituted
the potential falsifier. This again automatically creates an objective problem;
the so-called Duhem-problem: on which of the premisses to put the blame.
In Figure 1 this is represented by a decision triangle at the output side of
the theory-appraisal box. The minus and the plus signs indicate the possible
answers to the question, 'Should the blame for the falsification be put on
premisses other than the theory under test?' If this question is answered in
the negative, the theory is falsified - relative to the falsifying hypothesis,
that is, provided FH is not problematized. Since all empirical hypotheses are
fallible, revision of FH is also in principle possible. Falsification cannot be
conclusive since at least some of the premisses are not conclusive. Figure 1
renders this circumstance by means of an arrow from the 'revision clause'.
In Figure 1 two clear-cut cases are portrayed. In one, the falsification
attempt has succeeded - which automatically produces a new objective
problem: Pi has been transformed into Pi+ 1. With the new problem a new
turn in the research enterprise begins, an iteration of the feed-back loop that
has just been completed. But the new problem is more advanced, if only
because one of the inputs into the HG-station is now the knowledge gained
by the preceding investigation, including the falsification itself. This novel
knowledge can help to refine the theory falsified, or, if you please, help to
produce a refmed version of it. In the second case, the falsification attempt
has miscarried. Here the methodology advises us to regard Ti+1 as corrobo-
rated to some degree - the degree of corroboration depending upon the
severity of the testing - and to retain it until further notice, i.e., to continue
to test it, to generalize it - but above all to regard it as preferable to the pre-
decessor theory, at least in this respect.
1.5. Possible Repercussions on the 'Cosmological Hypotheses' on the World-
Picture Hypotheses
Figure 1 indicates also that the replacement of the predecessor by the suc-
cessor theory may have repercussions on our 'background knowledge.' It
may have philosophical implications which feed back to the 'cosmological
80 GERARD RADNITZKY
THEORY APPRAISAL
,/
"- '-
r ---.----- --<./'
;:.....
preference -........ '>
rule ,/
<Jl '- ,/
I V
.i
"BACKGROUND KNOWLEDGE"
cp-
x x x y
Fig. 2.
(This figure copyright © 1978, G. Radnitzky.)
PROGRESS AND RATIONALITY IN RESEARCH 83
leaves us with the options already mentioned above (in Section 1.4) as the
so-called Duhem-problem. We will postpone discussing it until Section 2.3.
made. Successful explanation is not a test of the theory. The idea of 'severity'
is built into the idea of a test: to constitute a test of the theory, the informa-
tion of the deduced hypothesis must go beyond what we claim to know on
the basis of background knowledge.
The intuitive idea of background knowledge is roughly that it is all puta-
tive knowledge available at a certain time. In Figure 1 this is represented by
the horizontal beam which is labelled in the right margin 'background knowl-
edge'. In Figure 1 various inputs from this background knowledge are por-
trayed, with the corresponding selection mechanisms - for in a particular
research enterprise only a small segment of 'what is known or assumed to be
known' will be relevant. Moreover what is judged relevant will vary with the
research tradition and of course with time, with the state of the discipline.
This maximal interpretation of 'background knowledge' is seen to be proble-
matic when we remember that one desideratum is 'boldness' of theories, i.e.,
that a theory should make predictions which are 'unlikely' in the light of our
expectations based on what we think we know. If it is successful, we may say
'based on what we thought we knew'. Perhaps no more should be required
than the following: background knowledge is the set of those theses which
(are relevant for the scientific problem at hand and which) there is at present
no reason to doubt. If more is required, viz. that we have reason not to doubt
them, we seem to get entangled in circularity: if we were to require that they
be theses that so far have a sufficiently high degree of corroboration, an expli-
cation of 'degree of corroboration' and hence of 'severity of testing' and
hence of 'background knowledge' would already have to be at our disposal.
While the 'maximal' interpretation of 'background knowledge' appears
problematic, the minimal interpretation is unproblematic. Testable conse-
quences are deduced from the theory in the presence of auxiliary hypotheses
and statements of initial conditions. The minimal construal of our intuitive
idea of background knowledge explicates it as the information of the addi-
tional premisses, which at present we have no reason to doubt and are neces-
sary for deducing testable consequences from the theory, i.e., for deducing
consequences that constitute at least potentially an answer to our current
scientific problems. 39 In Figures 1 and 2 background knowledge in this
minimal interpretation is depicted as the output from the various selectors
whose input comes from the beam labelled 'background knowledge'. The
requirements that the consequences not be deducible from the additional
premisses alone, Le. that the theory's empirical information go beyond that
of the additional premisses, is obviously needed since we want to test the
theory, to extract information from it and not from the other premisses,
PROGRESS AND RATIONALITY IN RESEARCH 85
2 this stage is represented by the box marked 'deducing'. Its output is a grow-
ing system of deductive arguments such that the premisses of each argument
have a common component: the theory (Ti+l in Figure 2). The conclusions
are hypotheses that describe a reproducible effect (Vorgang). The arguments
may be either explanatory or predictive. In Figure 2 one pattern T and A -+
Ho ,40 portrays an explanation; the other patterns predictions (HI," . ,
Hn)·
In an explanatory argument the conclusion is 'known', i.e. it has evidential
support independent of this argument. An interesting case is that in which,
although the pattern is still an explanatory one, some new knowledge is gen-
erated because the original explanandum gets corrected by the explanation.
Popper has emphasized this most significant phenomenon of the improve-
ment of an 'observational' law in explaining it with a more universal law or
theory since at least 1941,41 and has also pointed out its close link to the
idea of depth: " ... whenever ... a new theory of a higher level of universality
successfully explains some older theory by correcting it, then this is a sure
sign that the new theory has penetrated deeper than the old ones".42
This improvement by explanation is best dealt with by means of a simple
example. Let the explanandum E by Galileo's law of free fall, i.e. his conjec-
ture based on experiments and on many thought experiments that the accele-
ration of a freely falling body in a vacuum on earth is constant. Let Ho in
Figure 2 portray the Newtonian version of Galileo's law. Ho contradicts E.
But within a certain realm, for falls where the height above the surface of the
earth is negligible in comparison with the earth's radius, the numerical values
of E are so close to those of Ho that this constitutes a good reason for con-
sidering E mathematically (and empirically) as an approximation of H o , the
hypothesis deduced by means of Newton's theroy, or Ho as an improved
successor of E. Although we say that Newton's theory explains Galileo's,
what is 'assimilated' into Newton's theory as one of its consequences can only
be the improved successor hypothesis of E. This is still more obvious if we
consider Einstein's theory and Newton's. For a certain realm the consequences
deduced by means of Newton's theory may be considered mathematically
(and empirically) as approximations of the consequences deduced by means
of Einstein's. But this relation of approximation holds only for certain
deduced consequences of both theories. It would be inappropriate to take
this as a good reason for considering Newton's theory, its basic theses, as an
approximation to Einstein's since the latter introduces new concepts, new
ways of looking at the world that are incompatible with the Newtonian
perspective (Part I, Section 1.4).
PROGRESS AND RA TIONALITY IN RESEARCH 87
The hypotheses HI, ... ,Hn differ from Ho only in that when the deduc-
tions are made they do not possess any evidential support. They constitute
potential knowledge, so to speak. The respective deductive arguments yield
virtual predictions: if Ti+ 1 is a conjecture that has not yet been tested, the
consequences deduced by means of it cannot profit from the theory's degree
of corroboration. But in the present context the predictions function primarily
to test the theory and we are only secondarily interested in the pieces of
novel knowledge they may convey. Assuming that we do not have reasons now
to problematize the additional premisses, the negation of Hi, viz. J and not
P constitutes a potentially falsifying hypothesis for Ti+l.
How can we find out whether the negation of H can be generalized to a
falsifying hypothesis describing a reproducible effect? Even HI , describing
a reproducible effect, condenses too much empirical information to permit
testing it directly. Thus again from HI in the presence of statements of initial
conditions (singular sentences) consequences are deduced, hypotheses of such
a low degree of generality that we can directly confront them with the 'facts'.
We can call them 'data' sentences; they have singular form and describe an
individuated event. Such a data sentence is not a report of a perceptual
experience. Reports about perceptual experience (e.g. about my own or
NN's perceiving a certain pointer reading) do, as Popper long ago pointed
out, form part of our reasons for accepting a certain data sentence, i.e. for
seeing no point in questioning its correctness for the time being. Other parts
needed in these good reasons will involve a theory of the hardware instru-
ments used, of their functioning and the causal link between object of study
and measuring apparatus, a theory of perception, and a theory of communica-
tion, since more than one observer is involved and the findings have to be
formulated in language to be communicable at all. We shall return to this
presently.
At this point the process of deduction ends. By negating the deduced data
sentence and combining it with the statement of initial conditions (J & -,P)
we get a potential falsifier (Falsijikationsmoglichkeit) for H. The potential
falsifier describes an individuated instance of a kind of state of affairs
(Zustand) or process outlawed by the law hypothesis H. If even a single
individual belonging to its realm of application exemplifies that state of
affairs or effect, then the law hypothesis is falsified unless the blame is put on
a different premiss (or premisses). Although this is so in logic, in actual re-
search more is needed. Before returning to this we wish to make two com-
ments.
We speak of a theory, e.g. Ti+ 1 in Figure 2, when we assert that the
88 GERARD RADNITZKY
2.2.0. Comparing the negation of the deduced data sentence with the experi-
mentally obtained data sentence; the role of technical data-generating systems.
The logical structure of this component has been dealt with in Section 2.0.
From the logical point of view a singular data sentence, incompatible with
a data sentence deduced from Hl in the presence of statements of initial
conditions, J, if assumed true, permits the deduction of either not-H 1 or
not-J. However, in actual research more must be requested. First, of course,
the potential falsifier, the conjunction of the negation of the deduced data
sentence and the statement of initial conditions, has to be empirically tested.
As mentioned in Section 2.0, an experiment will yield a positive result,
e.g., J and P*, a statement of initial conditions together with a statement
describing the observed event; and if p* logically entails not-P (if I- p* ~ ,P),
P is falsified unless we go back to question the correctness of the statement of
initial conditions. Thus basically the test of the potential falsifier consists
in an attempt to falSify it by comparing it with an observed, an experimentally
established data sentence about the type of event in question. In Figure 2 the
SW box within the right box (empirical testing) depicts this comparison of
the negation of the deduced data sentence with the experimentally obtained
data sentence (where J is not problematized for the moment). The input into
this box is the result of our experiment (in Figure 2 d*). To get an experi-
mental result one needs hardware instruments, technical data-generating
systems. In Figure 2 these are represented by the box DGS. A technical data-
generating system is based upon some physical theory (Ty in Figure 2).
Hence in an obvious sense all data are 'theory-Iaden'to some extent. 43 But the
90 GERARD RADNITZKY
data are thereby made dependent not upon the theory under test, but on the
various theories used (not in this context tested) in this theory appraisal. The
theories used are theories underlying the construction and manipulation of
the hardware instruments as well as theories used as software, e.g., logico-
mathematical techniques (used mainly in the information-theoretical part of
the appraisal) or statistical theories used in connection with data-generation. 44
If only because the theories underlying the data-generation are in principle
fallible, data sentences that could not be produced without these theories
must themselves be fallible in principle. There can be no epistemically privileged
'basic' sentences.
2.2.1. At the output side of the 'comparison station' there is a decision point:
the decision concerns the answer to be given to the question, "Has the attempt
to falsify the potential falsifier (J & -P) succeeded?" If the answer is positive
(in Figure 2 indicated by a plUS-Sign at one end of the decision-triangle),
i.e., if the negation of the deduced data sentence has been falsified (J being
considered to be fulfIlled), this potential falsifier is to be rejected and the
theory has received corroboration. If, on the other hand, the potential falsifier
(in Figure 2, not-d 1) has withstood testing, then the researcher automatically
has the new problem of finding out whether or not the potential falsifier,
thus established, can be processed into a falSifying hypothesis (in Figure 2,
PPH)·
Before continuing, a remark may be in place about the accusation of
conventionalism often voiced in the literature with respect to the procedures
just outlined. The decisions involved are neither based merely on conventions
nor do they give rise to conventions. They are based on good reasons. In
principle any of the data sentences may be questioned any time. But it would
be stultifying to do so without good reason. For instance, one may decide to
stop re-checking a particular statement about initial conditions or re-checking
the experimental apparatus simply because there are at the moment no
reasons for supposing the statement to be incorrect or the apparatus to be
working improperly. This may turn out to be wrong - and may get corrected.
But here a convention has been established only in a Pickwickian sense. The
decisions are based on good reasons; they are controllable and corrigible and
if necessary often get corrected by the objective method of falsification. (As
was pointed out in Part I, this correction is objective since the outcome of a
properly conducted experiment cannot be changed by human intervention.)
Decisions are unavoidable in research. But if the moments of decision were
considered to be all that mattered in research, then this methodological view
PROGRESS AND RATIONALITY IN RESEARCH 91
2.2.2. Let us return to the problem we left: Can the potential falsifier be pro-
cessed into a falSifying hypothesis? (PFH in Figure 2.) First it is conjectured
that J together with the experimentally obtained data sentence, which con-
flicted with the data sentence deduced from HI, describes an instance of a
reproducible effect. This is again a move that can be labelled 'hypothesis
generation' (in Figure 2, the boxHFHG) whose output is more general than a
data sentence but less general than the hypothesis HI . It describes a repro-
ducible effect; it claims that for all occasions the conditional (J ~ P*) obtains.
It is a potentially falsifying hypothesis for HI and A. (In Figure 2, hi in the
upper middle of the right box.) In order to get it corroborated, if possible,
we have to test the conjecture, attempt to falsify it. In Figure 2 the testing
process is portrayed in the box HFHCE. For the purpose of testing, further,
additional data are requested from the data-generating system (arrow from
HFHCE to DGS). The data obtained constitute one of the inputs into the
testing station. At the output side is again a decision point. The minus sign at
one end of the triangle represents the case where the attempt to falsify h I has
miscarried. We have now a falsifying hypothesis proper, not only a potentially
falsifying hypothesis. This process can also be described in sociological terms
or, better, it has its counterpart in Popper's world-2: To secure the objective
status of the falsifying experiment, to secure intersubjectivity, other re-
searchers must be able to repeat the experiment with the same result. Other-
wise the scientific community will not accept the hypothesis as a falsifYing
hypothesis. But when a hypothesis has been accepted (pro tempore) as a
falsifying hypothesis, the severity of any future repetition of the same
empirical test will be zero.
A law hypothesis may get corrected in attempting to explain it if the con-
clusion of the explanatory argument can be considered an improved successor
of the original explanandum. Analogously for the falsification of a predicted
hypothesis: it may give rise to a refmement or, if you please, to an improved
successor of the predicted bypothesis, which was falsified. Again a very simple
92 GERARD RADNITZKY
mass in the sun. This good reason becomes almost compelling if, in addition
to this, the successor T* has been corroborated, has gotten independent
evidential support by making possible the deduction of novel knowledge. But
this development is not a necessary condition for having good reasons to
abandon T, i.e., for rejecting the rescue operation, which introduces A '. If A '
does remain ad hoc, then, even without any available successor theory, T
would have to be considered falsified, Le., to have some falsity content. But
since it has so many other successes to its credit it would still not be irrational
to continue to work with T or on refming T.
In Figure 2 the large triangle between the two boxes represents the deci-
sion point described by the Duhem-problem. Where to put the blame for a
falsification of a testable consequence is something only the researcher him-
self can decide. But methodology does have an uncompromising prohibition:
"If you decide to put the blame on the theory, you must never attempt to
repair the situation by reducing the empirical content of the theory. This is
forbidden by the master rule, the 'anti-conventionalist' rule which outlaws
any strategies that would immunize a theory against criticism." In Figure 2
the arrow from this rule to the 'Duhem-triangle' marks this guiding precept.
In addition, methodology tells the researcher that in constructing good reasons
for his decision everything depends upon whether or not the component
introduced in response to the challenge of the falsification - be it A' or T* -
remains ad hoc. If A' or T* has just been invented and lacks independent
evidential support, the ex ante hypothesis that it will or will not remain ad
hoc is a risky conjecture. The researcher himself is in the best position to
make such a conjecture.
2.4. The Situation after It Has Been Decided Whether or Not the Successor
Theory is Falsified, at Least in Its Present Form
If the successor theory is falsified, then this situation automatically creates
a new problem, a return to the problem that was the starting point of the
research enterprise in Figure 1, but now on a higher level. In Figure 2 this is
portrayed by the arrow which runs in the left margin from Tio+1 to Pi+ 1.
This turn brings us back to Figure 1, although Pi is now replaced by Pi+ 1. It
is needless to point out once again that falsification cannot be conclusive
since some of the premisses in the argument cannot be conclusive. It is fallible,
like all our empirical knowledge, yet it is perfectly objective.
If the falsification attempt has failed, the successor theory has thereby
been corroborated to the extent corresponding to the severity of the test
involved. Then the problem automatically arises of appraising the comparative
PROGRESS AND RATIONALITY IN RESEARCH 95
EPILOGUE
From the Quest for Certainty to the Search for Cognitive Progress:
The Non-Foundationalist, Meliorist View of Human Knowledge
With a brief glance backwards into the history of ideas, one can see in Popper
a Copernican revolution not only in methodology but also in our view of
human knowledge: the change from a justificationist to a non-justificationist,
fallibilist view of human knowledge. Impressed by the advent of the scienza
nuova and by the triumph of Newtonian physics, philosophers committed to
a foundationalist view of human knowledge saw their task in proving scientific
knowledge to be true. This quest was basically a theologoumenon: the idea of
scientific knowledge was modelled upon the idea of revealed knowledge,
which for the believer is infallible. This is reflected in the ordinary-language
use of 'to know' and is particularly clear in scientism. If a foundationalist
program is to be attractive, one needs an infallible source of knowledge. The
philosophers sought this either in an empiricist basis (e.g. Bacon) or in an
intellectualist foundation (e.g. Descartes).
Hume's insight that, even if a secure foundation were granted, since the
evidential basis is always finite, the justification of a general statement would
need an amplificatory move discredited the whole approach. A vain search
ensued for a Principle of Induction that might bridge the gap, and the strict
foundationalist was left with either infinite regress or resort to apriorism. This
intellectual experience opened the gates to irrationalism on a large scale, not
only in epistemology but also in political thinking - e.g. the disastrous
influence of Rousseau.
Kant attempted to rescue a place for rationality by means of apriorism, an
apriorism which in a way combined elements of both the above-mentioned
foundationalist scheme, with his synthetic a priori truth. But his examples
lost their status, some of them in consequence of scientific developments.
The logical empiricists a limine rejected the idea of a synthetic a priori.
They wanted to construct an inductive logic that could answer Hume's objec-
tions by yielding a measure of degree of inductive support applicable also to
96 GERARD RADNITZKY
general statements from a fmite evidential base. They were more impressed
by the way the crisis in mathematics had been resolved than by the crisis in
physics; and they took metamathematics as their model for philosophy of
science. They even flirted with the idea of finding a secure foundation in
observation sentences or observational predicates; and even if this founda-
tionalism has by and large been abandoned nowadays, their ideal of science
is clearly within the foundationalist line (cf. Part I, Section 1.1, 1.2).
Popper's work signified a Copernican revolution both in methodology and
in epistemology: he produced the first worked-out non-foundationalist view
of human knowledge. His criticism of both the solutions offered and the prob-
lems raised by logical empiricism has been dealt with in Part I, Section 1.3. A
critique of Kant's rescue attempt on Popperian lines would run: that the
synthetic a priori truths which for Kant were paradigmatic examples have
dissolved;48 that the question of whether there is a synthetic a priori needs
first a more precise distinction between a priori and a posteriori, which must
partly take recourse to psychology, and second a precise interpretation of an
alleged synthetic a priori sentence so that we could examine its status. But
the most decisive question is: is the conceptual structure one gets by crossing
'analytic/synthetic' with 'a priori/a posterion"' fruitful for methodology? That
does not appear to be the case - at least to this writer. Are the sentences
which Kant viewed as synthetic a priori truth 'conditions of the possibility' of
'experience'? To answer this query we would also have to clarify Kant's con-
cept of experience. However, we would transform the question into: Are
there conditions of the possibility of cognitive progress, and if so, which? A
Popperian would answer this last question in the affirmative, and would in
this sense be a Kantian. Our capacity for conjecturing creatively, for prolife-
rating proposals, as well as our capacity to learn from our errors- in sum the
critical attitude and the criticist method - he would see as conditions of the
possibility of cognitive progress, although not a guarantee of such progress.
After the foundering in principle of all foundationalist philosophizing, 49
how can rationality be restored to its rightful place? By the critical method.
Rationality can also be given a place in political activity by the method of
criticizing and of testing proposals. 50 This is of course a rejection of utopian
schemes referring to 'the totality' and ending in totalitarianism.
NOTES
Cf. Radnitzky (1978b), Section 1.2, and Radnitzky (1979b).
2 Radnitzky (1972) and Radnitzky (1974b) attempt to spell out this way oflooking at
research.
PROGRESS AND RATIONALITY IN RESEARCH 97
3 Cf. Watkins (1978) in Radnitzky and Andersson (1978); see also Radnitzky (1980).
4 This is the main thesis of Radnitzky (1968/1970, Vol. I); (see Radnitzky, 1972).
5 Since normative-evaluative utterances can be questioned in a way in which descriptive
statements cannot, this alone is a sufficient reason for the distinction between facts and
values, between 'Is' and 'Ought'. This is spelled out in Radnitzky (1979c).
6 Cf., e.g., Popper (1972), pp. 196-204, esp. p. 197.
7 Cf. Agassi (1975).
8 Cf., e.g., Popper (1972), pp. 204f; Popper (1963), p. 222; Popper (1975), p. 75.
9 Cf. Kant (1783), Section 57, in the edition Kant (1911), p. 352.
10 Cf., e.g., Popper (1963), p. 222; Popper (1972), esp. p. 118; Popper (1975), p. 75.
11 Cf. Popper (1963), p. 222.
12 This is spelled out in Radnitzky (1977), Section 6.
13 Cf., e.g., Griinbaum (1976) and Griinbaum (1978).
14 Cf. Andersson (1978) and comments on it in Radnitzky (1977), Section 1.34.
15 Cf. Radnitzky (1977), Section 1.34, on Popper's use of Tarski and also Section 1.2
on the good reasons for working with formalized languages.
16 Cf., e.g., Radnitzky (1968/1970), the chapter on Empirical Significance.
17 It cannot be influenced in the same way in which, say, a norm expressed by a state-
ment such as, e.g., 'It is forbidden to ... ' can be revised by human action. The connec-
tion with the distinction between 'Is' and 'Ought' has already been mentioned. The
importance of this distinction for liberal democracy, its role in the political philosophy
of Critical Rationalism, is examined in Radnitzky (1979c).
18 In Figure 2, X, DGS, together with the observer, all merge into a black box. For a
penetrating examination of this idea of the 'participation' of the observer in the 'crea-
tion' of reality cf. Jammer (1977).
19 For a brief critical examination of this trend the reader is referred to Jammer
(1977).
20 Cf. Heisenberg (1958).
21 Cf. Bunge (1977a), p. 151; see also Bunge (1973) and Bunge (1977b).
22 The left part of the box 'falsification attempts' in Figure 2.
23 Cr., e.g., Albert (1968/1975), pp. 11, 13, 15,24,27,31,35,56,61,72,170.
24 Cf. Hans Albert's critique: Albert (1975).
25 Cf., e.g., Popper (1972), and Popper and Eccles (1977).
26 Cf. Popper (1934-1959), p. 44 n. 1, and Popper (1934), p. 18.
27 Cf., e.g., Radnitzky (1974b), pp. 7f. The claim that the rules apply to all research
should not be taken to imply any form of a 'unity of science' thesis. We only wish to
imply that the sciences humaines - a very heterogeneous group - also contain certain
facets (explanation and criticism of explanations and descriptions) to which the model
of Figure 1 and 2 may be adapted. Again, to claim this is not to deny that the typical
humanities, the Geisteswissenschaften, e.g. philology, may have additional methodological
problems of their own not covered by these models: there is a great difference between
the task of 'giving an explanation of understanding' and the task of 'understanding expla-
nation'. One such peculiarity is the phenomenon of self-fulfilling (self-stultifying) proph-
ecy. Whenever a prediction refers to a process which may be influenced by the action of
those whose future behavior is predicted, communicating the prediction to these people
may influence the outcome. Hence, if it cannot be excluded that the predicted effect
may be due to a self-fulfilling prophecy, the prediction cannot constitute a test of the
98 GERARD RADNITZKY
theory by means of which the prediction was made, and hence a successful prediction
cannot corroborate the theory. The difference between, e.g., whether forecasting and
economic forecasting (if people believe the prediction of a bullish tendency on the stock
market, there will be a bullish tendency!) is a facet of the distinction between 'Is' and
'Ought' (cf. here too Radnitzky, 1979c. Section 2.1).
28 E.g. in Radnitzky (1972), (1974a), (1974b).
29 For example, if you have made the assumption that the real system studied suffi-
ciently approximates an isolated system, then you will suppose that a deterministic
system can be used to model it and it will be reasonable to require of your law hypo-
theses that they be univer8.l1 in form; if you regard it as an 'open system', you may
be willing to accept statistical laws. Such a 'metaphysical' posit will influence your
programmatic defmition of your discipline. E.g., your conception of what a human
being is will influence your view on whether the discipline of psychology should be
much like a natural science or should be a Geisteswissenschaft or a Handlungswissenschaft.
30 E.g., Popper (1972), p. 144.
31 Cf., e.g., Radnitzky (1972), Section 223, and Radnitzky (1974a), p. 86.
32 Cf., e.g., Popper (1972), pp. 160, 118f, 147.
33 Cf., e.g., Popper (1972), pp. 160, 118f, 147.
34 Popper's 'transference law', e.g., Popper (1972), p. 114.
35 Popper would be able to agree with this; e.g. (1972), p. 259 "growth of our knowl-
edge ... as consisting throughout of corrections al.J modifications of previous knOWl-
edge", or (the growth of knowledge) "is largely dominated by a tendency towards
increasing integration towards unified theories" (1972, p. 262), or p. 71 "the growth of
all knowledge consists in the modification of previous knowledge"; cf. also (1959,
~. 276). .
6 Usually rendered by the meta-sentence: I- IT &A -+ (I -+ P)).
37 This problem is dealt with in details in, e.g., Radnitzky (1979a), Section 2.2.
38 The inductivist's answer is 'always'. Because he wants to be able to stylize con-
ditionals with the connective for material implication, for' 'All swans are white' every-
thing that is not a non-white swan functions as a potential satisfier although intuitively
cases of non-swans are completely irrelevant for the 'credibility' or estimated truthlike-
ness of the conditional. Among other things, the inductivist's answer leads him to make
the paradoxical claim that the richer theory is easier to probabilify than the less rich
theory. Cf. Musgrave (1978).
39 Of course it would be pointless trivially to proliferate consequences by making use of
the peculiarities of the v-connective.
40 Expressed in the symbolism used in Section 2.0. Ho might be (/0 ""* Po).
41 Cf., e.g., Popper (1972), pp. 204f (first published in 1957).
42 Cf., e.g., Popper (1972), p. 202 or Popper (1975), p. 97.
43 Popper (1934/1959), p. 107, n. 2; Radnitzky (1974a), Section 2212 (p. 80) and n.
75 and Radnitzky (1974b), p. 24.
44 These 'auxiliary' theories or even the hardware itself can take the lead and research
can become 'governed' by the instruments instead of being oriented toward the original
scientific question. Such a problem shift, if temporary, may constitute an interlude with
a high growth rate of new knowledge, in which new hardware techniques and instru-
ments make possible the production of new knowledge, or new software, e.g. mathe-
matical techniques, make the deduction of novel conclusions possible. However, if it is
PROGRESS AND RATIONALITY IN RESEARCH 99
'totalized', i.e. if the instruments, hardware or software, take the lead for some time, the
ensuing problem shift deteriorates into a 'pathology' of the discipline concerned.
45 From a logical point of view there are three options; but in praxis the third option is
not interesting.
46 I-A'->-,A, I-[(T&A')->(J->P*)).
47 It can take quite a long time before such an issue can be decided. E.g., in the case of
the clash between Miller's experimental results about an alleged 'ether-effect' and a
system whose key component was Einstein's theory, it took about 25 years to find out
that one of the additional premisses used in Miller's falsification argument was falsified
in empirical testing (temperature variations in the apparatus).
4 8 As regards arithmetic, both logicist and formalist conceptions contradict Kant's
view; in physical geometry Euclid's system has lost its monopoly position; as regards
the propositions of 'reine Naturwissenschaft' (e.g. the principle of causality) the only
reasonable construal appears to be to interpret them as methodological advice.
49 In today's German philosophical scene the foundationalist approach takes various
forms. The so-called Erlangen School manifests the whole syndrome: a constructivistic
approach leading to the 'ortholanguage' (Orthosprache), 'protophysics' and a 'pragmatic
grounding' (pragmatische Begriindung) of sentences. According to K.-O. Apel's own
interpretation of his philosophy, it is a continuation of Kant in the pragmatics of lan-
guage and it attempts to provide a 'transcendental-pragmatic foundation' from the con-
ditions of the possibility of our ability to speak with each other. J. Habermas attempts
to reach the same goal with a 'quasi-transcendental/quasi-empirical' foundation, with
'compelling arguments' which could guarantee the sameness of opinions in the long run.
Thereby the idea of truth is played down and it is replaced by a consensus conception
of 'truth'. They all appear to commit the same basic error: discourses are taken to be the
forum for establishing truth and value (wertsch6pfende Instanz) - but discourses can
only be the forum for examining claims to truth or value (wertpriifende Instanz). This
would be the gist of a Popperian criticism of these contemporary trends in German
philosophy, which are again steeped in the foundationalist approach.
50 Cf., e.g., Radnitzky (1977).
BIBLIOGRAPHY
Aggasi, 1.: 1975, 'Between Metaphysics and Methodology', Poznan Studies in the Philos-
ophy of Science and the Humanities 1, 2-8.
Albert, H.: 1968, Traktat iiber kritische Vernunft (J. C. B. Mohr, Tiibingen, 1968; 3rd.
enl. ed. 1975).
Albert, H.: 1975, Transzendentale Traumereien. Karl-Otto Apels Sprachspiele und sein
hermeneutischer Gatt (Hoffmann & Campe, Hamburg).
Albert, H.: 1978, 'Science and the Search for Truth. Critical Rationalism and the
Methodology of Science', in Radnitzky and Andersson (1978).
Andersson, G.: 1978, 'Truth and Scepticism: The Problem of Verisimilitude', in
Radnitzkyand Andersson (1978).
Antiseri, D.: 1977 a, Epistem%gia e didattica delle scienze (Armando Armando, Rome).
Antiseri, D.: 1977b, 'Prova di una teoria ed educazione al risconoscimento dell'errore',
in Epistem%gia, metod%gia clinica e stona della scienza medica (Arti Grafiche
E. Cossidente, Rome), pp. 111-168.
100 GERARD RADNITZKY
Biihler, K.: 1978, Sprachtheorie (Gustav Fischer, Jena, 1934; repro UJlstein, Frankfurt,
1978).
Bunge, M.: 1972, 'On Method in the Philosophy of Science', in Dockx (1972), pp. 97-
120.
Bunge, M.: 1973,Philosophy of Physics (Reidel, Dordrecht).
Bunge, M.: 1977a, Foundations of Physics (Springer Verlag, Berlin).
Bunge, M.: 1977b, 'The Interpretation of Heisenberg's Inequalities', in Pfeifer (1977),
pp. 146-155.
Cohen, R., Feyerabend, P., and Wartofsky, M. (eds.): 1976, Essays in Memory of Imre
Lakatos. Boston Studies in the Philosophy of Science, Vo!. 39 (Reidel, Dordrecht).
Dockx, J. et al. 1972,De Iii methode (Office International de Librairie, BruxeJles).
Grmek, M.: 1976, 'Le role du hasard dans la genese des decouvertes scientifiques', Medi-
cina nei Secoli 13, 277-306.
Griinbaum, A.: 1976, 'Can a Theory Answer More Questions than One of its Rivals?',
British Journal for the Philosophy of Science 27, 1-23.
Griinbaum, A.: 1978, 'Popper versus Inductivism', in Radnitzky and Andersson (1978).
Harre, R. (ed.): 1975, Problems of Scientific Revolution. Progress and Obstacles to
Progress in the Sciences (Clarendon Press, Oxford).
Heisenberg, W.: 1958, 'The Representation of Nature in Contemporary Physics', Daedalus
87,95-108.
Jammer, M.: 1977, 'Physics and the Search for the Absolute', in The Search for Absolute
Values in a Changing World. Proceedings of the Sixth International Conference on
the Unity of the Sciences. San Francisco, Nov. 1977 (International Cultural Foun-
dation Press, New York, 1978).
Jammer, M.: 1979, 'A Consideration of the Philosophical Implications of the New
Physics', in Radnitzky and Andersson (1979).
Kant, I.: 1783, Prolegomena zu einer jeden kiinftigen Metaphysik, die als Wissenschaft
wird auftreten k6nnen; reprinted in Kants Gesammelte Schriften (Reimer, Berlin,
1911, Bd. 4).
Lakatos, I.: 1970, 'Falsification and the Methodology of Scientific Research Program-
mes', in Lakatos and Musgrave (1970), pp. 91-195.
Lakatos, I.: 1974, 'Popper on Demarcation and Induction', in Schilpp (1974), pp. 241-
273.
Lakatos, I. and Musgrave, A. (eds.): 1970, Criticism and the Growth of Knowledge
(Cambridge University Press, Cambridge).
Miller, D.: 1974, 'Popper's Qualitative Theory of Verisimilitude' , British Journal for the
Philosophy of Science 25, 166-177.
Miller, D.: 1977, 'Verisimilitude Redeflated', British Journal for the Philosophy of
Science 27, 378-381.
Musgrave, A.: 1974, 'The Objectivism of Popper's Epistemology', in Schilpp (1974), 1
560-596.
Musgrave, A.: 1978, 'Evidential Support, Falsification, Heuristics, and Anarchism', in
Radnitzky and Andersson (1978).
Pfeifer, H.: 1977, Denken und Umdenken. Zu Werk und Wirkung von Werner Heisenberg
(Piper, Munich).
Popper, K.: 1934-1959, The Logic of Scientific Discovery (Hutchinson, London);
revised trans!. of Logik der Forschung (Springer, Vienna, 1934).
PROGRESS AND RATIONALITY IN RESEARCH 101
Popper, K.: 1963, Conjectures and Refutations (Routledge & Kegan Paul, London).
Popper, K.: 1972, Objective Knowledge. An Evolutionary Approach (Oxford University
Press, London).
Popper, K.: 1975, 'The Rationality of Scientific Revolutions', in Harre (1975), pp. 72-
101.
Popper, K.: 1976, 'A Note on Verisimilitude', British Journal for the Philosophy of
Science 27, 147-160.
Popper, K. and Eccles, J.: 1977, The Self and its Brain. An Argument for Interactionism
(Springer International, Berlin).
Radnitzky, G.: 1968, Contemporary Schools of Metascience (Humanities, New York;
Esselte Studium Giiteborg, 1968; 2nd ed. 1970; enl. paper ed. Chicago, Gateway,
1973).
Radnitzky, G.: 1972, 'Towards a "Praxiological" Theory of Research', Systematics 10,
129-185 (repr. in enlarged 1973 ed. of Radnitzky, 1.968/70).
Radnitzky, G.: 1974a, 'From Logic of Science to Theory of Research', Communication
and Cognition 7, 61-124.
Radnitzky, G.: 1974b, Preconceptions in Research (Literary Services & Production,
London).
Radnitzky, G.: 1976, 'Popperian Philosophy of Science as an Antidote against Rela-
tivism', in Cohen et al. (1976), pp. 505-546.
Radnitzky, G.: 1977, 'Philo sophie und Wissenschaftstheorie zwischen Wittgenstein und
Popper' in ZeIger (1977) pp. 249-281.
Radnitzky, G.: 1978, 'The Boundaries of Science and Technology', in The Search for
Absolute Values in a Changing World. Proceedings of the Sixth International Con-
ference on the Unity of the Sciences (San Francisco, Nov. 1977) (International
Cultural Foundation Press, New York), (1978) Vol. 2 pp. 1007 -1036.
Radnitzky, G.: 1979a, 'Justifying a Theory vs Giving Good Reasons for Preferring a
Theory', in Radnitzky and Andersson (1979), pp. 213-256.
Radnitzky, G.: 1979b, 'Metodo', in Encyclopaedia of the Twentieth Century (Istituto
della Enciclopedia Italiana, Rome, 1979).
Radnitzky, G.: 1979c, 'Die Sein-5011en-Unterscheidung als Voraussetzung der liberalen
Demokratie', in Salamun (1979), pp. 459-493.
Radnitzky, G.: 1980, 'From Justifying a Theory to Comparing Theories and Selecting
Questions', Revue Intemationale de Philosophie 34, 179-228.
Radnitzky, G. and Andersson, G. (eds.): 1978, Progress and Rationality in Science.
Boston Studies in the Philosophy of Science, Vol. 58 (Reidel, Dordrecht).
Radnitzky, G. and Andersson, G. (eds.): 1979, The Structure and Development of
Science. Boston Studies in the Philosophy of Science, Vol. 59 (Reidel, Dordrecht).
Russo, E.: 1974, 'Typologie du progres des connaissances scientifiques', Revue des
questions scientifiques 145, 345-363,479-502.
Salamun, K. (ed.): 1979, Sozialphilosophie als Aufkliirung. Festschrift for Ernst Topitsch
(J. C. B. Mohr (Paul Siebeck), Ttibingen), 1979.
Schilpp, P. (ed.): 1974, The Philosophy of Karl Popper (The Library of Living Philoso-
phers. 2 vols.) (Open Court, LaSalle, Ill.).
Toulmin, S.: 1974, Human Understanding. Vol. 1: General Introduction and Part I
(Princeton University Press, Princeton).
102 GERARD RADNITZKY
Watkins, J.: 1970, 'Against "Normal Science" " in Lakatos and Musgrave (1970), pp.
25-38.
Watkins, J.: 1978, 'Corroboration and the Problem of Content-comparison', in Radnitzky
and Andersson (1978).
Zeiger, J. and Marek (eds.): 1978, Osterreichische Philosophie unil ihr EinflufJ auf das
analytische Denken der Gegenwart, Vol. I (Sonderband Conceptus, nos. 28-30).
JOSEPH AGASSI
The purpose of the present study is to illustrate the need to remodel our
studies of science, both philosophical and scientific, and together. I wish to
illustrate a point by an examination of a case history. Let me state, first, that
my history is going to be crude: the history I wish to oppose is crude because
it is a popular prejudice; the history I wish to present is crude because it is
only in outline there - I am no historian and I can do no better. Of course,
my view of the history of science may be a prejudice too: I leave it for others
to decide matters.
The thrust of my argument is this: the history of science as we are taught
it, is largely modelled after the history of physics. The history of physics as
we are taught it is the history of the validation of physics, and by way of
the validation of both Copernicanism and Newtonianism, more specifically
of Newtonianism with Copernicanism as a corollary to it. But what is New-
tonianism and how was it validated?
I shall come to all this soon. Let me now give a slight historical point that
I personally fmd intriguing. It is this. Today no physicist will include under
the heading of Newtonian physics the topic of the stability of the solar
system. Even historians of science hardly notice it. Yet it is at least a signifi-
cant link between the history of physics and the history of biology - via the
history of geology. It was the need to study the stability of the system that
has led to the need to study its origins, and it is the need to decide empirically
between competing hypotheses that led to the upheavals in geology that
preceded the Darwinian revolution. But let us leave the origins. It was one
of the revolutionary ideas of Alexandre Koyre that the stability of the solar
system was an essential topic of debate between the Newtonians and their
detractors, and it was therefore only after Laplace oiled the celestial clock-
work and put it on a firm foundation that he could tell Napoleon he could do
without its maker.
But was it so? Historians of science do not report the Laplacean proof of
the stability of the solar system. This may be because the proof is complex.
103
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 103-114.
Copyright © 1980 by D. Reidel Publishing Company.
104 JOSEPH AGASSI
But then they could do with an outline perhaps. Except that this might be
dangerous: after all, Jacques Hadamard did prove that Laplace's proof is
faulty in a sense. But I think things go deeper. Laplace's proof of the stability
of the solar system is very famous though hardly known. Yet Kant's proof
of the instability of the solar system is the other way round: it is far from
famous yet very well-known. That is to say, hardly anyone will be puzzled
if I told him that since gravitation causes tides and tides cause heat and heat
dissipates, and since energy conserves, gravitation dissipates. If anyone is told
that this proof is Kant's, he usually shows no interest. Not even in the face of
the fact that the law of conservation of energy is a mid-nineteenth century
discovery, namely, long after Kant's death. But what kind of history of science
do we have, where an erroneous proof is presented as correct, a seemingly
correct one suppressed and its foundations ascribed to later writers? It is, I
submit, a history of science devoid of problems. The reason so many histo-
rians of science suppress problems, I think, is that they are still doing mis-
sionary work, propaganda science. The result is a mess not only in history but
also in philosophy: science does not need propaganda, validation, excuses. At
least it should not repress the interesting problem: why was stability part of
Newtonianism that was later lost?
So much for the example. Let me repeat my thesis: the problem of valida-
tion of scienc.e or of a scientific hypothesis seems to me to be an archaism
sustained by the division of labor between philosophers and historians of
science: each of them relies on the other to validate their concern with valida-
tion. Let me now backtrack and start allover again, and slowly.
Before explaining why I think St. Robert, Cardinal Bellarrnino, is the man
who raised the problem of validation which I want to eliminate, I feel I must
introduce him again, and in a way shared by no historian of science known
to me. I am a profound admirer of the Cardinal, but I feel I cannot do him
justice unless I say at once I think that the legal murder of Giordano Bruno in
1600 is his almost exclusive guilt: he was ready to burn at the stake both
Bruno and Galileo; but he was not an evil man. As George Bernard Shaw said
of the legal murder of Joan of Are, tragedy is only possible when terrible
things are done by men of good intentions and hard thinking: a person shot
by a stupid criminal dies by a mishap akin to a natural disaster. Unlike a
stupid murderer, Bellarrnino cared for neither fame nor fortune: he cared, we
know full well, for the stability of the social order. And we all now agree that
THE PROBLEMS OF SCIENTIFIC VALIDATION 105
the facts validate - notice the word! - his thought that if enquiry is allowed
to roam free and discovery left unchecked, then soon the established social
order will be inquired into and questioned too, not only the established
physical doctrine; and soon after, he realized, discovery will bring forth social
ideas as well, not only new ideas about physical matters. Nor did he wish to
repress discovery and innovation. He only felt that Church hegemony about
such matters must remain quite unquestioned, and that the way to do so is by
guarding very jealously the need for each discovery to be validated by the
Church: no idea, he felt, should be allowed all freedom, unless it had the
Church's nihil obstat and imprimatur.
We know that this idea is a new idea, and we may safely ascribe it to the
not-so-saintly Saint Robert Cardinal Bellarmino. That it is a new idea is
declared by Galileo Galilei himself, in his celebrated samizdat essay, Letter
to Castelli, and its expansion, Letter to the Grand Duchess. Of course, we can
question Galileo's judgment. We can even say he is unreliable since he was an
interested party, especially since he accused the innovator - whom he leaves
nameless - of an ulterior motive, namely of the love of power. But I will
not go into that now: I assume that Galileo needed - and received - the
imprimatur only because of Bellarmino's say-so. I wish to stress that at least
Galileo saw here a major point of philosophic principle. He said he was pro-
posing to the Church of Rome the policy that it be quite indifferent to what
ideas a thinker endorses on matters of nature. The Lord wrote two books -
the Book of Nature and the Holy Writ - and so the two cannot be in any real
conflict. Hence, the guardians of the Holy Writ need not trouble themselves
with the Book of Nature.
Not only Galileo but also Bellarrnino saw this as the essential point, and
Bellarmino was willing to concede: unless we realize that Bellarmino was a
reasonable man we miss the point. Reasonably enough, Bellarmino felt that
things are less simple from the viewpoint of the Church than Galileo was
pretending (and others too, for example Kepler). Whereas Galileo, following
Pico della Mirandola's classic Oratio on the Dignity of Man, I suppose, quoted
Psalm 19 in defense of the study of the Book of Nature, Bellarmino too
quoted the same psalm to say that the sun revolves around the earth. Of
course, this can be reintegrated; but what should be reintegrated and what is
up to the official interpreter, namely the religious authority? Again, we must
admit, Bellarmino's idea was validated by facts: geology, biology, archeology,
all came in series and forced the scientific believers to decide between the
Book of Nature and the Holy Writ. Bellarmino was right in fearing that sci-
ence threatened his religion.
106 JOSEPH AGASSI
What was wrong with the development of the scientific movement is that it
developed its hagiography and had its saints and martyrs. And martyrs need
enemies. And so BelIarmino entered the history of science and his challenge
was taken up and answered. The answer is very very simple: Copemicanism
is indeed proven and GaIileo was right.
The answer, as I have just now stated it, is just terrible. It incorporates
all that is most erroneous and evil within science at its very best. First and
foremost, who is Bellarmino and why on earth should he be answered? To
THE PROBLEMS OF SCIENTIFIC VALIDATION 107
criterion of validation, not the whim of the scientist and less so the whim of
the Cardinal. What, then, is this criterion?
science. This was done by Michael Polanyi, who amongst scientists is better
known through the writings of Thomas S. Kuhn.
The alternatives to conservatism are two. First, radicalism, which can only
rest on a proper, fully attested criterion of validation existing and validated
already; second, reformism, which can rest on a proper and balanced attitude
towards the status quo. The radicalist should offer a validated criterion to
tell us when a theory or a hypothesis or an institution is properly validated,
and validation means one-hundred-per-cent-validation. Reformism may allow
for so-so validation, which means at times tolerating the status quo, but with
the help of a criterion telling us when. It is very important to notice that
Bellarmino's ploy was to demand total validation, and as long as his demand
is answered we are either conservative or radicalist, but any compromise
requires us to refuse to meet his demand.
majority say Galileo was absolutely right and the minority say the opposite,
so the majority say the phlogistonists were absolutely wrong, and the minor-
ity say they were absolutely right until one day the anti-phlogistonists were
absolutely right. But there is a difference here. The defenders of Bellarmino
say he was right because he was defending the status quo. They are conserva-
tives proper. The defenders of the phlogistonists pro-tern do not defend them
in the name of the status quo. Their argument is still radicalist, as I shall now
explain.
This is a strange case, a very subtle one, and one which is not easy to ana-
lyze - at least I have made mistakes about it. It is the case of the philosophy
known as instrumentalism and conventionalism - a philosophy shared by
some radicalists, some conservatives, and some reformers. Unlike the tradi-
tional radicalists, the instrumentalists allow validation to shift. Yet they do
require validation and do not rely on the status quo. The question when valid-
ity is withdrawn depends on the question, when is validity granted. Usually
validity is said to go with simplicity and so when a theory loses its simplicity
it may lose its validity to its simpler successor. There still is no room here for
a compromise.
What is a compromise, then? How can there be a principle of compromise
at all? Is compromise not between principle and reality?
Indeed, compromise is between principle and reality. Suppose validation
is only of principle. Then, in the life of compromise we do not act validly
but otherwise. How? What is this otherwise? When is this otherwise inevitable
realism and when is it the cowardice, the sluggishness, the ineffectiveness that
we all deplore? Suppose science is principle and the otherwise of the com-
promise something else. It follows that a man of science as a man of science
cannot compromise. Empirical experience shows this an error.
But what are the facts about validation? Do scientists in fact validate? If
so, all we need is a social scientist to make an empirical study and find how
scientists validate, and he should validate his discovery and close matters. This
may be difficult because we may need a historical perspective to differentiate
the work of a scientist as a scientist from his work as a person living in the
real world. Also, there are differences of opinion that have to be settled, and
this also requires time for a historical perspective. And then the historian of
science becomes our empirical social scientist who irons out the history of
science and who shows how scientists did validate their discoveries.
Thus, the historian of science eliminates the differences and the com-
promises - between schools, between the status quo and the innovation,
between principle and reality, between theory and observation.
112 JOSEPH AGASSI
All this must change, because such history of science is only post mortem
- on principle, since in life there is always dissent, doubt, and compromise.
When this kind of post mortem is presented as analysis in vivo it is sheer
poison.
There is, none the less, something missing in all that I have said thus far.
There is no doubt that we all take seriously some ideas and dismiss other
ideas off-hand. On what ground? There is no doubt that some debates go on
for millennia, some get settled. How? On what ground? Are the endless
debates not a waste of time?
Are these not problems of validation?
The question is ill put. There is no doubt that the Copernican views were
serious and could not be dismissed. Yet Bellarmino asked for a validation and
historians of science still debate, these days, the question, was Bellarmino
right. Hence these problems are not problems of validation in the traditional
sense. We can call them the new problems of validation. These are soluble,
and even with ease!
The first point to notice is that there is no fmal authority, neither Church
nor Reason: we are the sole judges, and we are fallible. But we have a limited
ability to say what we want and whether what we have is satisfactory, and to
what extent.
If we specify what we want, we can put this to critical debate. If we agree
about what we want, we may seek whatever satisfies these wants; for exam-
ple, we want a simple explanation of a given body of facts and we look for a
hypothesis that may serve the purpose. The more desiderata we have, the
more likely it may be that one hypothesis answers some, another answers
some, and we may have to make do.
In other words, we have to start not with validity but with minimum
necessary criteria and allow for diversity both of criteria and of possible solu-
tions by these criteria, in the light of these criteria. In other words, not what
view is reasonable, but what disagreement is reasonable, and what criteria for
these are reasonable. Obviously, the Copernicians disagreed with the Ptole-
means in a very reasonable way!
Karl Popper's philosophy goes a long way towards this solution, since he
stresses rejection, not acceptance. But he offers criteria that are obligatory and
which make the choice of the winning hypothesis unique; and so his highly
original view is still too much in line with traditional radicalist philosophy.
THE PROBLEMS OF SCIENTIFIC VALIDATION 113
to pay any attention to the status quo qua status quo. Here in the present
suggestion the status quo is left as utterly marginal. How we should approach
it depends on us and we may agree to handle it one way or another. For
example, clearly, when big science is involved, i.e. much money invested in
research, we may expect more attention to the status quo. This, said Imre
Lakatos, makes big science quite necessarily conservative. But in fact it is
not necessarily conservative: the criteria big science may have to meet - and
does at times - may force it to radicalism.
And with this I must conclude: the status quo, the politics of science -
external or internal - are not always important; and the problem of valida-
tion as classically understood - of criteria that forbid all concession to the
status quo - is therefore a great exaggeration. We all lead our lives in the way
we think best, yet we still differ from each other: some of us, however, are
open to others' opinions and to suggestions that may be improvements. More
abstractly, we have some disagreements that are silly, some where all parties
are intelligent and listen to each other. These facts must be taken care of by
any theory of science and by any progressive history of science.
JOHN D. NORTH
I have chosen to speak about the use of analogy in scientific argument, and I
ought to begin by clearing away a number of potential misconceptions: I
cannot completely ignore the historical origins of the word 'analogy', but the
time at my disposal is much too valuable to be spent in tracing its Aristotelian
pedigree in detail. On the other hand, I have no wish to do what so many
contemporary philosophers do, that is to say, treat the word as synonymous
with the word 'model.' What I shall say has much to do with the notion of
scientific model, but my perspective will be rather different from that of a
historian searching for the various uses of models in science. I repeat that my
concern is with analogical argument. As you will see in due course, the sorts
of argument that count as analogical are usually reckoned to be rather weak,
and even rather dangerous. There is no point in my trying to persuade you
that things- are otherwise. Mine will not be a history of 'positive science', in
Comte's sense, but a history of tentative science, and of certain methods of
conjecture. Analogy is the basis for much scientific conjecture, but even con-
jecture is an art, which can be done well, done rationally, that is, even though
it might prove in the end to have yielded a false conclusion.
This last remark might well seem highly paradoxical, but it is one that I
wish to emphasize, both because I believe it to be true, and because one's
attitude to it affects one's whole approach to the history of science. Many
historians - more in the past than in the present, I should say - are interested
only in scientific success, and in the gradual progress of mankind towards
the truth. Other historians, aware of the distorted image created by history
practised in that style, boast that they are equally concerned with scientific
error, falsehood, and misunderstanding. Some historians - even historians of
ideas - profess not to be interested in the quality of the arguments they
chronicle - 'That's not a historian's job', they say - but anyone who is so
interested must take into account that no-man's-land between truth and false-
hood, Le. those arguments that were reasonably based in the light of the
knowledge of the time, but in our own time are judged to have been mis-
taken. The past, after all, like the present, framed irrational as well as rational
truths, rational as well as irrational falsehoods. Not all nonsense is equally
foolish. To show what I mean, I shall first take a number of examples from
115
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 115-140.
Copyright © 1980 by D. Reidel Publishing Company.
116 JOHN D. NORTH
the work of Isaac Newton - a respectable enough scientist, you will agree. I
shall then look at the way in which the concept of scientific analogy devel-
oped in the middle of the nineteenth century, in the hands of Thomson and
Maxwell. I shall not have time to say much about the history of what one
might call the logic of analogy, but I will include a brief sketch of Mill's ideas
on the subject, if only to show that this was not abreast of the best developed
scientific uses of analogy at the time.
In June 1672, Newton wrote a long letter 1 to Henry Oldenburg, Secretary
of the Royal Society, concerning some objections raised by Robert Hooke 2
against Newton's work on colours and the refraction of light. I don't want to
be sidetracked into a discussion as to which of the two men was first to main-
tain that light is a periodic phenomenon: following E. T. Whittaker,3 it is
commonly said that Newton's letter of 1672 contains the first statement
''that homogeneous light is essentially periodic in its nature, and that differ-
ences of period correspond to differences of colour"; but this is somewhat
too generous. On the evidence, one could as easily ascribe the discovery of
periodicity to Hooke - as does Richard Westfall 4 - although for my own
part I think neither claim is particularly illuminating. I am content to frame
the discussion in the words used at the time. As Newton said, Hooke's hy-
pothesis was that the parts of bodies, when briskly agitated, excite vibrations
in the aether, and that these in due course, acting on the eye, cause us to have
the sensation of light, in much the same way as vibrations in the air cause a
sensation of sound, by acting on the organs of hearing. s The analogy is all I
want to consider, namely the analogy between sound and light. Newton does
not actually say he is arguing by analogy when he goes on to say that the
largest vibrations in the aether give rise to a sensation of red, and the shortest
a sensation of deep violet; but it is clear that he was indeed using an analogical
mode of reasoning. He goes on to draw the parallel: variation in the size of
the vibration of the air, he says, is responsible for variation in the tones in the
associated sound. There is no doubt that by 'size' ('depth' or 'bigness') he
means not our 'amplitude', but something like 'wavelength.' He certainly
thought of the vibrations as being longitudinal, rather than transverse. He
knew from his experiments that aether vibrations of various 'sizes', that is,
light of various colours, could be separated and recombined by refraction,
and he tried his hand - not very successfully in this letter of 1672 - at ex-
plaining the colours of thin plates and bubbles. His greatest concern with the
Hooke-Newton hypothesis (if I may call it thiS)6 was that waves or vibrations
in a fluid would be expected to spread out into the adjacent medium, rather
than be confmed to straight lines. For his own part, Newton believed he could
SCIENCE AND ANALOGY 117
his fancy." Newton believes that he can abstract from the alternative theories
a hard core of indisputable truth. He insists, in fact, that however we may
think of light, we are at least to think of it as a succession of rays, differing
from one another in such contingent circumstances as size, shape, and strength
- as do almost all things in nature. We are to think of light, moreover, as dis-
tinct from the vibrations of the aether, which - as we saw worrying Newton
earlier - have the unfortunate property of going round corners. We are to
think of light as being alternately reflected and transmitted by thin plates,
according to their thickness; and we are to suppose that just as light stimu-
lates the aether, so the interaction is mutual, and the aether refracts the light.
(The greater the density of the aether, the greater the refraction.) Reflection
is to be considered the result of secondary vibrations in the surface of the
reflecting body, some going into the aether within the body, and some being
returned to the aether outside it. 11 I will not elaborate further on the precise
mechanism suggested by Newton, or on the difficulties he encounters, or even
on the fact that he continues to draw analogies with sound vibrations. I wish
merely to point out that the model has changed in a rather subtle way. Des-
pite the change, the model is still compatible with Newton's experimental
findings, and although the old analogy between light and sound is no longer
as clear as it was, he continues to develop it.
Newton wants, he says, "to explain colours." The emphasis is mine, but the
phrase is Newton's. He supposes, he says, that just as bodies of various sizes,
densities and tensions 12 "by percussion or other action" excite sounds of
different tone, that is, vibrations of different wavelengths ('bignesses'), so
rays of light, by impinging on the aether both inside and outside bodies, 13
excite vibrations of different wavelengths in the aether. I leave aside the
physiological part of the explanation. The first point I want to make is that
the analogy is no longer quite as good as it was, for there is nothing carried
through the air, in the case of sound, analogous to the rays of light which
stimulate the aether on passing through it. My second observation is that
Newton has begun to take his analogy very seriously, talking as he does of
the 'Analogy of Nature' in the style of his later 'Third Rule of Philosophizing'
in the Principia. 14 Newton conjectures that just as harmony and discord of
sounds proceed from the ratios between their vibrations in the air, so may
the harmony and discord of colours proceed from the ratios of corresponding
vibrations in the aether.
Here is one argument by analogy, and the analogy is continued imme-
diately after. He describes how he and a friend independently divided the
spectrum of light from a prism into its seven component colours. (I assume
SCIENCE AND ANALOGY 119
that the friend was told to distinguish the seven colours named by Newton.
This, at least, is what his phraseology suggests.) He and the friend arrived at
much the same division. Newton took the mean of each pair of alternative
divisions, and found on measuring the length of the fmal divisions of the
image of the spectrum that they were in approximately the same ratios as the
divisions of a string capable of sounding the notes in an octave. 15
Newton was clearly very fond of this analogy, in its newly extended form,
for when he published his treatise Optieks in 1704 he repeated the material
on the octave of colour in much the same style. 16 What is more, when he
came to summarize his measurements of the diameters of what we now call
'Newton's rings',17 he again used the musical scale to do so. Remember that
at first he did not produce the rings by monochromatic light, and that the
rings were therefore coloured. 18 It was natural enough, under these circum-
stances, that he should extend his musical comparison. What he does is
calculate the thi~knesses of the wedge of air between the glasses at those
points where the rings are made by his seven spectral colours. He fmds that
these thicknesses are in the ratio of the lengths of a string yielding the notes
of the octave, raised to the power 2/3. 19 These thicknesses he subsequently
equated with what he calls "the Intervals of the following Fits of easy re-
flexion and easy Transmission."2o The explanation of the rings offered by
Newton on the basis of his theory of fits of easy reflexion and transmission
is remarkable and interesting, but does not concern the analogy with sound
I am now discussing. As far as I know, Newton does not develop the analogy
any further.
I have mentioned so many details in the course of my account of Newton's
analogy between light and sound that the shifting character of the analogy
has probably been lost to view. I will summarize the six examples I have now
given, three from 1672 and three from 1675:
N(I) Correspondences (some would call them analogies) are set up (or im-
plied) between the following concepts:
air (a 1 ); aether (a 2 ); vibration in a sounding body (b 1 ); vibration
in a luminous body (b 2 ); the tone of sound (e 1 ); the colour of
light (e 2 ); the sensation of sound (Sl); the sensation of light (S2);
vibration in the air (V1); vibration in the aether (v 2 );21
and also between the following:
causation of v 1 by b 1 , causation of v2 by b 2 ; causation of Sl by
v 1 , and ofs 2 byv 2 •
120 JOHN D. NORTH
N(2) The correspondences in N(1) are obviously still meant to hold, in addi-
tion to a correspondence between the two types of reflection, of sound and
of light. Newton now offers, not an argument, in the usual sense, but a state-
ment of a conceptual possibility. He can, he says, conceive a mixture of Vi,
selectively reflected, and he can (therefore) just as easily conceive a mixture
of v2 , selectively reflected.
Some would label N(2) a 'heuristic analogy.' It seems to me to be better
described as a new correspondence relation. It has illustrative value, but is
not an analogy in any of the senses I shall define later.
N(3) The same correspondences apply as before. Newton now says, in effect,
What I regard as two different sounds (e l , 8 1 , or vi?) I cannot regard as different degrees
of one fundamental sound. Therefore, what I regard as two different colours (e 2 , 8 2 , or
v2 ?) I cannot regard as different degrees of one fundamental colour.
Insofar as Newton here gives a reason for anything, the reason is a psycholog-
ical one. In fact the 'argument' could be looked upon as a statement of intent,
rather in the style of N(2), as to what theoretical concepts are to be utilized.
It is no doubt supposed that by showing them to be translatable, their plausi-
bility is increased. What is well worth noticing here is that Newton spent
much of his time in the documents of 1672 and 1675 - and of course else-
where - denying that he made any use of hypotheses. The correspondence
relations, as a whole, as well as the statements of what is to be taken as con-
ceivable, are, however, good examples of fallible hypotheses.
N(4) In the 1675 document, the correspondence relations are different from
those of 1672. As I have already explained, light rays are brought into the
aether side of the analogy, having no obvious counterpart in air. (In fact,
there is a difference between the functioning of the systems at this point, for,
as already noted, sound does not travel along straight 'rays' whereas light was
thought to do so.) The change does not affect the argument offered. There
are some new correspondences, namely between harmony of sounds (h 1) and
harmony of colours (h2), and also perhaps between ratios of properties of
SCIENCE AND ANALOGY 121
N(S) Here Newton attempts to determine what to him must have seemed a
significant property of colour. How is one colour separated from the rest? It
is almost always assumed by commentators on Newton that he took colour to
be subject to infmite gradation, but there is clearly a sense in which he wished
to preserve the traditional division of the spectrum into a limited number of
named colours.22 His reasoning is not altogether clear, but the four premises
seem to be these:
The conclusion is then that the divisions are at points dividing the spectrum
in exactly those ratios.
N(6) The reasoning follows closely that of N(S). The rings are coloured, and
by a suitable stretch of the imagination, take the place of the seven (much
purer) colours of the spectrum. Newton seems to have thought that in the
musical scale he had found an item of conceptual apparatus suitable for
investigating the phenomena of light and colour.
Electrostatics. Heat.
The electric field. An unequally heated body.
A dielectric medium. A body which conducts heat.
The electric potential at different The temperature at different
points of the field. points in the body.
The electromotive force which The flow of heat by conduction
tends to move positively elec- from places of higher to places
trified bodies from places of of lower temperature.
higher to places of lower po-
tential.
A conducting body. A perfect conductor of heat.
The positively electrified surface A surface through which heat
of a conductor. flows into the body.
The negatively electrified surface A surface through which heat
of a conductor. escapes from the body.
A positively electrified body. A source of heat.
A negatively electrified body. A sink of heat, that is, a place at
which heat disappears from
the body.
An equipotential surface. An isothermal surface.
A line or tube of induction. A line or tube of flow of heat.
Thomson's analogy with Fourier's theory of heat was only the first of several
he developed. They may be summarized in brief as follows:
T(1) 1841. The foregoing analogy with heat (this being the same as M(4)
below). Faraday does not seem to have known of the analogy until 1845. By
1850 at the latest, he had begun to make conceptual use of it in his formula-
tion of the notion of a field, with lines of force in empty space independent
of conductors, dielectrics, or magnets.
T(2) 1845. 31 Analogy between Coulomb's theory of electrical action at a
distance and Faraday's theory of action by contiguous particles in a con-
tinuous medium. The common formal element was a mathematical frame-
work in which Green's potential function played an important role.
T(3) 1845 (British Association meeting, Cambridge). Sketch of possible
analogies between optics, electricity and magnetism. This strongly influenced
the direction of Faraday's research. One consequence of Faraday'S exchanges
with Thomson was the discovery of the rotation of the plane of polarization
of light by magnetism. 32 Within a year or so Faraday had been led to formu-
late a number of new concepts, including that of diamagnetism, and that of
continuous and polarized lines of force capable of vibration, and thus of
transmitting optical 'forces'. He no longer needed his aether particles. 33
T(4) 1846. 34 Analogy between elastic solids and magnetic and electrical
SCIENCE AND ANALOGY 125
In searching for these notions I have come upon some ready made, which I have appro-
priated. Of these are Faraday's theory of polarity ... also his general notions about 'lines
of force' with the 'conducting power' of different media for them. Then comes your
allegorical representation of the case of electrified bodies by means of conductors of
heat ... Then Ampere's theory of closed galvanic circuits, then part of your allegory
about incompressible elastic solids and lastly the method of the last demonstration in
your R.S. paper on Magnetism. I have also been working at Weber's theory of Electro
Magnetism as a mathematical speculation which I do not believe but which ought to be
compared with others and certainly gives many true results at the expense of several
startling assumptions.
Now I have been planning and partly executing a system of propositions about lines
of force etc. which may be afterwards applied to Electricity, Heat or Magnetism or
Galvanism, but which is in itself a collection of purely geometrical truths embodied in
geometrical conceptions of lines, surfaces etc.
The first part of my design is to prove by popular, that is not professedly symbolic
reasoning, the most important propositions about V and about the solution of the
equation in the last page ... and to trace the lines of force and surfaces of equal V.42
On 14 February 1856, three days after Maxwell had read the second part
of his paper 'On Faraday's lines of force' to the Cambridge Philosophical
SOciety,43 he notes that he left the paper with Thomson, whom he asks to
return it, because he wishes to write up the second part 'On Faraday's elec-
trotonic state.' "I think I left an abstract too", he adds. 44 On an unspecified
date in the same month, Maxwell read an essay on analogy to the Apostle's
Club in Cambridge. 45 This light-hearted essay, in a flippant style characteristic
of university societies, adds nothing to the argume.nts offered in the scientific
paper, although it might well be used to settle a number of disputes over
Maxwell's early Weltanschauung. I am less concerned with this than with his
rather specific claims on behalf of analOgical arguments. I will begin with some
remarks made at the beginning of the paper 'On Faraday's lines of force.'
SCIENCE AND ANALOGY 127
Maxwell goes on to outline the following analogies (I will refer to that be-
tween the laws of science and the laws of number as M(l»:
M(2) That between light undergoing refraction and a particle moving in an
intense force-field.
M(3) That between light, the vibrations of an elastic medium (elasticity
being a sort of midwife); and electricity.
M(4) That between attraction at a distance (according to an inverse square
law) and the conduction of heat in uniform media. This is Thomson's first
analogy.48
M(5) That (which it is the purpose of the paper to explore) between a
system of electrical and magnetic poles, acting under an inverse square law,
and a field of incompressible fluid, moving within tubes directed along Fara-
day's lines of force. The lines are analogous to the streamlines in the fluid.
Maxwell's comments on these analogies are of some interest. It is said that
M(2) was "long believed to be the true explanation of the refraction of light",
and that ''we still fmd it useful in the solution of certain problems, in which we
employ it without danger, as an artificial method". I will anticipate a passage
128 JOHN D. NORTH
in which Maxwell says, in effect, how analogical argument does not oblige us
to accept the prior theory. He is now saying that we may continue to use an
analogy of highly restricted validity. We may use it for 'certain problems'; but
how the argument is to be kept under control, Maxwell does not say.
Of analogy M(3), Maxwell says that it extends further, and yet "is founded
only on a resemblance in form between the laws of light and those of vibra-
tions". (He here adds a sentence that I find confusing, which can be ig-
nored.)49 The drift of his meaning is plain as soon as he discusses analogy
M(4), which we might refer to as 'Thomson's first analogy':
The laws of the conduction of heat in uniform media appear at first sight among the
most different in their physical relations from those relating to attractions. The quan-
tities which enter into them are temperature, flow of heat, conductivity. The word force
is foreign to the subject. Yet we find that the mathematical laws of the uniform motion
of heat in homogeneous media are identical in form with those of attractions varying
inversely as the square of the distance. We have only to substitute source of heat for
centre of attraction, flow of heat for accelerating effect of attraction at any point, and
temperature for potential, and the solution of a problem in attractions is transformed
into that of a problem in heat. 50
When he said of M(3) that the resemblance was 'only one of form, Maxwell
was, even if only half consciously, making a distinction between this sort of
analogy and analogies in which the 'objects' in the two related domains so
closely resemble each other that the same word may even be used for both.
(An example is N(1), where vibrations (in air) correspond to vibrations (in
aether).) One may well ask about the danger that two domains will be con-
fused in a carefully prescribed scientific analogy. This is a genuine problem,
but it does not seem to have been what was most worrying to Maxwell, who
went on to say, in connection with M(4), that
the conduction of heat is supposed to proceed by an action between contiguous parts
of a medium, while the force of attraction is a relation between distant bodies, and yet,
if we knew nothing more than is expressed in the mathematical formulae, there would be
nothing to distinguish between the one set of phenomena and the other. 51
In other words, the phenomena are very different, and the formulae fail to
reveal the difference; and yet part of the value of the best analogical argument
is that it allows the mind "clear physical conceptions". 52 The stimulation of
mathematical ideas is only a part of what one should hope for:
It is true, that if we introduce other considerations and observe additional facts, the two
subjects will assume very different aspects, but the mathematical resemblance of some of
their laws will remain, and may still be made useful in exciting appropriate mathematical
ideas. 53
SCIENCE AND ANALOGY 129
examples, the rules governing terms in the prior domain will be a matter of
common sense and - one hopes - of common agreement. But the rules are
nevertheless there, to be taken into account in any fonnal rendering of a
logic of analogy .
In discussing Newton, I spelt out in some detail six analogical 'arguments',
whereas in Maxwell's case I have only hinted at the basis for such arguments.
Analogy M(l), as I have said, lends itself to no argument in particular. Anal-
ogies M(2) and M(3), providing the familiar corpuscle and wave explanations
of light phenomena, are no doubt so familiar that I need not describe them in
detail. I would like to point out that both M(2) and M(3) were, in a sense I
have explained, based on newly contrived analogues. 58 This applies also to
M(4) and M(5), the systems which are considerably more sophisticated than
M(2) and M(3). (Even so, I think it could be argued that M(3), namely
Huygens' wave analogy, contained what to fellow physicists was the least
familiar analogue of the four, at the time of framing.)
I should like to make a distinction here between two sorts of analogy.
Some of Maxwell's examples were, in a sense I have explained, based on
analogues specially contrived for the occasion. Huygens' wave analogy is
an example. How different are analogies with an artificial basis (models, in
one sense of that word) from those with a pre-established analogue - as in
Newton's case, where he did not have to invent a theory of sound? A pre-
liminary and obvious answer is that the first may be modified again and
again until it satisfies its creator, whereas with the second - something con-
ceived of as given - we are obliged to distinguish between so-called 'positive'
and 'negative' analogies, i.e. respects in which the analogues agree and dis-
agree. But matters are rarely so simple. We are reminded of the analogies
under headings N(2), N(3), and N(6). There we found Newton establishing
conceptual possibilities, rejecting and refming a concept, and confirming the
value of a concept. Newton was there, in fact, establishing in this way restric-
tive conditions of a sort which I mention later as having been obscured by
Mill in his analysis of the subject, conditions limiting the functioning of key
concepts (selective reflection, dilution of sound, octave division, and so on).
There is another side to this question of the difference between pre-
established and artificial analogues. I refer to the ontological problem. When
Maxwell said of his incompressible fluid that it was "merely a collection of
imaginary properties", 59 he was not saying anything likely to colour our
views of the real nature of the space occupied by electrical charges. This is
less true of Thomson's analogy (viz. M(4», although the influence of this was
oblique. Here is what Maxwell wrote in his Elementary Treatise on Electricity:
SCIENCE AND ANALOGY 131
We must bear in mind that at the time when Sir W. Thomson pointed out the analogy
between electrostatic and thermal phenomena men of science were as firmly convinced
that electric attraction was a direct action between distant bodies as that the conduction
of heat was the continuous flow of a material fluid through a solid body. The dissimilar-
ity, therefore, between the things themselves appeared far greater to the men of that
time than to the readers of this book, who, unless they have been previously instructed,
have not yet learned either that heat is a fluid or that electricity acts at a distance. 60
No one reading his work was likely to come away with the idea that heat is
really electricity, or that electricity is really heat ;61 and yet Thomson's paper
did persuade many that Faraday had been right to suggest that electrical
action was effected through a continuous medium. This was the outcome of
an interesting clash of paradigms. Throughout the eighteenth century (and up
to about 1820) Newtonian dynamics had been considered as an almost essen-
tial mode for physical science, and the successes of molecular physics, as
practised by such men as Laplace, Navier, Cauchy and Poisson, confirmed a
majority of the scientific community in their belief - a belief which lingered
on in England rather longer than it did on the continent of Europe. There it
had been challenged indirectly by Fourier's theory of heat. In commenting in
his second paper (first printed 1845)62 on the analogy between Fourier's
theory and Faraday'S theory of electrical action in a medium, Thomson hints
at the uneasy compatibility of action at a distance and contiguous action.
Since his style was mirrored in some degree by that of Maxwell's paper of
1855-6,63 I will quote a more extended passage than is necessary to illustrate
the new ontological situation:
Now the laws of motion for heat which Fourier lays down in his Theorie analytique de
la chaleur, are of that simple elementary kind which constitute a mathematical theory
properly so-called; and therefore, when we find corresponding laws to be true for the
phenomena presented by electrified bodies, we may make them the foundation of the
mathematical theory of electricity: and this may be done if we consider them merely as
actual truths, without adopting any physical hypothesis, although the idea they naturally
suggest is that of the propagation of some effect by means of the mutual action of con-
tiguous particles; just as Coulomb, although his laws naturally suggest the idea of mate-
rial particles attracting or repelling one another at a distance, most carefully avoids
making this a physical hypothesis, and confines himself to the consideration of the
mechanical effects which he observes and their necessary consequences.
All the views which Faraday has brought forward, and illustrated or demonstrated by
experiment, lead to this method of establishing the mathematical theory, and as far as
the analysis is concerned, it would, in most general propositions, be even more simple, if
possible, than that of Coulomb. (Of course, the analysis of particular problems would be
identical in the two methods.) It is thus that Faraday arrives at a knowledge of some of
the most important of the general theorems, which, from their nature, seemed destined
never to be perceived except as mathematical truths. 64
132 JOHN D. NORTH
F(a) ~ II (a) ,
F(a) ~ 12 (a), etc.
SCIENCE AND ANALOGY 133
light whatever on the sort of arguments I gave from Newton, Thomson, and
Maxwell.
Perhaps Mill would have claimed that his analogies extended to the case of
different domains, but I think not. At all events, this case is only found in
Mill, so far as I can see, under the rubric 'false analogy'.72 The examples given
are not such as to have gained much sympathy from his readers, involving as
they do numerology and Pythagorean harmony.
I have now considered excerpts from the writings of four men - Newton,
Thomson, Maxwell, and Mill - three of whom made important use of anal-
ogical argument in a scientific context, and two of whom wrote about the
theory of analogy. I have tried to avoid imposing my own logical views on the
historical material, and I hope the result was not too loosely shaped. The
subject of analogy is a large and difficult one, extending as it does into every
region of human activity. Analogies have two sides to their nature: they are
instruments of argument, prediction, and validation, and they are instruments
of cognitive meaning, understanding, formalization and classification. The
problem of meaning and categories is not easily disentangled from the prob-
lem of argument and law. Are the planets the same sort of thing as bodies in
free space? Newton said Yes, Descartes thought No. Are terrestrial motions
governed by the same laws as celestial? Aristotle said No, Galileo said Yes.
But what a thing is is obviously to a large extent decided by how a thing
behaves; and this is what scientific laws inform us about.
This contrast between problems of meaning and of argument is closely re-
lated to one of these dogmas of the logic of analogy which has been so often
repeated that it is frequently taken for granted. I am referring to the idea that
analogies can be easily divided into two classes - namely of so-called substan-
tial and formal sorts. A substantial analogy is supposedly one where there
is a correspondence of simple properties, while a formal analogy is taken to
involve a correspondence between relations, or, in a more sophisticated ver-
sion of the idea, between relations among constituents somehow stripped of
all their properties. Now it is very difficult to comment on this view, unless
we are clear as to the epistemology of the person who is proposing it, but
the view usually goes with the doctrine that linguistic conventions (and the
formulae of scientific theories) are somehow models of complex facts. The
'formal' relations are, it seems, regarded as though they were fixed for all
eternity - for otherwise, how can we be sure that phenomena which are now
(to take a crude case) explained predominantly in terms of properties will
not in future be better explained in terms mainly of formal relations? If I am
SCIENCE AND ANALOGY 135
have little or nothing to do with the problem set by those who, in the natural
sciences, have advanced their knowledge by analogical techniques that they
have seldom tried to justify. One often reads the platitude that analogical
arguments are inevitably limited in their scope, because what is radically new
is precisely that which cannot be accounted for in familiar terms. The whole
purpose of analogies, however, is to explore, and to explore in the hope that
what seems to be radically new will have unsuspected elements in common
with what is familiar.
Perhaps the sociologist will categorize this as reactionary thinking. Perhaps
the logician will dismiss it as invalid or illogical. It is the philosopher's job,
nevertheless, to offer an analysis of this very common mode of thought, and
if he can offer a satisfactory analysis of real historical examples, so much the
better. I hope that I have shown something of the way in which the use of
analogy matured between the seventeenth century and the nineteenth, and
how in Thomson and Maxwell there is a conscious awareness of the function
of a mathematical calculus as an intermediary between analogues - between,
that is to say, a theory and a model for that theory. And theirs was more than
an idle philosophical observation, for it suggested ways of applying a powerful
tool for conjecture and for the unification of the physical sciences.
NOTES
between fluid pressure, electrical potential, and temperature, he went on: "A fluid is
certainly a substance, heat is as certainly not a substance, so that though we may fmd
assistance from analogies of this kind in forming clear ideas of formal relations of elec-
trical quantities, we must be careful not to let the one or the other analogy suggest to us
that electricity is either a substance like water, or a state of agitation like heat". (My
italics.)
62 Paper II in Thomson, Reprint of Papers ... ; paper first appeared in Cambridge and
Dublin Math. Journal, November 1845, and then in an extended version in Phil. Mag.
1854.
63 That is, the paper from which I have quoted much already (,On Faraday's Lines of
Force'; see Note 39). It was read in two parts, in December 1855 and February 1856.
64 Thomson, Reprint of Papers . .. , p. 29.
65 Mill, System of Logic, III. xx. 2 (see Note 28 above).
66 He calls this 'the fact m'.
67 Loc. cit.
68 Cf. Ibid., III. xx. 3: "If, after much observation of B, we find that it agrees with
A in p out of 10 of its known properties, we may conclude with a probability of 9 to
1 that it will possess any given derivative property of A". Dissimilarities are said to
furnish counter-probabilities. Mill makes no attempt to decide whether some properties
may not be more fundamental than others; or whether there is any limit to the known
properties of a thing; or to their triviality.
69 Ibid., II. xx. 2. Further references to Mill are to this section, unless said to be other-
wise.
70 Cf. Ibid., V. v. 6: "It has to be shown that in the two cases asserted to be analogous,
the same law is really operating; that between the known resemblance and the inferred
one there is some connection by means of causation".
71 Mill wished to settle for nothing less than absolute truth in science. He was deeply
suspicious of hypotheses, which he admitted might be fruitful; but for this very reason
they might - if fruitful yet false - be an "impediment to the progress of real knowledge
by leading inquirers to restrict themselves arbitrarily to the particular hypothesis which
is most accredited at the time". The wave and emission theories of light are instanced as
"un susceptible of being ultimately brought to the test of actual induction", even though
they are not "worthy of entire disregard"!
72 Ibid., V. v. 6.
73 Bunge, M., Scientific Research, 2 vols., New York, Springer, 1967, Vol. 2: The
Search for Truth, esp. ch. 15; Metzger, H., Les concepts scientifiques, Paris, Alcan, 1926,
passim.
MARCELLO PERA
This essay intends to argue that induction is the method of scientific discovery
and that the current objections against the inductive method are not correct.
To this purpose I shall first specify the various meanings of 'method' and I
shall use these meanings as an Ariadne's thread in order to draw a map of the
problems of methodology and a model of inductive procedure (Section I).
Then I shall go on to show that the main arguments against induction put
forward by modern hypothetico-deductivists and by Popper are fallacious
(Sections II-III). Lastly, after rejecting the view of those who maintain that
a distinction should be made between discovery (by induction) of laws and
discovery (via hypotheses) of theories (Section IV), r shall try to prove my
thesis by showing that the act of conceiving or inventing a hypothesis is an
inductive inference from observational premises (Section V). I shall outline
also the advantages of such a view with respect to the hypothetico-deductivist
(or 'trial and error' or 'conjectures and refutations') approach.
It is not always recognized that the widely used expression 'scientific method
is a pollakos legomenon. In actual fact, it contains at least three different
explicanda which it is indispensable to distinguish.
In the first place, the scientific method is a procedure, a general strategy
that indicates an ordered sequence of moves (or steps) which the scientist has
to make (or go through) in order to reach the goal of his research. This is the
sense the word carries in the phrases 'deductive method', 'inductive method',
'hypothetico-deductive method', etc. Thus, to give an example, when Bernard
writes that "the experimental method is based on feeling, reason and experi-
ment, in that order" (1865, p. 57), he is using the term 'method' in this sense
and he therefore specifies, together with the ordered sequence of moves he
141
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 141-165.
Copyright © 1980 by D. Reidel Publishing Company.
142 MARCELLO PERA
essay deals with the frrst problem of methodology. The problem is currently
given little attention, as can easily be seen by comparing the amount of
discussion dedicated to it with that reserved for the second problem, which is
at present the focus of interest in the philosophy of science. 5 The explana-
tion for this is doubtless to be found in the persistent and almost uncontested
hegemony exercised by the current hypothetico-deductivist paradigm.
This paradigm is characterized by two theses: the anti-inductivist thesis,
according to which scientific discovery does not come about by induction
from facts, and the anti-continuist thesis, which holds that the process of
discovery is not unitary but rather is divided into two discontinuous thought
episodes, one being a-logical or pre-logical and belonging to the 'context of
discovery', and the other being, instead, logically or rationally reconstructible
and belonging to the successive 'context of justification'.
There are two versions of this paradigm, which we shall call the weak and
the strong versions. The weak version is the most widespread; it offers as the
explicatum of scientific procedure a model that can be represented by the
following schema:
(1) P ... H~Oc-+Hc·
This means that an investigation starts off with a problem P, introduces a
conjecture or hypothesis H through a mental jump (' .. .'), from this it deduces
('~') certain observational test statements Oc and then, if these statements
prove to be true, induces ('-+') the truth or probability ofthe hypothesis. 6
The strong version of the paradigm is Popper's. It offers a variant of
schema (1) where the last step has been cut off, i.e. the schema
(2) P ... H~Oc.
According to this variant 7 (which Popper expresses by the better-known
schema PI - IT - EE - P2, i.e. problem PI, tentative theory, elimination of
the errors, and problem P 2 ) the very argument by which the hypothesis is
tested or 'corroborated' is held to be completely devoid of induction. The
whole procedure is thought to consist of two types of attempts, an attempt
to guess H and an attempt to falsify it by Oc: the former being an intuitive
jump, the latter a deductive argument. Therefore, as Popper writes, "the
method of falsification presupposes no inductive inference, but only the
tautological transformations of deductive logic" (1959, p. 42).
As an alternative to the hypothetico-deductivist paradigm I will propose
as explicatum of scientific procedure the following schema:
(3) 0; -+ Hp ~ Oc -+ He.
144 MARCELLO PERA
This conception is quite widespread and just a few quotations will suffice
to document it.
According to Cohen and Nagel (1934, pp. 200-201):
We cannot take a single step forward in any inquiry unless we begin with a suggested
explanation or solution of the difficulty which originated it. Such tentative explana-
tions are suggested to us by something in the subject matter and by our previous knowl-
edge. When they are formulated as propositions, they are called hypotheses. The func-
tion of a hypothesis is to direct our search for the order among the facts.
Even more explicitly Hempel maintains that the viewpoint that sparks off
an investigation must be a specific hypothesis and not merely a problem. As
he puts it (1966, pp. 12-13):
Perhaps all that should be required in the fIrst phase is that all the relevant facts be
collected. But relevant to what? ... Let us suppose that the inquiry is concerned with a
specific problem. Should we not then begin by collecting all the facts - or better, all
available data - re~evant to that problem? This notion still makes no clear sense .... And
rightly so; for what particular sorts of data it is reasonable to collect is not determined
by the problem under study, but by a tentative answer to it that the investigator enter-
tains in the form of a conjecture or hypothesis .... Empirical 'facts' or fmdings, there-
fore, can be qualifIed as logically relevant or irrelevant only in reference to a given hypo-
thesis, but not in reference to a given problem .... Tentative hypotheses are needed to
give direction to a scientifIc investigation. Such hypotheses determine, among other
things, what data should be collected at a given point in a scientifIc investigation.
Finally, among the Popperians one may cite Medawar, who also holds that
hypotheses are indispensable to make observations meaningful:
We cannot browse over the fIeld of nature like cows at pasture .... Our observations no
longer range over the universe of observables: they are confmed to those that have a
bearing on the hypothesis under investigation (1969, p. 29 and 51).
146 MARCELLO PERA
Despite its apparent plausibility, I believe that it can be shown that this con-
ception - the conception according to which without hypotheses there can
be no relevant observations - is incorrect. To this end I will first avail myself
of a celebrated example of scientific research.
From September 1854 onwards - [Darwin writes in his Autobiography I - I devoted all
my time to arranging my huge pile of notes, to observing, and experimenting, in relation
to the transmutation of species .... After my return to England it appeared to me that
by following the example of Lyell in Geology, and by collecting all facts which bore in
any way on the variation of animals and plants under domestication and nature, some
light might perhaps be thrown on the whole subject .... I worked on true Baconian prin-
ciples, and without any theory collected facts on a wholesale scale, more especially with
respect to domesticated productions, by printed enquiries, by conversation with skilful
breeders and gardeners, and by extensive reading (1958, pp. 118-19).
Darwin's fame among the hypothetico-deductivists is above all linked to the
well-known slogan: "How odd it is that anyone should not see that all obser-
vation must be for or against some view if it is to be of any service!" 8 In the
light of this, the reference made in the above-quoted passage of the Auto-
biography to 'true Baconian principles' ought then to be explained as a
mishap or seen in the psychoanalytic terms of the unconscious persistence of
a "complex of the Lord Chancellor". 9
In actual fact what we are faced with here is neither a mishap nor a com-
plex. Admittedly, when Darwin set out on the Beagle journey gathering
together his 'huge pile of notes' and when he collected 'all facts' he thought
to be relevant to the 'transmutation of species' he was dealing with scientific
problems; but it is equally certain that he did not have the hypothesis of
natural selection in mind. If we use the term hypothesis in the same meaning
as the hypothetico-deductivists - namely as meaning a "suggested solution
of the difficulty which originated an inquiry" (Cohen-Nagel) or a "conjecture
for the solution of a problem" (Popper) or again as an "imaginative pre-
conception of what might be true" (Medawar) - then what can be said is
that Darwin's observations during the Beagle journey were not preceded by
any hypothesis and that he was observing in order to find a hypothesis rather
than to test a hypothesis.
Must we therefore conclude - as the hypothetico-deductivist would object
- that Darwin's mind was a tabula rasa or, to use Popperian imagery, an
empty bucket and that Darwin - as Medawar puts it - was browsing like cows
at pasture? Of course not! Darwin was extremely knowledgeable, he had vast
scientific learning, he had general ideas about how to solve his problems (one
need only think of his reference to the "example of Lyell in geology"). In a
word, he had a theoretical framework.
INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 147
It may seem that the distinction between theoretical framework and hypo-
thesis has merely circumvented the obstacle, transposing into psychological
and historical terms the problem raised by the hypothetico-deductivists,
which is really of a logical kind. For indeed, especially when one takes into
account that the theoretical framework of one era is constituted by the
theories that were hypotheses or tentative research programs in the preceding
era, one may legitimately ask: what precedes the theoretical framework?
Now if the answer is to be that the theoretical framework is derived from
previous observations, one falls back, by regression, into the empiricist pitfall
of pure observations; on the other hand, if this solution is rejected (as it must
be), then apparently one has to fall back onto the conception of the logical
priority of hypotheses over observations.
This is exactly what Popper claims. He maintains in the first place that a
theoretical viewpoint must always be presupposed in order for our observa-
tions to be meaningful.
148 MARCELLO PERA
no principled reason for rejecting the thesis that the former can be derived
from the latter through an inductive inference.
Before upholding this thesis, it is desirable, however, to take a look at a
more flexible conception than the hypothetico-deductivist position as regards
the logical structure of scientific discovery. I will call this conception 'qualified
inductivism'; it maintains that a distinction should be made between the
discovery of empirical laws and the discovery of theoretical laws.
J. S. Mill may be said to be the first of the qualified inductivists. As is well
known, according to Mill, wherever the plurality of causes and the intermixture
of effects make it impossible to investigate the coming into being of some
phenomenon by the direct method (that is, simply by applying the canons),
then this method must be replaced by the deductive or a priori method or by
the hypothetical method. The deductive method is a three-stage procedure:
the first stage is direct induction from observed phenomena in order to ascer-
tain the laws of causes, the second is calculation or ratiocination, to deter-
mine the effect produced by the combination of the causes, and the third is
empirical verification to establish that the conclusions calculated accord with
experience. The hypothetical method is similarly in three stages but with the
hypothetical formulation of the law instead of its inductive derivation.
According to Mill, therefore, the choice of the deductive or hypothetical
method imposes itself upon the scientist as a result of the degree of complex-
ity of the phenomena awaiting an explanation or, as he puts it"of "the very
nature of the case". This complexity has a nomologic concomitant: on
account of it, the laws of phenomena can no longer be formulated in terms of
observational predicates on a par with the laws obtainable by the canons, but
rather must be expressed in theoretical terms. Mill recognizes this explicitly
when, after having said that it is to the deductive method (and with all the
more reason, one may say to the hypothetical method) that "the human
mind is indebted for its most conspicuous triumphs in the investigation of
nature", he goes on to add: "to it we owe all the theories by which vast and
complicated phenomena are embraced under a few simple laws, which, con-
sidered as the laws of those great phenomena, could never have been detected
by their direct study" (1843, p. 304). Mill takes Newton's theory of gravity
to be one example of such theories.
This attitude, which holds that empirical laws are discoverable by induc-
tion while theoretical laws, on the contrary, are discoverable by hypotheses,
has been taken over, in a certainly much more elaborate and conscious
version than Mill's, by modern qualified inductivists, among whose ranks one
may place G. H. von Wright, J. P. Day, M. Bunge and R. Carnap.12
152 MARCELLO PER A
For his exemplary clarity on this question, Carnap himself can be picked
out as spokesman. In the Philosophical Foundations of Physics he writes:
Theoretical laws are, of course, more general than empirical laws. It is important to
understand, however, that theoretical laws cannot be arrived at simply by taking the
empirical laws, then generalizing a few steps further. How does a physicist arrive at an
empirical law? He observes certain events in nature. He notices a certain regularity. He
describes this regularity by making an inductive generalization (1966, p. 228).
On the other hand, Carnap enquires,
How can theoretical laws be discovered? We cannot say: 'Let's just collect more and
more data, then generalize beyond the empirical laws until we reach theoretical ones'.
No theoretical law was ever found that way. We observe stones and trees and flowers,
noting various regularities and describing them by empirical laws. But no matter how
long or how carefully we observe such things, we never reach a point at which we observe
a molecule. The term 'molecule' never arises as a result of observations. For this reason,
no amount of generalization from observations will ever produce a theory of molecular
processes. Such a theory must arise in another way. It is stated not as a generalization
of facts but as a hypothesis (1966, p. 230).
Making use of these argumentations, qualified inductivists agree with the old
inductivists in concluding that induction is operative in the discovery of
scientific laws, but agree with the anti-inductivists in maintaining that it is
operative only in the simplest, most elementary laws, those with the lowest
empirical level. From a historical point of view, induction is thus seen as
the childhood method of science. In principio erat inductio, deinde fit hy-
pothesis.
In order to evaluate this point of view, let us consider the following law-
like statements:
HI: 'Lions are carnivorous'.
H 2 : 'Magnets are composed of electric currents'.
As we have said, according to the qualified inductivist, HI is discoverable
by induction while H2 is not. And this is because the qualified inductivist
INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 153
contention one can in the first place invoke those existential hypotheses that
refer to the existence of observable entities. Bouvard, Bessel, Adams and Le-
verrier's hypothesis 'there exists a transuranic planet' was certainly put to the
test by means of implications and predictions. But there are also many hy-
potheses that, having the form of universal statements like those of the type
HI, are nevertheless tested in the same way as those of the type H 2 : thus
the hypothesis, 'migratory birds orientate themselves on the basis of the
celestial signs', is confirmed by logically deriving predictions of certain kinds
of behavior under certain kinds of situations.
By denying that laws and theories can be sharply differentiated according
to the method of testing them and above all according to how they are dis-
covered, we may seem to be upholding the anti-inductivist thesis: if, even
where it would appear to be legitimate to speak of mere generalization
processes, it turns out that the discovery of laws originates from an inven-
tion of hypotheses in the form of a conceptual colligation, one might well
conclude that discovery is not a matter of induction or reasoning. It appears,
however, that one can equally well arrive at the opposite conclusion; it is to
the legitimacy as well as to the advantages of this alternative point of view
that I will now turn.
and why he has conceived that idea, i.e. how and why he made that inference;
yet he does know what inference he has made: it is precisely the inference
through which he expresses, in the first place to himself and then to others,
the reasons why he thinks that the research problem is solvable, and the facts
explainable, in that particular way. There do not exist nor can there exist
two distinct thought episodes - flrst blind invention and then the interven-
tion of the plausibility considerations; there is rather a single argumentative
act: a hypothesis springs from the very same argument which provides the
initial reasons of its plausibility.
One example, among the many, will illustrate the adequacy of this induc-
tivist point of view.
On March 18, 1755 Benjamin Franklin sent the physician John Lining,
who had asked him how he had reached the hypothesis of the identity of
nature between electrical fluid and lightning, the following extract from the
records of his experiments:
Nov. 7, 1749. Electrical fluid agrees with lightning in these particulars: (1) Giving light.
(2) Colour of the light. (3) Crooked direction. (4) Swift motion. (5) Being conducted
by metals. (6) Crack or noise in exploding. (7) Subsisting in water or ice. (8) Rending
bodies it passes through. (9) Destroying animals. (10) Melting metals. (11) Firing inflam-
mable substances. (12) Sulphureous smell. The electric fluid is attracted by points. We
do not know whether this property is in lightning. But since they agree in all the par-
ticulars wherein we can already compare them, is it not probable they agree likewise in
this? Let the experiment be made (1755, p. 524).
I.et us indicate by F the electric fluid, by L the lightning and by PI, ... ,P13
the given properties; Franklin's reasoning can then be represented in the
following form:
Fhas the properties PI. Pl , •.• ,P1l • P 13
L has the properties PI. Pl , ••• ,P12
It is something like this. You are sitting working very hard, you have worked for a long
time trying to open a safe. Then some Joe comes along who knows nothing about what
you are doing, except that you are trying to open the safe. He says "Why don't you try
the combination 10.20. 30?" Because you are busy, you have tried a lot of things, maybe
you have already tried 10.20.30 (1965, p. 161).
the consequences of the guess ... and then we compare the result of the
computation to nature" (1965, p. 156), then we cannot excuse ourselves
from testing any suggestion whatsoever. Feynman should therefore not neglect
combination 10.20.30 either; if he does, it is evidently because he regards its
plausibility as too low: and after all we can hardly blame him for this, since
the one who has suggested it to him is any old Joe and not the famous Joe
the burglar wanted by the police on account of his unbridled love for safes.
But although it may be perfectly reasonable from our point of view, for
a consistent hypothetico-deductivist burglar the refusal to try a combination
is not legitimate; the same goes for a hypothetico-deductivist scientist: a
priori one guess is as good as another and no guess is more plausible than
another until it has been evaluated and hence submitted to a fIrst form of
test. Therefore the categorical imperative of the hypothetico-deductivist
is to test all conjectures. But 'all' conjectures are an infmite number; thus
the order cannot be obeyed: ad impossibilia nemo tenetur.
As an explanation of why in actual fact scientists do not get lost in this
theoretical infinity, Medawar, perhaps the only one to have clearly seen this
logical diffIculty, has advanced the following suggestion:
In real life, of course, just as the crudest inductive observations will always be limited
by some unspoken criterion of relevance, so also the hypotheses that enter our minds
will as a rule be plausible and not, as in theory they could be, idiotic. But this implies
the existence of some internal censorship which restricts hypotheses to those that are
not absurd, and the internal circuitry of this process is quite unknown. The critical
process in scientific reasoning is not therefore wholly logical in character, though it can
be made to appear so when we look back upon a completed episode of thought (Medawar,
1969, p. 53).
censorship' and at the same time to maintain, as Medawar does, that "the
process by which we come to formulate a hypothesis is not illogical but non-
logical, i.e. outside logic" (1969, p. 46).
In actual fact the 'internal censorship' is the Trojan horse unwisely allowed
through the walls of the hypothetico-deductivist conception. I believe that if
reasoning is not restrictively interpreted as mechanical reasoning and if it is
recognized that it does not exclude creativity or imagination or individual
talent, no relevant objection is left to the thesis that the inductive method is
the method of science. Perhaps the last remaining obstacle is only psycho-
logical resistance to the rejection of an old (and glorious) paradigm.
Translated by
RACHEL COSTA
NOTES
* I am extremely grateful to Professor Francesco Barone for having read a former ver-
sion of this paper and for having discussed the main thesis with me. My sincere thanks
also go to Professors John North and Carl Kordig for their stimulating comments and
suggestions.
1 See Bacon (1620), p. 109. "My way of discovering science goes far to level man's wits,
and leaves but little to individual excellence; because it performs everything by the
surest rules and demonstrations".
2 See Descartes (1628), pp. 371-72: "Per methodum autem intelligo regulas certas et
laciles quas quicumque exacte servaverit, nihil unquam falsum pro vero supponet, et
nullo mentis conatu inutiliter consumpto, sed gradatim semper augendo scientiam, per-
veniet ad veram cognitionem eorum omnium quorum erit capax" .... ["by a method I
understand certain and easy rules such that whoever has employed them exactly never
supposes anything false as true, and without uselessly consuming his mental effort but
rather always gradually increasing his knowledge, will arrive at a true cognition of all
those things of which he will be capable". 'Rules for the Direction of the Native Talents',
in Descartes, The Essential Writings, trans. J. J. Biom, New York, Harper and Row,
1977, pp. 22-98. Cf. p. 31.
3 Cf. Popper (1959), p. 49: "A methodology is a theory of rules of scientific method".
According to Lakatos (1971, p. 92), "modern methodologies or 'logics of discovery'
consist merely of a set of (possibly not even tightly knit, let alone mechanical) rules for
the appraisal of ready, articulated theories". Cf. also Radnitzky (1977), p. 1: "a meth-
odology is a system of rules or recommendations about how to act in certain types of
research situations in order to facilitate achieving the aim of this activity: scientific
progress". Very important on this and related topics is Radnitzky(1979).
4 The following defmition (Enciclopedia Mondadori delle Scienze, Milan, 1967, under
the heading 'Method') combines the second and the third meanings: "the scientific
method is the set of rules for the formation of concepts and the drawing of inferences,
together with the complex of techniques of tested observations that are used in the
acquisition of knowledge".
162 MARCELLO PERA
5 The current Methodenstreit between the school of Popper and the 'anarchist' school
or the trends of thought dominated by the latter concerns the very possibility or useful-
ness of invariant rules defining the 'game of science'. And indeed this is quite a trouble-
some point; in fact I think that the explication of each one of the three meanings of
'method' leads to worrying results that could be summarized in the following 'paradox
of the scientific method': "science is characterized by its method, but the characteriza-
tion of the scientific method destroys science". Despite this paradox I do not, however,
think that the best policy in matters of methodology is that of declaring oneself to be
'Against Method' even though this is undoubtedly the best way to shake up ingrained
manners of thinking and to call to the attention of lazy minds something that does con-
stitute a real problem. On the paradox of the scientific method see Pera (1978), Part I.
6 On the thesis of the 'mental jump' see Popper (1963), p. 46: "without waiting for
premises we jump to conclusions"; Wisdom (1952), p. 49: "a hypothesis is attained by
some mental jump".
7 The most complete and sophisticated version of the Popperian model of scientific
procedure is to be found in Radnitzky (1976) and (1977).
8 In his letter of 1861 to Henry Fawcett; cf. Darwin (1903), Vol. I, p. 195.
9 As always, a number of hypotheses can 'save' the same phenomena; in our case, even
the one - which happens to be ad hoc - advanced by J. Agassi (1975, p. 152) according
to which Darwin "collected while on the Beagle as much information as he could for an
obvious technical reason: he did not hope to arrange a second visit".
10 On this matter Z. Bechler's study (1974) is rather insightful. But it is likely that
Bechler would not be in agreement with the view that Newton's optical controversies
(especially with Huygens) derive from a conflict of epistemological assumptions in
the sense being used here. Cf. the following note.
11 In my (1978), Chapter II. 5, I have presented a more analytic classification and have
provided a variety of exemplifications regarding the various functions assumptions come
to have in scientific research. Referring to the meanings (1)-(4) as indicated, I speak,
respectively, of 'fundamental assumption', 'derivative assumptions', 'local assumptions'
and 'epistemological assumptions'. As far as their logical status is concerned, it is my
claim that they have the status of a priori principles that remain invariant throughout
whole historical eras (except the fust principle which is rigid) during which they leave
their imprint on research or provide the 'image of science'. I believe that the concepts
of 'paradigm' or of 'research programmes' would become more effective if they were
more fully articulated (especially the former, which is still very much in the same posi-
tion as Thomson's atomic model if one compares the latter with Rutherford's or Bohr's
models). If one had this more articulated reconstruction of the growth of scientific
knowledge, a revolution in the sense of Kuhn would, according to the ideas being enter-
tained here, be a drastic innovation undergone by the theoretical framework together
with the forsaking of some principle or accompanied by a different distribution of their
weight. For it appears that during scientific revolutions some principles - namely the
most central and protected ones, hence the ones which in a Kantian manner we may
consider to be constitutive - remain: thus, for instance, in astronomy the (derivative)
assumption of simplicity survived even after Kepler had abandoned the (local) assump-
tion of circularity of the orbits.
12 Cf. von Wright (1957), pp. 206-209; J. P. Day (1961), especially Chapter 5.2.4;
M. Bunge (1960); R. Camap (1966), Chapter 23.
INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 163
BIBLIOGRAPHY
Achinstein, P.: 1971, Law and Explanation (Clarendon Press, Oxford).
Agassi, J.: 1975, Science in Flux (D. Reidel, Dordrecht - Boston).
Bacon, F.: 1620, Novum Organon, in The Works of Francis Bacon, ed. by J. Spedding,
R. L. Ellis and D. D. Heath (Longmans and Co., London, Vol. 1, 1857).
Bechler, Z.: 1974, 'Newton's 1672 Optical Controversies. A Study in the Grammar of
Scientific Dissent', in Y. Elkana (ed.), The Interaction between Science and Philosophy
(Humanities Press, Atlantic Highlands, N. J.)
164 MARCELLO PERA
In the mechanism proposed by Charles Darwin to explain the origin and sub-
sequent evolution of living species, there are two essential factors: the chance
variation of hereditary characteristics and the selection by the environment of
those best suited for survival. These factors were already interpreted in Dar-
win's day as two main phases in the process by which an organism solves the
problem of its adaptation to environmental variations.
In Darwin we fmd the metaphor of an environment that 'selects' among
the diverse descendants of an individual or couple those endowed with char-
acteristics which are advantageous for their conservation and reproduction;
hence in about 1870 Samuel Butler introduced the metaphor of the organism
which 'solves the problem' of its own individual and species survival by bring-
ing about hereditable modifications in its own somatic structure or in its
behavior.
Today we include among these behavioral modifications all those that
involve the construction or improvement of exosomatic instruments (from
nests and incubators to traps and dams), these being more easily comparable
to the inventions of human technology and to their 'cultural' propagation.
The Darwinian metaphor no longer appears as such when it is made clear
that the environment is not to be understood as the anthropomorphized
counterpart of the farmers or breeders who effect an 'artificial' selection
among the natural products they deal with, but rather it is their work which
is to be seen as one of the 'natural' components of the environment in which
what are subsequently called 'domestic' animals and plants happen to live.
Similarly, the Butlerian metaphor is rendered more acceptable when it is
made clear that the subject of the natural actions of 'problem solving' is not
the single individual (as in Lamarck's two hypotheses of the voluntary
development of organs or 'habits' for the satisfaction of new necessities im-
posed by the environment, and of the hereditability of the characteristics so
acquired), but the whole evolving species to which that individual belongs.
The metaphor can then be applied also to the human species, whose
cultural progress is seen as the natural extension of the biological one, but
usually with the reservation that individual creative processes are really
individual, even if species-specific for homo sapiens, and that their 'mechanics',
167
MD. Grmek, R. S. Cohen, and G. amino (eds.), On Scientific Discovery, 167-177.
Copyright © 1980 by D. Reidel Publishing Company.
168 VITTORIO SOMENZI
no man could ever resolve, however intelligent and long-lived he might be.
Less removed from Campbell's radical position, which excludes true crea-
tivity in choices which are oriented by the information the individual starts
with, is the 'operant conditioning' interpretation of innovations - which are
cultural in an ethological sense - brought about through blind trials by
exemplars of different animal species and rapidly imitated by other members
of the collectivity to which these exemplars belong. From this point of view,
the greatest creativity seems to be that of the child, as it appears in the fIrst
phases of the acquisition of any language. Already in Piaget the Haeckelian
principle that ontogenesis recapitulates phylogenesis is used for a comparison
between the individual development of the child's thought with regard to
natural phenomena and the historical development of scientillc thought.
In more recent years Howard Gruber has pushed the analysis of creativity
farther in this direction of 'starting from scratch' and of exploratory curiosity,
both in the child and in the scientist, reaching the conclusion (implicitly in
agreement with those of evolutionary epistemology) that Darwin himself is
a typical example of the necessity of adding the adult factor of courage and
tenacity to the infantile factor of curiosity and creativity.
W. Koehler, too, was struck by the repetition in numerous historically
checkable cases of a link between the creativity of the scientist and a diver-
sion from his predominant concern, brought about by accidental circum-
stances like a walk or a trip, Archimedes' 'Eureka!' bath, and the less famous
half-sleep that permitted Otto Loewi twice to intuit the mechanism of
nervous action on the cardiac rhythm. But some British physicists confIrmed
Kohler's idea in such generalizing terms as to perhaps discourage further
analysis: "We often talk about the three B's: the Bus, the Bath and the Bed.
That is where the great discoveries are made in our science".
Systematic analyses were carried out later by William Dement and others
with the usual statistical criteria of experimental psychology, that is, they were
carried out on numerous subjects, but neither exceptionally gifted ones, nor
persons who were deeply motivated to solve in their sleep the uninteresting
problems put to them when they were awake. In the few positive cases found,
the result would seem to confIrm the hypothesis that the nocturnal inter-
ruption of the inhibition exercised by the dominant hemisphere on the minor
hemisphere favors the elaboration by the latter of an answer expressed in fIg-
urative terms to the problem put to the dominant hemisphere in verbal terms.
Translated by
R.MABERRY and A. TALIERCIQ
176 VITTORIO SOMENZI
BIBLIOGRAPHY
Ashby, W. R.: 1956, 'Design for an Intelligence Amplifier', in Automata Studies, ed. by
C. E. Shannon and J. McCarthy (Princeton University Press, Princeton).
Brown, J.: 1977, Mind, Brain and Consciousness (Academic Press, New York).
Campbell, D. T.: 1960, 'Blind Variation and Selective Survival as a General Strategy in
Knowledge-Processes', in Selforganizing Systems, ed. by M. C. Yovits and S. Cameron
(Pergamon Press, New York).
Campbell, D. T.: 1974a, 'Evolutionary Epistemology', in The Philosophy of Karl Popper,
ed. by P. A. Schilpp (Open Court, La Salle).
Campbell, D. T.: 1974b, 'Unjustified Variation and Selective Retention in Scientific
Discovery', in Studies in the Philosophy of Biology, ed. by F. J. Ayala and T. Dobz-
hansky (Macmillan, London).
Craik, K. J. W.: 1943, The Nature of Explanation (Cambridge University Press, Cam-
bridge).
Craik, K. J. W.: 1966, The Nature of Psychology, A Selection of Papers . .• , ed. by S. L.
Sherwood (Cambridge University Press, Cambridge).
Dement, W. C.: 1972, Some Must Watch and Some Must Sleep (Stanford Alumni Asso-
ciation, Stanford).
Eccles, J. C.: 1970, Facing Reality (Springer Verlag, Berlin, Heidelberg, New York).
Evans, R. 1.: 1975, Konrad Lorenz, the Man and His Ideas (Harcourt, Brace, Jovanovich,
New York and London).
Fogel, L. J., Owens, A. J., and Walsh, M. J.: 1966, Artificial Intelligence through Simu-
lated Evolution (Wiley, New York).
Gregory, R. L.: 1966, Eye and Brain (McGraw-Hill, New York).
Gregory, R. L.: 1980-81, 'Are Perceptions like Hypotheses in Science?', in Levels of
Reality (An International Symposium, Florence, September 1978), to be published
by Feltrinelli (Milan).
Grmek, M. D.: 1976, 'Le role du hasard dans 1a genese des decouvertes scientifiques',
Medicina nei secoli, No.2, 277-305.
Gruber, H. E.: 1973, 'Courage and Cognitive Growth in Children and Scientists', Piaget
in the Classroom, ed. by M. Schwebel and J. Raph (Basic Books, New York).
Gruber, H. E.: 1974, Darwin on Man (a Psychological Study of Scientific Creativity)
(E. P. Dutton, New York).
Hadamard, J.: 1945, The Psychology of Invention in the Mathematical Field (Princeton
University Press, Princeton).
Jaynes, J.: 1976, The Origins of Consciousness in the Breakdown of the Bicameral Mind
(Houghton Mifflin, Boston).
Kohler, W.: 1969, The Task of Gestalt Psychology (Princeton University Press, Princeton).
Lorenz, K.: 1973, Die Riickseite des Spiegels (Piper, Munich).
Mach, E.: 1943,Popular Scientific Lectures (Open Court, La Salle, Illinois).
Mahdihassan, S.: 1961, 'Keku1t!'s Dream of the Ouroboros and the Significance of This
Symbol', Scientia No.6, 187-195.
Medawar, P. B.: 1967, The Art of the Soluble (Methuen, London).
Newell, A., Shaw, J. C., and Simon, H. A.: 1963, 'The Processes of Creative Thinking', in
Contemporary Approaches to Creative Thinking, ed. by H. E. Gruber, G. Terrell, and
M. Wertheimer (Atherton Press, New York).
THE VIEWPOINT OF EVOLUTIONARY EPISTEMOLOGY 177
Piaget, J.: 1978, Behavior and Evolution (Pantheon Books, New York).
Poincare, H.: 1905,La valeur de la science (Flammarion, Paris).
Poincare, H.: 1908, Science et methode (Flammarion, Paris).
Popper, K.: 1963, Conjectures and Refutations (Routledge & Kegan Paul, London).
Popper, K.: 1968, The Logic of Scientific Discovery (Hutchinson, London).
Popper, K.: 1973, Objective Knowledge (Oxford University Press, Oxford).
Popper, K., and Eccles, J. c.: 1977, The Self and Its Brain (Springer, Berlin).
Pribram, K. H.: 1976, 'Language in a Sociobiological Frame', Annals of the New York
Academy of Sciences 280, 798-809.
Rumbaugh, D. M. (ed.): 1977, Language Learning by a Chimpanzee (Academic Press,
London, New York, San Francisco).
Simon, H. A.: 1966, 'Scientific Discovery and the Psychology of Problem Solving', in
Mind and Cosmos, ed. by R. G. Colodny (University of Pittsburg Press, Pittsburgh);
reprinted in Models of Discovery. Boston Studies in the Philosophy of Science,
Vol. 54 (Reidel Dordrecht, 1977), pp. 286-303.
Simon, H. A.: 1973, 'Does Scientific Discovery Have a Logic?' Philosophy of Science,
No.3, 471-480. (Reprinted in Models of Discovery. Boston Studies in the Philos·
ophy of Science, Vol. 54, Reidel, Dordrecht, 1977, pp. 326-337).
Toulmin, S.: 1971: 'Brain and Language: a Commentary', Synthese 22,369-395.
Toulmin, S.: 1972, Human Understanding, Vol. 1 (Clarendon Press, Oxford).
Wertheimer, M.: 1959,Productive Thinking (Harper, New York).
Young,1. Z.: 1951,Doubt and Certainty in Science (Clarendon Press, Oxford).
Young, J. Z.: 1978,Programmes of the Brain (Oxford University Press, Oxford).
Zaidel, E.: 1978a, 'Concepts of Cerebral Dominance in the Split Brain', in Cerebral
Correlates of Conscious Experience, ed. by Buser and Rougeul-Buser (Elsevier-North
Holland, Amsterdam).
Zaidel, E.: 1978b, 'Lexical Organization in the Right Hemisphere', Ibid.
DEREK DE SOLLA PRICE
Within the last few years a new level has been reached in a germinating field
of study that has developed from quantitative sociology of science, research
on science indicators statistics, general bibliometrics, and citation analysis.
From these advances I feel it is now possible to put forward, albeit tentatively
and with many reservations and much uncertainty at several places in the
technical detail, a comprehensive analytical theory of science. By this I mean
a conceptual framework that is consistent and extensive and can be related
both quantitatively and qualitatively to several places in the historical and
philosophical e~amination of science. It is the function of this paper to
explore the apparently far-reaching implications of this theory for the nature
of the concepts implied. This will be done in developing one topic and
discussing briefly two other separate researches in the field that display this
new approach to the explicand of 'Science of Science' or 'Science Studies',
viz:
(1) A study of the Ups and Downs in the History of Science and Tech-
nology 1 shows how quantification of almost anything in a time series can
throw light on what otherwise might be apparent only to a very competent
Toynbee-like historian. It also gives an objective method for commenting on
the periodization of science and suggests that the Scientific Revolution has
a precursor-like role and that the Industrial Revolution may be merely an
artifact of historiographical convenience.
(2) A study of Cumulative Advantage Processes by statistical mathe-
matics 2 gives results that go from a very simple probabilistic model to yield
quantitative laws in agreement with empirical evidence, and thereby explain
the peculiarities of all well-known regularities in scientometrics and biblio-
metrics. This gives a conceptual basis for the sociological hierarchies prevailing
in the scientific community.
(3) Recent discoveries by Griffith and Small show that science can be
mapped by using the technique of co-citation analysis. The resultant map is
surprisingly two-dimensional, and this indicates that irrespective of the map-
ping procedure, the structure of science can be ascertained in a very graphic
and provocative fashion. One important result is that the paper atoms cluster
into sub-discipline molecules each corresponding to an invisible college; there
179
M. D. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 179-189.
Copyright © 1980 by D. Reidel Publishing Company.
180 DEREK DE SOLLA PRICE
causes acting in concert, and these are seen against a background of random
fluctuation that is huge in the crude data. Statistically, as a rule of thumb,
if one generates data by counting a population of events sorted into pigeon-
holes, a count of N events implies a fluctuation of magnitude ±VJV. If one
has as many as 100 events in a group the random error is therefore 10% and
one cannot be sure of the reality of distinctions from one group to another
beyond such a limitation.
To avoid such problems, the present research demands only the minimum
from the data, the main trend of an overview of all scientific and technolog-
ical activity without any disaggregation by nation or by field. This ensures the
maximum number of events in each annual group, and to make the fluctua-
tions even less, we have combined many sources of data and applied well-
known smoothing procedures to the time series.
III. METHODOLOGY
The starting point was an examination and hand-count, page by page, through
a large number of chronologies and histories of science and technology,
recording each event that was given a precise or an approximate date. For
control purposes a similar but smaller study was made of general history. Ihe
year counts were then smoothed by taking a running weighted mean, the
weighting factors being given by cos 2 (90° xm/4), m = 0, ±l ,±2,±3) so-that
it ran for three years on either side of the target year. Zeros being absent
from the smoothed series, the logarithms of this mean were taken, a linear
regression against time computed, and the deviations from the regression
found. These deviations then gave for each source the amount by which it
exceeded or fell short of a linear increase in the logarithm, i.e. of regular
exponential growth which is the gross deterministic behavior. The deviations
from all sources were then averaged to give the results shown in Figure 2, and
this was further given a grand smoothing by a running weighted average, this
time taking in 10 years on either side of target, to give the fmal product
exhibited in Figure 1. The data was found to be sufficiently numerous for
some confidence only from 1500 to the present time, though any individual
source yielded information only over part of the entire range. A similar in-
vestigation for the General History control confirmed that the regularities of
Figure 1 were peculiar to the history of science and technology; systematic
fluctuations of this sort were far less well marked and quite different for the
general history events.
0.5
.....
00
N
COPERNICUS
SCIENTIFIC
REVOLUTION
)( 2.0 INDUSTRIAL
GALILEO
REVOLUTION
",
x 1.5-
"
steady
exponential 0 I . \ . \ '/ ., 1 /0: ~ {" ~ .y 'i'\ :/ \'. :I
growth
" "
BISMARK
-:- 1.5
BOER
CROMWELL
WAR
-:- 2.0 ININIININII
3D YEARS
WAR FRENCH (,
AMERICAN
REVOLUTIONS
-0.5+1;---;--~~--'----:-:~--"'----r----'---;------r--~
1500 1600 1700 1800 1900 2000
Fig. 1. Ups and downs in the pulse of science and technology.
0.5
x 2.0
x 1.5-'
o·
\
+1.5
+2.0 ~-
- 0.5 -l-,---;---r-----r----,----r---:-::-C-----r----:~--,---~;_-
1500 1600 1700 1800 1900 2000 ....00
w
Fig. 2. Ups and downs; fine detail.
184 DEREK DE SOLLA PRICE
IV. INTERPRETATION
V. HISTORIOGRAPHICAL IMPLICATIONS
It must be remembered that the indicator derived charts the secular and sys-
tematic swings of the exponential growth rate of science; it is not a measure
of the quantity of discovery and invention. It may be rather surprising that
over the entire range, the rate does not vary by more than about a factor of
two above or below the average long-term trend which we expect on theore-
tical grounds. Within these limits, I maintain the indicator gives one a measure
of scientific activity that agrees well with the historians' intuition of relatively
active and inactive periods.
Amongst the obvious expectations for eras of high activity one fmds very
clearly indicated the peaks of the scientific and the industrial revolutions as
well as localized outbursts in the times of Copernicus and Galileo (I use the
names only as surrogates for their periods). The major wars and social up-
heavals of history all correspond to localized troughs that are very clear.
Highly unexpected (at least to one historian) is the enormous trough
180
>-l
~
<lJ
::t:
CII t!1
>-
........ >
Z
UI
c: >
t"'
o
'.j:; 0<
<lJ >-l
,~
:::0
n
::l >
t"'
." a.
dQ' - >-l
0 ::t:
.;., ~
t!1
CII 0
..c ~
E 0<
~ 50
0
40 ."
U>
30 n
;;
20 Z
n
10 t!1
7"1.
1880 90 1900 10 20 30 40 50 60
WWI WWII
1913-19 1939-44
x2 in
3.6 years
100+---------------~----------------_+~~----~--~
10+---------------+----~~~~+---------------~--~
I I
1880 1890 1900 1910 1920 1930 1940 1950 1960
NOTES
BIBLIOGRAPHY
Pledge, H. T.: Science Since 1500, A Short History ofMathematics, Physics, Chemistry,
Biology (HMSO, London, 1939).
Walden, Paul: Chronologische Uebersichtstabellen zur Geschichte der Chemie von den
iiitesten Zeiten bis zur Gegenwart (Springer, Berlin, Goettingen, Heidelberg, 1952).
Williams, Neville: Chronology of the Modem World: 1763 to the Present Time (Mckay,
New York, 1966).
GABRIEL GOHAU
One can classify the whole range of pedagogical practices in two groupS:
on the one hand, those where the teacher alone speaks: I shall call these
expository methods: on the other, those where the teacher questions his
students: I shall call these interrogative methods.
Expository techniques are evidently those of text-books, as well as of
popularizing articles ... and lectures. In the classroom, they now meet with
a degree of disfavor. One has tended to label them 'dogmatic' since child
psychology has shown us that the young person is not an adult in miniature,
but someone needing to build his own knowledge progressively, by way of
successive re-adjustments.
Interrogative techniques came into being to respond to this need. There
the student is constantly solicited, so that the teacher is forced to follow, to a
certain extent, the course of his interlocutors' thought, with its hesitations,
deviations, even its temporary regressions. The archetype of these methods
is of course the Socratic dialogue, where the teacher manages, by way of
questions, to destroy the semblance of knowledge in the disciple, and replace
it by true knowledge.
This reference to Socratic maieutics shows that interrogative methods have
not waited for the development of child psychology to command the atten-
tion of the best pedagogues. Without aspiring to retrace the history of peda-
gogy, it may be said, very roughly, that the expository method seems to pre-
vail in periods when culture is being preserved, while the interrogative method
would be that of times of cultural creation, when previously acquired knowl-
edge is being questioned.
Thus, the Renaissance criticizes the dogmatic pedagogical methods of
the Middle Ages. To make a literary allusion, we can say that the expository
method is that of the first teacher of Rabelais's Gargantua, and the interroga-
tive method that of the second teacher. Rabelais jeers at the "great Sophist
doctor called Master Thubal Holophernes" who taught the alphabet so well
to his pupil that the latter "recited it by heart backwards". The second
teacher, Ponocrates, on the contrary, puts his pupil through all kinds of
exercises: for example, ''while passing through meadows or other grassy
places, inspecting the trees and the plants", he gathers some to study at
191
M. D. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 191-210.
Copyright © 1980 by D. Reidel Publishing Company.
192 GABRIEL GOHAU
home, and compare what he observes with what the ancients said about
them.l
I have chosen on purpose this short extract from Gargantua because
Rabelais alludes in it to the observation of nature (here plant matter). Now I
must speak of the pedagogy of science teaching, in which observation (as well
as experimentation) is a fundamental component. In this way the opposition
between the interrogative and the expository methods is governed in the fmal
analysis by an opposition between a process which leads to progressive dis-
covery by way of repeated observation and experiment, and a technique
whereby one recites the rational exposition constructed by the teacher, where
the observation of objects and phenomena is replaced by the reproduction of
schema. That is why this method is sometimes called 'bookish': for it pro-
ceeds, as books do, by way of exposition and schema.
But opposition is found as well at another level. The interrogative method
builds up knowledge, that is to say laws, theories, and general concepts, using
observations and experiments as a starting-point. It proceeds therefore by
induction, moving from the particular to the general. By the contrary method,
it is usual for the teacher, who unaided sets out knowledge, to start from
principles whose particular cases he deduces. If therefore he introduces
experiments into his teaching, he puts them at the end of the lesson as an
illustration of the theory. His method is deductive.
Finally, in science teaching, this distinction between induction and deduc-
tion is more fundamental than the exposition-dialogue opposition or the
opposition between a class without experiments and a class with experiments.
For today few teachers reject dialogue altogether, and all are obliged by min-
isterial instructions to experiment and to have their pupils handle apparatus
(during periods of practical work).
Because it aims at a construction based on the pupils' observations, the
inductive method may pride itself on following the thought of the young in
its development. Moreover it tries to reproduce the progress of science, mov-
ing from particular facts to laws and general theories: - a fact which wins it
a certain prestige in the field of science pedagogy. However, the relationship
between pedagogical method and the progress of science is not so Simple.
The place of induction in the progress of science has been discussed by
several of the previous speakers. I will limit myself to referring you to their
papers, adding that here I understand by induction, in a very wide sense, any
process of scientific discovery.
Induction, thus interpreted, evidently diverges from that of philosophers
(amplifying or generalizing induction). The logician R. Blanche 2 humorously
THE TEACHING OF THE SCIENCES 193
compares the first to a high jump (in the sense that it invents, and rises above
already known facts), the second to a long jump, for it simply extends what
is already established fact.
Thus defined, inductive methods tend to rediscover the progress of science,
of scientific discovery. They are therefore of greater concern to the historian
of science than deductive methods which, starting from principles, aim only
at the logical exposition of a theoretical edifice the construction of which is
supposedly completed.
However, their utilization poses an epistemological problem, for given a
state of ignorance on the science teacher's part vis-a.-vis the history and epis-
temology of his subject, inductive methods have led to absurdities in the
belief that it was enough to reverse the traditional deductive method, and
establish an active pedagogy of discovery. To demonstrate this, I shall ex-
amine the method used in the French teaching system since the end of the
war, and known by the name of the technique of rediscovery.
Proposed by Inspector-General Brunhold, this technique ties up with an
inductive tradition begun in the 1880's, at the time of the passage of the lay-
school laws, a fact which, in passing, emphasizes that a revival of inductive
methods accompanies periods of progress.
In a text of 1948,3 Charles Brunhold thus defines his objective: it is in
no way concerned with "repeating the historical progress of this or that
research", but only with "retracing, with the means at the teacher's disposal
today ( ... ) the results of such research". The author specified that if his
reflection had led him to opt for rediscovery rather than a stricter historical
method, it was because of "the necessity for many of our pupils, who intend
to take up scientific or technical careers, to absorb, prior to any specialized
training, the total picture of fundamental knowledge": something which, he
considered, would not have been possible with purely historical methods in a
study time of reasonable length.
In other words: 'rediscovery' aims at the acquisition, by means of the
'spirit' and 'method' of the history of science, of this 'indispensable knowl-
edge' which was previously instilled by expository processes. A praiseworthy
undertaking, if one can give the young the same learning in a manner more
formative for their minds, by following more closely their mental develop-
ment.
But the fundamental ambiguity of the method in question comes from the
fact that it believes it legitimate to erase the obscurities and gropings of his-
torical progress. "Everything holds together when the whole is built" said
Gaston Bachelard 4 to emphasize that after discovery all the uncertainties on
194 GABRIEL GOHAU
the path of research are wiped out by the one who knows. Then verification
(or confirmation) experiments, capable of testing the theory, can be set up.
This means that the path which leads from the hypothesis to its confirmation
is not the exact opposite of the one which leads from preliminary observa-
tions to the formulation of the hypothesis. We find once more an opposition,
typical in the history of science, between science-in-the-making and estab-
lished science. I would not insist on this point known to all of you if 'redis-
covery' were not ignorant of this primordial distinction.
For indeed, at the prompting of Brunhold, appointed Director of French
secondary education, teachers and text-books reversed the order of exposi-
tion of theories and experiments. This took place in such a way that an
experiment, which was used until then to illustrate a lesson, at the end of a
chapter, was suddenly placed in the introduction, the lesson deriving from an
'interpretation' of the handling of apparatus. To give an idea of the applica-
tion of this method, here is a short extract from the draft syllabus presented
in 1966 by the General Inspectorate of natural science. It will be seen that
when I state that the same experiment which formerly illustrated the lesson
today serves to introduce it, in no way do I exaggerate. I shall take as an
example lesson 20, entitled 'chlorophyllian syntheses'.
"Practical work:
Analyses of experiments showing
- that a green plant can grow in a purely mineral medium,
- that certain elements are necessary in this medium ( ... )
Experiments on an aquatic plant revealing a release of oxygen ( ... )
Experiments leading to relative measurement of the intensity of release. A
study of the influence of one factor on this intensity ( ... )
An expression in graph form of the results.
Lesson: chlorophyllian gaseous exchanges."
The final version made the wording much less rigid by eliminating practical
work-lesson division, and giving more freedom in the choice of experiments.
But the spirit remained the same. That is, in the General Inspectorate's intent,
the class is in particular invited to perform and analyze a series of complicated
experiments for the purpose of deducing from them the notion of chIoro-
phyllian exchanges. All this, last but not least, in two hours of practical work
and one hour of teaching.
A comparison between this extravagant list and the time allotted for
studying it would make one smile if it did not reveal a completely erroneous
idea of the scientific process. Thus, to imagine that a scientific notion can
THE TEACHING OF THE SCIENCES 195
to fulfil this verifying (or corroborating) function, he would have known that
he could not be satisfied with those coarse 'tinkering' manipulations con-
cerned only with illustrating his exposition. An example will help us to clarify
this point.
Not long ago the world of natural science teachers in French high schools
was swept by a protracted debate about the use of methylene blue to show
up incidences of tissue dehydrogenase. I leave on one side details on the
theses opposing each other. It is enough for our purpose to know that one of
my friends, a biochemist by profession, performed a careful series of experi-
ments to resolve the question at hand. His answer 8 was that methylene blue
is in fact discolored by cases of dehydrogenasis and therefore enables them to
be shown up. All teachers who used methylene blue to show that the first
stage of cellular oxidation is a case of dehydrogenasis were immediately
reassured: the experiment they used was conclusive.
So be it. Yet, if Claude Bourgeois's sensitive experiments had been found
necessary to establish a correct 'interpretation' of our summary classroom
experiment, was that fact not also proof that that coarse manipulation was,
by itself, not demonstrative? Now, when Bourgeois carried out his study, all
natural science teachers in France used that non-conclusive experiment to
allow their pupils to discover (!) the presence of enzymes extracting hydrogen
from cellular metabolites.
Whether the interpretation given was exact or not is of no interest. Cer-
tainly, it was and it is evidently all the better. But, had it been false, our
pupils would none the less have 'rediscovered' the existence of the dehydro-
genases. Classroom experimenting, therefore, proves absolutely nothing,
because it is non-'falsifying', since the teacher teaches only exact theories -
or at least those which pass as such at the level of secondary school education.
It is therefore out of the question that the experiment should 'fail'. Let us
imagine that the experiment leads to a negative result: it is so crudely per-
formed that this is frequently the case. What does the teacher do? He evi-
dently will not refute the theory he believes to be exact, but blame the
experiment which has 'failed'.
Consequently, the experiment cannot, by rights, falsify the theory. Its
testing value is strictly worthless. It holds no information on the theory. The
pupil is deceived if he is allowed to believe that he has proved - or worse:
that he has discovered - the theory. When the experiment yields the hoped-
for result he is skilfully guided towards the correct interpretation ... which
he must underline in red. When it gives another result, one cannot help con-
fessing to the deception, while stating all the same, the 'correct' conclusion,
198 GABRIEL GOHAU
away they put before our eyes experiments done with a view to proving these
laws after they had been stated. This is doubly incorrect.
On the one hand, the experiment is 'parachuted'. Like Athena, it springs
fully armed from the teacher's brain. The traditional expository method is
called dogmatic, but is one less dogmatic when one explains an experiment
than when one explains a theory? In both cases, the teacher presents a fully
worked-out knowledge, sparing his pupils the effort of analysis which allowed
the progressive discovery of that knowledge.
On the other hand, the pupil is allowed to believe that the implication
uniting theory to experiment is reciprocal: which is nonsensical from the
point of view of formal logic. The experiment, a particular fact, is the con-
sequence of the theory, a general statement, but the reverse is obviously
absurd.
Finally, the method of rediscovery is altogether dogmatic, but insidiously
so: it imposes theory as much as the expository method. Even more so, since
it claims to 'deduce' it from an experiment the pupils have seen. Cannot its
hidden dogmatism, its cryptodogmatism, be called bi-dogmatism since, not
content with imposing the teacher's knowledge, it claims to base it on experi-
mentation?
The danger of this technique lies in the fact that the pupil, when he learns
later on that the knowledge instilled by his teachers was imperfect, will be
tempted to lose all belief in science, and to reject the whole scientific edifice.
He will perhaps take refuge in parapsychology or will throw himself into the
arms of one of those sects which flourish today. We should fear that, dis-
appointed with the false picture of science we give them, our pupils turn
towards hoaxers or quacks.
This severe criticism of the present state of inductive techniques may
appear excessive, especially for those who still live under the yoke of exposi-
tory methods. But precisely, the lesson must be learned without indulgence,
from a failure out of which French pedagogy has not yet found its way. Now
let us try to show what would be a veritable pedagogy of discovery which
would not turn its back on the history of science.
If one wants a pupil to discover, one must put him in a context of genuine
research. In the pedagogy of rediscovery, the situation of 'researchers' is
altogether artificial: thanks to a few summary experiments brutally hurled
before their eyes, they are required to refind a major scientific theory, which
took centuries of groping to work out. What is absurd, is not to believe the
pupil capable of performing research, but to think that in a few hours he will
rediscover a great concept of biology or physics.
200 GABRIEL GOHAU
The pedagogy of rediscovery was on the right path, but its ambition was
excessive. The Nuffield Foundation project, tried out in Great Britain,has far
better gauged possibilities of the young in the field of experimental research.
This project provides for 16-18 year old ('A level') students to devote a
tenth of their study time to 'a project or independent work' which is a
veritable program of experimental research. 9
The student may tackle a new problem, but most often he will be asked
to modify a technique he has already used and whose results he has studied.
For example, he will apply to a different animal a method of analysis studied
previously in class. He will be able to compare his results with those of the
bibliography. This type of research is clearly relevant to what Kuhn calls
'normal science': applying an experimental technique that one has mastered
to new materia1. 10
This research evidently requires only a modest amount of initiative. How-
ever, it offers original results, as opposed to rediscovery which simply repeats
things long known. From this point of view, the independent work of the
Nuffield project is far more satisfying for the student's mind than the pseudo-
discoveries of so-called inductive methods, for if the research involved is
modest, the discovery is real.
Mter all, this is basically quite normal, for the newest discoveries do not as
a rule start from fundamental questions. When he turned his attention to the
rotatory power of tartrates, Pasteur did not realize that he was going to lay
the foundations of a new science. Major discoveries are often solutions to
small problems ... which one has been able to carry out to the end. 'Normal
science' is the only one to have a real existence. What Kuhn calls the change of
'paradigm' has an existence only retrospectively. Pasteur introduced such a
change through his perseverance in following up the implications of his discov-
ery of molecular dissymmetry. At what moment his research comes out of the
framework of normal science is not easy to say, and in any case can be known
only after the event, when the consequences of his discoveries are known.
Finally, the pedagogy of rediscovery approaches problems inside out: it
believes that, under the pretext that a few experiments become a posteriori
crucial, one has the right to isolate them, to extract them out of their tissue
of obscure, stubborn researches and to reduce all scientific activity to those
turning points of its history. But if one wishes the pupil really to discover,
to produce an original result, it is impossible to choose a crucial experiment
for him: the already quoted example of Claude Bernard has shown us ade-
quately that an experiment is crucial only if one has in mind the new idea
one wishes to test.
THE TEACHING OF THE SCIENCES 201
The authors of the Nuffield project achieved what Brunhold had not
dared: a dissociation of the acquisition of scientific knowledge from appren-
ticeship in its method. A piece of research conducted by the pupil cannot aim
at discovering a major concept or theory. The two activities are quite different
if not opposed. They must be separated.
In the case of young children (5 to 13 years) this same project makes
generous allowance for student activity and initiative. 11 A few French teach-
ing establishments are experimenting in France in the same direction, under
the guidance of V. Host and J. P. Astolfi (in biology). The aim of these
scientific activities is to create an experimental attitude in the young. The
main point is that the students be brought to ask each other questions, then
to try to resolve them by themselves through experimentation. There is no
thought of teaching them science, but only of making them understand what
constitutes research.
But naturally this dissociation leaves untouched the problem of the
acquisition of knowledge. One may suggest placing side by side an active
apprenticeship in the spirit of research and a scientific teaching of the tradi-
tional type, presenting science in its most rational form. It seems to me that
under such a system one would lose part of the benefit of inductive methods.
It is possible to present acquired knowledge in a new way which avoids the
artifices of rediscovery techniques.
The great inconveniences of deductive methods is their dogmatic nature:
scientific knowledge is from the start presented under its present guise of
perfection. There is lacking in them, if not real induction, at least progres-
sive construction, either following the history of the discoveries, or finding
again the most general concepts starting from partial knowledge.
Genetic psychology has shown us that the pupil needed to assimilate
knowledge, to digest it in some way. Now deductive methods present it in a
form ... which is indigestible. We have a mania for wanting to teach from the
start our most abstract conceptions. Would it not be possible to distinguish
levels in the acquisition of knowledge? Respiration is first an alternative
movement, then it is a gaseous exchange; then it is transportation of gas,
fmally, it becomes an exchange of electrons.
The advantage of drawing out progressively the abstract concepts is
evidently to make their acquisition easier, by taking them to pieces. But it
also consists in associating one's pupils with the way they were worked out.
Certain general concepts appear as means of unifying fields of science until
then independent: the theory of evolution is of this kind. In a well-conducted
lesson the pupils will be brought, not to discover these unifying theories, but
202 GABRIEL GOHAU
the basics of knowledge. But the paths of apprenticeship must not be too
direct: one understands well only what one teaches.
Let me specify that when I speak of a book as the main instrument of
information (alongside the teacher and personal observation), I do not mean
to exclude modern aids to the transmission of knowledge. We do not live in
Rabelais's time, and our pupils can learn by way of records, the cinema or
television, as well as books. I have absolutely no intention of neglecting
audiovisual techniques, and I do not share the scorn with which too many
teachers still treat them.
However, the cinema and television must be dealt with in the same way as
books, and all I have said previously applies to them. In other words, teaching
will not be modernized because the teacher or the textbook have been re-
placed by a television screen which will broadcast knowledge in just as
dogmatic a form. (Even more dogmatic, for it suppresses all dialogue and
increases the distance between the one teaching and the one taught: a
mediocre teacher speaking in front of thirty listeners is often more effective
than an excellent teacher on the air reaching (?) millions of viewers.)
The question of knowing how a television program can replace a formal
lecture i.s outside my topic, since ... we are considering how to do away with
the formal lecture. Vis-a-vis our problem (how the pupil can build up his
knowledge with the help of documents), audiovisual techniques have the
same status as books. What applies to the one extends to the others. A tele-
vision program that has been too well processed is no more valuable than a
textbook or a formal lecture.
I said earlier that to understand a problem, one must teach it. Bf'cause
to assimilate processed knowledge, one must oneself have practiced its pro-
cessing. The scientist who hears a colleague's paper (or reads his article)
understands it better than the public because he discovers what lies behind
it: he can picture the experimentation and the vicissitudes which have led to
the stated result.
This remark can be extended to the book: to understand a textbook one
must be capable of writing it. And with regard to television: to watch a pro-
gram with awareness, one must have produced some. Learning by pictures
is first of all learning what a picture is. The pupil can therefore build up his
scientific learning ACTNELY by other means than mere experimentation.
However, if he learns his lesson from books or audiovisual techniques,
if he initiates himself to experimentation through personal research, does
he not lose practically all prolonged contact with his teacher and peers
alike? Techniques of rediscovery have accustomed us to classes by dialogue,
204 GABRIEL GOHAU
where the teacher directs and controls the pupils' instruction. Is all dialogue
impossible?
Undoubtedly the dialogue of our so-called 'active' classes is illusory. It
resembles certain Platonic dialogues where the pupil is content to punctuate
the master's discourse with approving remarks: 'as you say', 'certainly', 'you
are right'.12 G. Leroy, a Belgian educationalist, in an excellent little book 13
gives examples of the kind of dialogue in use in our classes. He shows that the
teacher's skill consists in asking sufficiently precise questions - they are called
'closed' - so that the pupil necessarily gives the expected answer. When a
mathematics teacher says: 'I write a2 and a X a. What does a2 stand for?',
one can only reply: a X a.
According to Leroy, there are four or five times as many narrow ques-
tions as broad ones. Moreover, even when the question is apparently open,
the teacher often expects only one answer. Thus, the teacher who asked,
'How does the compression of a gas occur?' wanted the answer, 'through
liquefaction'. Pupils gave various answers the least logical of which was
certainly not that of the boy who simply said, 'With a pump,!14
However, in spite of these criticisms, the fact remains that.dialogue exists.
Should one not proceed for dialogue in the same way as for experimentation
and induction, the two other characteristics of the method of rediscovery:
preserve, but appreciably improve them?
Now dialogue, in a veritable pedagogy of rediscovery, would have a funda-
mental role to play: to allow the expression of the pupils' 'representations'.
Teachers forget that pupils are not virgin wax on which are marked all the
impressions inscribed thereon. J. Piaget's school has shown for a long time
now, that the minds of the young traverse, mutatis mutandis, the important
stages of the history of thought, from the first civilizations to contemporary
science. Such misapprehension introduces a fundamental ambiguity in the
teacher-pupil dialogue, for the pupil burdens with affective relationships and
egocentric representations what the teacher expresses at a purely rational
level.
In other words, the aim of dialogue ought to be to rid children's minds of
what Bachelard called epistemological obstacles. 15 Teachers believe they
form their pupils' minds by loading them with the results of science. But they
are unaware that these minds are already full of pre-scientific representations,
magical conceptions, phantasms which encumber them. Let us begin by rid-
ding them of these. Let us perform what Bachelard calls a 'psychoanalysis of
the objective mind'. For the dialogue I propose is not essentially different
from that which develops between the psychoanalyst and his patient.
THE TEACHING OF THE SCIENCES 205
natural body is simpler than an industrial product? Now these are the obstacles
which have impeded the birth of modern chemistry.
To dispel that sloth, that clumsiness of thought, in the varied forms of
resistance to scientific progress, which one comes up against in every genera-
tion in the course of its apprenticeship, that slowness of mind which causes
them to confuse things which look alike, and find simple what is familiar:
such could be the task of a history of those obstacles. But the historian would
have to be certain of the collaboration of psychogeneticists and teachers who
would help him detect the same obstacles in the manner of thinking of young
people, so as the better to recognize them.
Once such obstacles are identified, solutions would have to be invented to
overcome them. For the teacher is not a collector of obstacles. Certainly, to
bring to the young person's level of consciousness the complex of analogies,
of false images which delay his acquisition of scientific knowledge, would
already be something decisive. Clumsiness of thought is all the more active
when it remains unconscious. However, the history of science offers us also a
choice method of overcoming them by highlighting the ways in which one
clears obstacles. Bachelard called these 'epistemological acts' .16
I am going to try to give an example, relating to the cellular theory.
Cellular division allows us to understand the transition from the single cell
which is the egg, to that complex assembly of cells which is a pluricellular
creature. Now the mechanism of this division has been understood only
belatedly in relation to the elaboration of the cellular theory. To explain the
development of the living creature, biologists for a long time preserved old
ideas inherited from the eighteenth century. Around 1830, the question
asked was whether cells are formed by the enlargement of a pre-existing germ
or whether new cells are incorporated through feeding and conveyed by the
blood. The slothful mind prefers to believe in their pre-existence rather than
explain the genesis of cells (or else he evokes a mysterious spontaneous
organization, starting from a homogeneous matter).
But in the same period (1800-1940) 'lower' animals were studied actively
and, among other things, animal colonies were discovered in the hydrozoa
and the tunicates .... 17 Now the organism is after all only a colony of cells,
and cellular multiplication is not far removed from the burgeoning of the
colony from the original polyp. Again in the same period it is noticed that
ringed worms (Annelida) are formed from an undivided larva by the sprout-
ing of successive rings. Cannot these observations, so much easier to make
than that of cellular division, have provided a model for it? Do they not
constitute the way towards the epistemological act which overcame the
208 GABRIEL GOHAU
obstacle of the pre-existing germ? Could they not be used to help young
people to understand cellular division? The study of such questions would
help a new pedagogy of the sciences to be thought out.
In being satisfied with turning upside down the traditional method, and
transforming the experiment of illustration into a pseudo-experiment of dis-
covery, the pedagogy of 'rediscovery' has thought itself capable of reforming
science teaching without the expense of methodological and epistemological
reflection. The result is disappointing. Experiments are performed but in a
totally artificial way, since the theories are hidden in the experiments and
smuggled in. The pupil's mental development is followed, but answers are
whispered to him . . . to such a point that some teachers write them in
advance in their lesson notes. Experimentation and dialogue, each one emptied
of substance, have brought under suspicion any recourse to books. At a time
when television gets ready to supplant books, teachers are still in the 'pre-
Gutenberg' era.
An authentic pedagogy of discovery is indispensable, but it would require
teachers to be trained as researchers. How can one initiate young people in
experimental research if one has had only a purely theoretical education?
It would also call for the history of science to help them in the development
of a history of epistemological obstacles. Since this meeting provides me with
an international audience, I especially want my appeal to be heard.
To conclude on a modest note, I shall say that the pedagogy of discovery
should aim more at strengthening than at instilling the spirit of research. For
it is not certain that a taste for enquiry, an intuition for hypothesis, the
genius of demonstrative experiment are entirely shaped by school exercises.
My friend Evry Schatzman, a French astrophysicist, wrote recently that to
form a scientific mind, "early contact with experimentation is very important.
A very great number of experiments may be conducted by the children them-
selves". After quoting a few examples, he added that "this can take place in
the school setting but also within the home environment". Later, he also
expressed the belief that "the spirit of research is not acquired through
contact with a teacher", for the latter "can beget neither curiosity nor a
taste for discovery". 18
To add a polemical touch, I would readily declare that one would be
satisfied if the school preserved the spirit of discovery, for very often, alas!,
present-day teaching sterilizes the child's curiosity. May masters and parents
promote the flowering of this curiosity, may they goad it on at every turn
and not only during science lessons.
Does not the unravelling of detective stories encourage the maturing of the
THE TEACHING OF THE SCIENCES 209
scientific mind? Some of Edgar Allan Poe's short stories - I am thinking pre-
cisely of The Murders in the Rue Morgue 19 apply the procedure of science
exactly to the successful pursuit of a search for a murderer: observation aimed
at finding clues, the working-out of a hypothesis in breach of common sense
(which expects a murderer to be a human being, when the murderer is in fact
here an orang-utang), then verification of that hypothesis (through a small
advertisement ).
After all, what if the scientific spirit were to be formed by the reading of
good detective stories?
Translated by
MARGARET ROUSSEL
NOTES
Rabelais, F., Gargantua. 1534. [The Histories of Gargantua and Pantagruel, trans. by
J. M. Cohen, Penguin, Baltimore, 1969.]
2 Blanche, R., L'induction scientifique et les lois naturelles (Presses Universitaires de
France, Paris, 1975).
3 Brunhold, C., Esquisse d'une pedagogie de fa redecouverte dans f'enseignement des
sciences (Paris, 1948).
4 Bachelard, G., Le rationalisme applique, (Presses Universitaires de France, Paris,
1949).
5 On logical empiricism one may consult the following:
Camap, R., 'The Aim of Inductive Logic', in Nagel, Suppes, Tarski (eds.), Logic, Meth·
odology and Philosophy of Science (Stanford University Press, Stanford, 1962) and
Hempel, C. G., Philosophy of Natural Science (Prentice Hall, Englewood Cliffs, 1966).
6 Popper, K. R., The Logic of Scientific Discovery (Hutchinson, London, 1959).
7 Grmek, M. D., Raisonnement experimental et recherches toxicologiques chez Qaude
Bernard (LibI. Droz, Geneva, 1973).
8 Bourgeois, C., 'Utilisation du bleu de methyJene pour mettre en evidence les deshy-
drogenases respiratoires', Bull. A.P.B. G. (Paris, 1966).
9 Nuffield A·Level Biological science, 14 volumes in all; on independent work: Triker,
B. J. K. and Dowdeswell, W. H. Projects in Biological Science (Penguin Books, Har-
mondsworth,1970).
10 Kuhn, T., The Structure of Scientific Revolutions (University of Chicago Press,
Chicago, 1962).
11 cr. Nuffield Junior Science (Collins, London, 1967). Emphasis is put on 'problem
solving'. "The role of the school is not to teach results but to use the scientific process
of research as an educative tool".
12 Plato, The Early Dialogues; for example, The first Alcibiades.
13 Leroy, G., Le dialogue en education (Presses Universitaires de France, Paris, 1970).
14 Lazerges, G., 'Techniques de la classe', Special number of the Bulletin de f'enseigne·
ment public au Maroc, May 1956; also Paris, 1959.
210 GABRIEL GOHAU
Does vocation exist within naturalistic knowledge? It seems so, if one goes
through the autobiographies of Einstein, Freud and Jung. Einstein was born
in 1879, and when he was close to seventy, he wrote a long, autobiographical
essay for the volume that Schilpp's 'Library of Living Philosophers' 1 devoted
to him. He asserts that he has no doubts ("es ist mir nicht zweifelhaft") that
our thinking goes on for the most part unexpressed in words and uncon-
sciously. Otherwise, our wonder about some experiences could not be ex-
plained. "This 'wondering' ["dies 'sich wundern'''] seems to occur when an
experience comes into conflict with a world of concepts which is already
sufficiently fIxed in US".2
A wonder of such nature [Einstein continues] I experienced as a child of four or five
years, when my father showed me a compass. That this needle behaved in such a deter-
mined way did not at all fit into the nature of events, which could not find a place in the
unconscious world of concepts (effect connected with direct 'touch') .... I can still
remember - or at least believe I can remember - that this experience ['Erlebnis'] made
a deep and lasting impression upon me. Something deeply hidden had to be behind
things. 3
Even beasts of prey would lose their voraciousness if they were forced to
devour continuously under the threat of the whip - it is Einstein speaking. In
211
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 211-226_
Copyright © 1980 by D. Reidel Publishing Company.
212 VINCENZO CAPPELLETTI
Young Einstein compared actual situations with each other, as he had done
when his father had shown him a compass with the needle moving with no one
'touching' it: and the comparison became a problem, a problem of identity,
analogy or difference of situations, which are similar in some aspects or op-
posite in others. The 'fundamental' physical concept that sixteen-year-old
Einstein was considering was the 'field': it surely revived the memory of the
compass needle that he had seen when he was four and brought into conscious-
ness; but in the unconscious the memory of a childhood observation stated in
rational terms represented the dynamic factor. And also the other experience
of the boy Einstein - his contact with geometry and its proofs - influenced
the genesis of relativity. Einstein asserts that he had been looking for a 'uni-
versal formal principle' , like the principle of thermodynamics, which states the
impossibility of constructing a perpetuum mobile, supposing the laws of na-
ture to be what they are. An essential step towards axiomatics, towards defini-
tions that, like the geometric ones, represent an absolute beginning: a cogni-
tive stage later examined in Geometry and Experience, published (in German)
in 1921. 7 The principle of special relativity, of value equal to the above-men-
tioned principle of thermodynamics, is the following: "The laws of physics
are invariant with respect to the Lorentz-transformations (for the transition
from one inertial system to any other arbitrarily chosen system of inertia)". 8
In essence, a basic principle of natural laws, like that of thermodynamics,
excludes the existence of perpetual motion. But let us pick up the thread of
our argument and try to draw a conclusion. Einstein's autobiography shows a
singular continuity between scientific observations of childhood and adoles-
cence, and the discoveries of adulthood. The premise on which this continuity
is based is wonder: the habit of asking why and how each phenomenon
diverges from what is traditionally or customarily considered lawful, regular.
DISCOVERY AND VOCATION 213
Freud's autobiographical writings number more than one. 'On the History
of the Psychoanalytic Movement'9 starts with the years following his M.D.
degree, that is, his journey to Paris in October 1885, his friendship with
Breuer and subsequent estrangement, and his first works. So far, nothing of
interest for us. The Autobiographical Study, dated 1925,10 however, includes
some interesting suggestions. He entered gymnasium, the 'Sperl Gymnasium'
in Vienna, when he was nine, and in the following seven years, he says
(actually only six, according to the investigations of his biographers who
succeeded in checking such minute details) he was at the top of his class. No
precocious disposition towards medicine.
I was moved, rather, by a sort of curiosity, which was, however, directed more towards
human concerns than towards natural objects, nor had I grasped the importance of
observation as one of the best means of gratifying it. 11
alienation while playing with his playmates that "alienated me from myself',
surely due to difficulties connected with symbolism, when coming from a
private sphere of meanings and going back to a public one of intersubjective
references. And yet another, the experience of fire, of its ineluctable 'burn-
ing', relived thirty years later, coming back to the same places, in the garden
of his father's house. At school, the contact with a teacher for whom
algebra was a perfectly natural affair, to be taken for granted, whereas I didn't even
know what numbers really were. They were not flowers, not animals, not fossils; they
were nothing that could be imagined, mere quantities that resulted from counting. 1S
Wonder, vocation for knowledge, feeling for the difference between under-
standing and learning: these are the characteristic features of the first contact
with science - or with experience lived through as science - of three great
discoveries of reality. Let us immediately make one observation: during the
same years in which Einstein and Freud, as children, were opening their
minds, the one to 'Wunder' and the other to 'Ahnung einer Aufgabe', 'grown
up' science vindicated the necessity of these fundamental situations for the
knowing subject. Ernst Mach, one of the protagonists of scientific innovation
at the end of the nineteenth century, offers an example of this. The Science
of Mechanics published in 1883 19 is contemporary with little Einstein's
wonder at the compass needle moving without being touched. It is an extrin-
sic circumstance, which will acquire significance in what we are about to say.
With Mach science became historiography in order to regain its own origins
DISCOVER Y AND VOCA nON 215
Let us turn to the analogies between Mach's Science of Mechanics and the
autobiographical accounts mentioned earlier. In Mach, too, wonder has a
precise cognitive relevance:
But the aim of my whole book is to convince the reader that we cannot make up proper-
ties of nature with the help of self-evident suppositions, but that these suppositions must
be taken from experience. 24
This is the clearest programmatic statement to be found in the whole of
Mach's work. Sensation reveals the world, which, being what in fact is, is
undeducible. Sensation precedes analysis, since the world is given before it
can be asserted. Static moment, acceleration, inertia, the impossibility of
perpetual motion: all of these are mechanical qualities, which cannot be
deduced, which become engraved in our representations. They are "a treasure-
store which is ever close at hand and of which only the smallest portion is
embodied in clear, articulate thought".25 And how about wonder? Wonder is
the appearance, in the cognitive process, of novelty and unpredictability of
nature. Mach quotes Aristotle, not the Aristotle of 'thaumazesthai',26 but the
author of the mechanical writings, explaining why the lever appears wonderful:
If a thing takes place whereof the cause be not apparent, even though it be in accordance
with nature, it appears wonderful (. . .). Such are the instances in which small things
overcome great things, small weights heavy weights, and incidentally all the problems
that go by the name of 'mechanical'. (... ) To the apories (contradictions) of this charac-
ter belong those that appertain to the lever. 27
The lever contains, embodies in itself one of the world's peculiarities, the
static moment:
As a matter of fact, the assumption that the equilibrium-disturbing effect of a weight
P at a distance L from the axis of rotation is measured by the product P' L (the so-called
statical moment) is more or less covertly or tacitly introduced by Archimedes and all his
successors.2 8
adhered throughout to mechanics as the secured basis of physics. It was Ernst Mach who,
in his History of Mechanics, shook this dogmatic faith; this book exercised a profound
influence upon me in this regard while I was a student. I see Mach's greatness in his
incorruptible skepticism and independence; in my younger years, however, Mach's
epistemological position also influenced me very greatly, a position which today appears
to me to be essentially untenable. For he did not place in the correct light the essentially
constructive and speculative nature of thought and more especially of scientific thought;
in consequence of which he condemned theory on precisely those points where its
constructive-speculative character unconceivably comes to light, as for example in the
kinetic atomic theory. 30
There is no place in Mach for thought: and, in fact, the third component of
his construction, 'connection', appears as weak, groundless, with respect to
'sensation' and 'element'. Nonetheless, for Mach, also scientific reasoning
teaches how to read, how to interpret the sensation: and in the course of this
reading, as we saw, wonder fades into evidence. The latter, since it is capable
of absorbing any novelty, is before and beyond the marvelous and causes
its disappearance. Evidence is the organ of the 'great whole'. "Galileo and
Huygens used to constantly alternate the consideration of the particular
phenomenon and the great whole", he writes in a passage of the Science of
Mechanics, which is reminiscent of a passage in Plato's Phaedrus concerning
Hippocrates' method. Sensism becOlp.es a program, a password, the equiva-
lent to idealism.
Mach's epistemology and the embryonic pedagogy of science found in
the Science of Mechanics or derivable from it, assumed a psychology. Other-
wise, the postulate of sensation would have remained hazy and, above all,
devoid of connections with experimental research. Well, Mach was a psy-
chophysicist, the author of Lessons of Psychophysics, 31 convinced supporter
of Fechner's theory of parallelism, but not of his work on the unique, un-
knowable substance of which physical and psychical facts would be attributes.
"My natural bent for the study of these questions" - he noted from Prague
in November 1885, presenting The Analysis of Sensations - "received the
strongest stimulus twenty-five years ago from Fechner's Elemente der Psy-
chophysik (Leipzig 1860)".32 And in the preface to the fourth edition, in
November 1902:
If there is" no essential difference between the physical and the psychical, we shall hope
to trace the same exact connection, which we seek in everything that is physical, in the
relations between the physical and the psychical also. 33
if elements exist, and not the element, if the structure of the world is rich
and different, then quality must exist besides quantity, the 'what' as well as
the 'how much'. And here we have, in addition to the 'exact connection' -
that is the 'connection' quantitatively connotated - another relationship
among phenomena: 'symmetry', which is discussed in one of the Popular
Scientific Lectures. 34 Mach is always ingenious and undetermined, and his
theoretical wavering is the expression of his need to connect different cate-
gories: but the frequent lack of connection favors the partial interpretations
- either phenomenalistic or reductionistic, subjectivistic or prerelativistic -
of his work. At the end of his life he, too, considered himself as a man of nov-
elties not arranged in a system, as subject of a vocation towards innovation:
Through the constant autoanalysis or criticism of myself since my younger years, I was
oriented towards the direction that today is called relativistic, and I could have brought
forward all these things; instead, subordinating my own ideas, I first of all tried to attain
an overall view on the future and to surpass the limits of the past, eluding the sphere of
influence that naturally the great thinkers create around their names. This is the reason
why I was concerned also with general problems of the physiology of sense and with
psychology. After all, we live only once, and I wanted, within my own limits, to have
as much of the world as possible; I could not fill up my life with only one thought, to
save my strengths. 35
But we should not forget that the background of Mach's psychology still
remained psychophysics. And the historian, always curious about details, can
confirm it, on the basis of still another circumstance: Mach's friendship with
Joseph Breuer, the first to dissent from Freudian dynamic psychology.
The Analysis of Sensations sets Breuer's figure in a light different from
that of Ernest Jones' Freudian 'biography'. Breuer was a neurophysiologist
of great merit and marked originality: his research on the function of the
otoliths in the vestibular apparatus were prompted by sagacious hypotheses
and by a physicalistic presupposition which, through Ernst Wilhelm Briicke,
physiologist in Vienna, was connected with the school of Milller, and in par-
ticular with du Bois-Reymond and Helmholtz. In 1875 Mach had published
Grundlinien der Lehre von den Bewegungsempfindungen,36 and Breuer's
investigations were the development of a subject well known to him:
Breuer, in a later piece of research, has made it probable that the sensations of progressive
acceleration vanish very much more quickly than those of angular acceleration and that
perhaps the organ of the former, at any rate in human beings, is atrophied. Further,
Breuer finds that, except for the semicircular canals, B, the otolitic apparatus, 0, with
its plane of sliding corresponding to the planes of the semicircular canals, is the only organ
adapted to the signalizing of progressive accelerations and position simultaneously.37
220 VINCENZO CAPPELLETTI
In the paper on 'Theory' which he (Breuer) contributed to the Studien, one notes the
fundamental importance he attached to the idea that the basis of hysteria was an ab-
normal excitability of the nervous system, so that an excess of free energy, that could
not be disposed of, was available for conversion into somatic symptoms. 38
Psychophysics, however, was and still is inadequate to sustain both the epis-
temological and the pedagogic constructions: even more so if sensation, per-
ception are regarded not as moments in the process of knowing, but as the
matrix and single normative source of knowledge. We recalled that the per-
ception of the motion of a body untouched by anyone impressed the mind of
little Einstein and awakened a 'wonder' that would vanish only through the
field theory implicit in the general theory of relativity. There is continuity
between perceiving and knowing in Einstein, but also in Mach as in the fol-
lowing passage from the Science of Mechanics:
DISCOVERY AND VOCATION 221
Still, great as the importance of instinctive knowledge may be, for discovery, we must
not, from our point of view, rest content with the recognition of its authority. We must
inquire, on the contrary: under what conditions could the instinctive knowledge in
question have originated? We then ordinarily find that the very principle, to establish
which we had recourse to instinctive knowledge, constitutes in its turn the fundamental
condition of the origin of that knowledge. And this is quite obvious and natural. Our
instinctive knowledge leads us to the principle which explains that knowledge itself, and
which is in its turn also corroborated by the existence of that knowledge, which is a
separate fact by itself.41
Mach the epistemologist was more open, less unilateral than Mach the psy-
chologist, and did not hesitate to acknowledge that sensation is already
reason. Mach and Freud were ignorant of each other - on Freud's side the
only hint of some relevance is in his letter to Fliess of 12 June 1900 - but
the memory, perhaps remote, of his great contemporary can be found in a
very beautiful autobiographical passage that Mach wrote in his last years, in
which he speaks of 'constant autoanalysis'. We already quoted it, and we refer
again to the citation. 42 Owing to Mach's waverings, Ehrenfels - who asserted
psychical 'Gestaltqualitdten' (form qualities) that were after all quite different
from Wertheimer's perceptual 'Gestalt' - could indicate as their source
Mach's Analysis. 43
Other science creators also show the same precocious inclination towards
the cognitive activity which in their lives soon assumed the centrality and
intensity of a vocation. Scientific vocation expresses itself in attitudes that
both epistemology and historiography regard as necessary for the growth of
the doctrinal organism: wonder at experience, the understanding of the
postulate objectivity of knowledge, the devotion to research. To become
objective knowledge, the ideal moment of vocation must promote the inter-
pretation of the perceptive datum. To favor this process is the task of peda-
gogy of science, which is also a pedagogy of perception.
First of all, perception must be attributed to the human subject, and must
be seen as physis and logos, both necessary. If pe{ception lacked the connec-
tion with nature, man would never refer it to reality in itself, with an attitude
which, if it is a naive identification of what is sensed with the other than the
self, it still originates from a legitimate inference from the sensed to the
existing, from the self to reality. On the other hand, if perceiving were not
thought, perception would never become wonder, that is, a problem. Wonder,
problem, resulting from the presence, and not the absence of conformity.
Let us be more explicit. Uttle Einstein, before the magnetic needle moving
without anyone touching it, had to perceive not only a phenomenon different
222 VINCENZO CAPPELLETTI
from others which are omnipresent, but also the doubt as to whether that
omnipresence of touching could be the formal cause of motion: doubt, his
own and any other analogous one, presumed the dialectic between the
all explaining thought and one principle that can explain only something.
Einstein must have been struck by the intuition that 'to be cause' and 'to
touch' are different principles, the former being more general and necessary
than the latter. He who marvels at something doubts traditional explanations,
not the existence of an explanation. Wonder is expectation: the expectation
to understand experience, as in Jung, or, as in Stevin, the expectation to
restore lawfulness.
However, perception, to be at the same time nature and thought, must
be more than quantity, or, using different terms, more than energy, motion.
There has never been a single attempt to give a quantitative interpretation
of thinking, which, on the contrary, is entirely qualitative, or, to specify
the polysemanticity of the term 'quality', structured. The rule organizing
the parts, which we call structure, produces judgment as unity of subject,
predicate and verb, or of terms with the relations connecting them. In
Identite et realite, Emile Meyerson effectively stated the priority of 'com-
prehension' over 'extension'. Each rational act unifies the diverse: and in
doing so, it assimilates diversity to identity. But the unity-identity must be
ascertained in the thing: otherwise, the legitimacy of the act, with which
sensation is seen and interpreted in terms of objectivity and world, would
cease. This is the inadequacy of psychophysics, and the reason for the revi-
sion done by Fechner in his last years. A historian of psychology, E. G.
Boring, distinguished a 'psychology of content' and a 'psychology of act',44
and included Fechner and Wundt in the former, Brentano, Meinong and
Ehrenfels in the latter. But the psychology of the last thirty years of the
nineteenth century can be better distinguished in psychophysics and psy-
chology of representative activity. Act and content cannot be heterogeneous
and Boring's scheme is liable to render them such. If the act and the content
of the psychical process do not fmd a way of corresponding to each other,
then the intentional relation to the object, in which Brentano perspicaciously
detected the peculiarity of the psychical, fails.45
Those were wonderful days, beginning in 1916, when for hours and hours I was fortunate
enough to sit with Einstein, alone in his study, and hear from him the story of the dra-
matic developments which culminated in the theory of relativity. During those long
discussions I questioned Einstein in great detail about the concrete events in his thought.
He described them to me, not in generalities, but in a discussion of the genesis of each
question. 50
Here comes again the thought experiment of running after a light beam,
here we find the consideration of the Michelson-Morley experiment, the need
to define simultaneity, the intuition of invariance, the identification of the
invariant with the velocity of light, the relativistic postulate. Perception - the
child's perception of the magnet and others, that, perhaps, have not been
recalled - as thought: dense, unexpressed but effective. And thought as
perception. In 1918 Eddington reported to the Royal Society that the devia-
tion (from a straight line) of the light coming from a star, due to the Sun in
the relativistic hypothesis, had been confirmed within the limits of experi-
mental accuracy.
Vocation for knowledge, and rational construction: these make a discovery
surpass the range of chance and transmute it into necessity, necessity of being
224 VINCENZO CAPPELLETTI
so for a hypothetical world which could identify with the existing world.
Pedagogy is applied to vocations so that they become constructions, to the
invisible deus in nobis in order to achieve the deus inter nos, knowable by all
people and recognized as such. There must be a school willing to be a school
of science, indeed a society which sees in the birth and development of scien-
tific vocations an element capable of granting its meaning and continuance.
Not all the great creators of science wrote their memoirs: but the lack of
an autobiography can be compensated by biography, if it is compiled keeping
in mind that it should not be a collection of events and circumstances but a
cognitive integration of the scientific document. Historiography is and has
long been active: Gediichtnissreden, etoges, obituaries, even precede the
historiography of ideas in the Machian sense, and their importance and abun-
dance of intrinsic value goes far beyond the colorless sequences of names,
dates, discoveries that were so frequent not so long ago. Biography is more of
a guesswork than autobiography, but it is better protected against the risk of
'protective memories', and more generally, of self-deception. Together with
the essay, biographical historiography competes for the space of the novel,
which is not exhausted but threatened by the possible languishing of the
imagination devising it.
Biography, or else, biographical historiography, has already proved, even
without the verification of the autobiographical account, that in many cases
vocation lies behind the scientific discovery. The discovery which is reflected
in axiomatics - it is now necessary to differentiate it from discovery with
no further connotation - is always coming from a distance: its trace is also
the trace of memory and the word through which it expresses itself is like a
Janus-faced image looking at the same time at things and their origin, in the
microcosm of the human subject.
Translated by
FRANCESCA PARDI LEVI
NOTES
6 Ibid., p. 53.
7 A. Einstein, Geometrie und Erfahrung (Springer, Berlin, 1921); Engl. transl., 'Geome-
try and Experience', in Sidelights on Relativity, II (Methuen, London, 1922).
8 A. Einstein, 'Autobiographical Notes', p. 57.
9 S. Freud, 'Zur Geschichte der psychoanalytischen Bewegung', in Gesammelte Werke
[referred to hereafter as G. W.), 18 vols. (Imago, London, 1940-1968). Vol. 10, pp. 43-
113. Translated as 'On the History of the Psychoanalytic Movement', in The Standard
Edition of the Complete Psychological Works [referred to hereafter as S.E.), ed. by
James Strachey, 24 vols. (Hogarth Press, London, 1953-1974). Vol. 14, pp. 7-66.
10 S. Freud, 'Selbstdarstellung', in G. w., Vol. 14, pp. 31-96 (S.E., Vol. 20, pp. 1-74).
11 S. Freud, 'An Autobiographical Study', S.E., Vol. 20, p. 8.
12 S. Freud, 'Zur Psychologie der Gymnasiasten', in G. W., Vol. 10, pp. 203-207 (S.E.,
Vol. 13, pp. 239-244).
13 Ibid., p. 205 (S.E., p. 242).
14 Erinnerungen, Triiume, Gedanken by C. G. Jung, ed. by A. Jaffe (Zurich-Stuttgart,
1962); Engl. trans!., Memories, Dreams, Reflections (W. Collins Sons & Co., Glasgow,
1977); French transl., Ma vie (Paris, 1966); Ital. transl., Ricordi, sogni, riflessioni di
C. G. lung (Milan, 1978). (The following notes refer to the English edition.)
15 Ibid., p. 43.
16 Ibid., p. 45.
17 Ibid., p. 47.
18 Ibid.
19 ~. Mach, Die Mechanik in ihrer Entwicklung historisch-kritisch dargestellt (Brock-
haus, Leipzig, 1883); seven editions were published while the author was still alive, the
last one of which in 1912. Afterwards, there were two posthumous editions, one in
1921, edited by J. Petzoldt, and one in 1933, edited by L. Mach; English transl. by
Thomas J. McCormack, with the author's approval: The Science of Mechanics: A Critical
and Historical Account of Its Development (Open Court Publ. Co., La Salle, Ill., 1960 6 )
(the following notes refer to this English edition, unless otherwise stated).
20 E. du Bois-Reymond, Uber die Grenzen des Naturerkennens - Die sieben Weltriithsel
(Veit, Leipzig, 1891).
21 E. Mach, Die Analyse der Empfindungen und das Verhiiltnis des Physischen zum
Psychischen (Fischer, Jena, 1896). The last edition, while the author was still alive, is
dated 1911. English transl., The Analysis of Sensations, and the Relation of the Physical
to the Psychical, translated from the first German edition by C. M. Williams; revised and
supplemented from the fifth German edition by S. Waterlow (Dover, New York, 1959):
the following notes refer to this English edition.
22 Ibid., p. 30, Note 1.
23 Ibid., pp. 29-30.
24 E. Mach, The Science of Mechanics, p. 27.
25 Ibid., p. 37.
26 Met. I, 2.
27 E. Mach, The Science of Mechanics, p. 13; the Aristotelian passage is in Mechanica
847 abo
28 E. Mach, The Science of Mechanics, p. 19.
29 Ibid., pp. 40-41.
30 A. Einstein, Autobiographical Notes, p. 21.
226 VINCENZO CAPPELLETTI
CASE STUDIES
JACQUES ROGER
If discovery is the strong beat of scientific life and the instrument of scientific
progress, it is not easy to defme or explain. The very word, at least in several
European languages (de-couvrir, s-coprire, dis-cover, ent-decken) is deceptive,
for it indicates the simple act of moving aside an obstacle, of removing the
veil which 'covered' a truth that lay at hand, and which was simply waiting
to be at last looked at: a defmition more tempting in anatomy than elsewhere
and yet disputable.
There is no question here of reviewing all the possible kinds of discovery in
the very wide variety of their epistemological nature and their historical
circumstances. I wish to study only two precise discoveries, their genesis,
their realization and their destiny. I am here concerned with an account I
have already written! and which I now take up from a different perspective.
These two discoveries are (1) that of the ovarian vesicles and the phenomenon
of ovulation in vivipara, and (2) that of the spermatozoa. The first is attri-
buted to the Dutch anatomist, Regnier de Graaf, and dated 1672; the second,
to the Dutch microscopist Antoni van Leeuwenhoek, and dated 1677. They
were made in the same scientific field - at least according to modern de-
finitions - in the same country, and within an interval of five years. These
external similarities allow us to isolate more easily the differences and try to
account for them.
Let us briefly recall the facts. Until about the middle of the seventeenth
century two theories on the generation of animals and man shared the
allegiance of the scientific world. According to the first, which comes from
Aristotle, the male provides with his semen the 'form' of the being to come,
while the female, with the menstrual blood, provides the 'matter' of the
embryo. According to the second theory, which comes from Hippocrates,
male and female produce a liquid semen, the mixture of which, or 'conceptus'
forms the embryo. In the middle of the seventeenth century, Descartes
presented a 'corpuscularist' and mechanicist version of the theory of the two
semina (De [ormato [oetu, published posthumously in 1664), while William
Harvey gave a fresh lustre to the Aristotelian theory (Exercitationes de gener-
atione animalium, 1651).
However, Harvey corrects Aristotle in that he does not believe in the role
229
M. D. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 229-237.
Copyright © 1980 by D. Reidel Publishing Company.
230 JACQUES ROGER
of the menstrual blood. According to him the male semen fertilizes the uterus
of the viviparous female, and this uterus produces an egg, exactly as the ovary
of the ovipara does. On the one hand, therefore, the so-called 'female testi-
cles' are only lymphatic glands and play no part in the act of generation. On
the other hand, however, vivipara are born from an egg as ovipara are, which
justifies the well-known formula inscribed in the frontispiece of Exercita-
tiones: "Ex ovo omnia." For Harvey viviparous females have eggs but no
ovaries.
Harvey had hardly any following in his refusal to assign a role to the
'female testicles', but the idea of the egg of vivipara was very successful from
1660, each exponent, however, putting forward his personal opinion as to
the origin and formation of that egg. Things became clearer in 1667, with
the contribution made by the great Danish anatomist, Niels Stensen, alias
Steno. Steno, who, like Harvey, first thought that the egg was formed in the
uterus, discovered eggs in the female testicles of a dog-fish - a species known
since Aristotle's time as ovo-viviparous. He immediately concluded "that
women's testicles are analogous to the ovary" and produce eggs which then
pass into the uterus. While Steno continued his observations (which were
published only in 1675), a Dutch anatomist, Jan van Horne, published in
1668 a short treatise on the male genital organs, to which he added a note on
the female organs, where he stated: "Female testicles correspond to the ovary
in the ovipara, given that they contain perfect eggs, full of body-fluid
[humeur] and wrapped in a thin membrane." Let us note that three years
before, in a short anatomical treatise, van Horne had already spoken of those
ovarian vesicles and of "the dual duct whose purpose is to evacuate the body-
fluid which is prepared thereiIl.. " In 1668, therefore, he had not just dis-
covered them - they had, in fact, already been described by Vesalius,
Falloppio and others - but, after taking them for reservoirs of female semen,
he had read Steno and decided to call these vesicles eggs. Van Horne died in
1670, before being able to write the great treatise he had been contemplating
on these questions.
From 1670 onwards, work was done nearly everywhere on the ovarian
vesicles. In 1671, the Dutchman Kerckring published a treatise of a few pages,
Anthropogeniae Ichnographia, where he produced, among other things, the
picture of a human embryo already well-formed - too well-formed - which
he claimed to have discovered in an egg three days after conception. In Paris
and in London, eggs were discovered in cows' ovaries. At the end of the year,
Regnier de Graaf very briefly set out anew van Horne's ideas which he
claimed as his own. In short, everyone by then believed in the eggs ofvivipara,
TWO SCIENTIFIC DISCOVERIES 231
but no one had yet proved anything about them. As Gallois wrote at that
time in the Journal des Savants (21 March 1672), two things are certain:
there are vesicles in female testicles and what encloses the embryo in the
womb has the shape of an egg: but "it remains to be seen if those vesicles,
which are attached to the female body, detach themselves, and if that kind of
egg where the embryo takes shape is one of those detached vesicles."
The reply to these questions had in fact just been given by Regnier de
Graaf in his treatise: De mulientm organis generationi inservientibus. Thanks
to methodical observations on female rabbits, de Graaf established that after
copulation, ovarian vesicles change into yellow bodies. Furthermore, he had
managed to discover developing eggs in the Fallopian tubes, which developed
in equal number to the yellow bodies in the corresponding ovary. It was thus
shown that each vesicle was an egg, the shedding of which left a scar in the
form of a yellow body, that the female testicles were ovaries and that the
Fallopian tubes conducted the detached eggs from the ovary to the uterus.
The 'egg doctrine', as it was then called, had thus firmly been established. As
early as 1679, according to the Journal des Savants, scarcely favorable as it
was to the idea, it had become "something so common at present that there is
practically no new Philosopher who does not accept it today." The existence
of eggs in viviparous females was no longer seriously doubted in time to
come.
Let me end this first historical account with two remarks. First, that
Regnier de Graaf's demonstration fell far short of the success that his meth-
odological rigor ought to have garnered, even if that rigor (I shall return
to the point) is imperfect. In the immediate spread of the new doctrine,
Kerckring's loud assertions or William Harvey's remote authority, had far
more weight. No doubt, later, but only later, things became clearer. The
second remark is that, as we know, ovarian vesicles or de Graafs follicles, are
not themselves the eggs of vivipara: they hold eggs which are released at the
moment of ovulation. As early as 1681, Malpighi had an inkling of this mis-
take and even believed, but wrongly, that he had discovered the veritable
little egg, which he christened ovule. All through the eighteenth century, the
ovule escaped the attention of the world of research, and the 'egg doctrine'
was none the worse for it. We know that it was only in 1827 that von Baer
discovered the ovule inside de Graafs follicle.
Just as the discovery of the egg of vivipara was a long and complicated pro-
cess, that of the spermatozoa was a simple and quick one. It is described in
a letter which van Leeuwenhoek sent in November 1677 to Lord Brouncker,
232 JACQUES ROGER
In the light of these two historical accounts, we can make a few reflections.
The flISt will be to note the difficulty of saying who has made a discovery.
It has been shown that it was almost impossible to say who discovered
oxygen. Let us take here the simpler of our two examples. Who discovered
the spermatozoon? First possible reply: the student Ham, who was the
first to see it. But Ham did not see a spermatozoon: he saw a microscopic
animalcule born, according to him, of putrefaction. Second answer: van
Leeuwenhoek, who observed systematically the sperm of various animals
and found everywhere spermatic animalcules, which he took for the active
agent in generation. But he believed that the animalcule contained the pre-
formed embryo. In French, the very word spermatozo'ide appeared only
in the nineteenth century, progressively taking over from kindred names:
TWO SCIENTIFIC DISCOVERIES 233
But then the world of the microscopists, which developed during the last
quarter of the seventeenth century, did not wholly partake in the spirit of the
time. Let us note right now that the microscope, practically a contemporary
of Galileo's telescope, remained unused for a long time. When it began to be
used, it was used by specialists of the technique itself, who examined with
their instrument the most varied objects. Van Leeuwenhoek's observations
236 JACQUES ROGER
were typical of such a disorder, which was already found in Hooke. At that
time, however, the microscope was above all a 'great purveyor of marvels.' It
showed the unexpected, the improbable, something which could not have
been discovered by analogical reasoning. It accompanied or encouraged in the
observer the idea that nothing is impossible, and that the human mind cannot
conceive, much less pretend to order, the infmite variety, the infmite richness
of creation. In this sense the discovery of spermatozoa was not a chance
phenomenon. It was prepared, or at least made pOSSible, by the existence of a
certain mentality in the particular scientific sector where it happened.
Such a particular state of mind was scarcely in keeping with the one pre-
vailing at the time. More precisely, it heralded that of the naturalists primarily
dedicated to observation, so numerous in the first half of the eighteenth
century. In this sense the micrographers of the 1670's were witnesses to the
European 'crise de conscience. ' They came a little early, and in spite of the
curiosity they aroused, van Leeuwenhoek's work suffered because of this
fact. The discovery of the spennatozoon, like many others, suffered in parti-
cular from the difficulties of verification. Many people, who had never
handled a microscope, debated about it. But once the fact was established,
interpretation presented far more important difficulties. It was in conformity
to the spirit of the time that van Leeuwenhoek imagined an embryo pre-
fonned in the spennatozoon. But that animalcule, which one saw being born
and dying, presented no real guarantee of safety to a pre-existing structure.
Besides, its movement led it to be considered an animal. By what extra-
ordinary 'metamorphosis' could that animal become a man, at a time when
'equivocal generations' had gone out of fashion? Finally, the enonnous
number of spennatozoa, the immense majority of which were condemned to
a purposeless death, was in too violent contradiction of the idea of a thrifty
nature and a wise God, creator of a well-ordered universe. As a contemporary
said: 'that makes a good deal of seed lost!'
It is therefore not surprising that the two discoveries, coming to light in
different areas of research, placed in a totally different relationship vis-a-vis
their interpretation, and therefore the spirit of the time, coming within the
purview of two different epistemological types, had opposite destinies. One is
allowed to think that many discoveries have a position between these two
extremes, but that the history of their genesis and success lends itself to the
same kind of analysis.
Translated by
MARGARET ROUSSEL
TWO SCIENTIFIC DISCOVERIES 237
NOTE
1 Cf. Roger, J., Les sciences de la vie dans la pensee fram,aise du XV/lIe siecle: la
generation des animaux de Descartes a l'Encyclopedie (Colin, Paris, 1971), esp. the 2nd
part, Chapter II, pp. 255-323.
RICHARD TOELLNER
case history. This is, however, what happens in a history of discoveries, where
the discovery as such is reduced to a date, to one station on the road to one's
own preconceived aim. 2
When modern history of science asks for new discoveries, inventions, and
insights, in short, for the latest news in the history of science, then it is not
interested in new facts as mere facts, then it is not primarily concerned with
the question, by whom, when, and where something new was discovered, but
it is interested in the question in which historical context this new fact could
emerge as a new fact, what the historical prerequisites of this discovery were,
in what way and under which historical conditions it could come into effect.
Modern history of science asks, as does the topic of the second course of the
International School of History of Science, for the 'Logical, Psychological,
Cultural and Social Aspects of the Scientific Discovery. '
But it is only the enquiry into all these single aspects and the adequate
linking of each result in research into one over·all picture that transform the
isolated date of chronological science, the pure fact of case history into an
event of history of science. Only the presentation of a discovery as a histor·
ical scientific event, into which the patterns of ideas, the content and aims of
preliminary science have blended as well as the cultural and social conditions
of today's life, only such a presentation that has explored all this and takes it
into consideration, will truly be able to offer any information on the specifi·
cally 'new' aspect of a scientific discovery, can really describe how a new
insight developed, in which way and by which means it came into effect or
had no effect at all.
It is only against the background of Scientifically immanent as well as
sociocultural factors that the figure of the discoverer and his psychology can
claim the importance that is due to it as the most important factor of the
historical event of the discovery. In this context I fully agree with Bernard
Cohen, when he says:
It is my thesis that in studying science, just as in every other area of creative activity, the
historian must take account of the special qualities of individual genius, that an aware·
ness of the temperament and personality of each scientist is for the historian of equal if
not greater importance than the general character of the age and the particular environ-
ment of ideas in which the scientist worked. 3
There are, of course, discoveries in the history of science that have become
without any doubt the basis for today's scientific view of the world, and their
effect is such an overwhelming one that traditional history of science seems
to be correct in its opinion that only the discoveries really matter and that
DISCOVERY OF THE CIRCULATION OF THE BLOOD 241
the historical conditions for their development and their influence are of
minor importance. To name a few examples, one can refer to discoveries
which now constitute our physical conception of the world linked to the
names of Copernicus, Galilei, Kepler and Newton. No doubt that, as far as
medicine is concerned, such an exceptional discovery, on which a completely
new physiology and medicine was based, must be the discovery of the circula-
tion of the blood by William Harvey.
It is not by chance that all the examples mentioned are taken from the
particular era when ancient-medieval science was replaced by modem science.
The eighteenth century in which the new science had its breakthrough under-
stood this process as "an emancipation of science from the chains of author-
ity", as a farewell of science to blind faith in authority, as "a departure of
man from self-inflicted infancy" as Kant's classical remark goes,4 as "a
liberation of scientific research from any preconceived views, from ideological
dogmatic Authority" as the physicist and historian of science Walther Gerlach
called it. 5 The new basic postulate of science is now: Scientific statements are
no longer to be verified, validated and authenticated by calling on the
authority of the Ancients, but by referring to autonomous reason and its
deductions on the one hand and to autonomous experience, that is, one's
own observation, research and experiment on the other.
It may be easily understood that with such a deep change in the structures
of thinking of science it is not the mere results of this historical process as
pura facta that constitute the major value of appreciation for modem science
but rather the historical conditions of their possibilities, their complicated
genesis and the way they come into effect.
I thus propose to demonstrate this general process by dealing with
the special case of Harvey with its logical and psychological implications
that illustrate the dispute Harvey had not directly but indirectly with
Descartes about the nature of cardiac motion. This debate represents only
a minor episode in the history of seventeenth century medicine and has
not hitherto received much attention in the Harveyan or Cartesian litera-
ture. 6
However, problems and insights may emerge from the study of a single
historical event which could have a more general and lasting character and
may still affect us today. In this sense, we shall attempt to illustrate how
Descartes and Harvey, two great scientists, deal with experience and reason
for the correct description and explanation of a physiological phenomenon.
Furthermore, we will also try to show how the dispute between the founders
of modem philosophy and physiology heralds the great controversy over the
242 RICHARD TOELLNER
I Harvey's Position
Harvey was a contemporary of Francis Bacon and Galileo Galilei, men whose
names we link with the introduction of the empirical and inductive method,
and thus the beginnings of modern scientific research. Although Harvey must
have encountered Galileo at Padua and, as a court physician to James I, could
not have ignored the activities of Lord Bacon, there is no proof that they ever
influenced him directly .1 5
The appreciation of experiments, observations and experience was, there-
fore, a general characteristic of a period in this history of science deeply
polarized by the followers and detractors of Aristotle. Harvey's preferences
for the 'sensus' (perception) and 'autopsia' (personal observation) as con-
trasted with 'ratiocinium' (the fruits of reason) are not so much characteristic
of the man himself but rather of his time. Harvey's confidence in experience
was therefore more a contemporary presupposition than a self-acquired
insight which allowed him to become the immortal discoverer of the circula-
tion.
The discovery was aided by Harvey's preference for anatomical studies.
These investigations were carried out during his student days in Cambridge
and Padua, and continued until his death. From the notes which Harvey made
and the manuscripts of his anatomical lectures - in 1615 he held a chair of
anatomy and surgery in London - the stages of his discovery become
apparent.
At first, Harvey carried out careful anatomical observations of the venous
valves which his teacher, Fabricius ab Aquapendente, had described in Padua.
244 RICHARD TOELLNER
occupies a special place among all other movements. Moreover, life is a dis-
tinctive process within the sub1unar, terrestrial sphere, characterized by birth
and death, growth and decay, change and motion, all of which must be under-
stood and derived from first philosophical principles. Thus, as the sun's move-
ments and warmth are the source of all motions, mixtures and transforma-
tions of the macrocosrnic elements, the heart with its contractions and beat is
the fountain, origin and highest principle of the body or microcosm.
Therefore, Harvey can state:
Therein, by the natural, powerful, fiery heat, a sort of store of life, it [the blood 1 is re-
liquefied and becomes impregnated with spirits and (if I may so style it) sweetness. From
the heart it is redistributed. And all these happenings are dependent upon the pulsatile
movement of the heart. This organ deserves to be styled the starting point of life and the
sun of our microcosm just as much as the sun deserves to be styled the heart of the
world. For it is by the heart's vigorous beat that the blood is moved, perfected,
activated, and protected from injury and decay.18
We know from Harvey's notes that he was convinced as early as 1616 "that a
steady flow of blood takes place in a circular manner with the help of a
cardiac pu1sation."19 However, only twelve years later did he finally decide
to acquaint the public with his theory. Although repeated observations and
experiments had convinced him of the correctness of his conclusions - they
coincided with the principles of Aristotelian-Galenic medicine - Harvey fore-
saw the effects which his new theory was to have. Indeed one of his early
adherents in Germany, Paul Marquard Schlegel, wrote that
the unheard of scheme which was directed against the general concepts prevailing for
centuries brought about a great commotion. There were hardly any physicians who, after
hearing of Harvey's discovery, did not decry the work as complete nonsense and urge its
banning from the schools. 20
confusion and fear; one representative of the French landed gentry immedi-
ately greeted the book as "by far the most important and useful discovery in
medicine.,,21 The author of this laudatory statement was the Sieur du Perron,
Rene Descartes, born in Touraine in 1596, who had just emigrated to Holland
when the De motu cordis appeared.
Descartes had gone to a country in which both economy and science were
thriving, and tolerance as well as peace prevailed. There, at his leisure and in
the seclusion and safety of that land, he proposed to do those things which he
had planned as a young man. In his search for truth, Descartes declared:
I intended to spend a great deal of time in necessary preparations in order to eradicate
from my mind all previously adopted and detrimental convictions. Moreover, I wanted
to collect a lot of experiences as material for subsequent conclusions, always practicing
my self-prescribed method so as to acquire greater skills in using it. 22
and breathing. Finally, it is endowed with all our functions which we presume are
derived from matter and the disposition of the various organs. 29
C. THE CONTROVERSY
Thus, the question concerning the correct description of cardiac motions and
their cause becomes the cornerstone for the entire Cartesian physiology.
Descartes himself declares
that it is of the greatest importance to ascertain the true origin of the cardiac motions.
Without them, it would be impossible to learn anything concerning physiology since all
other bodily functions depend on the movement of the heart. 3S
Indeed, the entire question becomes the touchstone for Descartes' philos-
ophy. "I explain ... everything that belongs to the motions of the heart
quite differently than Harvey", Descartes announces on February 9, 1639, to
his close friend Mersenne, "but if what I have written is false, I will concede
that the rest of my philosophy is likewise an error ."36 What these sentences
reflect is the hitherto overlooked seriousness with which Descartes viewed his
so-called 'fables' or hypotheses, and the reason for his stubborn opposition to
Harvey's ideas on cardiac motion which were based on observation and
experiment.
If one compares the Cartesian and Harveyan schemas of cardiac movements
250 RICHARD TOELLNER
as well as their origin and effects with those valid today, Harvey's theory
emerges as the correct one in all essential aspects, while Descartes' ideas
were all erroneous. Such a fact constitutes a devastating indictment of the
Cartesian hypothesis. Now, if one asks why Harvey arrived at the correct
solution while Descartes blundered, the answer will have to be sought in the
relationship between 'experimentum et ratio.'
blood and the controversy regarding the nature of cardiac motion) in connec-
tion with our overall subject (logical and psychological aspects of discovery).
(1) Harvey was only able to make his discovery because he was completely
rooted in the old way of thinking. Coming from ancient-medieval philosophy,
Aristotelian physics and Galenic physiology, he and his contemporaries could
easily regard two important preconditions for his doctrine of circulation as
self-evident and thus posing no problems:
(a) the theory of circulation as such,
(b) the active contraction of the heart as a logical sequence of its facultas
movendi which as vis pulsifica, is a substantial form of the heart.
(2) Harvey's method of careful anatomical research and observing phy-
siological experiments was nothing new in his time; the empirical research of
anatomical structures and - in initial stages - of physiological functions is
the general characteristic of Renaissance medicine and science.
(3) New is the result of his hypothesis, that ensues conclUSively from the
observed phenomena: the quick circulation of blood in the body. He is well
aware of the novelty of this result. But apparently this does not make him
proud, but hesitant and uncertain, for this result does not fit at all into his
traditional conception of physiology. The function of a quick circulation of
blood in the body is not only utterly obscure to him, but it also contradicts
the functions ascribed to it by the old physiology in all major aspects.
(4) It was only when Riolan pointed out the undisputable dilemma and
showed that his doctrine of the circulation of the blood would completely
jeopardize the old physiology and medicine that Harvey tried to develop new
ideas about the function of circulation. It is, however, much more important
that he insisted on his observation being correct and that he defended it
against the reproach of having contradicted the authority of the Ancients by
declaring it to be a statement of nature itself. Nature is thus made the
supreme court of appeal "for nothing is older and of greater authority than
nature."
(5) Harvey can only justify his method and its result by referring to the
authority inherent in the old way of thinking. He plays off this authority of
the Ancients against the even older authority of nature. Nature gives him the
certainty necessary to believe in his fmdings. These can only be legitimized
by the old way of thinking.
Descartes is quite different. He achieves certainty about his cognition by
his new "method to guide the thoughts in a safe way." The new self-assurance
of the powers of reason had now discovered the general questionability of
every old authority as well as of the sensory cognition of nature and now
254 RICHARD TOELLNER
developed 'more geometrico' a new system of the world that is for the
organism a machine theory of the living, a mechanical physiology. Starting
from a new method of thinking Descartes arrives at a new pattern of thinking.
He realizes that this innovation is a radical one and is proud of it. The
authority of the Ancients and of nature have to yield to the authority of
reason.
What fascinates Descartes when contemplating Harvey's theory of circula-
tion is the possibility of a mechanistic interpretation. Anatomical and physio-
logical observations confirm his machine theory. The doctrine of circulation
fits into his system and it is not by pure chance that the recognition of the
doctrine of circulation and the establishment of a mechanistic physiology are
closely connected.
The one thing about Harvey's doctrine of circulation that had to displease
Descartes was Harvey's assumption about the motor of this circulation. The
active contraction of the heart could only be explained and understood by
using the old pattern of thinking of Galenic physiology and Aristotelian
physics. This is why he had to explain cardiac motion by a new mechanistic
theory. History proved, however, that his theory and observations were
wrong. His self-assured reason had led him astray; Harvey was right in the
end.
The clock of historical reason runs in a different way than Descartes could
have dreamed of. It is our task to investigate it if we wish to contribute to the
self-knowledge of science.
Translated by
I. MAGNUS and G. B. RISSE
NOTES
G. Lehner (Ehrenwirth, Munich, 1965), pp. 121-135: "Mit der Befreiung der Natur-
forschung von jeder Voreingenommenheit, von weltanschaulich-dogmatischer AutoriHit
... wird die neue Naturwissenschaft groB". (pp. 125-126)
6 Cf. Toellner, R., 'The Controversy between Descartes and Harvey regarding the Nature
of Cardiac Motions', in Science, Medicine and Society in the Renaissance. Essays to
Honor Walter Pagel, ed. by A. G. Debus (Science History Publications, New York, 1972),
Vol. 2, pp. 73-89; Bernoulli, R., 'Descartes' Grundgedanken in medizin-historischer
Sicht', Gesnerus 35 (1978),44-53.
7 Rothschuh, K. E., 'Die Entwicklung der Kreislauflehre im Anschluss an William
Harvey. Ein Beispiel der "autokatalytischen Problementfaltung", in den Erfahrungs-
wissenschaften', Klinische Wochenschrift 35 (1957), 605-612. Cf. p. 611.
8 Leibniz and Haller said in criticism of Descartes that he tended constantly toward the
speculative. "Mons. des Cartes n'est tombe icy dans l'erreur que par ce qu'i! se fioit trop
11 ses pensees." (Leibniz, G. W., 'Discours de metaphysique', in Die philosophischen
Schriften, ed. by C. I. Gerhardt (Weidmannsche Buchhandlung, Berlin, 1875-1890).
Vol. 4 (1880), p. 443. [" ... Descartes has fallen into error here only because he had too
much confidence in his thoughts ... ", 'Discourse on Metaphysics', in Leibniz, Philosoph-
ical Papers and Letters, trans. and ed. by L. E. Loemker. 2nd ed. (Reidel, Dordrecht,
1969), pp. 303-330. Cf. p. 315.
"Deux autres romans physiologiques de Descartes demontrent qu'on peut connoitre
la bonne methode de rechercher la verite, et suivre celle qui lui est la plus contraire."
(Haller, A., Art. 'Physiologie' in Supplement Ii l'Encyclopedie by Diderot and D'Alem-
bert, Vol. 4, Amsterdam, 1777, 349.) [Two other physiological works of Descartes show
that one can know the right method of looking for the truth and follow the one most
contrary to iLl
9 Brecht, B., Leben des Galilei (Suhrkamp, Berlin, 1957), Act 4, scene 8, pp. 57 -73.
(The Life of Galileo, trans. by D. I. Vesey (Methuen, London, 1963).
10 Kepler, J., Johannes Kepler in seinen Briefen, ed. by M. Caspar and W. von Dyk.
2 vols. (Oldenbourg, Munich, 1930), Vol. 1, p. 353.
11 Cf. Olschki, L., Geschichte der neusprachlichen wissenschaftlichen Literatur. 3 vols.
(Winter, Heidelberg, 1919-1927), Vol. 3: Galileo und seine Zeit, 1927, p. 220, Note 3.
12 Harvey, W., Exercitationes duae anatomicae de circulatione sanguinis, ed. by K. J.
Franklin (Blackwell, Oxford, 1958), p. 137ff.: "Nulla alia certior demonstratio ad fidem
faciendam adduci poterit, quam sensus et Cti'TO\jJi.c,".
13 "Aut approbatas opiniones relinquere indignum putent; et per tot saecula traditam
disciplinam, veterumque auctoritatem, in dubium vocari nefas putent; his omnibus
respondeam, facta manifesta sensui nullas opiniones, naturae opera nullam antiquitatem
morari: natura enim nilii! antiquius, majorisve auctoritatis." Ibid., p. 136.
14 Cf. Lesky, E., 'Harvey und Aristoteles', Sudhoffs Archiv 41 (1957), 289-316 and
349-378. Pagel, W., William Harvey's Biological Ideas. Selected Aspects and Historical
Background (Karger, Basel and New York, 1967), pp. 28-47.
15 Pagel,op. cit., pp. 20-23.
16 Harvey, W., Exercitatio anatomica de motu cordis et sanguinis in animalibus (Frank-
furt, 1628), ed. by K. J. Franklin, (Blackwell, Oxford, 1957), p. 165: "Coepi ego met
mecum cogitare, an motionem quandam quasi in circulo haberet: quam postea veram
esse reperi, et sanguinem e corde per arterias in habitum corporis et omnes partes
protrudi et impelli a sinistri cordis ventriculi pulsu, quemadmodum in pulmones per
256 RICHARD TOELLNER
venam arteriosam a dextri; et rursus per venas in venam cavam et usque ad auriculam
dextram remeare, quemadmodum ex pulmonibus per arteriam dictam venosam ad sinis-
trum ventriculum ... ".
17 "Quem motum circularem eo pacto nominare liceat, quo Aristoteles aerem et pluviam
circularem superiorum motum aemulari dixit." Ibid.
18 "Ibi calore naturali, potenti, fervido, tanquam vitae thesauro, denuo colliquatur,
spiritibus et (ut ita dicam) balsamo praegnans; inde rursus dispensatur: et haec omnia a
motu et pulsu cordis dependere. Ita cor principium vitae et sol microcosmi, ut propor-
tionabiliter sol cor mundi, appeUari meretur; cujus virtute et pulsu sanguis movetur,
perficitur, vegetatur, et a corruptione et grumefactione vindicatur." Ibid.
19 "Unde perpetuum sanguinis motum in circulo fieri pulsu cordis." Harvey, W., Prae-
lectiones anatomie universalis. De musculis, ed. by G. Whitteridge (Livingston, Edinburgh
and London, 1964), Fol. 80 v.
20 Schlegel, P. M., De sanguinis motu commentatio, in qua praecipue in Joh. Riolani,
V. C. sententiam inquiritur (Hamburg, 1650), Praef., I.
21 Descartes to Beverwick, 5 July 1643: "Quippe, quamuis circa sanguinis circulationem
cum Heruaeo plane consentiam, ipsumque vt praestantissimi illius inuenti, quo nuUum
maius & vtilius in medicinii esse puto, primum auctorem suspiciam, tamen circa motum
cordis omnino ab eo dissentio". Descartes, Oeuvres, published by Charles Adam and
Paul Tannery, 13 Vols. (Cerf, Paris), 1897-1913, Vol. 4 (1901), p. 4. [Indeed, although
I clearly agree with Harvey about the circulation of the blood, and believe him to be
the first author of that most excellent discovery, which I think is greater and more useful
than anything else in medicine, I nevertheless disagree with him completely about the
movement of the heart.)
22 "et que ie n'eusse, auparauant, employe beau coup de terns a m'y preparer, tant en
deracinant de mon esprit toutes les mauuaises opinions que i'y auois receues auant ce
terns la, qu'en faisant amas de plusieurs experiences, pour estre apres in matiere de
mes raisonnemens, & en m'exerc:;ant tousiours en la Methode que ie m'estois prescrite,
affin de m'y affermir de plus en plus." Descartes, Oeuvres 6 (1902), 22 (Discours de la
Methode, 1637, referred to hereafter as Discours).
23 For many similar judgments, HaUer cited here: "Des D(escartes) Vorurtheile kosteten
im das Leben, er woUte im Seitenstiche sich nicht die Ader Mnen lassen", G6ttingische
Gelehrte Anzeigen, 1773, Zugabe, 371.
24 "Mais apres que i'eu employe quelques annees a estudier ainsi dans Ie liure du monde,
& a tascher d'acquerir quelque experience, ie pris vn iour resolution d'estudier aussy en
moymesme, & d'employer toutes les forces de mon esprit a choysir Ie chemins que ie
deuois suiure." Descartes, Oeuvres 6 (1902), 10 (Discours, 1637). ["But after I had
spent a few years studying the book of the world, and in trying to acquire some experi-
ence, one day I made a resolution to also study of myself and to employ aU the forces
of my wits in choosing the paths I ought to foUow". 'Discourse Concerning the Method',
in Descartes, The Essential Writings, trans. by J. J. Blom (Harper, New York, 1977),
pp. 114-164. Cf. p. 120. Referred to hereafter as Discourse.)
25 "Nempe quidquid hactenus ut maxime verum admisi, vel a sensibus, vel per sensus
accepi; hos autem interdum fallere deprehendi, ac prudentiae est nunquam illis plane
confidere qui nos vel semel deceperunt." Descartes, Oeuvres 7 (1904), 18 (Meditationes,
1641).
26 "Nihil nisi punctum petebat Archimedes, quod esset firmum & immobile, ut integram
DISCOVERY OF THE CIRCULATION OF THE BLOOD 257
terram loco dimoveret; magna quoque speranda sunt, si vel minimum quid invenero quod
certum sit & inconcussum." Ibid. 24 (Meditationes, 1641).
27 Subtitle of Discours de la Methode: Pour bien conduire sa raison, et chercher la verite
dans les sciences.
28 Descartes said of the foundation of his physics: "elles m'ont fait voir qu'il est pos-
sible de paruenir a des connoissances qui soient fort vtiles a la vie, & qu'au lieu de cete
Philosophie speculatiue, qu'on enseigne dans les escholes, on en peut trouuer vne prati-
que, par laquelle connoissant la force & les actions du feu, de l'eau, de I'air, des astres,
des cieux, & de to us les autres cors qui nous enuironnent, aussy distinctement que nous
connoissons les diuers mestiers de nos artisans, nous les pourrions employer en mesme
fa«:;on a tous les vsages ausquels ils sont propres, & ainsi nous rendre comme maistres &
possesseurs de la Nature." Descartes, Oeuvres 6 (1902), 61-62 (Discours, 1637). ["For
these notions have made me see that it is possible to reach various knowledge very useful
to life, and that, instead of the speculative philosophy taught in the schools, one can find
another practical philosophy by which we can know - as distinctly as we now know the
different professions of our artisans - the force and action of fue, water, air, the stars,
the heavens, and all the other bodies that surround us; and thus, in the same way that we
employ artisans, we could also employ all these things for all the uses for which they are
fit, thereby rendering ourselves like masters and possessors of nature." Discourse, op. cit.,
p.154.)
29 "Ie suppose que Ie Corps n'est autre chose qu'vne statuti ou machine de terre, que
Dieu forme tout expres, pour la rendre la plus semblable it no us qu'il est possible: en
sorte que, non seulement il luy donne au dehors la couleur & la figure de tous nos mem-
bres, mais aussi qu'il met au dedans toutes les pieces qui sont requises pour faire qu'elle
marche, qu'elle mange, qu'elle respire, & en fin qu'elle imite toutes celles de nos fonc-
tions qui peuuent estre imaginees proceder de la matiere, & ne dependre que de la dis-
position des organes." Descartes, Oeuvres 11 (1909), 120 (Traite de [,Homme, 1633).
30 Cf. Descartes, Oeuvres, VI, 46-53; XI, 123, 127; XI, 228-245. Rothschuh, K. E.,
'Rene Descartes und die Theorie der Lebenserscheinungen', Sudhoffs Archiv 50 (1966),
25-42.
31 "ie me contentay de supposer que Dieu formast Ie cors d'vn homme, entierement
semblable a l'vn des nostres, tant en la figure exterieure de ses membres qu'en la con-
formation interieure de ses organes, sans Ie composer d'autre matiere que de celle que
i'auois descrite, & sans mettre en luy, au commencement, aucune arne raisonnable, ny
aucune autre chose pour y seruir d'ame vegetante ou sensitiue, sinon qu'il excitast en
son coeur vn de ces feux sans lumiere, que i'auois desia expJiquez, & que ie ne conceuois
point d'autre nature que celuy qui echaufe Ie foin, 10rsqu'on I'a renferme auant qu'il
fust sec, ou qui fait bouillir les vins nouueaux, lorsqu'on les laisse cuuer sur la rape."
Descartes, Oeuvres 6 (1902),45-46 (Discours, 1637). ["I contented myself with assum-
ing that God formed the body of man entirely similar to our body, as much in the
exterior figure of its members as in the conformity of its interior organs; and that he
composed this body only from the matter I had described, and without placing in it, in
the beginning, any reasoning soul, or any other thing to serve therein as a vegetative or
sensitive soul, except that he excited in its heart one of those fires without light that I
had already explained and conceived as having the same nature as the fue that heats hay
when one has stored it away before it was dry, or that activates new wines when one
allows them to ferment over sediment". Discourse, op. cit., p. 143.)
258 RICHARD TOELLNER
40 "Ad eos qui circulationem repudiant, quia neque efficientem, neque finalem causam
vident, cui bono? (de quo adhuc nihil adjunxi ... ): Prius in confesso esse debet, quod
sit, antequam propter quid, inquirendum." Ibid., pp. 135-136. ["With regard to those
who repudiate the circulation because they see neither its efficient nor its fmal cause, I
have to date added no reply to their query 'Who benefits?' ... First, one ought to admit
what should be investigated rather than the reason for such further study." HarveY,De
circulatione (Oxford, 1968), pp. 44-45.
LUIGI BELLONI
hospital back in May 1838 while dissecting the body of a female peasant who
had died from croupous pneumonia. Dubini made his discovery public in
April 1843 6 and designated this new worm as the Agchylostoma [sic!]
duodenale, a designation that underlined its hooked mouth and 'habitat' in
the human organism. In this work Dubini emphasized the very high incidence
of this worm, which "even if it has not been detected or described by other
workers, is to be found in at least twenty corpses out of a hundred that are
dissected in order to fmd it." So, the problem then became to discover the
"reasons that have delayed its discovery up to now"; and it is worthwhile
giving the entire explanation of this as left by Dubini:
For a few years now, namely since most recent anatomo-pathological studies undertaken
in cases of typhoid fever or tuberculosis with intestinal ulcerations,? we have been dis-
secting the intestine to investigate the state of the mucous membrane. Doctors of old,
who either observed the intestine intact or did not always open it completely, could have
no knowledge of a worm that was in any case so small as to be undetectable to fingers
that had squeezed the initial intestine loops by chance. However, let us admit it, modern
practice was not the most suited to detect it either, as doctors usually opened the in-
testine and rinsed it with abundant water before examining the inner membrane. In this
way its mucous product was washed off and settled to the bottom of the wateI together
with the helminths [or intestinal worms) housed therein. Finally, the considerable
yellow-pink or ash-grey, and less-than-transparent mucus in which they are usually
wrapped, coupled with then small size, undoubtedly constituted a considerable obstacle
even for the most watchful observers. Yet, it is sufficient to separate the intestine from
the mesentery, segment by segment, an arm's length at a time, thereby unfolding its
loops; cut it lengthwise by means of an enterotomy [incision into the intestine); lay it
over one of the corpse's thighs and carefully examine the mucus contained therein. This
mucus should then be removed to expose the underlying membrane by using the back of
the same instrument, no washing being necessary. When, in the aforementioned circum-
stances, considerable thick, ash-grey or reddish mucus exists in the duodenum or
jejunum, it will not be difficult, provided it is carefully observed, to see small worms
wrapped in it, curled very much around themselves, transparent for the front quarter of
their length, yellowish, reddish and brown for then hind three quarters and marked in
the middle by a black spot which is the stomach that is normally filled with blackish
matter.
This passage is indicative both of the reasons that prevented earlier anatomists
from observing the worm and of the reasons that led Dubini away from attri-
buting any pathogenic action to it. The worm is, in fact, usually found free in
the duodenum-jejunum lumen and coated with a thick and "mostly yellow-
pink colored'" mucus, which is present in the lumen itself in greater
quantities. Sometimes the worm 'adheres' to the mucous membrane "by
means of its hook-shaped oral extremity." The mucous membrane "may
appear as normally constituted, i.e., slate-color-like with many block or red
264 LUIGI BELLONI
dots or also simply veined." The meaning of this dotting, for us quite
obvious, was the source of deep and prolonged concern for Dubini, who was
partially misled by cases in which the dots were either not accompanied by
worm-infestation or were localized in other areas that were not the Ancy-
lostoma's 'habitat' (as, for example, the case in which "the dotting was con-
fined to the elliptical plates of the ileum"). In one case, "this dotting was
found in the jejunum and there amidst the minute slate-colored dots, striae of
red points could be seen, which evidently had given rise to a color deeper
than slate in other parts of the mucous membrane."s In his Entozoografia 9
Dubini also recalls these 'slate-colored dots' which "seem to have originally
been red dots."
Dubini then pointed out - both in the worm and in the human organ
holding it - various indications of a hematophagous· aggressiveness of the
worm towards the organ, but preferred to adopt a very cautious attitude to-
ward the hypothesis of the hookworm's pathogenic activity. This perplexity
was increased both by the very high frequency (20% incidence at least) with
which the worm appeared in deliberately undertaken autopsies, as well as
by the presence of other lesions that might serve as causa mortis. What is
more, "the variety of diseases that caused those individuals to die prevents us
from establishing a relationship between the character of the disease and the
presence of the worm. I could only observe that deteriorated constitutions,
the cachectic [related to a general lack of nutrition and wasting occurring in
the course of a chronic disease or emotional disturbance] , the diarrheic, the
emaciated, or the anasarcous [a generalized inf:tltration of edema fluid into
subcutaneous connective tissue] were those in which it was found more
often." Dubini concludes that the presence of the worms "certainly cannot
be indifferent to the organism, even though - as is for trichocephalus [a
genus of worms related to the trichina worm] - they do not give rise to parti-
cular symptoms."
Extremely suggestive is the anatomo-clinical parallel relationship that
Dubini established 10 at the end of an autopsy he carried out on 5 July 1843,
in which the most conventional of anaemic patterns appeared - starting from
a 'tiger heart' - in a woman "not drained of blood by natural or artificial
losses" and carrier of "100 or more agchilostomas" [sic!].
The 'agchilostoma' -later changed to 'ancylostoma' - was to become the
object of special treatment in Dubini's parasitology treatise, Entozoografia
umana per servire di complemento agli studj d'anatomia patologica 11
published in 1850. This is now a classic treatise thanks also to its fifteen
copper plates, that were reprinted in Edoardo Perroncito's treatise, I parassiti
DISCOVERY OF DUODENAL ANCYLOSTOMA 265
against homeopathic medicine. Side by side with this main delirium, Raiberti
did not forget either neomesmerism nor hydropathy, while he considered two
other therapeutic methods as eclipsed. He wrote:
Our age - I am referring only to those systems within the reach of popular intelligence
and favor - has witnessed the birth, life and death of two crazy methods of healing all
diseases: Bucellati's vermifuge method and Le Roi's purgation method. The former
(Bucellati), having grown up in our lands, was not brilliant enough to be known abroad
and so his reputation was confmed to our city. The other (Leroi), because he came
from Paris, was known over the whole of Europe as if he were a fashion plate. Both of
them had their followers, both boasted of miracles, and had some martyrs. Hence they
passed on, as all errors do, leaving no traces other than a number of victims and a few
epigrams.
These cautious words are Dubini's19 and, to a great extent could also be
adapted for the case of hookworm, especially since this worm was found free
in the intestinal lumen, more or less like an ascarid. As we know today, this is
a typical post-mortem fmding - migration of the worms - but at that time
it was the decisive element that put Dubini off the track from attributing a
definite pathogenic action to the worm he had discovered.
DISCOVERY OF DUODENAL ANCYLOSTOMA 267
Dubini published his discovery in April 1843. Two years later, in 1845, it
was highlighted by the zoologist, Karl-Theodor-Ernst von Siebold (1804-
85),20 in the course of a summary review of the progress made in the field of
helminthology in the previous two years. In 1847, Franz Pruner (Pruner Bey,
1808-82)21 mentioned the high incidence of hookworm in corpse dissections
carried out in Egypt where, unlike Lombardy, corpses could be dissected
immediately after death. The parasite adheres to the mucous membrane of
the duodenum by means of the hooks on its sucking apparatus thereby giving
rise to ecchymosis [black-and-blue spot] .
After moving to Egypt in 1850, Wilhelm Griesinger (1817-68)22 and
Theodor Bilharz (1825-62)23 went even further. The enterorrhagia [intes-
tinal hemorrhage] that accompanied ecchymosis in one corpse dissected on
17 April 1852 emphasized the analogy between hookworms and leeches. The
so-called Egyptian chlorosis [synonym for ancylostomiasis, i.e., hookworm
disease] , a severe and fatal cachectic disease, widely endemic in Egypt, may be
considered as a 'hookworm-originated disease' (,Anchylostomenkrankheit').
The Egyptian hookworm firmly anchors itself to the mucous membrane
of the duodenum and sucks abundant quantities of blood, gradually putting
the infested patient into an acute state of anaemia until he dies ('progressive
pernicious anaemia'). The Lombard hookworm, on the other hand, is much
less dangerous and usually prefers to remain free in the intestinal lumen, as
Dubini had discovered it. Basically, this was the opinion of Giacomo Sangalli
(1821-97),24 Professor of Pathological Anatomy at the University of Pavia,
who stated in 1866: "for some years I have found the hookworm in nearly
half the corpses dissected in my school". 2S This same opinion was set forth
again by Sangalli 26 some ten years later in 1876.
The Brazilian hookworm showed itself to be just as aggressive as the
Egyptian as its blood sucking causes the so-called intertropical anemia or
tropical chlorosis. In these cases, autopsy findings were quite like those for
Egyptian chlorosis, as Otto Wucherer (1820-73)27 found in Bahia, from
1865 on.
On 20 May 1877, Prospero Sonsino,28 a Florentine physician who prac-
ticed in Egypt, informed the Societa Medico-Fisica Fiorentina about the
observations he had made in connection with hookworm chlorosis following
Griesinger's thesis during autopsies performed in Cairo. On 30 September
of the same year, Carlo Morelli,29 influenced by Sonsino, found in Florence
a hookworm in the corpse of a woman who had died from 'progressive
pernicious anemia', thereby finding the hookworm in Italy for the first time
outside of Lombardy_
268 LUIGI BELLONI
So, "Egyptian chlorosis occurs among humans in Italy too, seeing as the
hookworm sucks blood in our country as well", Grassi stated in a paper dated
the following November in 1878.37
His work with the Parona brothers had been essential to this end and had
to be improved upon by therapeutic evidence which would be the only proof
to confirm the cause and etiological relationship between hookworm and
anemia. Anemia had to be treated by administering an anthelmintic drug
capable of ridding the patient of hookworm. In the hands of Giovan Cosimo
Bonomo (1687)38 topical treatment, aimed at killing the acarus played an
important role in asserting the acarus-based etiology of scabies. 39
attribute of mine workers: the so-called 'miner's cachexy'. Disabled for work,
many sick miners returned home and went to their local hospitals. Cachectic
veteran miners from the tunnel were also hospitalized in the general internal
medicine clinic of Turin directed by Concato. Subjected to copromicroscope
examination by Perroncito,44 they were shown to be hookworm carriers.
This clinical diagnosis was confirmed by a post-mortem fmding in
February 1882. The case was a severely cachectic miner who had died quickly
as a result of serofibrinous peritonitis follOwing an intraperitoneal transfusion
of defibrinated blood,45 attempted 'in extremis.' During the course of the
autopsy, solemnly performed by Professor of Pathological Anatomy,
Francesco-Vittorio Colomiatti (1848-1883), in the presence of a large and
authoritative audience,
more than 1500 hookworms were found in the duodenum and jejunum, many of whom
exhibited their hematophagic properties very clearly. Their intestines contained various
quantities of very well-preserved erythrocytes [mature red blood cell].
To a certain extent their concern reflects our own for the partially unknown,
environmental conditions, to which astronauts are subjected today.
DISCOVERY OF DUODENAL ANCYLOSTOMA 273
Translated by
R. MABERRY and A. TALIERCIO
NOTES
* All medical definitions have been taken from Stedman's Medical Dictionary, 22nd ed.
(Williams and Wilkins, Baltimore, 1972).
1 L. Belloni, 'La scoperta dell'Ankylostoma duodenale', Gesnerus 19 (1962), 101-118;
'La medicina a Milano dal Settecento aI1915', in Stona di Milano (Fondazione Treccani
degli Alfieri, Milan, 1962), Vol. 16, pp. 991-997. The exhaustive Bibliography of Hook-
worm Disease (Rockefeller Foundation, International Health Board, New York, 1922)
is particularly worthy of mention.
2 A. Cazzaniga, La grande crisi della medicina italiana nel primo Ottocento (Hoepli,
Milano, 1951), cf. p. 122. This important historical study was also issued in instalments
in Vols. 4, 5 and 6 (1948-50) of Castalia (Milan).
3 E. H. Ackerknecht, Medicine at the Paris Hospital 1794-1848 (Johns Hopkins Uni-
versity Press, Baltimore, 1967).
4 L. Belloni, 'Per la storia del cuore tigrato. I contributi portati nell'Ospedale Maggiore
da A. Gambarini (1838) e A. Dubini (1843-44)', L 'Ospedale Maggiore 44 (1956),
252-8.
5 In Gazz. med. di Milano 6 (1847), 319-22, 328-31 and 388. Shortly after its publi-
cation, this paper was published both as an abstract of 30 pages (Milan 1847), and in a
"second edition with many additions and remarks" in Enciclopedia anatomica (G.
Antonelli, Venice) together with a reprint of Trattato di antropotomla.
6 'Nuovo verme intestinale umano (Agchylostoma duodenale), costituente un sesto
genere di Nematoidei proprii dell'uomo',Ann. universali di med. 106 (1843), 5-l3.
'Notizia di un nuovo verme degli intestini umani costituente un settimo genere di
Nematoidei proprii dell'uomo', Gazz. med. di Milano 2 (1843), 153-54.
7 Dubini is referring to studies carried out by Paris writers on tuberculosis, abdominal
typhus (typhoid fever) and intestinal ulcers of the two diseases. See Ackerknecht's work
mentioned in Note 3.
8 'Alcune avvertenze di anatomia patologica medica', Ann. universali di med. 113
(1845), 268-71.
9 p. 109. See Note 11.
10 A. Dubini, 'Abito cereo: palpitazioni di cuore con rumore di soffio perisistolico:
tensione e dolore all'epigastrio: lingua pallida, liscia e asciutta: poi subdelirio, stupore,
e nell'ultimo giorno 6 respirazioni per minuto: battendo ancora il polso 78 volte nello
DISCOVERY OF DUODENAL ANCYLOSTOMA 275
16 1. Castaldi, 'II contributo di Angelo Dubini aile conoscenze causali della scabbia e
l'importanza della sua Entozoografia per la Storia della microbiologia', Riv. di storia d.
scienze med. e natur. 28 (1937), 204-8.
17 A. Dubini, Entozoografia umana (Milano, 1850), p. 390.
18 E. Benassi, 'Un medico sedicente novatore del primo Ottocento: il Dott. Bucellati e i
suoi successi terapeutici', in Minerva Med. 42 (1951) 220-225 of the 'Parte varia.'
19 A. Dubini, Entozoografia umana (Milano, 1850), p. 277.
20 'Bericht liber die Leistungen im Gebiete der Helminthologie wahrend des Jahres 1843
und 1844', Archiv for Naturgeschichte 2 (1845), 202-255; see especially pp. 220-221.
21 Die Krankheiten des Orient's vom Standpunkte der vergleichenden Nosologie be-
trachtet von Dr. F. Pruner (Palm and Enke, Erlangen, 1847), pp. 244-245: "Selten
offnet man eine Leiche in Aegypten ohne Individuen von einer oder auch mehreren
Arten zu finden. Der Ascaris lumbr., Ascaris und Oxyuris vermicularis und Trichoceph-
alus dispar finden sich zu Haufen schon in den Gediirmen der Kinder. Unter den
Erwachsenen sind es besonders die kachektischen, wassersiichtigen und skrofulosen
Subjekte, we1che ausserdem an Anchylostoma duodenale im Zwolffingerdarme leiden,
wo dieser Parasit seinen vierfacherigen Saugeriissel mit 40 Haken an die Schleimhaut
heftet, wodurch er Ecchymosen veranlasst." [One seldom opens a corpse in Egypt
without finding worms of one or more kinds: ascaris lumbr., ascaris, oxyuris vermicularis
and trichocephalus dispar are found in swarms even in the intestines of children. In
adults it is especially the cachectic, dropsical and scrofulous who suffer from ancy-
c1ostoma duodenale in the duodenum where this parasite sticks its four-sided proboscis
with forty hooks into the mucous membrane where it causes ecchymosis.]
22 'Klinische und anatomische Beobachtungen uber die Krankheiten in Egypten', in
Archiv for physiologische Heilkunde, issued in many instalments, see particularly Vol.
13 (1854), pp. 555-561: 'Anchylostomenkrankheit und Chlorose.' - Please refer also
to Griesinger's own short communication mentioned in Note 27.
23 'Ein Beitrag zur Helminthographia humana, aus brieflichen Mittheilungen des Dr.
Bilharz in Cairo, nebst Bemerkungen von Prof. C. Th. v. Siebold in Breslau', Zeit. t
wissensch. Zoologie 4 (1852), 53-76 and plate V. See also p. 56: "Die Stelle, wo ein
solcher Strongylus sass, ist durch eine linsengrosse Ecchymose bezeichnet, in deren Mitte
ein weisser Fleck von Stecknadelgrosse bemerkbar. Dieser weisse Fleck ist in der Mitte
durchbohrt von einem nadeldicken bis in das submucose Bindegewebe dringenden
Loche. Manchmal zeigt die Schleimhaut flache Erhabenheiten von Linsengrosse und livid
braunrother Farbe, welche eine zwischen Tunica mucosa und muscularis im Bindegewebe
befindliche, mit Blut geflillte Hohle und darin zusammengeringelt den lebenden von Blut
voll gesogenen Wurm (bald ein Mannchen, bald ein Weibchen) enthalten." [The place
where such a strongyle was fixed is indicated by a lentil-sized ecchymosis in whose center
a white spot the size of a pinhead can be seen. This white spot is pierced in the middle by
a hole the thickness of a needle, which penetrates into the submucous connective tissue.
Sometimes the mucous membrane shows smooth protuberances the size of a lentil and a
livid reddish brown color where there is a blood-filled cavity (between the mucous tunic
and the muscle of the connective tissue) containing a curled-up living worm (sometimes
male, sometimes female) sucked full of blood.]
24 G. Sangalli, 'Geografia elmintologica: Anchilostoma e trichina', Giomale di anatomia
e fisiologia pat%gica 3 (1866), 100-106.
25 "More often than in the duodenum, I would find it in the first few loops of the
DISCOVER Y OF DUODENAL ANCYLOSTOMA 277
jejunum, where it would lie stuck to the mucous membrane by means of a thick and
copious greyish mucus, that is always to be found in helminthiasis of this kind." The
worm would be simply stuck to the mucous membrane and not fixed to it by its perioral
hooks, as Bilharz found in Egyptian chlorotic patients. "Bilharz attributes the chlorosis
to the sad effects of this hemorrhage, chlorosis being very frequent in Egypt. In our
countries this phenomenon due to the effects of hookworm has not yet been observed,
but I was able to convince myself of such a possibility. In fact the worm sometimes
adheres loosely to the mucous membrane with its mouth; and sometimes, on removing
it, a reddish spot is observable. This is due to the blood spreading in the mucous
membrane provoked by the worm's biting. Once I found a large ecchymosis in the
mucous membrane and the hookworm firmly flxed in the middle of it by means of its
mouth."
26 G. Sangalli, 'Sopra alcuni punti controversi di elmintologia' Mem. d. R. lstit.
Lombardo (Classe di scienze matem. e natur.) 13 (1876), 349-61: " ... I am not able to
give a reasonable explanation for the important fact that the worm, equipped with a
four-hooked mouth both in Egypt and in Italy, reveals in Egypt its fatal property (where
it is widespread) by thrusting its head into the mucous membrane of the intestine where
it lives, thereby causing hemorrhages, and consequently serious illnesses, while, in our
country, it only exceptionally exhibits a similar behavior and is, therefore, virtually
harmless." At this point, it may be appropriate to point out that, when using the term
'virtually harmless', Sangalli meant the case of a seriously anemic, 57 year old male
subject who, once autopsied, revealed 700 hookworms without any other detectable
recent or past perforation of the intestinal mucous membrane: "Yet, I admit that the
anemic condition exhibited by that corpse, and therefore the cause of the patient's
death, would not have been clear to me, had I not admitted some fatal influence of the
worms on the intestine, hence faulty chylification and hematopoiesis." In some rare
cases, then, among the many in which it is hosted in the human organism, the hookworm
has an anemia-producing action, that could, however, be ascribed to altered intestinal
absorption, rather than to a blood~ucking mechanism. We know that the longer the time
gap between patient death and corpse dissection, the more numerous are the worms
which detach themselves from the mucous membrane and are free in the intestine lumen.
On the basis of this observation, we can now explain a phenomenon that at that time
was construed as a difference in hematophagic aggressivity between the Italian hook-
worm and the Egyptian one.
See also Sangalli's 'Annotazioni critiche sull'anchilostoma duodenale', Rendiconti d.
R. lstit. Lombardo 11 (1878),460-467.
27 'Ueber die Anchylostomenkrankheit, tropische Chlorose oder tropische Hypoamie',
Deutsches Archiv fill' klinische Medicin 10 (1872), 379-400. - The importance of
Wucherer's research had already been highlighted by W. Griesinger, 'Das Wesen der tropi-
schen Chlorose', Al'chiv der Heilkunde 7 (1866), 381.
28 P. Sonsino, 'L'Anchilostoma duodenale in relazione coll'anemia progressiva per-
niciosa',Imparziale [Florence] 18 (1878),227-234.
29 C. Morelli, 'Intorno ad un caso d'anemia progressiva con Anchilostoma duodenale',
Lo Sperimentale 41 (1878),27-39.
30 R. Leuckart, Die mensch lichen Parasiten und die von ihnen hel'iUhrenden Kl'ank-
heiten. 2 vols. (Winter, Leipzig-Heidelberg, 1876), Vol. 2, pp. 410-460 and 480.
31 In honor of Giuseppe Balsamo-Crivelli (1800-1874), Professor of Natural History
278 LUIGI BELLONI
Gottardo al Sempione. Una questione risolta (C. Pasta, Turin 1909). - La maladie des
mineurs du St. Gothard au Simp/on. Une question resolue. Appendice (Etablissement
typographique nationale, Turin, 1912).
45 Lava, 'La trasfusione del sangue per la cavita del peritoneo', in L 'osservatore. Gazzetta
delle Cliniche. Organo utI d. Soc. di med. e chir. di Torino 16 (1880),81-82.
46 Many case histories in favor of thymol therapy were produced by B. Graziadei, 'II
Timolo nella cura dell'anchilostomanemia', Gior. d. R. Accad. di medicina di Torino
30 (1882), 821-855.
47 'L'anemia al traforo del Gottardo dal punto di vista igienico e clinico', Giorn. d. Soc.
Ital. d'Igiene 2 (1880), 276-346.
48 For the first cases studied by E. Parona in the Varese Hospital, see his 'L'anchilo-
stomiasi e la malattia dei minatori del Gottardo', Annali universali di Med. e chir. 253
(1880), 177 -202 and 464.
49 On the fiftieth anniversary of his death, Giordano was commemorated on June 28,
1965 at Palermo University by Giovanni Frada and Giuseppe Pontieri. See Folia medica
48 (1965), fasc. 10.
so E. Parona, 'L'anchilostomiasi nelle zolfare di Sicilia', Annali universali di med. e chir.,
277 (1886), 464-468.
51 From Giorn. d. R. Accad. di medicina di Torino 30 (1882), 90: "On behalf of Dr.
Giordano of Lercara, Perroncito wishes to communicate that Giordano has found hook-
worms in the sulfur mine workers suffering from anemia and that upon subjecting the
hookworm to male fern treatment, the patient recovers from the anemia." (From the
minutes of the meeting of 24 February 1882.)
SALVO D'AGOSTINO
I. INTRODUCTION
systems allowed Weber to conclude that Cw had also the meaning of conver-
sion factor between units of current in the two systems. 7
In more general terms Weber's procedure can be summarized thus: once
the idea is accepted that the written algebraic expressions of physical laws
are valid within a specific system of units, any difference in the value of
constants is a mere metrological 'accident'. However, it is this very 'accident'
which allows one to measure the value of Cw in the form of a 'conversion
factor', a ratio between the units of the same physical quantity in the two
systems.
Through this method, Weber, together with Rudolf Kohlrausch, an expert
experimentalist, succeeded in surmounting the difficulty of measuring Cwo
Towards this goal he also invented a new instrument, the ballistic galvano-
meter.
Depending on the choice of units, the result of the measurement 8 was:
Cw = 439450.10 6 mm sec-I (static and electrodynamic units, respectively).
Also, Cw = 310740.10 6 mm sec-I (static and electromagnetic units, respec-
tively).
The second value was equal to the speed of light as Weber and Kohlrausch
soon realized, but they attributed no particular significance to this new result,
which they considered 9 as a mere coincidence.
About five years later, Maxwell considered this equality as one of the most
important elements supporting his electromagnetic theory of light.
This point underlines the weight that research programs have in the evalua-
tion of empirical data. The idea that electricity could consist of particles and
involve remote forces as one of its fundamental features, oriented Weber's
theory and his evaluation of the meaning of Cw and implied a sharp differen-
tiation between light (already interpreted within the context of a quasi-elastic
wave theory) and electric phenomena. This differentiation tended to reject or
dismiss as irrelevant any evidence that the two might be connected through
Cwo
Yet Weber's conception of Cw influenced later developments in physics
for its other aspects. His theory introduced the concept of a characteristic
velocity in electrodynamics and this concept set a precedent lO that Einstein
certainly could not have ignored in his well-known statement that a limit
velocity exists in electrodynamics and in physics in general.
It needs to be emphasized that Weber's result stems from a metrological
approach to electrodynamics and physics.
History of physics seems to have overlooked that it is at this point in the
mid-1800's that metrology becomes an essential component of physics. The
286 SALVO D'AGOSTINO
manner in which physics laws are written influences physics theory as 'syntax'
conditions 'semantics' in general.
The historical event we have dealt with so far now enables us to discuss
further the problem regarding the empirical contents of a theory, with partic-
ular reference to the case at hand.
Both Weber and Maxwell accepted the usual electromagnetic phenomena
and experiments of their times as a phenomenological basis for their theories.
One of these is the experiment which charges and discharges a capacitor
suitably inserted in a circuit (the so-called 'electric transient' experiment)
and the typically uneven distribution of electrostatic and electromagnetic
forces in various parts of the circuit. Probably both scientists would have
agreed that the electromagnetic forces indicated by the rotation of a magnetic
needle set up by a current flowing in the wire are correlated with the variation
DISCOVERY OF THE VELOCITY OF LIGHT 289
of the density of charges upon the capacitor plates indicated by the displace-
ment of the leaves of an electroscope and consequently with the variation
in electric forces. In order to quantify these experiments and observations
Maxwell and Weber probably would also have agreed on methods for measur-
ing such forces, on the criteria for defining charge and force units. These
methods and criteria involved Ohm's and Kirchhoff's laws. Therefore they
would have agreed on a whole series of strictly correlated experimental
methods and laws, identifiable with the 'overlap area' of the two theories.
It would be just as wrong to consider this 'overlap area' as consisting of
mere observational contents as it would be inaccurate to consider it as being
purely theoretical.
While theories exist in this area, they are not comparable, however, with
Weber's and Maxwell's two rival theories. The theoretical contents of the
'overlap' may be located closer to the level of the 'observables' analyzed by
Margenau. 2o In the same sense, these contents may be considered as consisting
of weaker theories, or instrumental theories according to Bunge. 21 In other
words one could affirm that this overlap also contains a theoretical com-
ponent because it is made up of a whole complex of procedures and rules and
therefore its contents can be defined 'quasi-empirical'. 'Overlap' assertions
are not theory-neutral yet their theoretical contents are not of the same level
as those of Weber's and Maxwell's theories.
This point should be further analyzed. One should explain how the 'over-
lap' contents, substantially those on which Weber's discovery is based, can be
included in two such radically different theories.
In this analysis I have been substantially guided by K. F. Schaffner's con-
siderations in his 'Outlines of a Comparative Theory Evaluation'. 22
According to Schaffner, a scientific theory consists of statements and rules
of different kinds. Some statements concern non-logical primitive terms
whose Significance is related to previous theories (Le. to preliminary, or back-
ground knowledge). This is defmitely true if we look at a word like 'molecule'
as Maxwell used it in his kinetic theory; or 'electric current' as it was used in
its early meaning (according to Ampere). Besides, there are some statements
which establish how entities and/or processes described by axioms of the
theory are in the end related to the realm of the observable. For purposes of
convenience and in accordance with Schaffner's usage, let us refer to both
statements as Type-C statements (sometimes also referred to as correspond-
ence rules).
According to Schaffner, Type-O statements, on the other hand, describe
inter-subjective experiences such as, 'the appearance of light and dark fringes
290 SALVO D'AGOSTINO
Schaffner, this characteristic of the 'overlap area' has been connected with
the role of Type-O statements.
Translated by
JUDITH A. BOFFA
- ee' [ 1 1
r' - -Cw'
where e and e' are the charges and r is their distance. When the relative motion is uni-
form this breaks down to the simpler form:
ee' [ 1 dr']
-;z 1- Cw' --;uz .
It is evident that Cw represents the velocity dr/dt for which the force becomes null.
The force reverse its sign for dr/dt > Cwo
4 E. Whittaker, A History of the Theories of Aether and Electricity (Nelson, London,
1951),2 vols., Vol. 1, p. 201. Salvo D'Agostino, 'La scoperta di una velocita quasi uguale
alla velocita della luce nell'elettrodinamica di W. Weber (1804-1891)" Physis 3-4
(1976),297-318.
5 Whittaker,op. cit., p. 201.
6 W. Weber and R. Kohlrausch, 'Ueber die Elektricitiitsmenge, welche bei galvanischen
DISCOVERY OF THE VELOCITY OF LIGHT 293
Strom en durch den Querschnitt der Kette fliesst' (1856), Weber, op. cit., pp. 597 -608;
Wiederkehr,op. cit., pp. 140, 141; D'Agostino, op. cit., p. 309.
7 D'Agostino,op. cit., pp. 305-306.
8 Wiederkehr,op. cit., pp. 140-141.
9 Weber, 'Elektrodynamische Massbestimmungen insbesondere elektrische Schwingun-
gen', Werke, Vol. 4, pp. 107-241; p. 157ff.; D'Agostino, op. cit., p. 312.
10 D'Agostino,op. cit., p. 310.
11 J. C. Maxwell praises Weber's Electrodynamic Theory in his 1855 'On Faraday's
Lines of Force' as "a professedly physical theory ... which is so elegant, so mathema-
tical, etc.... " In 'A Dynamical Theory of the Electromagnetic Field' (1864), the theory
as developed by W. Weber and C. Neumann is considered "exceedingly ingenious and
wonderfully comprehensive ... it has served to guide the speculations of one who has
made so great an advance in the practical part of electric science, both by introducing
a consistent system of units in electrical measurement, and by actually determining
electrical quantities with an accuracy hitherto unknown".
12 R. Sviedrys, 'Physical Laboratories in Britain', in Historical Studies in the Physical
Sciences, Vol. 7, pp. 405-436, p. 425. S. D'Agostino 'Esperimento e teo ria nell'opera
di Maxwell', Scientia 113 (1978),453-467, p. 454. Also: D'Agostino 'Experiment and
Theory in Maxwell's Work' (English Translation), ibid.
13 J. C. Maxwell,A Treatise on Electricity and Magnetism, 3rd ed. Preliminary, Sections
1-6; also Sections 620-628.
14 Ibid., Section 786.
15 Ibid., Section 787.
16 Another evidence is, for Maxwell, the transversality of electromagnetic waves.
Treatise, Section 791. The equality between the square root of the dielectric constant
for melted paraffin, on one side, and its index of refraction, on the other (the so-called
Maxwell's Law) is not considered by Maxwell satisfactorily confrrmed by the existing
experiments. Treatise, Section 789.
17 See for details: "D'Agostino 'Esperimento e teoria ... ' ".
18 A related consideration to this aspect was given by Maxwell in a section of his
Treatise (Sections 769-770).
19 A. Sommerfeld, Physikalische Zeitschrift 36 (1935), 814, 820; D'Agostino, 'Esperi-
mento e teoria', p. 463.
20 Henry Margenau, The Nature of Physical Reality (McGraw-Hill, New York, 1950),
especially: Chapters 6, 7.
21 Mario Bunge, Philosophy of Physics (Reidel, Dordrecht, 1973), pp. 39-179.
22 Schaffner,op. cit., pp. 311-330. Schaffner's arguments are exemplified in his essay
by the Lorentz-Einstein theory of Relativity and the Michelson-Morley experiment. In
this paper I use 'overlap' in the same sense as Schaffner.
HOWARD E. GRUBER
often deliberately done their utmost to expunge meaning from the material
they present to the persons whose mental processes they are studying: non-
sense syllables or strings of unconnected words to be remembered; problems
to be solved having nothing to do with the prior knowledge or real concerns
of the person - these are still the commonplaces of experimental psychology.
Even when more meaningful contents are considered, as in various attempts
at 'content analysis', the approach taken is often statistical in nature - sub-
stituting the counting of the occurrences of ideas for a thorough and reflec-
tive examination of the structure of the particular ideas under investigation
as seen within the framework of the larger structure of the individual's
knowledge.
When we turn to the history of science, the picture seems, at least to this
occasional visitor in that domain, exactly the opposite. Maximal attention is
given to the contents of thought, and almost none of all to the process. More-
over, the most powerful vector is toward treating thought as highly specialized.
Monographic treatments tend not to deal with the person's thinking as a
whole, but to treat of particular achievements in isolation from each other.
Consider one example, which I choose not because it is poor, but because
it is excellent work, Camille Limoges' La selection naturelle: etude sur fa
premiere constitution d'un concept (1837-1859).2 Limoges' examination of
Darwin's initial construction of the theory of evolution through natural
selection is restricted to Darwin's 'Transmutation Notebooks'. Limoges does
not take up the interplay of the contents of these notebooks with the note-
books on 'Man, Mind, and Materialism', which Darwin kept during the very
same period of time. 3
Thus, the prevailing tendencies in both fields are expressions of three
questionable assumptions: (1) the contents and processes of thought can be
understood in isolation from each other; (2) diverse intellectual processes
can be studied in isolation from each other; and (3) individual ideas can be
studied in isolation from other ideas.
In recent years, efforts to consider the structure of knowledge and the
processes of thought together have grown in several quarters. The work of
Jean Piaget and other exponents of genetic epistemology is a foremost
example. At the level of concern for biological adaptation, Piaget's work has
always been highly process-oriented. But in his extended researches on
children's thinking, it is probably fair to say that, unlike other psychologists,
Piaget has been primarily concerned with the contents and structure of
knowledge. Only in recent years, especially under the leadership of Barbel
Inhelder, has the Geneva school devoted much attention to specifiable thought
COG NITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 297
of well-defined variables. But our task goes far beyond that. We want to
understand how these variables function in relation to each other, in some
coherent system, such as a solar system of many bodies, an individual organ-
ism, an ecospace, or a society. Such systems may be thought of as organized
structures: no functional relationships occur outside such structures, all
functional relationships occur within such structures. See Figure I for a
simple example.
Sense of Proportion
In trying to understand the case, we have a chance of being led to attend to
factors that are important in at least that case. This means that we may pro-
tect ourselves from wasting a career on something that is never important.
Moreover, attending to that factor in just the ways necessary for understand-
ing the one case may lead to understanding how that factor is important in
other cases. (But not necessarily in the same way, as I will discuss below.)
else's. Would that make them any the less interesting? Would we have to turn
studies of such individuals out of the scientific fold? Or, on the other hand,
do we need all the tools of science to understand these exceptional events?
tears over this loss. There does remain something constructive that we can do,
a point I shall expand in the second half of this paper.
What Kind of Generality Can We Expect from the Case Study Method?
We take specialization so much for granted that we forget that even in the
mind of the expert, the specialist, there is some organization of knowledge as
a whole - first of all, his own knowledge and secondly, his conception of
human knowledge in general. These organizations guide the work of the
individual, and their psychogenesis ought to be of far greater interest to us
than has been the case until now. In childhood, both in spontaneous play and
in organized education, specialization is absent - making its appearance
mainly at a point approximately the equivalent of advanced university train-
ing. In adolescence, especially in creative adolescents, there is commonly a
period of romantic, often passionate search for cosmic truths, for some
framework to understand the meaning of life, to justify the impending
specialization that the young person sees looming before him, to find his own
identity by placing his own future in some context. 9 Then, of course, in life
as we know it, disciplined specialization, careerism, and the pursuit of realistic
goals - all take hold of the person. Later in life, often enough, this narrowing
ebbs and gives way to a renewed search for meaning and coherence. In a
happy few this search is active throughout the life span, and it has been often
said (although not yet well documented) that great innovations are made by
those who bring together previously disparate disciplines - Le., by those who
do not accept the alienated condition of science as a world of non-communi-
cating villages.
In many scientific lives then, and not necessarily only those attaining
heights of great eminence, the appearance of narrow specialization is only a
moment in a life-history that is really far more complex. To understand all
this movement we need to look carefully at the individual's organization of
knowledge and at his organization of purposes. We need to consider them as
they develop over the life-history.
We may also choose to look at these organizations of knowledge and
purpose which are shared by individuals in face-to-face scientific communities
such as research teams, in 'invisible colleges', or in paradigmatic 'villages' of
the kind Derek de Solla Price discusses in his paper for this volume. But no
matter which focus we choose, it is plausible that some of the most interest-
ing intellectual phenomena we will wish to understand are unique events.
To pursue this goal we will have to struggle toward a method for con-
structing what may be called the theory of the case. We will need to develop
the necessary structural and systemic acumen. We will of course be guided by,
and in the course of our work will hope to contribute to, some very general
ideas about such structures and systems. The kind of generalization we hope
for, then, may not be about the objective world, so much as it is about our
304 HOWARD E. GRUBER
evolving way of describing it. Even if every innovative event in the history
of science were incurably unique, our ways of describing these events would
have something in common, we would ask similar questions about them all.
Imagine, for example, a motor mechanic in a world inhabited by a multitude
of ingenious inventors. He wants to understand how each new machine
works. But they are all different. Suppose each one uses a different fuel. He
may still ask, "How does this one get its energy?" The ways in which the
machines are alike may be of some interest, but such generalizations would
not satisfy our mechanic. He would want to know, 'How does this one work?'
That is the kind of question and the kind of answer we may hope for in the
study of unique systems producing new knowledge.
Related Methods
By now the reader may very well be asking: 'Isn't he flogging a dead horse?
Isn't he merely making a plea that we go on doing what we already do? What
does he want?'
To answer these questions I will now briefly survey a number of related
approaches, each of which might seem to produce cognitive case studies,
but does not.
I begin with Stillman Drake's critique of the historiography of science,
in which he voices concerns similar to my own.l0 First, he characterizes
recent shifts in historiography of science as a change from maximizing the
role of a few great revolutionary pioneers to minimizing the role of the
individual in history. In dealing with someone seemingly great, the strategy
of the new style is to "attribute as much of his thought as possible to his
predecessors, and to grant as little as possible to his own originality ..." and
thus to "discern the great underlying causes... [of] the slow, uneven but
continuous emergence of modern science from ancient philosophy" .11 Then
Drake points out the danger of error in using the historical context to explain
an individual's thought without deep study of the documentation of that
person's life: the "concatenation of ideas within an individual mind mayor
may not be identical with one selected as characteristic of a given society" .12
He suggests that a careless importation of our own general concerns into our
description of the thought of the individual being studied may lead to serious
mistakes. For example, if we are interested in the role of extraphysical ideas
in Galileo's thinking, we need not necessarily look only or even chiefly to
philosophy; in Galileo's case musical theory may have been more important,
a point which Drake develops in some detail. His father was a musician and
COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 305
both father and son were interested in musical theory. In sum, Drake charac-
terizes history of science as having been the study of the development of
disembodied ideas, detached from the individuals who think them. He makes
a plea for psychologically oriented biographical work, and it is clear that he
does not mean psychoanalytic biography so much as the kind of cognitive
case study that is the subject of the present paper. He explains that he has
deliberately chosen the title, Galileo Studies, Personality, Tradition· and
Revolution, rather than Koyre's 'Galilean Studies', to bring home the point
that the thinking person is at the center of his concerns.
In spite of this orientation, it seems to me that Drake himself has not
come very far in carrying out such a program. His own work, seminal as it is,
is only a set of sketches moving in the direction of a cognitive case study. The
situation, then, is that in the case of Galileo, a figure of the first order of
magnitude, and an unusually well-documented case, nothing like a cognitive
case study has yet appeared.
I have often asked myself, how could it happen that a novice and an
amateur like myself might have the opportunity to write the first article
on the thousands of pages of notebooks Charles Darwin kept during the
Beagle voyage? How could it happen that Darwin's transmutation notebooks
and his notebooks on man, mind, and materialism lay almost unexploited in
the Cambridge University Library until a psychologist knowing little about
the history of science came along? And now, how does it happen that my
student, Martha Moore-Russell, has a similar opportunity to open up the
almost virginal notebooks of John Locke? If it is true in the case of such
figures, there must be many others waiting. Somehow, our disciplines,
psychology and the history of science, have not evolved in such a way as to
pay close attention to the inner workings - in all their richness and in their
complex development - of creative minds, even when the documentation is
available. That we do not yet have the conceptual tools to do this work, and
must improvise them as we go along, means that, for some reason, we have
not been asking the kind of questions that generate cognitive case studies.
A similar situation obtains in other fields. When I read biographies, such as
Clark's recent excellent biographies of Einstein and of Russell, I am invariably
disappointed at how little attempt is made to examine the actual thought pro-
cesses of a person whose greatest distinction lay in those thoughts. How indeed
could Clark address himself to the thinking that went into Principia Mathe-
matica in the scant twenty-two pages he devotes to that chapter in Russell's
life? It would seem as though the biographer is all too easily distracted or
bemused by 'objective' and external events in the life of the thinking person,
306 HOWARD E. GRUBER
incident in Darwin's later life (the relation between Darwin and Butler began
with Butler's letter to Darwin in 1863, when the latter was 54 years old) is
seized upon; some psychosexual connection is made between a highly specu-
lative reconstruction of the two men's infancies on the one hand and the
vicissitudes of their later relationship on the other; and the whole is served up
as an explanation of Darwin's creative work. All without opening up a single
notebook from the massive documentation left behind by Darwin, and worse
still, all without any serious discussion of the contents of Darwin's (or for
that matter Butler's) thinking.
I do not mean by these remarks to pass any judgment on the general
relation between neurosis and creativity. I only question how work of this
sort can ever be expected to shed any light on the actual thought processes
involved in creative scientific work. Suppose we grant that Darwin's almost
cosmological aspirations and his seeming interest in the obscure origins of
species can plausibly be explained as expressions of yearning for the lost
mother (she died when he was ten years old) and curiosity about his own
origins. This does little to explain how his thinking developed, how he went
from one idea to another.
If we try to explain the theory of evolution itself as an expression of
personality, we stumble on the difficult fact that two men as different as
Alfred Russel Wallace and Charles Darwin - different in class background,
family life, temperament, and even philosophical point of view - hit upon
approximately the same theory. What we need to do is to show how each
individual, through his own struggles, arrived at a viable intellectual adapta-
tion. Darwin and Wallace travelled different roads for most of their lives, the
same theoretical road for a while, and then their paths diverged (Wallace, a
spiritualist, rejected the idea of the origin of human mentality through na-
tural evolutionary processes). If personality is relatively fixed, which parts of
these two men's life cycles were expressed in their works? In short, a theory
that explains thought as the re-enactment of early experiences cannot explain
creative work; such work has as its predominant characteristic, not re-enact-
ment but purposeful growth.
Perhaps a more promising approach is to suggest that every individual
contains several potential or virtual personalities: rather than pre-existing
and causing the person's work, they are brought alive and even generated
as needed by the specific situations in which the individual finds himself.
These situations are in good part determined by the fortunes of the work,
so central in the lives of the kind of people we are discussing. Just as a brave
act may make a man brave (Mao Tse Tung?), a successful theoretical effort
308 HOWARD E. GRUBER
may make a person more reflective. Thus, an evolving systems approach takes
a special view of the relation between the person and his or her work: both
evolve continuously throughout the life history, and they co-exist in a fruit·
ful dialectical relationship which is the motor of growth. This is not the
position taken in psychoanalytically oriented studies: neither Freud's work
on Leonardo, nor Erikson's seminal study of Luther, 18 nor Frank Manuel's
fascinating psychobiography of Issac Newton. 19 These works all have a com·
mon logical structure: personality is the explanans and creativity is the
explanandum.
These efforts represent a search for causal laws, while the approach to the
psychology of creative scientific thought that we are proposing is systemic.
Just as we are sceptical of the value of searching for one great moment of
insight, or intellectual conception, we are sceptical of the value of searching
for the causes of adult creativity in one infantile experience. There is not
some one special moment in the functioning of an evolving system that
explains all later moments. Each moment prepares the way for a new set of
possibilities, and in the system's interaction with its milieu, a new choice is
made, and so on indefmitely.
If the reader will retrace his path a few pages he will note that I spoke
of Darwin's seeming preoccupation with origins. In reality, Darwin made
little or no effort to trace out the particular origins of any organism or
group. He was preoccupied with describing the function of the system of
nature in order to account for its perpetual originality. Marc Bloch wrote
disparagingly of the 'idol of origins'. It cannot be thought that this great
historian was uninterested in the past, but from the evolving systems point
of view the past has no special privileges.
a major part of the task, but in the long run we face the more difficult one of
reconstructing the underlying structure giving rise to it.
Behavior is Choice
the analogy of a chart room. Within any given chart, there is a marvelous
order, and what is more, a point-for-point correspondence between the chart
and some natural object, such as an island, or a continent, or a starscape. We
would not lightly change the relationships so carefully depicted within any
chart. For example, we would not, without considerable thought about our
map of the earth, interchange Greenland and Australia. Between charts, how-
ever, there is no such stable ordering. If we are investigating a biogeographical
hypothesis, we may overlay maps of the distributions of species on a map
depicting climatic changes. If we are investigating a certain geological hypoth-
esis, we do not hesitate to bring the coastlines of distant continents together.
I do not, of course, mean to suggest that cognitive structures are all maplike
in the sense of being spatialized representations of spatial phenomena.
Sometimes our charts may be ordered merely to make them accessible for
any purpose, as alphabetically. At other times they may be ordered for a
particular purpose. We might be tempted to speak of a collection of 'local'
organizations. But some of the representations we are discussing are hardly
local - they may be quite inclusive, even cosmic. It is better to speak of
special organizations, each one a representation constructed for some purpose.
A first phase, then, in deepening our grasp of the organization of knowl-
edge and purpose is to make a set of such 'charts'. Or at least to develop ways
of making them as needed. As I have argued, our approach to the problem of
representation must be pluralistic. We are not looking for one master chart
containing the individual's entire knowledge and purposes. In a recent exhibi-
tion at the Museum of Modern Art in New York City, a quarter of Tokyo,
Shinjuku, was depicted by means of more than 12 'experience maps' (geo-
graphical trails of different individuals, odor distributions, affective quality,
etc.). No one of them could be taken as capturing the 'essence' of Shinjuku.
Rather, the ensemble of maps represented some part of the complexity of a
few individuals' experience of that quarter. This approach corresponds to
trends in the field of information processing and computer simulation of
cognitive processes, where such representations are called 'frames' or 'scenes'
deliberately in order to emphasize their multiplicity.2o
We can expect the number of possible representations to be indefinitely
large: from the interplay of already existing representations of substantive
knowledge, procedural knowledge, and purposes a new representation can be
generated at any time. For example, the words 'take a closer look at that!'
describe an intention to enlarge a representation and to map more detail
onto it. From recent work on imagery, we know that the subjective 'size' of a
representation is a meaningful variable that has an experimentally measurable
COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 311
time the person has completed the feasible tasks entailed in the enterprise at
one point in time, a new collection of tasks has emerged to replenish the
inventory of work to be done.
Second, I want to distinguish between enterprises, tasks, and problems.
Not all tasks are problems. Again the ditch: for most farmers, digging a ditch
is not usually a problem, although admittedly sometimes a problem may be
encountered, such as a big rock. A creative life is one in which a group of
enterprises are organized in a structure that permits the individual to achieve
a creative end, or a series of them. Each enterprise entails a group of tasks,
some of which turn out to be problematical. In some cases a task may be
undertaken specifically in order to expose any problems it may involve, and
in the solution of those problems the enterprise as a whole will be stressed
and shaped. Solving a series of problems does not necessarily make a creative
life, as any crossword puzzle addict knows. The problems solved must be
organized as members of a coherent enterprise, leading to some novel and
effective product.
Third, I want to distinguish between enterprise and paradigm. In Kuhn's22
usage, a paradigm is a highly evolved, shared way of working, which only
emerges when considerable progress has been made along certain lines, and
when a socialized pattern of education, communication, and criticism has
developed. The enterprises of which I speak may be paradigmatic, pre-
paradigmatic, or in some obstinate cases even post-paradigmatic. Most likely,
in a creative life, some enterprises are shared and others quite unique to the
individual. I would add, however, that even if every enterprise in a given
person's network was a shared paradigm, the network as a whole would cer-
tainly be patterned and orchestrated in a unique way by that person, and
creative results might ensue.
Journa[24 of the voyage. The diary and Journal are important because there
we see much of Darwin's imagery, and also because his observations and
reflections on homo sapiens wherever he went in his circumnavigation are the
beginning of another major branch of the network, his concern for the evolu-
tion of psychological processes in homo sapiens and other animals.)
The geological enterprise took definite shape during the Beagle voyage and
remained at least sporadically active throughout his lifetime, or from 1831-
1882. The biological enterprise, although having somewhat earlier beginnings
in his boyhood, grows more slowly during the five years of the voyage. By the
end of the voyage or shortly thereafter, both these major branches have been
transformed so that they are part of one much larger enterprise, the search
for one unifying theory of evolution and evolutionary approach to all of
nature; Darwin's interest in the human species and all other enterprises are
incorporated in this all-embracing enterprise.
From 1838 on it is clear that we must speak of three major branches of
the network: geology, biology, and psychology (or better, perhaps, sciences
of man). Darwin's quasi-literary interests recede into the background, but
do not disappear, giving rise to the much re-worked second edition of the
Journal,25 to his biography of his grandfather Erasmus Darwin,26 and to his
own autobiography;27 these writings are spread out over some 40 years.
Darwin's network of enterprise is characterized by a very high order of
continuity. All branchings are foreshadowed early, before they become mani-
fest as distinct enterprises. Once a branch is begun, it never becomes per-
manently dormant again.
I will give one example of the emergence of a 'minor' branch. During the
voyage, in 1835, Darwin became interested in the way in which organisms
transform the physical characteristics of the earth, the theme that life makes
land. At that early time, he worked out his highly successful theory of the
formation of coral reefs through the action of the coral organism, or rather,
through its interaction with a group of geological and climatic processes.
Shortly after the voyage, in May 1837, he presented this theory to the London
Geological Society. In November of 1837, among his manifold activities, he
read another paper before the Society, this time on ''worms forming mould,"
as he wrote in his personal journal. 28 His interest in worms continued through-
out his life, and in 1881 he published his book, The Formation of Vegetable
Mould through the Action of Worms, with Observations on Their Habits. 29 As
is evident from the title, he used this work to report on aspects of two differ-
ent branches of his network of enterprise - on the one hand, his peculiar
blending of geological and biological considerations in the "life makes land"
314 HOWARD E. GRUBER
motif; on the other hand, his prolonged search for rudiments of intelligent
behavior in lower organisms.
In addition to its continuity, Darwin's network of enterprises is also
remarkable for its great coherence. From 1838 on, almost everything he does
is relevant to this evolution enterprise. Nevertheless, he conducted many of
his enterprises in such a way that they would be valid contributions to the
science of his day, independent of (and for a long time, without mentioning)
their relevance to his views on evolution.
As I have described in Darwin on Man, he was a pastmaster at the separa-
tion of issues when this would serve his purposes. His eight-year study of
barnacles was certainly guided by evolutionary theory, but when he published
his four classic monographs on the subject, he did not refer to this theoretical
base. On the other hand, when he wrote the Origin of Species 30 he exploited
the material he had so laboriously worked up. Because he could not solve
fundamental genetic questions, he managed to separate the issue of the cause
of variation from the issue of the role of variation in evolution. Because he
felt the time was not ripe in 1859, when he wrote and published the Origin,
he managed to separate the question of human origins from evolution in
general. He kept his silence for twelve years, publishing the Descent of Man 31
in 1871.
Without question, then, Darwin's clear grasp of his own network of enter-
prise permitted him to plan his work purposefully, to concentrate his thinking
on different subjects in a flexible and adaptive way, and to time his publica-
tions strategically. But the protracted parallel activity of different branches of
the network also serves other functions, perhaps more profoundly related to
the process of creative work. When diverse activities are coeval they have a
greater chance of influencing and enriching each other. In Piaget's language,
the mutual assimilation of schemes is one of the major mechanisms of intel-
lectual growth. Resuming work on a lapsed enterprise permits the magnifica-
tion of this effect: a technique or a style of thought developed in one context
may be seen as useful in another. In Darwin's case, as Ghiselin and I have
independently pointed out, his formulation of the theory of coral reefs bears
a striking formal resemblance to the theory of natural selection, as he con-
structed it three years later. 32 The transfer of ideas from one domain to
another is facilitated by the organization of work in separate enterprises
which can lie fallow and be re-activated in complex temporal patterns. In
general I believe it gives a more faithful picture of Darwin's thinking to em-
phasize his purposeful, conscious control of his own intellectual work, but
it should not surprise us to find major exceptions. The above example is a
COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 315
striking case. The parallels are very close, and even the language is sometimes
almost identical, but nowhere does Darwin indicate any conscious awareness
of the similarity between the two theories.
In their seminal work, Plans and the Structure of Behavior, Miller, Galanter,
and Pribram 33 begin by introducing the distinction between Image and Plan.
Images have to do with the organization of knowledge and plans refer to the
organization of behavior. Almost the entire book is about plans: "A Plan is
any hierarchical process in the organism that can control the order in which
a sequence of operations is to be performed".34 When one part of a plan is
completed, the system as a whole (i.e., the larger plan of which it is a part)
is so organized as to guide the inception of the next part of the plan (or
subroutine, or schema - terms that are practically equivalent to plan). In an
important sense, the plan is conceived of as closed - it has a built-in mecha-
nism for detecting when execution up to some criterion has been completed,
and at each level of the hierarchy the same closure characteristic can be
found. Thus, plans that exist control behavior, but plans are not in themselves
generative. Nor are plans equivalent to purposes: rather, plans are what we
must have in order to carry out our purposes.
In contrast, the concept of enterprise is open-ended and generative. The
network of enterprises describes the individual's organization of purposes. Of
course, procedural knowledge (plans) is embedded in every enterprise. But
when the running off of a plan comes up against obstacles, new procedures
must be invented. How the individual decides whether to struggle with such
difficulties or to shift to some other activity is regulated by the organization
of purposes as a whole.
Images of Wide Scope
Although Miller et al. were aware of the need for a system capturing the inter-
play between knowledge and action, they abandoned the Image and pursued
Plans. I believe this may have been because they conceived of the 'image of
the world' as singular, as one great organization embracing all the individual's
knowledge. This is probably a self-defeating idea, unrealizable and well worth
abandoning. But in the years since this work appeared great effort has been
expended on ways of representing knowledge. As I have indicated above, I
believe that a pluralistic approach is most promising. Each individual must
have at his disposal a number of modalities of representation. Systems of
laws, taxonomic systems, and thematic repertoires - such as those explored
by Gerald Holton 35 - are all pertinent. In the present essay I take up only
one such idea, images of wide scope.
316 HOWARD E. GRUBER
anything else, nor is it often conveniently like anything else in all the ways we
might need for some particular scientific purpose. For this reason, a versatile
repertoire of images is valuable for exploring the properties of the phenom-
enon that interests us. When we have done this exploration, we may know
enough about the phenomenon to construct a model. Because the model is
completely uncoupled from any reality, we can build into it any properties
we like, thereby coupling it as tightly as possible to our knowledge of the
phenomenon or process in question. In the cognitive economy as a whole,
each modality of representation has its contribution to make.
Among all his metaphors, Darwin's image of the tree of nature as an irregularly branch-
ing tree certainly deserves pride of place. It appears early in the B-notebook ... and is
then quickly redrawn to bring out Darwin's thought more precisely. Over the years,
Darwin drew a number of tree diagrams, both trying to perfect it and to penetrate it -
to learn what his own imagery could tell him. In a highly formalized version, the tree
diagram is the only figure in the Origin, and Darwin refers to it over and over, through-
out the book.42
The tree diagram helped Darwin see and formulate a number of points. It
depicts speciation and evolutionary divergence. It explains why gaps (so-called
'missing links') among contemporaneous species are not an argument against
evolution, since, in a branching structure, continuity in evolutionary time can
be associated with great gaps among contemporaries. Darwin used the tree
image to explore and expound his profound conviction that living nature is
irregular. Viewed in another way, any tree diagram is a model of exponential
growth; combined with a constraint, such as a limit on the number of organ-
isms, a formal principle of selection necessarily follows. By formal I mean
COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 319
NOTES
* I am grateful to Martha E. Moore-Russell and to Doris Wallace for their helpful com-
ments on a draft of this paper.
1 Bloch, M., The Historian's Craft (Random House, New York, 1953), p. 194.
2 Limoges, C., La selection naturelle: etude sur la premiere constitution d'un concept
(1837-1859) (Presses Universitaires de France, Paris, 1970).
3 Gruber, H. E., Darwin on Man: A Psychological Study of Scientific Creativity; to-
gether with Darwin's Early and Unpublished Notebooks, transcribed and annotated by
Paul H. Barrett (E. P. Dutton, New York, 1974; 2nd ed. University of Chicago Press,
Chicago, 1980).
4 See for example, Gillithon, Ch., Decalages et seriation (Archives de Psychologie,
Monograph No.3, 44, 1976).
COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 321
5 Newell, A., 'You Can't Play 20 Questions with Nature and Win', in Visual Information
Processing, Chase, W. G. (ed.) (Academic Press, New York, 1973).
6 Moore-Russell, M. E., 'John Locke: The Development of a Philosopher as a Person-
in-Society', Proceedings of the Sixth Annual Symposium of the Jean Piaget Society
(Philadelphia, 1976).
7 Psychologists attempting to fractionate creativity into component abilities have la-
belled one factor 'divergent thinking', the ability to generate a large number of potential
solutions to a problem, usually a simple problem such as, 'Think of as many uses as
possible for a brick'. See Guilford, J. P., The Nature of Human Intelligence (McGraw-
Hill, New York, 1967).
8 Employed with sensitivity, statistical tools might be helpful in detecting unusual
events, so that those of us who wish could pursue the rare bird wherever it flies, and
study it close at hand.
9 See Erikson, E. H., Childhood and Society, second edition (W. W. Norton, New York,
1963); Inhelder, B. and Piaget, J., The Growth of Logical Thinking from Childhood to
Adolescence (Basic Books, New York, 1958); Gruber, H. E. and Voneche, J. J., 'Re-
flexions sur les operations formelles de la pensee', Archives de Psychologie 44 (1976),
45-55.
10 Drake, S., Galileo Studies: Personality, Tradition, and Revolution (University of
Michigan Press, Ann Arbor. Mich., 1970).
11 Ibid., p. 3.
12 Ibid.,p.5.
13 Kramer, H., 'Writing Writers' Lives', New York Times Book Review, May 8, 1977,
p.3.
14 Steinberg, M., in New York Times Book Review, December 25, 1977, p. 21, review
of Solomon, Maynard, Beethoven (Schirmer/Macmillan, New York, 1977).
15 Friedliinder, S., Histoire et psychanalyse: essai sur les possibilites et les Iimites de la
psychohistoire (Seuil, Paris, 1975).
16 Freud, S., Leonardo da Vinci and a Memory of His Childhood (W. W. Norton, New
York, 1964) (originally pub!. 1910).
17 Greenacre, P., The Quest for the Father: A Study of the Darwin-Butler Controversy,
as a Contribution to the Understanding of the Creative Individual (International Univer-
sities Press, New York, 1963).
18 Erikson, E. H., Young Man Luther: A Study in Psychoanalysis and History (W. W.
Norton, New York, 1958).
19 Manuel, F. E.,A Portrait of Isaac Newton (Belknap Press of Harvard University Press,
Cambridge, Mass., 1968).
20 See for example: Minsky, M. and Papert, S.,Artificial Intelligence (Condon Lectures,
Oregon State System of Higher Education: Eugene, Oregon, 1973).
21 See for example: Kosslyn, S. M., Murphy, G. L., Bemesderfer, M. E., and Feinstein,
K. J., 'Category and Continuum in Mental Comparisons', Journal of Experimental
Psychology: General 106 (1977), 341-375.
22 Kuhn, T. S., The Structure of Scientific Revolutions, 2nd edition enlarged (The
University of Chicago Press, Chicago, 1970).
23 Darwin, C., Charles Darwin's Diary of the Voyage of H.M.S. 'Beagle' (edited from the
MS by Nora Barlow) (Cambridge University Press, Cambridge, 1934).
24 Darwin, C., Journal of Researches into the Geology and Natural History of the
322 HOWARD E. GRUBER
323
M. D. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 323-325.
Copyright © 1980 by D. Reidel Publishing Company.
324 BIOGRAPHICAL NOTES
JOHN NORTH is Professor of the History of Philosophy and the Exact Sciences at the
Rijksuniversiteit Groningen (The Netherlands). He is Editor of the Archives lntema-
tionales d'Histoire des Sciences, the journal of the Academie Intemationale d'Histoire
des Sciences, of which he is a member. His publications include The Measure of the
Universe (Oxford U.P., 1965) and Richard of Wallingford (Oxford U.P., 3 vols, 1976).
DEREK DE SOLLA PRICE is Avalon Professor of the History of Science at Yale Univer-
sity. In 1971, when the International Council for Science Policy Studies was founded, he
became its first president. In 1976 he was the recipient of the Leonardo da Vinci medal,
the major award of the Society for the History of Technology. He has published nearly
200 scientific papers and six books, including: Science since Babylon, enlarged edition,
1975; Little Science, Big Science, 1964; Gears from the Greeks: The Antikythera
Mechanism - a Calendar Computer from ca. 80 B. c., 1975; and was editor of Science,
Technology and Society: A Cross·disciplinary Perspective, 1977.
Editors:
ROBERT S. COHEN and MARX W. WARTOFSKY
(Boston University)
1. Marx W. Wartofsky (ed.), Proceedings of the Boston Colloquium for the Philosophy
of Science 1961-1962. 1963.
2. Robert S. Cohen and Marx W. Wartofsky (eds.), In Honor of Philipp Frank. 1965.
3. Robert S. Cohen and Marx W. Wartofsky (eds.), Proceedings of the Boston Collo-
quium for the Philosophy of Science 1964-1966. In Memory of Norwood Russell
Hanson. 1967.
4. Robert S. Cohen and Marx W. Wartofsky (eds.), Proceedings of the Boston Collo-
quium for the Philosophy of Science 1966-1968. 1969.
5. Robert S. Cohen and Marx W. Wartofsky (eds.), Proceedings of the Boston Collo-
quium for the Philosophy of Science 1966-1968. 1969.
6. Robert S. Cohen and Raymond J. Seeger (eds.), Ernst Mach: Physicist and Philos-
opher. 1970.
7. Milic Capek, Bergson and Modern Physics. 1971.
8. Roger C. Buck and Robert S. Cohen (eds.), PSA 1970. In Memory of Rudolf
Carnap. 1971.
9. A. A. Zinov'ev, Foundations of the Logical Theory of Scientific Knowledge
(Complex Logic). (Revised and enlarged English edition with an appendix by
G. A. Smirnov, E. A. Sidorenka, A. M. Fedina, and L. A. Bobrova.) 1973.
10. Ladislav Tondl, Scientific Procedures. 1973.
11. R. J. Seeger and Robert S. Cohen (eds.), Philosophical Foundations of Science. 1974.
12. Adolf Griinbaum, Philosophical Problems of Space and Time. (Second, enlarged
edition.) 1973.
13. Robert S. Cohen and Marx W. Wartofsky (eds.), Logical and Epistemological
Studies in Contemporary Physics. 1973.
14. Robert S. Cohen and Marx W. Wartofsky (eds.), Methodological and Historical
Essays in the Natural and Social Sciences. Proceedings of the Boston Colloquium
for the Philosophy of Science 1969·1972. 1974.
15. Robert S. Cohen, J. J. Stachel and Marx W. Wartofsky (eds.), For Dirk Struik.
Scientific, Historical and Political Essays in Honor of Dirk Struik. 1974.
16. Norman Geschwind, Selected Papers on Language and the Brain. 1974.
18. Peter Mittelstaedt, Philosophical Problems of Modern Physics. 1976.
19. Henry Mehlberg, Time, Causality, and the Quantum Theory (2 vols.). 1980.
20. Kenneth F. Schaffner and Robert S. Cohen (eds.), Proceedings of the 1972 Biennial
Meeting, Philosophy of Science Association. 1974.
21. R. S. Cohen and 1. 1. Stachel (eds.), Selected Papers of Leon Rosenfeld. 1978.
22. Milic Capek (ed.), The Concepts of Space and Time. Their Structure and Their
Development. 1976.
23. Marjorie Grene, The Understanding of Nature. Essays in the Philosophy of Biology.
1974.
24. Don Ihde, Technics and Praxis. A Philosophy of Technology. 1978.
25. Jaakko Hintikka and Unto Remes, The Method of Analysis. Its Geometrical Origin
and Its General Significance. 1974.
26. John Emery Murdoch and Edith Dudley Sylla, The Cultural Context of Medieval
Learning. 1975.
27. Marjorie Grene and Everett Mendelsohn (eds.), Topics in the Philosophy of Biology.
1976.
28. Joseph Agassi, Science in Flux. 1975.
29. Jerzy 1. Wiatr (ed.), Polish Essays in the Methodology of the Social Sciences. 1979.
32. R. S. Cohen, C. A. Hooker, A. C. Michalos, and J. W. van Evra (eds.), PSA 1974:
Proceedings of the 1974 Biennial Meeting of the Philosophy of Science Association.
1976.
33. Gerald Holton and William Blanpied (eds.), Science and Its Public: The Changing
Relationship. 1976.
34. Mirko D. Grmek (ed.), On Scientific Discovery. 1980.
35. Stefan Amsterdamski, Between Experience and Metaphysics. Philosophical Problems
of the Evolu tion of Science. 1975.
36. Mihailo Markovic and Gajo Petrovic (eds.), Praxis. Yugoslav Essays in the Philosophy
and Methodology of the Social Sciences. 1979.
37. Hermann von Helmholtz: Epistemological Writings. The Paul Hertz/Moritz Schlick
Centenary Edition of 1921 with Notes and Commentary by the Editors. (Newly
translated by Malcolm F. Lowe. Edited, with an Introduction and Bibliography,
by Robert S. Cohen and Yehuda Elkana.) 1977.
38. R. M. Martin, Pragmatics, Truth, and Language. 1979.
39. R. S. Cohen, P. K. Feyerabend, and M. W. Wartofsky (eds.), Essays in Memory
of Imre Lakatos. 1976.
42. Humberto R. Maturana and Francisco J. Varela, Autopoiesis and Cognition.
The Realization of the Living. 1980.
43. A. Kasher (ed.), Language in Focus: Foundations, Methods and Systems. Essays
Dedicated to Yehoshua Bar-Hillel. 1976.
46. Peter L. Kapitza, Experiment, Theory, Practice. 1980.
47. Maria L. Dalla Chiara (ed.),Italian Studies in the Philosophy of Science. 1980.
48. Marx W. Wartofsky, Models: Representation and the Scientific Understanding.
1979.
50. Yehuda Fried and Joseph Agassi, Paranoia: A Study in Diagnosis. 1976.
51. Kurt H. Wolff, Surrender and Catch: Experience and Inquiry Today. 1976.
52. Karel Kosik, Dialectics of the Concrete. 1976.
53. Nelson Goodman, The Structure of Appearance. (Third edition.) 1977.
54. Herbert A. Simon, Models ofDiscovery and Other Topics in the Methods of Science.
1977.
55. Morris Lazerowitz, The Language of Philosophy. Freud and Wittgenstein. 1977.
56. Thomas Nickles (ed.), Scientific Discovery, Logic, and Rationality. 1980.
57. Joseph Margolis, Persons and Minds. The Prospects of Nonreductive Materialism.
1977.
58. Gerard Radnitzky and Gunnar Andersson (eds.), Progress and Rationality in Science.
1978.
59. Gerard Radnitzky and Gunnar Andersson (eds.), The Structure and Development
of Science. 1979.
60. Thomas Nickles (ed.), Scientific Discovery: Case Studies. 1980.
61. Maurice A. Finocchiaro, Galileo and the Art of Reasoning. 1980.