$ULVWRWOH V5HYLHZRIWKH3UHVRFUDWLFV,V$ULVWRWOH)LQDOO\

D+LVWRULDQRI3KLORVRSK\"
&DWKHULQH&ROOREHUW

Journal of the History of Philosophy, Volume 40, Number 3, July 2002,

pp. 281-295 (Article)

3XEOLVKHGE\-RKQV+RSNLQV8QLYHUVLW\3UHVV
DOI: 10.1353/hph.2002.0047

For additional information about this article

http://muse.jhu.edu/journals/hph/summary/v040/40.3collobert.html

Access provided by Florida Atlantic University (19 Jul 2015 11:48 GMT)
 A R I S T O T L E ’ S R E V I E W O F T H E P R E S O C R AT I C S 281

Aristotle’s Review of the

Presocratics: Is Aristotle Finally a
Historian of Philosophy?1
CATHERINE COLLOBERT*

“JUST AS INEXPERIENCED SOLDIERS IN FIGHTS, rushing forward from all sides, often strike
fine blows, but without knowledge, so they do not seem to understand what they
say” (Met. 985a13–16). This negative judgment of Aristotle about his predeces-
sors has been the object of numerous controversies, which could be summarized
by the following question: was Aristotle writing philosophy or history of philoso-
phy when he set out the doctrines of his predecessors?
This controversy was reopened2 by H. Cherniss,3 according to whom the Aris-
totelian review of the Presocratics rests upon, on the one hand, the firm belief of
Aristotle that each Presocratic groped around the truth and, on the other hand,
the fact that Aristotle judged their mistakes and successes in the light of his own
system, understood as the final realization of previous doctrines.4 W. K. C. Guthrie

I would like to thank J. Brunschwig, M. Conche, and L.-A. Dorion for their helpful comments on
an earlier version of this paper.
1
I do not want to suggest that there is no development in Aristotle’s thought. I am thus not
assuming a unitarian standpoint. However, as will become clear, the statements I quote, even if they
belong to works from different periods of time, are consonant in many respects.
2
As R. Weil writes, “The most diverse and contradictory judgments have been held on Aristotle as
historian, and this since Antiquity” (Aristote et l’histoire, Essai sur la “Politique” [Paris: Klincksieck, 1960],
87, my translation).
3
H. Cherniss, Aristotle’s Criticism of Presocratic Philosophy (Baltimore: The Johns Hopkins Press,
1935; New York: Octagon Books, 1964).
4
J. Burnet among others holds the following opinion: “He is convinced that his own philosophy
accomplishes what all previous philosophers had aimed at, and their systems are therefore regarded
as ‘lisping’ attempts to formulate it” (Early Greek Philosophy, 4th ed. [New York: The Meridian Library,
1959], 31). Cherniss’s analysis is taken up again by J. B. McDiarmid, “Theophrastus on the Presocratic
Causes,” Harvard Studies in Classical Philology 61 (1953). To answer Cherniss’s critique, S. Mansion
distinguished between the faithful testimony and the use of predecessors’ doctrines for his own philo-
sophical purposes. We thus have to differentiate between criticism and historical account (“Le rôle de
l’exposé et de la critique des philosophies antérieures chez Aristote,” in Aristote et les problèmes de méthode
[Louvain-Paris: Publications Universitaires/B. Nauwelaerts, 1961]).

* Catherine Collobert is Associate Professor of Philosophy at the University of Ottawa.

Journal of the History of Philosophy, vol. 40, no. 3 (2002) 281–95

[281]
282 JOURNAL OF THE HISTORY OF PHILOSOPHY 40:3 JULY 2002
5
has replied—renewing Jaeger’s arguments —to Cherniss’s thesis, by arguing that
Metaphysics, books 1 and 2 show that Aristotle respected the first thinkers and had
“a stronger historical sense than most original philosophers possess.”6 More re-
cently, H.-G. Gadamer has stated that Aristotle and Plato as well provide “the sole
philosophical access to an interpretation of the Presocratics.”7
To answer the question if Aristotle is a historian of philosophy requires, as a
preliminary, a definition of ‘history of philosophy.’ Now, neither Cherniss nor
Guthrie raises the question of the nature of this history—their discussion dealing
mainly with the reliability of Aristotle’s testimony. Cherniss refuses to bestow on
Aristotle the status of historian because he constantly misinterpreted, even be-
trayed, the Presocratics by a willful lack of exactness. The betrayal is a consequence
of a freedom taken by Aristotle toward the texts, which is a corollary of a philo-
sophical but not a historical goal. On the contrary, if one supposes with Guthrie
that Aristotle is a historian of philosophy, one has nevertheless to admit that he is
one in a very special sense. Indeed he does not respect any of the rules accepted
nowadays. It has been a long time, as J. Brunschwig points out without any irony,
since “Aristotle summarized Empedocles in eight lines, ending his summary by
adding that ‘one can realize that in reading his poem.’”8
The discussion concerning Aristotle’s historical method results from his sys-
tematic review of the forerunners’ doctrines, of which Aristotle is not really the
father. Indeed, as J. Mansfeld reminds us, “the rudimentary beginnings of the
historiography of Greek philosophy may be dated to the period of the Sophists.”9
In order to understand the nature of Aristotle’s discussion of the Presocratics,
one has first to understand the methodological role of his review of the Presocratics’
doctrines. In other words, one must determine whether this review is to be con-
sidered as an act of history of philosophy, consisting in a report and an explana-
tion, i.e., an interpretation of these doctrines, or as being an opportunity to phi-
losophize. Many points are at stake. First, this issue is interesting for historians of
Presocratic philosophy (who are compelled to question the authenticity of the
fragments quoted by Aristotle), and second, for those who think about methods
used in the history of philosophy, and finally, for those who raise questions about
the relationship between philosophy and its history.
In order to clarify the notion of history of philosophy, one could borrow a
definition from M. Gueroult:
It is a fact that a historical succession of doctrines exists. If one considers it from a strictly
historical point of view, this succession has to appear as a succession of events that are

5
See W. Jaeger, Aristotle, Fundamentals of the History of his Development, trans. R. Robinson, 2nd ed.
(Oxford: Clarendon Press, 1948), 324 ff., mainly 336.
6
W. K. C. Guthrie, “Aristotle as Historian,” Journal of Hellenic Studies 77 (1957): 35–41; Studies in
Presocratic Philosophy, D. Furley and R. E. Allen, eds. (London-New York: Routledge and Kegan Paul,
Humanities Press, 1970).
7
H.-G. Gadamer, The Beginning of Philosophy, trans. R. Coltman (New York: The Continuum Pub-
lishing Company, 1998), 10.
8
J. Brunschwig, “Faire de l’histoire de la philosophie aujourd’hui,” Bulletin de la soc. franç. de
philosophie 4 (1976); Nos grecs et leurs modernes, B. Cassin, ed. (Paris: Seuil, 1992), 88.
9
J. Mansfeld, “Aristotle, Plato, and the Preplatonic Doxography and Chronography,” in
Storiographia e dossographia nella filosophia antica (Torino: Tirrenia Stampatori, 1986); Studies in the His-
toriography of Greek Philosophy (Assen/Maastricht: Van Gorcum, 1990), 27.
 A R I S T O T L E ’ S R E V I E W O F T H E P R E S O C R AT I C S 283
subject to the laws of a historical determinism, explainable by the play of contingent and
individual causes, where the past is evoked only as an explanatory factor for the present.
The succession of doctrines will thus constitute historical data.10

Gueroult deduces elsewhere11 the existence de facto of history of philosophy from

the existence of studies and researches on past philosophies. History of philoso-
phy implies first that philosophy has a past to examine. Previous doctrines are set
out again and can be an object for a philosophical investigation. The object of
history of philosophy is thus past doctrines, always valuable as material, to borrow
a phrase from Gueroult, for “a possible polemic”12—history and polemic being
inseparable in the philosophical field.
From that perspective, the reasons why, according to Gueroult, Aristotle is “the
founder of ancient philosophical historiography”13 are exactly those that led
Cherniss to refuse him this very status.14 Gueroult claims that Aristotle “interprets
the whole past of philosophy in terms of both his system of causes and the passage
from potentiality to actuality. He establishes the value of all previous doctrines
and justifies their opposition as conflicts of partial truths of which his doctrine
ensures the synthesis.”15 History of philosophy, as practiced by Aristotle, entails a
progress from a partial to a synthesized truth, which is the fulfillment of these
doctrines.16 Aristotle would thus have written a history of philosophy, which
amounts to a critical review based upon an idea of progress as a sign of a historical
realization.
Progress is inseparable from a process according to which one event results
from another. A previous event “prepares and heralds the following one.”17 With-
out going deeply into an analysis, which would lead us too far afield, we have
nevertheless to be more specific about this issue. The possibility of progress is
inseparable from the idea of finality. Finality (tevlo~) is interdependent with per-
fectibility, which can be defined as the possibility of change from a worse state to
a better one. The perfectible (teleiwtikovn) is what can become perfect (tevleio~),
that is to say terminated, fulfilled. The best is that “which is for the sake of a future
situation.”18
10
M. Gueroult, Dianoématique, Philosophie de l’histoire de la philosophie, Livre II (Paris: Aubier-
Montaigne, 1979), 34, my translation.
11
“Therefore, to say that history of philosophy actually exists means precisely that for a long time
studies and investigations of past philosophies have existed” (“Le problème de la légitimité de l’histoire
de la philosophie,” in La philosophie de l’histoire de la philosophie [Paris-Rome: Vrin, 1956], 47, my trans-
lation).
12
Gueroult, ibid., 49, my translation.
13
Gueroult, ibid., 53, my translation.
14
That is not surprising, since, according to Gueroult, the history of philosophy is par excellence
philosophy.
15
Gueroult, op. cit., my translation.
16
This judgment is based upon that of Cicero: “And so Aristotle in upbraiding the philosophers
of old for thinking, according to him, that thanks to their genius philosophy had reached perfection,
says that they have been guilty of extreme folly or boastfulness; all the same he adds that he saw that,
as a consequence of the great advance made in few years, philosophy would be absolutely complete”
(Tusculanae Disputationes, III, xxviii, 69, trans. J. E. King [London-Cambridge: Harvard University Press,
1950]).
17
Gueroult, “Les postulats de la philosophie de l’histoire,” Revue de Métaphysique et de Morale 91
(1986): 437, my translation.
18
D. M. Balme, Aristotle’s Use of the Teleological Explanation. Inaugural Lecture at Queen Mary
College (London, 1965), 4.
284 JOURNAL OF THE HISTORY OF PHILOSOPHY 40:3 JULY 2002
The teleological process is described thus by Aristotle: first, there is an initial
moment followed by a moment of development or growth that continues until
the moment of akme, which is the end and the goal of the process (Phys. 194a30).
Birth is the crucial moment without which nothing can develop or be accom-
plished. Every art, whatever it is, is the fruit of an initial discovery. However, begin-
nings are always imperfect, although fundamental, since discovery is an inaugural
gesture:
It is true that if there had been no Timotheus we should have been without much of our lyric
poetry; but if there had been no Phrynis there would have been no Timotheus. The same
holds good of those who have expressed views about the truth; for from the better thinkers
we have inherited certain opinions while the others have been responsible (ai[tioi) for the
appearance of the better thinkers. (Met. 993b11–19, trans. W. D. Ross)19

What comes before is cause for what comes afterward. “For Aristotle,” W.
Kullman writes, “the previous stages as prior are indispensable requirements for
the realization of an end, this end being a sufficient reason for the preliminary
stages.”20 Thus, if Timotheus was able to develop melody, it is thanks to Phrynis,
his master from whom he learned musical art.21 The first condition of progress in
art is nature. For instance, iambic verse was discovered by nature which “found
the appropriate verse” (aujth; hJ fuvs i~ to; oijkei'on mevtron eu|ren) (Poet. 1449a25).
Besides, if philosophers have been led to think of other causes than only the ma-
terial cause, it is because the “thing itself has shown them the way and forced
them to seek” (Met. 984a21). Moreover, nature rules over progress, insofar as the
source of every human activity is human nature. This is why Aristotle can think of
the city-state as a living body, as an organism (Pol. 1302b35)22 and this also ex-
plains Aristotle’s sketch of the progressive evolution of poetical art, from its origin
to its telos, in Poetics, chapter 4.23 “It seems that Poetics was born because of two
causes which are natural” (1448b4). One is a tendency to imitation, the other to
pleasure given by this imitation. The former cause is here understood as the point
of departure from which an evolution is possible, i.e., the effective cause; the
latter as the end (goal and term) of an evolution, i.e., the final cause. Aristotle
endeavors, as C. Thomas Powell states, “to discover the archaioi, or origins, of a
phenomenon such that its present form can be seen as arising from these origins;
thus is the telos of a phenomenon discerned.”24
Incidentally, an interesting parallel can be established between the progress of
art and that of philosophy as it is presented by Aristotle in the first book of Meta-
physics. Two causes are equally appealed to: universal desire of knowledge, and
wonder, which is concomitant with the former. The general movement of philoso-
19
See Guthrie, “Aristotle as a Historian,” 243.
20
“Concepts of the Final Cause,” in Aristotle on Nature and Living Things, A. Gotthelf, ed. (Pitts-
burgh-Bristol: Mathesis Publications-Bristol Classical Press, 1985), 170.
21
This is why “the most ancient is the most venerable” (Met. 983b32). A text from the Sophistical
Refutations offers a similar analysis (see 183b18–24).
22
J. Fergusson writes, “There is in fact a clear biological basis to Aristotle’s political thought”
(“Teleology in Aristotle’s Politics,” in Aristotle on Nature and Living Things, 261).
23
As Balme states, the process is explained by the result, “the end-product” and the contrary
(“Aristotle’s use,” 11).
24
C. Th. Powell, “Why Aristotle has no history of philosophy,” History of Philosophy Quarterly 43
(1987): 353.
 A R I S T O T L E ’ S R E V I E W O F T H E P R E S O C R AT I C S 285
phy as introduced in Metaphysics, book 1, like all activities, is goal-directed. Since
philosophy is defined as the “search for first principles and causes,” its develop-
ment is merged with that of discoveries of causes, in which every philosopher has
participated. However, truth being non-monolithic, some contributed to it more
than others. Empedocles’ philosophy, because it is capable of explaining more
points in defining more causes, is superior to Anaxagoras’s philosophy (see Met.
984a12–13),25 which is indeed prior to the former. The superiority of a doctrine
or an art rests upon its proximity to the end: it is more fulfilled. Success is mea-
sured in relation to the end.
Though philosophy came from myth, it is superior to the latter. Philosophy is
not a creation ex nihilo.26 In fact, it is rare that there is “nothing previous” (Soph.
Ref. 184a1–b1). Poets such as Homer or Hesiod have also philosophized to some
extent, by searching for the origin or principle of everything. Homer’s inferiority
to Thales is located by Aristotle not in the type of causality he employs but in his
failure to name the principle. They both gave water as principle,27 whatever it is
supposed to be: natural stuff or divinity (Met. 982b19). As J. Mansfeld writes, “as
soon as the science of Miletus is called ‘philosophy,’ the discerning of ‘philosophi-
cal’ elements in myth proves feasible […] Historiography when practiced in this
way, of course remains a sort of inverted teleological account.”28 Even if Thales is
an arbitrary beginning,29 this beginning reveals the Aristotelian view of history of
philosophy as a teleological process, from its birth to its akme. This is why Aristotle,
in introducing the theory of causality, shows that it is not a major innovation,
since it is present in a rough state in his forerunners’ philosophy. As Cherniss
points out, what is prominent is that “the tendency to develop the ‘necessary ante-
cedents’ or ‘necessary consequences’ of an early statement so as to reconstruct
the original scope of the doctrine discussed and its intended meaning is one of
Aristotle’s favorite methods.”30
In grasping the common roots of myth and philosophy, Aristotle emphasizes
the sketchiness of the beginnings. Unlike Thales, Homer did not clearly name
the first cause. To clarity and distinctness are opposed obscurity and stammering,
which belong to the beginnings. Democritus expressed his thought more subtly

25
For this controversial issue, see D. O’Brien, “The relation of Anaxagoras and Empedocles,”
Journal of Hellenic Studies 88 (1968).
26
Philosophy comes from myth, but myth comes also from philosophy, according to Mansfeld,
who bases his statement on the cyclical view of time affirmed in Met. XII: “[W]hat we have in Hesiod
and others is what survived of the ideas of the Aristotle of the lost civilization that preceded our own”
(“Myth, Science, Philosophy: A Question of Origins,” in Studies in the Historiography of Greek Philosophy
[Assen/Maastricht: Van Gorcum, 1990], 53).
27
This text can be brought closer to Theaet. 152e and Crat. 402b–c. Regarding this issue, see B.
Snell, “Der Nachrichten über die Lehren des Thales und die Anfänge der griechischen Philosophie
und Literaturgeschichte,” in Gesammelte Schriften (Göttingen: Vandenhoeck u. Ruprecht, 1966) and
Mansfeld’s discussion “Cratylus 402a–c: Plato or Hippias,” in Studies in the Historiography of Greek Philoso-
phy (Assen/Maastricht: Van Gorcum, 1990).
28
Mansfeld, “Myth, Science, Philosophy,” 58.
29
A need for beginnings is also expressed when Aristotle traces back the development of rheto-
ric. According to Cicero, “Aristotle brought together in a single compilation the ancient writers on the
art of rhetoric, going right back to their founder and inventor, Tisias” (De Inventione II ii 6–7, trans. J.
Barnes and G. Lawrence).
30
H. Cherniss, Aristotle’s Criticism, 355.
286 JOURNAL OF THE HISTORY OF PHILOSOPHY 40:3 JULY 2002
(glafurwtevrw~) than his forerunner Anaxagoras (On the Soul, 405a8). “The crite-
ria notion of clarity,” Mansfeld states rightly, “is used in two ways. First, it serves to
distinguish philosophy from poetry, or myth, or unscientific notions generally.
Second, it serves to distinguish accomplished philosophy from primitive (or rela-
tively primitive) philosophy.”31 A rough sketch is the price to pay for the begin-
nings; Aristotle expresses this by using different expressions: vaguely (ejpipolaivw~:
Met. 987a23), more obscurely (morucwvteron: 987a11). However, Anaxagoras, in
stating that “the separation will never end, has spoken without knowing (oujk
eijdovtw~), but he has spoken rightly (ojrqw`~)” (Phys. 188a6).
The mere statement of the truth is not sufficient, but has to be motivated, i.e.,
justified. Justification is the guarantee of a clear and distinct statement. Of
Anaxagoras, Aristotle adds that “If one were to follow him up, piecing together
what he means, he would perhaps be seen to be somewhat modern in his view”
(Met. 989b5, trans. W. D. Ross). The novelty is here truly related to the truth,
which can only appear with distinctness (989b8). The teleological process is an
evolution toward clarity (safw̃~) and distinctness. Obscurity or lack of clarity,
strictly speaking, is not to be understood as a real reproach that Aristotle directs
at his forerunners, since, as Cherniss states, he “believed with apparent sincerity
that the fundamentals of his system were vaguely and with confusion recognized
by previous thinkers.”32 His critical stance is justified by his teleological concep-
tion of philosophy. Trial and error that are associated with the beginnings, in this
regard, are the result of perfectibility.
The teleological process is thus one of an increasing clarification (a[dhlon ver-
sus ajpofhvnasqai) that occurs in the course of time. In tracing the development of
philosophy, Aristotle shows that it becomes more precise and refined over time.
Thus, from the first cause discovered by Thales, material cause, passing through
the moving cause discovered by Anaxagoras and Empedocles, to the formal and
final causes, the search for causes ends with Aristotle, who establishes the com-
plete theory. The progress of philosophy merges with that of truth, since the end
of philosophy is truth (qewrhtikh̃~ me;n ga;r tevlo~ ajlhvqeia: Met. 993b22).33 How-
ever, “truth is, as it seems, like the proverbial door: who could miss it? As such, this
study is easy. But the fact that we are unable to possess a particular truth although
we possess the whole shows how difficult the undertaking is” (Met. 993b4). What
is offered at first is the whole. A discourse that deals with the whole is less subject
to error. The larger the target is, the easier it is to hit: generality is the enemy of
precision. The process of clarification, being a process of precision and distinct-
ness, requires that we deal with specific and narrow issues. In this perspective, the
general movement toward truth is to be understood as a process of specialization
that leads Aristotle to distinguish between different topoi to address particular
problems.34

31
Mansfeld, “Myth, Science, Philosophy,” 43.
32
Ibid., 42.
33
See also Nic. Eth. 1139a20.
34
Aristotle writes, “To put the matter briefly, there are three classes of propositions and prob-
lems. Some are ethical, some physical and some logical propositions” (Top. 105b19–21, trans. E. S.
Forster).
 A R I S T O T L E ’ S R E V I E W O F T H E P R E S O C R AT I C S 287
The Aristotelian picture of the development of philosophy rests upon a teleo-
logical analysis, which is similar to that underlying the development of poetical
art. Aristotle appears truly as standing on the shoulders of his predecessors (see
Met. 993b12–15). He placed himself in their wake, knowing to some extent that
he is only an instrument of the teleological process of philosophy.
History of previous philosophers is decisive. It is a legacy received and reported
by Aristotle, in the form of monographs—such as those devoted to Pythagoreans
(now lost)—that must have constituted a documentary collection and a basis for
reflection.35 This is why, while he sketches their philosophy, he also points out
that he had “presented it elsewhere, more accurately, in another work” (Met.
986a13). History as a collection of previous doctrines is thus one way to reach
truths by distinguishing between error and truth within these doctrines. History
allows for corrections—a necessary act to achieve the end, which is truth. The
practice of history is thus related to the teleological characteristic of every human
production.
The union of partial truths is precisely the deposit named history that Aristotle
claims to be useful (ti prou[rgou: Met. 983b5) to his research. Because philosophy
is conceived as developing teleologically, its history cannot be a mere doxography.36
“A mere doxography” is a selected and classified collection of doctrines under-
taken for the sake of conservation or of an external purpose, as M. Frede claims.37
Frede propounds an idea according to which doxography as an actualization of
ancient thought is a means for contemporary discussions. The idea can be ap-
plied to some extent to Aristotle’s history of philosophy, except that these discus-
sions are to be understood within a teleological perspective.
Aristotle is interested in the theses of his forerunners insofar as he saw them,
as Frede states, as parts of the development of philosophy. With this in mind,
Mansfeld writes, “I know that there is no doxography in the proper sense of the
word in Aristotle or Plato, or even in Theophrastus, because the doxai at issue are
presented from a systematical point of view in order to further discussion of prob-
lems of a systematical nature.”38 A history of philosophy supposes, even implicitly,
a philosophy of history, which is, for Aristotle, a teleological conception of
philosophy’s development. Therefore, the collection is undertaken not for its own
sake but for the sake of the progress of philosophy. As Cherniss states, “[E]ach of
his predecessors was aiming at the goal represented by his own system.”39 The
teleological conception is thus the philosophical basis on which to select and or-

35
As P. Moraux claims, there is a possibility that Aristotle wrote the monographs that are mainly
known from the list of Diogenes Laertius (Lives of the Philosophers V, 22–7). For further details, see
Moraux, Les listes anciennes des ouvrages d’Aristote (Louvain: Editions universitaires de Louvain, 1951).
36
As Mansfeld and D. T. Runia recall, “The term ‘doxography’ is a neologism with a shorter
history than is generally recognized. It was invented by Hermann Diels in the late nineteenth century
as the designation of what he believed to be a specific genre of ancient writings” (Aëtiana, The Sources,
vol. 1 [Leiden-New York-Köln: E. J. Brill, 1997], xiii).
37
“Doxographie, historiographie philosophique et historiographie historique de la philosophie,”
Revue de Métaphysique et de Morale 3 (1992).
38
“Aristotle, Plato, and the Preplatonic Doxography and Chronography,” in Storiografia e
Dossographia nella Filosofia Antica (Torino, 1986), 24.
39
Cherniss, Aristotle’s Criticism, 356.
40
See Frede, op. cit., 316.
288 JOURNAL OF THE HISTORY OF PHILOSOPHY 40:3 JULY 2002
40
ganize the doctrines of the Aristotelian forerunners. Nonetheless, all this does
not imply that Aristotle transmitted false information about his predecessors. In
order for the teleological process to make sense, it must have a certain degree of
accuracy—even if this accuracy is far from the standards of modern philology,
with its control procedures and ambition of objectivity. However, philology is no
longer the rigorous science that Diels conceived.41 Betrayal begins, as A. Laks
points out, “when the distance between use and reconstitution is abolished.”42 In
what way does Aristotle use and reconstitute the doctrines of his forerunners?
The way of collecting the previous doctrines is rarely chronological. The his-
tory of doctrines, as written by Aristotle, often appears less subject to a chrono-
logical than to a rational order. Does it mean that history intervenes “to fulfill the
a priori framework prepared by the philosophical reason,”43 as P. Aubenque claims,
or that this kind of review has no historical value, as Cherniss asserts, because “the
doctrines of his predecessors were materials to be remolded for his own purpose”?44
Aubenque asserts that “If we consider history as going from the past to the future,
we see only a blind accumulation of materials; if on the contrary we look from the
present to the past, these materials acquire their meaning as materials for a con-
struction.”45 History is thus the instrument by means of which the teleological
character of philosophy can be brought to its full expression.
The teleological dimension of philosophy has two consequences for research.
It gives opportunities first of avoiding the search for what has already been found,
and second of relying on relevant, i.e., true statements. It is thus not necessary to
discuss again what has already been asserted or what is obvious. Aristotle thinks of
numerous statements acknowledged by everyone and valid for everyone, i.e., uni-
versal and taken for granted. “This is why one must use what has been discovered
adequately and searches only for what has been left aside” (Pol. 1329b34). Previ-
ous discoveries help present reflections in permitting one not to linger on prob-
lems already solved, but to investigate further. The lack of agreement among phi-
losophers on many issues does not exclude that some problems are solved.
Moreover, it means that inquiry should be pursued. When Aristotle claims that
his philosophy has its roots in Empedocles’ or Anaxagoras’s philosophies, he ac-
knowledges a true point within their doctrines. An example is when he asserts,
regarding the name of ‘ether,’ that “it has been passed down to the present time
by the ancients, who thought of it in the same way as we do” (Cael. 270b18, trans.
W. K. C. Guthrie).
The notion of past achievements is inseparable from the teleological process
of human productions. In fact, progress has a sense only if it is possible to con-
sider results of its earlier stages as taken for granted. The state of knowledge is at
the beginning temporary: it is modified as research goes on. Humans learn from
experience, observation, and recording of facts (i.e., from history), as Xenophanes
has already stated (DK 21b18). This is why when Aristotle deals with the nature of

41
See Aëtiana, ch. 2.
42
“Histoire critique et doxographie,” Les Etudes philosophiques 4 (1999): 474.
43
P. Aubenque, Le problème de l’être chez Aristote, 4th ed. (Paris: Puf, 1977), 90.
44
Aristotle’s Criticism, 347.
45
Aubenque, op. cit., 91.
 A R I S T O T L E ’ S R E V I E W O F T H E P R E S O C R AT I C S 289
soul, he indicates that “[F]or our study of soul it is necessary, while formulating
the problems of which in our further advance we are to find the solutions, to call
into council the views of those of our predecessors who have declared any opin-
ion on this subject, in order that we may profit by whatever is sound in their
suggestions and avoid their errors” (On the Soul 403b20–23, trans. J. A. Smith).
There are two ways of practicing history of philosophy that ensure a speedier
development of philosophy in order to reach its akme sooner: first, to put into
perspective conflicting points of view in order to try to find a solution; second, to
review the general movement of a philosophical question ending with its final
solution, i.e., to confirm a theory. The former, displaying given doctrines regard-
ing a specific question as contemporaries, is a synchronic ordering. The latter,
unfolding doctrines over time, is a diachronic ordering. The first mode empha-
sizes conflicts, while the second indicates their solution. Aristotle uses both of
them inasmuch as they both participate in the unfolding of the final solution,
which is, as Mansfeld recalls, the solution of Aristotle: “He must have believed
that his own almost finished system approximated the final goal very closely.”46
The diachronic ordering is predominant in the first book of Metaphysics. It
displays over time the teleological process of philosophy, as we have seen, as a
progressive clarification. If the synchronic presentation appears to some extent as
chaotic, or at least if there is no possible hierarchy of doctrines, as Aubenque
notes, it is just because this presentation is concerned with a specific point of a
doctrine, i.e., a precise problem. In this regard, Aristotle acknowledges that a
philosopher can be closer to the truth than his successors regarding a particular
issue. For example, Empedocles’ thought is said to be superior to Anaxagoras’s
regarding his theory of the four elements or that of the alternation of movement
and rest, but it is nevertheless inferior to it regarding the possibility of explaining
alteration.47 Anaxagoras’s solution of the problem of alteration is closer to the
truth than that of Empedocles, because the latter simply cannot explain it. The
truth, while it is considered according to a particular point of view, is discovered
through a process, which is not completely linear. It is however possible to bring
out main themes within the evolution of philosophy, contrary to what Aubenque
thinks, when he writes that “there is no progress from one solution to another but
a sort of timeless system of points of view, where any idea of genealogy is absent.”48
When Aristotle displays the different conceptions of infinity held by his prede-
cessors, he first introduces Plato and the Pythagoreans, and not Anaximander,
who is quoted en passant (Phys. 203b13). Aristotle divides the doctrines into two
groups: on the one hand Plato and the Pythagoreans, who thought of the infinite
in the same way, i.e., in Aristotelian language, as a substance; on the other hand
the physiologoi who thought of it rightly (eujlovgw~) as a principle (Phys. 203b4).
The latter group is more accurate than the former. Aristotle analyzes the number

46
Mansfeld, “Myth, Science, Philosophy,” 48. Cicero, paraphrasing Aristotle, wrote: “ . . . since in
a few years a great advance has been made, philosophy will in a short time be brought to completion ”
(F 53 R Tusculanae disputationes, iii xxviii 69, trans. J. Barnes and G. Lawrence).
47
See O’Brien, op. cit.
48
Aubenque, Le problème de l’être, 90.
290 JOURNAL OF THE HISTORY OF PHILOSOPHY 40:3 JULY 2002
of principles, in the first book of Physics, with the same method of division
(diaivresi~): on the one side, the monists; on the other side, the pluralists.49 The
parallel evolution of these two currents of thought indicates indeed that there is,
as Aubenque writes, no “homogeneous time where the final moment would be
necessarily favored over previous ones because it would contain all of them.”50
However, this does not quash the idea of progress. Lines of thought coexist in
a specific moment of the process, but only one line will be drawn later. This is
what happens regarding the number of causes, since the pluralist solution is cor-
rect according to Aristotle.51 The collecting of doctrines is a reconstruction whose
goal is to sort out truth from falsehood. Such collecting is performed synchronically,
for the chronological presentation of opinions regarding, for example, the soul
would not be only useless but also impossible in virtue of the nature of soul: what
the soul is is not easy to grasp (On the Soul 402a10).
According to a diachronic logic, Aristotle has to organize elements of truth
that he encounters in his predecessors’ writings, since he notes a dispersal that is
the result of the fact that everyone considers the issue not in its entirety but par-
tially (see On Gen. and Corr. 323b15). What is “first obvious and clear, are muddled
things (sugkec umevna)” (Phys. 184a22). The goal of the synchronic presentation
is to set out the web in order to see how to disentangle it.
For those who want to get clear of difficulties (eujporh`sai), it is advantageous to state the
difficulties well (diaporh`sai), for a clear view (eujporiva) implies the solution of the previ-
ous difficulties, and it is not possible to untie a knot which one does not know. But the
difficulty of our thinking points to a knot in the object; for in so far as our thought is in
difficulties, it is in like case with those who are tied up; for in either case it is impossible to
go forward. (Met. 995a27–33, trans. W. D. Ross, slightly modified)

Aporia is thus the sign of a knot that one can untie only by going deeply into
different points of view regarding an issue. However, this supposes that the data
are organized in a conflicting way. To be able to untie a knot, one must see it as a
knot to begin with.
The collecting of doctrines thus consists in regarding a specific issue as orga-
nized according to affinities of thought that exclude a general philosophical filia-
tion. The idea of filiation here entails that a given doctrine comes from a previous
one and supports its main assertions. Therefore, there is an agreement among
these doctrines that can be grouped genealogically. This idea of filiation does not
fit Aristotle’s presentation of his forerunners. In a synchronic presentation, filia-
tion and genealogy are not relevant at all because they deal with a specific prob-
lem. On the question of the soul, Aristotle could not present Plato as the succes-
sor of Pythagoreans, since Plato rejects the theory of the soul-harmony; he can,
on the contrary, associate them regarding the question of the infinite.52 This is
moreover what Aristotle acknowledges himself when he sketches Plato’s biogra-

49
New divisions will be possible from this pair as Mansfeld has shown (“Myth, Science, Philoso-
phy,” 29 ff.).
50
Ibid., 91.
51
See Phys. I 7–8.
52
Thus there is no inconsistency in the method of the Stagirite here, unless one considers that
every doctrinal display must respect a rigorous chronological order that would presuppose in turn a
form of a historical necessity unknown to Aristotle.
 A R I S T O T L E ’ S R E V I E W O F T H E P R E S O C R AT I C S 291
phy: “after the systems we have named [Pythagoreans] came (ejpegevneto) the phi-
losophy of Plato, which in most respects followed these thinkers, but had pecu-
liarities that distinguished it from the philosophy of the Italians” (Met. 987a29–
32, trans. W. D. Ross).
Historia, as the collection of information and assembling of data about a prob-
lem, is the gathering of positions (thesis, antithesis) waiting for a dialectical treat-
ment.53 The synchronic display corresponds to this last function, i.e., putting side
by side contrary doctrines so that they appear to be contemporary. This display
rests on a principle identified in the analytic reading54 proposed by P. Engel,55
namely the principle of atomicity. In fact Aristotle isolates the concepts, theses,
and problems of his forerunners from their context to discuss them. “But first let
us run over the theories (uJpolhvyei~) of others, since to expound one theory is to
raise the difficulties involved in its contrary” (Cael. 279b6–13, trans. W. K. C.
Guthrie). The proofs of a thesis are the problems of the antithesis. This sort of
problem is dialectical, first, because it deals with properties of being, which are
always subject to controversies, and second, because it can be solved. One can
affirm either the thesis or the antithesis since evidence sustains the former or the
latter as well. However, to acknowledge the truth of the thesis means to acknowl-
edge the error of the antithesis.56 One who wants to sustain, for example, that the
cosmos is ungenerated must take into account the arguments of those who main-
tain the contrary and must consider these arguments as problems to be over-
come.57
Here historia encounters dialectic, as Mansfeld claims rightly, “We ought also
to select from written disquisitions and make up descriptions of each class of sub-
ject, putting them in separate lists, for example, about ‘the good,’ beginning with
the essence” (Top. 105b13–16).58 As Brunschwig points out, “one does not cross

53
See Mansfeld, “Physicai doxai et Problemata physica d’Aristote à Aétius (et au-delà),” Revue de
métaphysique et de morale 3 (1992).
54
This connection has already been made by Barnes, who states that the Anglo-American (ana-
lytical) style has been practiced by “Martin Heidegger, by Franz Brentano, by Thomas Aquinas . . . by
Aristotle himself, who in addressing any problem would collect his forerunners’ opinions, put them to
test and extract from them his own Aristotelian philosophy” (“Aristote chez les Anglophones,” Critique
399–400 [1980]: 708, my translation).
55
P. Engel, “Retour aval,” Les Etudes philosophiques 4 (1999): 455.
56
“The defender can hope nothing better than to thwart his partner’s undertaking; he will
correctly defend his thesis p if and only if the limited set of premises proposed and granted by the
questioner do not lead validly to the conclusion not-p” (Brunschwig, “Faire de l’histoire de la philosophie
aujourd’hui,” 241, my translation).
57
J. D. G. Evans writes, “Aristotle’s discussions of these [philosophical] problems move uncom-
promisingly beyond the difficulties which are constituted by the assemblage of inherited and conflict-
ing views. Aristotle says that the views are all thereby shown to be wrong; and he attempts to provide an
account of his own that avoids their defects” (“Dialectic, Contradiction, and Paraconsistency in Aristotle,”
in From Puzzles to Principles?: Essays on Aristotle’s Dialectic, M. Sim, ed. [Lanham, MD: Lexington Books,
1999], 143–5).
58
He adds, “Without doubt, he used such schemes to suit his own dialectical convenience, refin-
ing the classifications and adding to or modifying and rearranging the material” (Mansfeld, “Aristotle,
Plato, and the Preplatonic Doxography and Chronography,” 25). In the same spirit, R. Smith states,
“A compilation of the opinions of others is one of the components of the dialectical method” (“Dia-
lectic and Method in Aristotle,” in From Puzzles to Principles?, 43).
292 JOURNAL OF THE HISTORY OF PHILOSOPHY 40:3 JULY 2002
swords with just anyone”; this is why one has to “admit into the dialectical arena
only opinions reasonably defensible.”59 Opinions that can be put in conflict and
that, consequently, can constitute a dialectical problem are those that have “au-
thorized supporters,” to borrow the phrase from Brunschwig.60
One has to select among opinions and to choose those that deserve a dialecti-
cal treatment, that is, those that can be transformed into a dialectical problem.
Among these opinions, also called “conceptions” or “assumptions” (uJpolhvyei~)
by Aristotle, can be those of philosophers. In submitting doctrinal elements to a
dialectical examination, Aristotle reserves the right to arrange them in a way that
emphasizes their conflicts.61 Opinions that are not “dialectizable” (Top. I, 10) are
first those that are contradicted by experience—an appeal to evidence is thus
enough to refute them without the necessity to argue further;62 and second, those
that involve no conflict. “For, no one in his senses would make a proposition of
what would be held by nobody nor make a proposition of what is obvious to every-
one. There is no difficulty for the latter, and no one would accept the former”
(Top. 104a5–8). Received opinions, i.e., those assented by every one, are taken for
granted. This is why there is no reason to subject them to a dialectical treatment.
Thus, they cannot constitute a dialectical problem. Immediate clarity does not
produce any discussion. This is why a thesis or a dialectical problem expresses a
conflict between different opinions (Top. 104b32–33) that is also between “para-
doxical thoughts” (104b20).
Conflict is the first step or, more precisely, the means toward clarification.
Aristotle holds that everyone has a certain built-in grasp of a little of the truth, and we should
therefore treat the opinions of the many, as well as those of the wise, with respect, in need
of correction and clarification rather than refutation and rejection. Aristotle is not simply
advocating a dialectical method in philosophy but giving us a reason for doing so.63

One has to provoke a conflict: “To be a good investigator a man must be alive to
the objections inherent in the genus of his subject” (Cael. 294b11–13, trans. W. K.
C. Guthrie). It is a question of directing the thought for fear of failure (Met.
995a35). Dialectic, thanks to its “diaporetic” function, offers such a direction.
The dialectic movement goes from embarrassment to clarity, i.e., to solution. From
aporia to euporia, the movement of thought progresses by confrontation. This

59
Brunschwig, “Dialectique et ontologie,” Revue philosophique de la France et de l’étranger 154 (1964):
188, my translation.
60
Brunschwig, “Remarques sur la communication de R. Bolton,” in Biologie, Logique et Métaphysique
chez Aristote, D. Devereux and P. Pellegrin, eds. (Paris: CNRS, 1990), 247.
61
Gueroult writes, “Thus, the elaboration of the history of doctrines by dialectic is conceived as
the only method apt to make progress in science” (“Logique, argumentation et histoire de la philosophie
chez Aristote,” in La théorie de l’argumentation, perspectives et applications [Louvain-Paris: Nauwelaerts,
1967], 435, my translation).
62
Phenomena are used as evidence and examples (marturivoi~ kai; paradeivgmasi: Eud. Eth.
1216b26–27, fanero;n: Phys. 193a4), the rejection of which is considered a violence to the truth (Met.
1082b4). This rejection produces fictions (plasmatw`de~: 1082b3). Thus, the dialectical examination
is also not useful here (Top. 105b1). See for this issue G. E. L. Owen, “TIQENAI TA FAINOMENA,” in
Aristote et les problèmes de méthode (Louvain-Paris: Publications Universitaires/B. Nauwelaerts, 1961) and
M. Nussbaum, “Saving Aristotle’s Appearances,” in Language and Logos, M. Schofield and M. Nussbaum,
eds. (New York: Cambridge University Press, 1982).
63
R. Smith, “Dialectic and Method in Aristotle,” 42.
 A R I S T O T L E ’ S R E V I E W O F T H E P R E S O C R AT I C S 293
movement also goes from conjecture to truth: “a dialectical problem is an investi-
gation leading to choice and avoidance or to truth and knowledge” (Top. 104b1–
2, trans. E. S. Forster). Because it takes a progressive effort to reach the truth, for a
while we have to be satisfied with conjecture. An opinion is conjectural as long as
it is not subjected to a dialectical examination by which it will be transformed into
certainty.
This movement from conjecture to certainty requires arbitration: “at the same
time, things which will be said to have more certainty(pista;) for those who will
hear the trial of opposite theses; for one needs to be arbitrator (diaithta;~) and
not opponents (ajntidivkou~) for those who want to judge (krivnei) the truth in a
satisfactory way . . .” (Cael. 279b8–13). The notion of arbitration indicates that it
is a matter of decision, which has to be done according to precise rules64 in the
examination of different arguments. One of the senses of diaitw` is “to judge as
arbitrator” a quarrel or a conflict. To arbitrate a conflict is not to reconcile the two
parts, but to be capable of judging, i.e., to separate, distinguish, or decide what is
true. Dialectic is thus this judicial activity within which truth and error are dis-
criminated (katoyovmeqa tajlhqev~ te kai; to; yeu`do~: Top. 101a37, and On the Soul
403b23). It is a matter of judging the truth.65 This judicial activity corresponds to
another principle of the analytic reading: the principle of argumentation. It re-
quires us to discuss our predecessors’ arguments, as if one enters in a dialogue
with them, and to correct them if possible.66 “For such a process one must possess
a certain natural ability, and real ability consists in being able correctly to choose
the true and avoid the false” (Top. 163b14–15, trans. E. S. Forster).
Another trial was conducted many centuries later on the philosophical scene
that was of a different, more conciliatory nature. The relevance of the joust and
its stake, truth, disappeared together. “The dialectical arena” in which opponents
would fight was judged to be merely useless as it was only a matter of making these
opponents recognize the “uselessness of their quarrel.”67 The aporetic position is,
according to Kant, an illusion that, as such, cannot be solved in the dialectical
way. Opposition as well as conflict is devoid of virtue. A contrario, Aristotle ac-
knowledges the “diaporetic” virtue of dialectic within which confrontation ends
and truth appears. Yet, if truth can be constrained, it can also compel. Truth com-
pels, to some extent, the judgment to be pronounced and thus assures its fairness.
Truth is constraining (Met. 984b10); this is why one reaches it in most cases (Rhet.
I.1.11). Things, according to Aristotle, are given; it just remains to us to discover
them, since the inference from being to knowing is right. In this regard, it is a
matter of improving imperfect discoveries, i.e., of perfecting them.
64
Aristotle defines these rules in Top. 159a40.
65
Moraux writes, “ . . . when able to see the difficulties raised by conflicting points of view, we can
more easily distinguish truth from falsehood; if we see clearly the consequences that follow from the
opposed theses, we have only to choose the one we think to be true” (“La joute dialectique d’après le
VIIIe livre des Topiques,” in Aristotle on Dialectic, The Topics, G. E. L. Owen, ed. [Oxford: The Clarendon
Press, 1968], 290). In the same perspective, R. Bolton states: “dialectic, in certain forms, can draw on
what may reasonably be claimed to be true” (“The Epistemological Basis of Aristotelian Dialectic,” in
Biologie, Logique et Métaphysique chez Aristote, D. Devereux and P. Pellegrin, eds. (Paris: CNRS, 1990).
66
Engel, “Retour aval,” 455.
67
The Critique of Pure Reason, trans. F. M. Müller (New York: Macmillan, 1966), 306.
68
As Bolton recalls, “in Topics I.2 Aristotle distinguishes the gymnastic use of dialectic from two
other distinct uses: the universal use, described at the beginning of the Rhetoric, in ordinary discussion
294 JOURNAL OF THE HISTORY OF PHILOSOPHY 40:3 JULY 2002
According to Aristotle, his predecessors dealt with fundamental questions, which
deserve to be renewed and specified, i.e., corrected68 in order to reach the truth
(metabibavzonte~ o{ ti a]n mh; kalw``~ faivnwntai levgein hJmi`n: Top. 101a34). Since the
truth cannot be discovered in its entirety at first, true points need to be enlight-
ened, to be rendered more explicit. In this respect, a restatement is necessary
because of the lack of distinctness and differentiation (ajpokekrimevnon). This cor-
responds to another principle of the analytic reading: the principle of translat-
ability according to which one should translate into one’s own words the concepts
of one’s predecessors.69
The possibility of transforming a falsehood into a truth is based on a definition
of error as an imperfect utterance. This imperfection has the following two as-
pects. First, error leads, independently of the will of its author, to a truth: it im-
plies, in that case, an implicit thought, which will be true. D. O’Brien states that,
“Aristotle takes the view that Anaxagoras’s precosmic rest, although false in itself,
at least shows recognition of the principle that movement and order are insepa-
rable, so that if there is no movement there could be no ordered universe.”70
Second, error is considered as an imprecise and obscure statement of a truth; this
leads Aristotle to say that one has to follow Empedocles according to his thought,
but not according to his obscure or indistinct discourses (yellivzetai) (Met. 985a6).
We have already seen three analytic principles that may be applied to the Aris-
totelian reading of the Presocratics. We could continue by quoting the main dif-
ference laid down by Engel between the analytical and the historical approaches.
The former, as Engel points out, favors upstream ascent of a contemporary issue
to its antecedents, whereas the latter favors the study of the emergence of prob-
lems from downstream to upstream.71 To the latter corresponds what we call the
diachronic display of doctrines (mainly, books 1 and 2 of Metaphysics); to the former,
synchronic display (which mainly makes up the first books of the Aristotelian
treatises). Both of these displays are guided by the same end: the truth. However,
we have to add that Aristotle’s method is, if not ruled by, at least interdependent
with his teleological conception of philosophy, of which dialectic constitutes one
of the most important techniques.
The historia of the Presocratics is ruled by the idea of culmination, as Aubenque
points out: “What was only isolated stammering becomes a contribution to a philo-
sophical thought moving toward its accomplishment.”72 Aristotle did not write a
history of philosophy in a modern sense or at least in a “continental” sense when
he transmitted the thoughts of his predecessors. For this reason, one can say with
U. Wilamowitz that “one does not have to blame the historian Aristotle, because

or conversation in order to correct the mistakes of others and submit to the same process ourselves”
(101a27, 30–34; cf. Rhet I.1 1355a24–29) (“The Epistemological Basis of Aristotelian Dialectic,” in
Biologie, Logique et Métaphysique chez Aristote, D. Devereux and P. Pellegrin, eds. [Paris: CNRS, 1990]).
69
Engel, op. cit.
70
O’Brien, “The Relation of Anaxagoras and Empedocles,” 100, explaining On the Heaven 300b16–
301a22.
71
Engel, “Retour aval,” 460.
72
Aubenque, Le problème de l’être, 77.
73
Wilamowitz, Aristoteles und Athen (1893; repr. Zürich: Weidmann, 1966), 308. This judgement
is held again by Gadamer: “Aristotle does not wish to write history any more than Plato does” (The
Beginning of Philosophy, 72).
 A R I S T O T L E ’ S R E V I E W O F T H E P R E S O C R AT I C S 295
73
Aristotle never was nor wanted to be an historian.” We would rather say that
Aristotle could not be a historian, as we now understand this position, since he
probably did not understand the difference between philosophy and history of
philosophy inasmuch as history of philosophy is understood as dialectic. In order
to be an efficient agent of progress in philosophy,74 Aristotle creates, on the one
hand, a method that allows for an easier way toward the end and, on the other
hand, he traces the different stages of the teleological process, showing how the
previous stages have prepared the present one. The method by which the perfect-
ing of philosophy is not any more left to arbitrary chance or to more or less bril-
liant philosophers is the dialectic of which Aristotle proudly claims to be the dis-
coverer (Soph. Ref. 184a2). By this method, philosophy not only knows itself as a
direction toward truth, but it is also henceforth capable of moving in this direc-
tion and escaping possible errors. It is from this perspective that dialectic has the
double function of examination and inquiry75 (Top. 159a33).
The progress of philosophy, in which every philosopher participates more or
less actively or skillfully—whether or not a significant stage in the teleological
process—must be understood from the point of view of the lack of distinction
between philosophy and history of philosophy. Doctrines of the predecessors are
for Aristotle data for diachronic and synchronic examinations, and, as such, are
the object of a philosophical reconstruction that is a means of philosophizing.

74
Evans writes rightly, “It follows that dialectical exploration of conflicting views is no mere orna-
mental preface to real philosophizing, but instead is an ineliminable aspect of rational inquiry and
intellectual advance” (“Dialectic, Contradiction, and Paraconsistency in Aristotle,” 143).
75
By its peirastic feature, as Aubenque points out, dialectic is “an instrument of progress” (“La
dialectique chez Aristote,” in L’Attualità della problematica aristotelica, Atti del convegno franco-italiano
su Aristoteles [Padova: Antenore, 1970], 12). Concerning the main features of peirastic, see Bolton,
“The Epistemological Basis of Aristotelian Dialectic,” in Biologie, Logique et Métaphysique chez Aristote, D.
Devereux and P. Pellegrin, eds. (Paris: CNRS, 1990).