Professional Documents
Culture Documents
Dtoken
Dtoken
JAMES G. LENNOX
University of Pittsburgh
www.cambridge.org
Information on this title: www.cambridge.org/9780521193979
doi: 10.1017/9781139047982
© Cambridge University Press 2021
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2021
A catalogue record for this publication is available from the British Library.
Library of Congress Cataloging-in-Publication Data
names: Lennox, James G., author.
title: Aristotle on inquiry : erotetic frameworks and domain-specific norms /
James G. Lennox, University of Pittsburgh.
description: Cambridge, United Kingdom : New York, NY, USA : Cambridge University
Press, 2020. | Includes bibliographical references and index.
identifiers: lccn 2020034989 (print) | lccn 2020034990 (ebook) | isbn 9780521193979
(hardback) | isbn 9781139047982 (ebook)
subjects: lcsh: Aristotle. | Inquiry (Theory of knowledge)
classification: lcc b485 .L46 2020 (print) | lcc b485 (ebook) | ddc 121/.6–dc23
LC record available at https://lccn.loc.gov/2020034989
LC ebook record available at https://lccn.loc.gov/2020034990
isbn 978-0-521-19397-9 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy of
URLs for external or third-party internet websites referred to in this publication
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.
Introduction 1
vii
Bibliography 293
Index Locorum 307
General Index 314
ix
This book has had a long gestation period, during which I have accumu-
lated many debts. Throughout the process, Patricia Lennox has supported
my work in every way imaginable – not the least of which is filling life
beyond my work with joy. Early on in the process I invited Alan Code to
Pittsburgh for an extended visit for the sole purpose of discussing this
project, and those discussions were critical in helping me frame it.
Discussions over meals and long walks with my former colleague James
Allen and with Allan Gotthelf, David Charles, Philip van der Eijk, and
István Bodnar also were helpful at this early stage. At a somewhat later
stage, Mariska Leunissen and Joe Karbowski read through a preliminary
draft and discussed difficulties of overall structure with me – the two-part
structure of the final version I owe to them. In a similar vein, Aryeh
Kosman and Joel Yurdin visited Pittsburgh and met with Jon Butacci
and me for discussions of how Aristotle sees experience in relation to
inquiry, which helped to alter the trajectory of my work at a crucial point.
While the manuscript was developing, a number of people did me the
(helpful) honor of organizing workshops where I could present and discuss
the ideas here presented. Klaus Corcilius organized one such event at the
University of California, Berkeley, March 6–7, 2015, during which
I presented an overview lecture and took part in four seminars organized
around specific chapters. I received helpful comments from Timothy
Clarke, Lucas Angioni, Michel Curbellier, and of course Klaus Corcilius
on that occasion.
Victor Caston organized a similar event at the University of Michigan,
Ann Arbor, April 16–17, 2015, during which I received helpful suggestions
from Victor, Richard Janko, Emily Katz, and James Joyce. During that
visit I also had two memorable conversations over coffee about the project
with Victor. My dear friend Giovanni Camardi and his colleague
Giovanna Giardina invited me to give two lectures based on material
from three chapters at University of Catania, Sicily, May 11–12, 2015,
xi
Aristotle
APo. Posterior Analytics (Analytica Posteriora)
APr. Prior Analytics (Analytica Priora)
Cat. Categories (Categoriae)
Cael. On the Heavens (De caelo)
de An. On the Soul (De anima)
EE Eudemian Ethics (Ethica Eudemia)
EN Nicomachean Ethics (Ethica Nicomachea)
GA On the Generation of Animals (De generatione animalium)
GC On Generation and Corruption (De generatione et corru-
ptione)
HA History of Animals (Historia animalium)
IA On the Locomotion of Animals (De incessu animalium)
Juv. On Youth and Old Age (De juventute et senectute)
Long. On Length and Shortness of Life (De longitudine et brevi-
tate vitae)
MA On the Movement of Animals (De motu animalium)
Mem. On Memory (De memoria et reminiscentia)
Metaph. Metaphysics (Metaphysica)
Mete. Meteorology (Meteorologica)
PA On the Parts of Animals (De partibus animalium)
Ph. Physics (Physica)
PN Parva naturalia
Po. Poetics (Poetica)
Pol. Politics (Politica)
Pr. Problems (Problemata)
Resp. On Respiration (De respiratione)
Rh. Rhetoric (Rhetorica)
Sens. On Perception and Perceptibles (De sensu et sensibilibus)
xv
Dante refers to Aristotle as “il maestro di color che sanno,” the master of
those who know.1 But Aristotle typically refers to the works that have come
down to us as ‘inquiries’ or ‘investigations,’ and in the pages that follow,
I will make a case for Aristotle as “il maestro di color che cercano,” the
master of those who inquire. The chapters to follow attempt to get clarity
on Aristotle’s conception of inquiry, insofar as the goal of inquiry is
scientific knowledge (ἐπιστήμη).2 Does Aristotle see inquiry, as he clearly
sees explanation, as a process constrained by epistemic norms – norms of
inquiry, as I am calling them? That is, given that Aristotle has clearly
articulated ideas about what the goal of scientific inquiry looks like, does he
also have clearly articulated norms that must be adhered to if one is to
achieve that goal?
Typically, when scholars ponder where to find Aristotle’s views on that
topic, they turn to the second book of the Posterior Analytics (APo.), which
begins by characterizing four different objects of inquiry “equal in number
to things we know” (ἴσα τὸν ἀριθμὸν ὅσαπερ ἐπιστάμεθα, 89b23-24). That
discussion operates, however, on a rarefied plane of abstraction – so
rarefied that Aristotle is at times happy to exemplify the same philosophical
thesis by means of examples drawn from geometry and botany. What that
discussion provides, I argue in the chapters that follow, is an erotetic
framework – a ‘logic of questions and answers,’ as it were – for any inquiry
aiming at achieving scientific knowledge. Every such inquiry must operate
within the general guidelines provided by this framework, which in turn is
1
Inferno, Canto 4, line 131 (Durling 1996, 77).
2
How best to translate this term is a matter of longstanding dispute. A tradition going back at least to
Kosman 1973, and endorsed and further defended by Burnyeat 1981, would render it ‘understanding’
and stress the centrality of the capacity to explain or demonstrate to its possession. Since Aristotle
uses a variety of terms to designate weaker and stronger forms of knowing, another option,
challenged by Burnyeat, is ‘scientific knowledge,’ to indicate that ἐπιστήμη, at least in its unqualified
form, is the sort of knowledge that one who has mastered a science has.
3
To anticipate a thought some readers might be having at this point, the Historia animalium (HA)
decidedly is not an exception. On the place of the HA in Aristotle’s zoological inquiries, see Balme
1987a, Gotthelf 1988 [2012, ch. 14], and Lennox 1991 [2001b, ch. 2].
4
One could also take the attitude that this is all that is of philosophical significance anyway – whether
Aristotle followed his own norms is perhaps an interesting historical question, but irrelevant to their
epistemological virtues. For reasons that I will allow to gradually reveal themselves as the argument
progresses, that is not the attitude I adopt here.
It has been said, then, what is the nature of the knowledge being
sought, and what is the object toward which our inquiry and our
entire methodos should aim.
Aristotle, Metaph. A.2, 983a21-231
1
Τίς μὲν οὖν ἡ φύσις τῆς ἐπιστήμης τῆς ζητουμένης, εἴρηται, καὶ τίς ὁ σκοπὸς οὗ δεῖ τυγχάνειν τὴν
ζήτησιν καὶ τὴν ὅλην μέθοδον. The entirety of Chapter 3 will be devoted to exploring the meaning
and uses of methodos in the Aristotelian corpus, and until that exploration is completed, I will simply
transliterate the word. However, to aid the reader, I note here that the term has two distinct but
closely related uses: to refer to an investigation or inquiry in pursuit of knowledge (the apparent
meaning here), or to refer to the way of carrying out such an inquiry, a method. These two uses
mirror two distinct uses of the root term ὁδός, which can refer either to a path along which a journey
proceeds or to the journey itself.
11
1.1 Introduction
One theme that unifies the first book of Aristotle’s Posterior Analytics is
that of specifying necessary and sufficient conditions for having know-
ledge in the strongest sense, unqualified scientific knowledge (ἐπιστήμη
ἁπλῶς; APo. i.1, 71a28-29; 2, 71b9, 15), the knowledge we have of some-
thing through demonstration, characterized as a syllogism productive of
scientific knowledge (συλλογισμὸν ἐπιστημονικόν).2 Aristotle does not
deny that there are less robust cognitive states that can lay legitimate
claim to knowledge (γνῶσις), and he insists that knowledge in the robust
sense must somehow arise out of cognitive states that fail to satisfy the
conditions for unqualified scientific knowledge (71a17-30). But he pur-
posefully puts off discussion of a question that is an immediate conse-
quence of that insistence: How does one move from less robust, prior
knowledge to those principles that are needed for demonstrative,
2
APo. i.2, 71b18. The question of the extent to which the theory of demonstration in APo. is dependent
on the theory of categorical syllogistic developed in APr. has long been debated (on which see Barnes
1981, which does a fine job of reviewing the history of this debate as well as defending one specific
answer). Since συλλογισμός just means ‘reasoning’ outside the context of the theory of the ‘syllogis-
tic’ developed in APr., how one should translate this expression depends on where one stands in that
debate. Barnes 1993 translates “scientific deduction,” but in his notes (93) expresses a preference for
‘deduction productive of understanding’; Irwin and Fine (1995, 40) have opted for “deduction
expressing knowledge”; Ross (1949, 507) and Mignucci (2007, 7) go respectively for “scientific
syllogism” and “sullogismo scientifico” without comment.
3
Thus in the first chapter he reminds his readers that if this is not in some sense true, the puzzle raised
in the Meno will arise, that we will either learn nothing or what we already know (APo. i.1, 71a29-30).
As McKirahan (1992, 23) notes, this follows Aristotle’s assertion that we think we have ἐπιστήμη of
a fact when we know its cause as its cause and that it could not be otherwise than it is – and thus it
entails that the fact to be known must be the necessary conclusion of a deduction and that the
premises must somehow identify, and be known to identify, the cause of the fact. In context, then, it
is a more significant claim than it may at first appear.
4
Any translation of νοῦς presupposes answers to at least two interpretive questions nicely distin-
guished by Barnes 1993, 259–271 (these two are in fact related to his first and third puzzles about the
chapter stated on page 259): Does Aristotle use it to refer abstractly to a faculty for acquiring
knowledge, or to the state of being in possession of principles? And does Aristotle take what it
possesses to be propositional or conceptual in character? I take νοῦς, at least here, to refer to a faculty
that holds knowledge of principles in conceptual form, and take ‘reason’ to be a preferable translation
to other ‘nonintuitionist’ options such as ‘thought,’ ‘intellect,’ ‘mind,’ or ‘comprehension.’ A more
complicated defense of this choice will emerge as we proceed.
5
For different attempts to clarify this relationship between essence-identifying definitions and causes
within the Posterior Analytics see Charles 2000, Ferejohn 2013, and most recently Bronstein 2016, esp.
Part ii.
6
Lennox 2014a provides a full defense of that claim, though the outlines of that defense can be found
in Lennox 1987, 1991, and 1994 (reprinted as chapters 1–3 of Lennox 2001b).
7
“If, then, we have no other kind besides scientific knowledge [that grasps] truth, reason (νοῦς) would
be the principle of scientific knowledge” (APo. ii.19, 100b14-15).
8
“It is quite clear that it is necessary for us to come to know the primaries (τὰ πρῶτα) by induction
(ἐπαγωγῇ); for it is in this way that perception implants the universal” (APo. ii.19, 100b3-5). “For we
are persuaded in all cases either through deduction (διὰ συλλογισμοῦ) or from induction (ἐξ
ἐπαγωγῆς)” (APr. ii.23, 68b13-14).
9
By ‘empiricism’ here I do not mean adherence to the philosophical views of the ‘empirical school’ in
Hellenistic medicine and philosophy, let alone to those of the British Empiricists. I have in mind
only the idea that perceptually grounded experience is the sole source of our knowledge of the world.
10
There are certain parallels to this idea in John Norton’s forthcoming The Material Theory of
Induction, a preprint of which is available at www.pitt.edu/~jdnorton/papers/material_theory/mat
erial.html. The norms of inquiry I discuss are not formal or algorithmic, and thus ‘material’ in
Norton’s sense; but in the view I am defending, Aristotle has both general, epistemic norms of
inquiry shaped by his understanding of the goal of scientific inquiry, demonstrative knowledge; as
well as domain-specific norms, which reflect differences in the objects of investigation and our access
to those objects.
13
That example is chosen for two reasons: (i) one can compare the many factual claims about elephants
in HA with the very long explanation (focused on the trunk, but covering pretty much everything
Aristotle knew) of many differentiating features of the elephant in PA ii.16; and (ii) there is a rich
paper by Allan Gotthelf (Gotthelf 1997b, reprinted in Gotthelf 2012, ch. 8) that unravels the
complicated explanatory threads in PA ii.16 and identifies which features are causally/explanatorily
fundamental.
14
I am using ‘domain’ rather than ‘science’ because, as we will see, there are distinct norms governing
inquiries within what Aristotle thinks of as a single science: the central theme of Chapter 4, in fact, is
that Aristotle thinks of the inquiries presented in his Physics, Meteorology, De caelo, De partibus
animalium, De generatione animalium, and a number of other treatises as contributions to a single
science of nature.
15
In this respect both Irwin and Frede appear to assume a view about the impossibility of
grounding knowledge in perception like that articulated by Donald Davidson): “The relation
between a sensation and a belief cannot be logical, since sensations are not beliefs or other
propositional attitudes. What then is the relation? The answer is, I think, obvious: the relation
is causal. Sensations cause some beliefs and in this sense are the basis or ground of those
beliefs. But a causal explanation of a belief does not show how or why the belief is justified”
(Davidson 2009, 127).
16
See Burnyeat 1981, 131–132 and note 59. In that note, Burnyeat describes Irwin as providing “a very
clear statement of the interpretation of the An. Post. programme in terms of knowledge and
justification which I am opposing.”
17
That is, it is inductive insofar as it is an inference from the explananda of a science to the best
explanans. But contemporary philosophers who defend this viewpoint see induction as
a confirmation procedure rather than as a discovery procedure, which is not how Aristotle understands
induction.
18
Recently, in as yet unpublished work, Bolton appears to be moving toward those who see Aristotle
adopting a form of justification of first principles akin to inference to the best explanation (Bolton,
in Salmieri forthcoming).
19
This is occasionally explicit: “Aristotle . . . argues that dialectic and empirical inquiry are simply
means of discovery, not essential for justification” (141). He cites no text where Aristotle says this,
and footnotes Sidgwick’s Methods of Ethics. An excellent way into the historical and philosophical
background to this distinction is Schickore and Steinle (2006).
20
Another contemporary philosophical assumption plays a role in Irwin’s argument, the assumption
that ‘perception is theory-laden’: “Aristotle should have seen that this second objection to dialectic
[that there is no assurance of the objective correctness of appearances] works equally well against
empirical inquiry . . .” [137] “. . . alleged non-inferential justifications are either inferential or not
justifications” [142].
21
For example, Ph. i.2, 184b25-185a20; cf. APo. i.12, 77a40-b15.
22
A point argued in detail in Bolton 1991, 13–22.
23
There is evidence that Frede thought this distinction was an innovation of Plato’s, and that Aristotle
follows him in this innovation (Frede 1987b, 3–8). For a classic statement of the problem, see
McDowell 1994.
26
There is a sense in which a weak form of demonstration (which I dubbed A-explanations in Lennox
1987) does play a role in induction, as I have argued before and will rely on here. One step, discussed
extensively in APo. ii.14–17, toward identifying ‘convertible’ or ‘coextensive’ universals is to search
for middle terms that connect nonconvertible universals: for example, by seeking a ‘reason why’ all
pigeons, sparrows, and ducks are feathered, you might notice that they are all forms of bird and
explain that they have feathers because all birds do. This aids you in identifying the convertible
relationship between feathers and birds – and, crucially, any other features that are convertible with
birds and thus with feathers (cf. Lennox 1991 and 1994 [reprinted in Lennox 2001b, chs. 1 and 3];
Lennox 2014a). But this is clearly not what Kosman has in mind, since he is discussing the
explanatory use of what are in fact first principles.
27
And the contrast in the following remark is also suggestive: “Hence, as he sees the problem of our
grasp of first principles, the difficulty is not a lack of evidence to transform inductive belief into
certain knowledge. That inductive belief is already knowledge (γνῶσις). What it is not yet is
understanding” (1981, 131).
1.5.1 Interlude
Before considering the next and final option,29 I would like to pause here
and draw attention to something common to many of the positions looked
at so far. Prior to Charles 2000, there was a tendency to elide the distinc-
tion between Aristotle’s views about the ascent from perception of particu-
lars to knowledge of universals and his views about how we grasp first
principles; and with that elision, a tendency to focus an inordinate amount
of attention on the very last chapter of APo. ii. Though there was literature
28
I say “mistakenly” because a scientifically minded individual would already have good reason to
reject that answer, given that leaves have those properties during the seasons when they are not being
shed. “Seasonally” becomes important, for the question is not why such trees shed their leaves, but
why they do so at certain times and not others. And that leads a curious investigator to inquire into
what else changes at the precise time that the leaves drop.
29
This interlude was added after conversations with David Charles and Marko Malink convinced me
that the norms of inquiry that are my focus are of epistemic significance whether or not they are
sufficient for the identification of first principles qua first principles.
30
And it is telling that in an otherwise systematic study of APo. (Bronstein 2016) David Bronstein
jumps from a detailed and enlightening discussion of ii.13 directly to ii.19 – chapters 14–18 are the
only chapters in book ii that are not discussed.
31
However, in “Aristotle on Knowledge” (Bolton and Code 2012, 55–78) the view defended sounds
more like an IBE account. An alternative defense of Aristotle’s empiricist foundationalism is to be
found in Marc Gasser-Wingate’s forthcoming book, entitled Aristotle’s Empiricism.
32
A convenient list of such inductive norms (or ‘rules’ as he calls them), with historical examples of
their application, can be found in chapter 6 of John Herschel’s wonderful A Preliminary Discourse on
the Study of Natural Philosophy (1830). An inexpensive facsimile of the first edition was produced by
The University of Chicago Press in 1987.
33
For example, here are the closing words of Bolton’s essay: “However crude his description or use of
the method may be his endeavours belong, in spirit, with those which we now think of as clearly
scientific” (Bolton 1991, 29).
34
καὶ γὰρ ἡ φαντασία καὶ ἡ αἴσθησις τὴν αὐτὴν τῷ νῷ χώραν ἔχουσιν· κριτικὰ γὰρ πάντα,
διαφέρουσιν δὲ κατὰ τὰς εἰρημένας ἐν ἄλλοις διαφοράς. As Nussbaum notes (1978, 334), the
reference is likely to de An. iii.3, but this view of perception as allied with reason as
a ‘discriminative faculty’ (κριτικόν) is expressed or implied throughout the corpus, including the
opening paragraph of the Metaphysics, which notes that we are especially fond of vision among the
senses because “it reveals many differences” (980a27). It is also found in de An. iii.9, 432a15-16,
where the soul of animals is said to be defined by two capacities, one of which is the ‘discriminative’
capacity, which is the function of both thought and perception.
An Erotetic Framework
The Posterior Analytics on Inquiry
1
διαφέρουσι δ’ οἱ λόγοι περὶ ἑκάστην μέθοδον, οἵ τε φιλοσόφως λεγόμενοι καὶ μὴ φιλοσόφως. διόπερ
καὶ τῶν πολιτικῶν οὐ χρὴ νομίζειν περίεργον εἶναι τὴν τοιαύτην θεωρίαν, δι’ ἧς οὐ μόνον τὸ τί
φανερόν, ἀλλὰ καὶ τὸ διὰ τί. φιλόσοφον γὰρ τὸ τοιοῦτον περὶ ἑκάστην μέθοδον.
37
4
This language is borrowed from work in erotetic logic (the logic of questions and answers; see Prior
and Prior 1955, Belnap and Steel 1976). That work was motivated by the idea of a distinctive logic of
inquiry reflecting the peculiar syntax and semantics associated with the language of asking and
answering questions, and derives its name from the Greek verb ἐρωτάω. The extent to which
Aristotle’s epistemology is shaped by his view that all knowledge arises in the form of answers to
questions is, I think, underappreciated.
5
How should one read the “or” (ἢ) in this phrase? Is it the “or” of explication (“i.e.”), of qualification
(“or rather”), or of alternative (“this or that”)? In fact, this passage forces us to face this question not
once but four times. This is not the place to defend an answer, but I am reading all of them as
presenting alternatives, though not always exclusive alternatives.
6
Cf. 90a14-15: “In all these cases it is apparent that ‘what it is’ and ‘why it is’ are the same.”
7
For a detailed consideration of these chapters, see Bronstein 2016, 144–147; Byrne 1997, 147–156;
Charles 2000, ch. 7.
8
As he puts it, such a definition “will clearly be like a demonstration of what something is, differing in
arrangement from a demonstration” (Barnes trans.; APo. ii.10, 94a1-2). Charles 2013 explores the
metaphysical underpinnings of the relationship between definition and causal explanation in APo. ii.
9
Ackrill 1981; Bolton 1976; Bolton 1987; Bronstein 2016, chs. 9–10; Charles 2000, chs. 2, 10; Goldin
1996; McKirahan 1992.
10
It is worth noting, however, the total absence of mathematical examples in these chapters. There is
a parenthetical reference to the practices of arithmeticians in chapter 9, but without an example.
Indeed, from this point on the ratio of examples from natural science to those from mathematics is
quite high, in contrast to book i. See Wians 1996b, 131–150 (in Wians ed. 1996a).
11
Twice in earlier chapters he has used a thought experiment to make this point: “If we were on the
moon, we would not have to inquire either whether the eclipse was occurring or why – both would
be clear simultaneously. For from the perceiving of it the universal would become known to us”
(APo. ii.2, 90a26-29; cf. APo. i.31, 87b39-88a4). Recall that chapter 8 begins by noting that it is
possible for the fact and the reason why to be discovered simultaneously in certain cases (93a17-18).
12
As emphasized by Kleiner 1988. The ‘aporia of the Meno’ is explicitly referred to at APo. i.1, 71a29 (as
is the doctrine that learning is recollection at APr. ii.21, 67a21). But it is also clearly in the
background of the aporia that introduces APo. ii.19. Cf. Bronstein 2016, 229–234. Charles 2010a,
ch. 3 (2010b), 8 (2010c); Ferejohn 2013, ch. 3.
13
By ‘partial universal’ I refer to attributes that Aristotle describes as ‘belonging to all (of a kind) but
extending beyond’ (e.g., the attribute ‘interior angles equal to two right angles’ belongs to all scalene
triangles, but extends to all triangles). The important defense (at APo. i.24, 85a13-86a30) of the
superiority of universal (katholou) to partial (kata meros) demonstration is much easier to compre-
hend once one realizes that the Greek often translated ‘universal’ (katholou) literally means ‘accord-
ing to the whole,’ and that the Greek for ‘partial’ (kata meros) literally means ‘according to the part.’
The argument in APo. i.24 is in defense of the superiority of demonstrating that an attribute belongs
commensurately to its subject versus demonstrating that it belongs to all the members of a subset of
that subject.
14
And thus, while I will note disagreements with the discussion of these chapters in Charles 2000,
chapters 8 and 9, we agree that chapter 13 provides a context for the discussion of problems in
chapters 14–18. It is useful to be reminded that the chapter divisions were introduced in the
Renaissance and have no ancient authority. What is puzzling about these summary remarks is
that they do not mention chapters 11 and 12, the chapters that explore causality in the context of
syllogistic demonstration. Interestingly, while Bronstein 2016 has an excellent discussion of chapter
13, there is no sustained discussion of chapters 14–18.
15
And compare: “For it is not what is reputable that we count as a principle, but rather what is primary
in the kind with which the proof is concerned – and not every truth is appropriate (oikeion)” (APo.
i.6, 74b24-26).
16
There is an excellent discussion of these passages in Ferejohn 2013, 81–97. He refers to these ‘primary
universals’ as Cathólic Predications. My only significant disagreement with that discussion is
Ferejohn’s insistence (84–86) that the “qua itself” condition is extensional, like the “belongs to
all” condition, rather than intensional, like the per se condition. Cf. Hasper 2006.
17
The connection between being prepared for demonstration and these two properties is made
especially clear at APo. i.9, 75b37-76a9.
18
I am avoiding the common practice of translating πρῶτον καθόλου, which appears twice here (and
again at 74a11 and a12), conjunctively (“primitively and universally,” as do Barnes and Tredennick),
since it obscures the fact that Aristotle is characterizing a type of universal predication that is primary
and distinguishing it from what he sometimes refers to as what holds κατὰ παντός (cf. i.4, 73b25-28;
74a10-11) – that is, distinguishing convertible from nonconvertible universal predication. The
common mistake, for which Aristotle is identifying three distinct sources, is of thinking that
a merely extensional, nonconvertible universal predication is a primary, convertible universal
predication.
19
Hasper 2006 provides a detailed analysis of this chapter. For a good discussion of problems in
translating this passage see McKirahan (1992, 172–173); I don’t find Barnes’ emendation at 74a38
necessary (1994, ad loc.). It will be noted that the passive participle of the verb ἀφαιρέω appears twice
to describe the process of cognitively ‘removing’ features to identify which is the first that, being
removed, gives you a subject to which 2R no longer belongs. That is significant, since the expression
often translated ‘abstraction’ (ἐξ ἀφαιρέσεως), based on the abstract noun formed from this verb, is
used by Aristotle to refer to the process by which mathematicians ignore or ‘subtract’ the materials
and processes of nature in order to focus on a natural object’s mathematical features, qua
mathematical.
20
People familiar with Bacon’s Novum Organum will see obvious affinities to the description in that
work of procedures for creating ‘tables of presence and absence’ (NO ii.11, 12).
21
See notes 17 and 18, above.
22
This fact is rarely taken note of by commentators; but see Ferejohn 2013, 115–120.
23
cf. APo. i.6, 74b5-12, 74b27-32.
24
On the concept of a problem in Aristotle, see Lennox 2001b, 13–15, 48–53, 76–91; and Byrne 1997,
55–80. I discuss the question of how Aristotle’s use of the concept of problêma in the Analytics and
the Aristotelian Problemata Physica are related in Lennox 2015, 36–58 in Mayhew (ed.) 2015.
25
The chapter division, which occurs halfway through this passage, makes very little sense. Bekker
quite rightly does not have a paragraph break in his text.
26
Indeed, in the “Problems” tradition, which may have been inspired by these chapters of the APo.,
this becomes a trope: virtually every chapter of the 38 books of Problems (now generally assumed to
be composed by later members of Aristotle’s school rather than by Aristotle), which run from 859b1-
967b26 in the Bekker edition of the Aristotelian Corpus, begin with the expression dia ti (why?).
I have made some tentative suggestions about how these books may be connected to the distinction,
in Greek geometry, between theorems and problems in Lennox 1995 (reprinted in Lennox 2001b,
chapter 3).
30
I refer to these as “A-explanations” in Lennox (1987 [2001b, ch. 1]); McKirahan as “application
arguments” (1992, 177–178). The kind (γένος) is in the middle term position, so that you explain
why some form of the kind has a property by noting that it is a property common to the kind.
31
See Lennox (1994, 64 n. 21 [2001b, 82 n. 20]) and (1987, 90–99 [2001b, 7–15]).
32
Aristotle actually would say “heart or heart analogue.”
33
Our term for the group under discussion is ‘ruminant,’ but the Greeks had no common name for
this group.
34
Cf. Gotthelf (1987, 178–185 [2012, 164–171]) and Lennox (1991, 275–276 [2001b, 49–50]).
35
This is precisely the issue in terms of which certain counter-examples were posed to the deductive-
nomological account of explanation initially proposed by Carl Hempel and Paul Oppenheim in
1948 (for a nice summary of the examples and the issue at stake, see Salmon 1990, 46–50). For
example (one due originally to Sylvan Bromberger [1962], though according to Salmon, 189 n. 12,
this precise example never appeared in publication), given sunlight shining on a flagpole of height
h at a given angle a, there will be a shadow of length l. You can deduce the length of the shadow from
the height of the flagpole and vice versa. But since the interaction of the light and flagpole determine
the length of the shadow, while the shadow is not similarly implicated in determining the height of
the flagpole, there seems to be more to explanation than mere deducibility. Aristotle, as we will see,
discusses a number of similar examples.
36
This is an evolutionary remnant of the shell possessed by the common ancestor of present-day
cuttlefish; cf. HA iv.1, 524b23-30, iv.7, 532b1.
37
“There is even an analogue to the spines of fish in the soft-bodied animals; for example, in the
cuttlefish there is a part called the ‘sêpion,’ in the squids what is called the ‘sword’” (PA ii.8,
654a19-22).
38
One thing that is relevant to determining whether these chapters of APo. set the agenda for the
tradition represented in the thirty-eight-book collection of problems is whether there are consistent
attempts in those books to discover these sorts of relationships among initially distinct problems.
39
A scholium to Euclid, Elements 5 says that the general theory of proportion presented there is due to
Eudoxus. See Heath (1931, 190–191).
40
For a detailed discussion of this example, both in APo. i.5 and ii.17, see Hasper 2006, 262–269.
41
However, it remains unclear how different these cases are – it may be nothing more than a matter of
whether one is focused on the identity of the middle term or of the proposed problem. Hasper 2006
suggests that in the ‘proportionals’ case the general demonstration is applicable to the various cases
without the cases being generically identical, a solution that may be applicable here as well.
42
As I will discuss in Chapter 6, PA i.1, 639a19-b3 distinguishes attributes, such as forms of locomo-
tion, that are generically alike in animals but that “differ by a difference in form” from features, such
as sleep and respiration, that are the same across kinds.
43
A work attributed to Aristotle on this subject is preserved only in a thirteenth-century Latin
translation, Liber Aristotelis de inundatione Nili. For a thorough discussion of the history of this
text, who the original translator may have been, and its authorship, see Beullens (2014). The
connection between this manuscript and the passage we are discussing is by no means direct. The
problem that De inundatione Nili discusses and presents an explanation for is why the Nile floods in
the summer rather than the winter, whereas in our passage the problem is why the Nile increases in
flow level at the end of each month.
44
Barnes translates the verbs as simple futures, but since this in the midst of stating an aporia and this
is the apodosis of a conditional, it is preferable to make the modal force of the potential optative
explicit. It is also important to note that the verb is δείκνυμι, to show or prove, not ἀποδείκνυμι, the
verb he uses for scientific demonstration. It is possible to construct a proof with either the cause or
the effect in the middle term position, but only one of those proofs will be a demonstration, in which
the middle term identifies the cause.
45
The proof that has the more familiar, noncausal middle term does not meet the criteria for
unqualified demonstration as he has defined it in APo. i.2–6; but there are a number of such uses
of ἀπόδειξις in the Posterior Analytics (e.g., i.13, 78a30, a36; i.24, 85a13-17, a30), and they often seem
to refer to cases where, at a certain stage of inquiry, one could think that one had produced an
unqualified demonstration.
48
It is important to note here that near the end of this discussion, at 99a30-b7, Aristotle considers
whether the same effect could have a different cause in different kinds – for example, could the
attribute of being long-lived be due to different causes in four-legged land animals versus birds
(99b4-7)? Though it appears he leans toward supposing that they are not really the same effect if the
kinds are truly distinct, he leaves the question open.
49
As Charles stresses (2000, 206–207), Aristotle’s question is whether there is something about broad-
leafed trees as such that produces leaf loss. Even if there is, that does not rule out other ‘incidental’
causes of leaf loss, such as disease or poisoning.
2.4 Conclusion
The primary purpose of this chapter has been to make the case for APo. ii
providing what I’ve been calling an erotetic framework for scientific
inquiry, that is, inquiry aimed at scientific knowledge. The framework
provides norms for how inquiries ought to be ordered and for how we
ought (and ought not) to think of the relationship between definition- and
demonstration-oriented inquiries – that is, between the search for essences
and the search for causes. I’ve focused special attention on the ‘Problems’-
oriented chapters (ii.14–18), highlighting the way in which they give us
a glimpse into Aristotle’s views about the process of establishing the sort of
factual knowledge that is poised for causal understanding, including the
role that weaker forms of explanation can play in that process.
Throughout, Aristotle highlights the framework status of the discussion
by repeatedly exemplifying his philosophical point with examples drawn
from widely disparate areas of knowledge. But in the very generality of its
results lie the book’s limitations.
The first principles of a science, we are told in the last chapter of the
Posterior Analytics, are achieved by a process called induction (ἐπαγωγή).51
I conclude this chapter by drawing attention to what is not discussed in
APo. ii.14–18 that one might expect to be, especially in chapters leading up
50
This interdependency of definition and causal demonstration is one of the primary themes of
Charles (2000); it is discussed in the context of these examples on 209–214.
51
Cf. EN vi.3, 1139b25-31.
52
It is easy to miss this because we fall into the habit of supplying plausible answers from common
sense or from his other treatises.
A Discourse on μέθοδος
Every art and every methodos, and likewise every action and decision,
seems to aim at some good: hence it has been well said that the good is
that at which all things aim.
(EN i.1, 1094a1-3)1
1
Πᾶσα τέχνη καὶ πᾶσα μέθοδος, ὁμοίως δὲ πρᾶξίς τε καὶ προαίρεσις, ἀγαθοῦ τινὸς
ἐφίεσθαι δοκεῖ· διὸ καλῶς ἀπεφήναντο τἀγαθόν, οὗ πάντ’ ἐφίεται.
65
2
This progression is somewhat surprising, since on certain epistemological assumptions you might
expect the progression to move from less to more abstract, and Socrates recommends beginning with
pure mathematics and moving toward what appears to be the application of the theory of ratios and
proportions to music and of solid geometry to the heavens. This does allow him, however, to
disparage a particular way of conceiving harmonics and astronomy as being about appearances.
3
The Proem (DK 28 B1) of Parmenides’ poem depicts a chariot’s journey on a “celebrated road” (ἐς
ὁδὸν . . . πολύφημον), maidens leading the way (κοῦραι δ’ ὁδὸν ἡγεμόνευον) to a meeting with
a goddess, a road upon which right and justice sent him. DK 28 B7 famously warns against traveling
the way of much experience (πολύπειρον ὁδὸν) and DK 28 B8 begins by asserting that only one story
remains about a way (μόνος δ’ ἔτι μῦθος ὁδοῖο): that it is.
4
Note too the periphrastic construction at 243d7, δεῖν ποιεῖσθαι τὴν μέθοδον ἡμᾶς, likely modeled on
the quite common expression ὁδὸν ποιεῖσθαι, to make a journey.
5
Compare Plt. 286d5-9: “Again in relation to the inquiry into something problematic, reason
commands that finding the answer easily and quickly should be regarded as secondary, not primary;
while it commands us to honor most of all and primarily the procedure itself of being able to divide
according to forms (πολὺ δὲ μάλιστα καὶ πρῶτον τὴν μέθοδον αὐτὴν τιμᾶν τοῦ κατ’ εἴδη δυνατὸν
εἶναι διαιρεῖν).”
6
On the endless debate about which, if any, work in the extant Hippocratic corpus is being referred to
in the Phaedrus, cf. Smith 1979, 44–50, Joly 1983, 28–47.
7
And it is an ambiguity that is rooted in this concept’s original context: recall that a ὁδός can refer
either to the path or road one takes to a goal or to the activity of traveling along that path (LSJ s.v.
ὁδός, i and ii).
8
There is no direct evidence that Aristotle was familiar with this work; in fact, the only direct evidence
that Aristotle was familiar with any work in the Hippocratic corpus is a lengthy quotation in HA
iii.3, 512b13-513a8, of a text which appears twice in Hippocratic works, once in chapter 9 of De natura
ossium and once in chapter 11 of De natura hominis. In some manuscripts, Aristotle attributes the
quotation to Πολύβιος, which is the name of Hippocrates’ son-in-law; in the majority, however, it is
attributed to Πόλυβος.
9
The line numbers follow Heiberg 1927.
10
Cf. VM 8, 9: αὕτη ἡ τέχνη πᾶσα ἡ ἰητρικὴ τῇ αὐτῇ ὁδῷ ζητεομένη εὑρίσκοιτο ἄν.
11
For a fine, detailed comparison of this passage with similar discussions of number, weight, and
measure in relation to goal-directed activity in Plato’s dialogues, see Festugière 1948, 41–43, note 41.
Perhaps the most interesting comparison drawn by Festugière is to the discussion of different forms
of knowledge in Philebus 55c-56e, where crafts achieve greater precision and are more scientific to the
extent that they make use of calculation, measure, and weighing (ἀριθμητικὴν . . . καὶ μετρητικὴν καὶ
στατικήν), in the absence of which there is nothing but “conjecture and the training of our senses
through experience and routine” (Frede 1993). Socrates goes on to say it is the latter that is typical of
medicine (56b1), but building, through the use of measures and instruments, achieves greater
precision and craftsmanship (b4-6). See Schiefsky 2005, 13–25 for a valuable discussion of the
concept of ἀκρίβεια in relation to τέχνη in this work and its connection to the discussion of these
concepts in the Philebus, as well as an interesting discussion of the varying attitudes in the
Hippocratic corpus toward natural inquiry and Socrates’ attitude as described in the Phaedo.
12
There is also a reference to this work in Simplicius’ commentary on the Categories. Cf. Rose 1886,
xix, fr. 116, l. 11.
13
Chapters 8 and 9 will examine Aristotle’s deliberate practice of borrowing the results of certain
inquiries to aid in advancing others. It turns out that such borrowings are also constrained and
guided by epistemic norms.
14
One central purpose of this chapter is to have, by its conclusion, a number of concrete examples
before us of what he has in mind.
16
In this context, the term logos seems to refer to the plan or organization of development.
17
For examples of δύναμις used in this way, see 729b6, b28, 730a2, a14; for ἀρχὴ τῆς κινήσεως, see
729a10-11, 729b13-14, 28, 730a27-28.
18
“The source comes to be first. And this is the heart in the blooded animals, while in others its
analogue, as has been said repeatedly” (GA ii.5, 741b15-17).
19
For further discussion of the argument in this text, see Leunissen 2018, 69–74. Book iii, which
discusses development in egg-laying and grub-producing animals (and finally those generated
spontaneously), focuses most of its attention on the development of the egg and its parts (where
eggs are involved) rather than on the development of the parts of the animal itself. I thank an
anonymous referee for insisting on this.
20
οὐ λίαν ἐμπειρικῶς ἔχοντες τῶν συμβαινόντων (742a17).
21
τὸ δὲ πρότερον ἤδη πολλαχῶς ἐστιν.
22
For more on this topic, see Lennox 2018 in Falcon and Lefebvre, eds. 2018, 249–272.
23
The day-by-day development is described at HA vi.3, 561a4-562a22.
24
It is worth noting that in order to overcome the problem of temporal unfolding he also had to
overcome the observational problem – in this case opening the egg in such a way as not to destroy
the embryonic chick inside.
25
For an illuminating and more detailed discussion of this passage, see Leunissen 2018, 69–74.
26
Here I suppose the use of this (present sense) perfect form of the verb φύω is tied more closely than
usual to its semantic roots, where it refers to processes of budding, growth, and development. Cf.
743a21-23 where the warmth that produces flesh and bone during embryogenesis is said to produce,
not any chance part at any chance place or time, but τὸ πεφυκὸς, ᾗ πέφυκε, ὅτε πέφυκεν – that
which is natural, where it naturally belongs, at the naturally appropriate time.
27
See, for example, Metaph. Δ.11; Θ.8; and Γ.2, 1005a14-18, which lists priority and posteriority among
attributes to be studied by the science of being qua being.
28
For contrasting accounts of this argument see the essays by Charles and Broadie in Lennox and
Bolton, eds. 2010, chapters 8 and 9; and for arguments on how to understand this notion of priority
in being, see Peramatzis 2011, 278–286; Beere 2009, 285–304; and Witt 1994, in Scaltsas, Charles, and
Gill 1994, 215–228.
29
There is in fact Platonic background as well: twice in the biological works (PA i.1, 640a15-19 and GA
v.1, 778b2-7, which is apparently referring back to the PA i.1 passage) Aristotle echoes (though
without explicit acknowledgment) Phlb. 54a5-11, c9-11, stating that “generation is for the sake of
being, rather than being for the sake of generation.”
30
The last few lines in this passage are essential, since in Greek this claim would be as paradoxical as it
would be in English if I were to say “the end is a beginning.” I’ve softened the paradoxical feel by
translating archê as ‘origin,’ but the word can be used for a first principle in mathematics, a primary
thing, the beginning of a process, and an authority (the masculine nouns ἀρχός and ἄρχων refer to
military and political leaders).
31
Recall, this is a critical feature of the erotetic framework of the Posterior Analytics.
according to a determinate program: “in this way the capacity of the nutritive soul, just as it produces
growth from the food in the animals and plants themselves later on, using as tools heat and cold (for its
motion is in these, and each thing comes to be by a certain logos), so too that which comes to be by
nature is constituted in this way from the beginning. For it is the same matter by which it grows and
from which it is first constituted, so that the productive capacity <for growth> is also the same as the
initial one (though greater in strength). So if this is the nutritive soul, it is also that which generates
(740b30-37; Gotthelf translation with modifications; emphasis added).
35
It is unclear both what the reference of “these” (τούτοις) is and how to interpret the dative case
ending. Peck 1942, Platt 1912, and Lanza & Vegetti 1971 all translate it as referring to the two previous
items in the list. Peck (1942, 213, note d) has a footnote that shows a concern about the plural,
perhaps because Aristotle had previously (742a32-33) connected the useful only to the end.
36
τὸ γὰρ τέλος ἐνίων μὲν ὕστερον ἐνίων δὲ πρότερον. Peck translates: “For the End, though it comes
after some of them, is prior to others.” But so translated, it is unclear what the antecedent of “them”
and “others” is.
37
On which see Code 1997, esp. 134–136; Charles 2012, 227–266 (in Shields, ed. 2012); and the
upcoming discussion in Chapter 6.
38
On the relationship between this kind of experience and the knowledge of the craftsman, see Charles
2000, 151–161.
39
This phrase has been variously translated: Balme 1992, “as it were a sort of educatedness”; Lennox
2001a, “a certain sort of educatedness”; Lanza and Vegetti 1971, “una sorta di cultura”; Ogle 1882, “a
kind of educational acquaintance”; Peck 1961, “what is roughly described as an educated person’s
competence.” There are two distinct problems: first, there is no word in English that captures all of
the nuances of the Greek word παιδεία; second, the above translations show the various options of
combining the relative οἷον with the enclitic indefinite τινά. Balme (“as it were”) and Peck (“roughly
described”) take it adverbially (LSJ s.v. οἷος v.2.d); Lennox, Vegetti, and Ogle take it as a general
relative with τινά, indicating an unspecified type of παιδεία. Given the meaning specified here and
in related passages throughout the corpus, the adverbial rendering is inappropriate: it is a form of
genuine discernment that Aristotle has in mind, which some people have regardless of the domain
under consideration, and others have about a specific subject. Thanks to an anonymous reader for
pressing me about the meaning of this phrase.
40
There is one occurrence in the Physiognomica, one in a fragment attributed to a work called On
Divisions, and three in fragments of On Philosophy (which also contains a reference to the Μεθοδικά
that, as we’ve already seen, is referred to in the Rhetoric).
41
It is not always clear that what our corpus treats as a single treatise was so considered by Aristotle.
A good example, to be discussed in the Chapter 4, is the occurrence of μέθοδος at the beginning of
Physics iii.
42
For similar uses in Top. and SE, see Top. i.6, 102b36-103a1; SE 11, 171b11-22.
43
Cf. Top. i.3, 101b5 (the first line of ch. 3) and Top. i.4, 101b11 (the first line of ch. 4) for other
examples of the term used in this way.
44
The word rendered here as ‘toward,’ πρὸς, can also be translated ‘in relation to.’ On the controver-
sies surrounding these sentences, see the judicious remarks in Smith 1997, 52–54.
45
Allan 1961; Anagnostopoulos 2009b; Barnes 1980; Burnyeat 1980; Irwin 1981, 1988; Owen 1961;
Kraut 2006b; Reeve 2012; Salmieri 2009; Whiting 2001. A recent exception is Karbowski 2019, esp.
88–106.
46
Joe Karbowski first drew this passage to my attention and I have benefited from numerous
conversations with him about it. On this passage in particular, see Karbowski 2019, 110–120.
47
With this passage compare Pol. iii.8, 1279b12-15: καὶ γὰρ ἔχει τινὰς ἀπορίας, τῷ δὲ περὶ ἑκάστην
μέθοδον φιλοσοφοῦντι καὶ μὴ μόνον ἀποβλέποντι πρὸς τὸ πράττειν οἰκεῖόν ἐστι τὸ μὴ παρορᾶν
μηδέ τι καταλείπειν, ἀλλὰ δηλοῦν τὴν περὶ ἕκαστον ἀλήθειαν.
48
The Solomon translation in the Revised Oxford Translation, “appropriate to that study,” has more
of the flavor I am suggesting. For similar uses of μέθοδος in Aristotle’s ethical treatises, see EN i.1,
1094a1-3; i.2, 1094b9-12; i.7, 1098a25-29; v.1, 1129a3-6.
49
This expression, and one with a virtually identical sense, κατὰ τὸν ὑφηγημένον τρόπον (cf. Mete. i.1,
339a6-7), appear a number of times in the corpus. See, for example, GA iii.9, 758a28-29 (discussed in
Leunissen 2018, 62) and my discussion in Chapter 4.
3.7 Conclusion
Understanding the dual roles of the concept of methodos in the Aristotelian
corpus is a key to understanding Aristotle’s views about scientific inquiry,
and specifically his views about the way in which distinctive scientific
inquiries are governed by norms that are domain-specific, appropriate for
and distinctive to that domain. By exploring its use across the corpus as
a whole we have seen how it is deployed and why it so often appears in the
introductory ‘methodogical’ passages of his investigations. Moreover, it
appears to be part of a network of concepts, such as paideia, zêtêsis, and
episkepsis, related to the question of how one ought to pursue one’s inquiry
so as to optimize the chances of acquiring scientific knowledge. From this
point forward, I will use ‘method’ to translate it when Aristotle is discuss-
ing the proper way to carry out an inquiry, and ‘methodical inquiry’ when
Aristotle is referring to a domain of inquiry, viewed from the perspective of
the norms governing its practices.
The focus of the remaining chapters of this book will be on Aristotle’s
scientific inquiry into the natural world. I will conclude Part i by
Natural Science
Many Inquiries, One Science
Wherefore one needs to have been educated about how one ought to
accept each subject, since it is absurd to inquire at the same time about
scientific knowledge and the way of acquiring it; but neither one is
easy to grasp. One ought not to demand mathematical accuracy in all
cases, but only in those subjects without matter; for which reason the
[mathematical] approach is not a natural one, for presumably nature
in all cases possesses matter. For which reason one ought first to
investigate what nature is; for in this way it will also be clear with
which things natural science is concerned, and whether it is for one
science or more than one to study the causes and principles.
(Metaph. α.3, 995a12-20)1
1
διὸ δεῖ πεπαιδεῦσθαι πῶς ἕκαστα ἀποδεκτέον, ὡς ἄτοπον ἅμα ζητεῖν ἐπιστήμην καὶ τρόπον
ἐπιστήμης· ἔστι δ’ οὐδὲ θάτερον ῥᾴδιον λαβεῖν. τὴν δ’ ἀκριβολογίαν τὴν μαθηματικὴν οὐκ ἐν
ἅπασιν ἀπαιτητέον, ἀλλ’ ἐν τοῖς μὴ ἔχουσιν ὕλην. διόπερ οὐ φυσικὸς ὁ τρόπος· ἅπασα γὰρ ἴσως
ἡ φύσις ἔχει ὕλην. διὸ σκεπτέον πρῶτον τί ἐστιν ἡ φύσις· οὕτω γὰρ καὶ περὶ τίνων ἡ φυσικὴ δῆλον
ἔσται καὶ εἰ μιᾶς ἐπιστήμης ἢ πλειόνων τὰ αἴτια καὶ τὰς ἀρχὰς θεωρῆσαί ἐστιν.
99
4.1 Introduction
Much like Aristotle’s use of the concept methodos, our current concept of
‘science’ is used in two closely related but distinct senses: to refer to the
structured knowledge that has been achieved about a domain of investiga-
tion, and to refer to the norm-governed investigation of that domain. This
duality naturally leads to two distinct ways of conceiving of the unity of
a science. Thinking of a science as a body of knowledge, we can ask if the
knowledge we have of a domain constitutes a structured unity, the unity
perhaps provided by explanatory relationships among propositions or
models, or classificatory relationships among concepts or theories. Of the
biological sciences, for example, we might ask whether there is one bio-
logical theory that unifies (perhaps by means of a single pattern of explan-
ation, or by means of a unifying concept or network of concepts) all our
knowledge of the living world. Thinking of a science as an inquiry carried
out according to certain distinctive practices, on the other hand, we might
ask whether there is a single set of methods or norms specific to the domain
in question that, if followed, will continuously advance our understanding
of that domain.2
I have been insisting that it is important to keep a distinction of that sort
in mind in considering Aristotle’s corpus as well, although the conceptual
map is drawn in a very different way. The Greek term that I have been
2
There is an extensive literature in philosophy of science on just these questions. To select a few,
staking out different positions: Cartwright 2000, Dupré 1993, Kitcher 1993, Mitchell 2003,
Rosenberg 1994, Schaffner 1993.
3
In EN vi it is one of a number of cognitive virtues, all of which are ‘possessions’ (hexeis), that is habits
acquired through practice or teaching. Strictly speaking, the account given here only applies to
unqualified scientific knowledge, (epistêmê haplôs, APo. i.2, 71b9), since epistêmê is used, even within
the Analytics, in a less restricted fashion – for example, when he distinguishes epistêmê of the fact from
epistêmê of the reasoned fact (APo. i.13, 78a22).
4
In works such as Burnyeat 1981 and Salmieri 2013, there have been valuable studies of Aristotle’s
network of concepts referring to knowledge (especially gignoskein, gnorizein, epistasthai, and eidenai).
One consequence of the research in which I’ve been engaged in writing this book is the realization
that there needs to be a similar exploration of this network of concepts referring to inquiry.
5
This is a topic that will be explored in a variety of ways in later chapters. But to point to what I have in
mind, some of these words come from the context of seeking and searching (zêtêsis, skepsis), some
from that of a route or path to a goal (hodos, methodos,), and some from that of observing or studying
(historia, theoria).
6
A different approach to the question of the unity of Aristotle’s diverse explorations of nature can be
found in Falcon 2005, driven more by ontological concerns about causation than the epistemological
issues on which I am focused.
7
A point nicely stressed in the opening pages of Wilson 2000.
8
Though there is reason to be skeptical about the writings of Empedocles, Anaxagoras, οr
Democritus having titles in our sense; cf. Schmalzriedt 1970. The reference to the Timaeus in
Diogenes Laertius (D.L. iii.60: Τίμαιος ἢ περὶ φύσεως) suggests it too was considered
a contribution to this tradition, and it does share the attribute of being a single, all-encompassing
treatise on every aspect of the natural world, from the elements to the heavenly spheres. Moreover,
Aristotle treats it in the same way as previous investigations of nature. For example, at de An. i.3,
406b25-28 he says: “And in the same way [as Democritus] the Timaeus also gives a natural account
(φυσιολογεῖ) of how the soul moves the body; for it is by itself being moved that it moves the body,
on account of being intertwined with it” (cf. Sens. 2., 437b9-14, where the theory of vision in the
Timaeus is compared to that of Empedocles). And Ph. i.4 opens with a general categorization of
earlier investigators of nature, and at 187a12-23, Plato is cited as an example (albeit an odd example)
of the first group discussed.
9
I am setting aside here the development, recognized by both Plato and Aristotle, of astronomy,
optics, mechanics, and harmonics as distinct mathematical sciences, as well as the developments in
medicine reflected in the earliest material in the Hippocratic corpus. I do discuss Aristotle’s views
about how the science of nature is related to mathematics in Lennox 1986, and to medicine in
Lennox 2005.
10
For paradigmatic examples, cf. PA i.1, 639a1-12; MA 1, 698a1-10; GA i.1, 715a1-18; de An. i.1, 402a1-
11 – and the passage we will consider in some detail shortly, Mete. i.1, 338a20-339a10.
11
And, in this respect, the science of nature is in stark contrast with mathematics. See, for example,
APo. i.14, 79a18-20: αἵ τε γὰρ μαθηματικαὶ τῶν ἐπιστημῶν . . . οἷον ἀριθμητικὴ καὶ γεωμετρία καὶ
ὀπτική. Cf. Metaph. Γ.1, 1003a25-6; Ε.1, 1026a25-28; Μ.3, 1078a33, b2.
12
μέρος τῆς μεθόδου ταυτῆς, 338a25. This is a good example of the way in which methodos can serve to
pick out a domain of inquiry in a way that stresses the mode of inquiry specific to that domain, as
discussed in Chapter 3. This passage says very little about what unites these as parts of this methodos;
but these are all natural inquiries, and we will explore what is involved in the methodos of nature in
Chapter 5.
13
See, for example, Metaph. Ε.1, 1025b19, 26, 1026a6, 12; Cael. i.1, 268a1, on which see Falcon, 2005,
31–48.
14
Throughout I will translate οὐσία as ‘substantial being,’ the adjective in recognition of the time-
honored, though in my view confusing, habit of translating it ‘substance,’ the nominal participle in
recognition of the word’s actual roots. It is an abstract noun derived from οὖσα, the feminine
participle of εἶναι, to be. A literal but inelegant translation would be ‘beingness.’
15
For a valuable discussion of just how puzzling, see Shields 2012, 343–371. See Frede 1987a, 81–95,
following up an earlier suggestion by Patzig, for one sort of solution.
16
Cf. Ross 1924: 1, 354. Ross prints ὡς οὐ χωριστὴν μόνον following mss. E and T and omits the
preceding comma of the Oxford Classical Text. But the addition of ὡς is not necessary. I am
following Alexander’s understanding of the passage, taking μόνον adverbially (LSJ, 1145 cites this
passage for this use of the neuter μόνον). On its own it is difficult to understand, but the next few
lines make Aristotle’s point clear.
17
Ross 1924: 1, 354; Kirwan 1971, 185; Peramatzis 2011, 100.
18
δεῖ δὲ τὸ τί ἦν εἶναι καὶ τὸν λόγον πῶς ἐστὶ μὴ λανθάνειν, ὡς ἄνευ γε τούτου τὸ ζητεῖν μηδέν ἐστι
ποιεῖν.
19
We may, perhaps, consider the fact that he mentions noses first in his list of nonuniform animal
parts that are ‘snub-like’ as an instance of Aristotelian humor.
20
δῆλον πῶς δεῖ ἐν τοῖς φυσικοῖς τὸ τί ἐστι ζητεῖν καὶ ὁρίζεσθαι (1026a4-5).
21
Among more recent discussions, see Balme 1987b, Code 2010, Ferejohn 1994, Frede and Patzig 1988,
Gill 1989, Lennox 2008a, and Peramatzis 2011.
22
For more detail on the dependence of the biological account of uniform parts on Mete. iv, see
Lennox 2014d, 272–305. Chapter 9 further develops the ideas in that paper.
23
It will be noted that the characterization of the accomplishment of Ph. viii.6 is in terms of the
erotetic framework set out in APo. ii – first it was established that there is eternal motion, and then it
was determined what it is. A number of the chapters in Primavesi and Rapp, eds. 2020 discuss the
many interdependencies of MA with other natural inquiries. Besides the Physics, the questions taken
up in the last six chapters are intertwined with those of de An. iii and PN – in fact, there are good
reasons (beginning with its last sentence) to think MA was conceived as part of the sequence of
studies that make up PN, on which see the Introduction to Laks and Rashed 2004, and the
Introductions to Primavesi and Rapp, eds. 2020.
24
Wilson 2000, 47–52. 25 Cf. GA v.4, 784b8-9.
30
The Greek is κατὰ τὸν ὑφηγημένον τρόπον; a comparison with Pol. i.1, 1252a17-18 and GA iii.8,
758a28-29, where the expression κατὰ τὴν ὑφηγημένην μέθοδον is used, suggests that in such
contexts the two phrases are more or less interchangeable. The verb ὑφηγέομαι has the general
sense of leading, guiding, or showing the way, and the participial form thus needs to be rendered in
a way that captures the implicitly normative force it has in these passages.
31
Burnyeat hints at this concern in a footnote to his 2004 essay, 13 n. 16: “By saying ‘a different set of
norms’ I don’t mean to imply that the norms are incompatible in any way.”
32
Burnyeat’s discussions do, however, make one questionable assumption: that the closing reference to
a study of animals and plants implies that De anima is part of this organized plan for the study of
nature. I discuss this further in Chapters 5 and 7.
33
The issue of how to make use of cross-references in investigating such questions is complex, but
I think it is helped by making a number of distinctions. First, one needs to distinguish references that
are imbedded in discussion and argumentation (such as the forward references in PA iv.10 to the
arguments and definitions provided by GA i.17–23, discussed on p. 110) from those that are
essentially isolated and which could be removed from the text without significant damage to
argument or exposition (e.g., PA i.5, 646a2-4; iv.14, 697b27-30; MA 11, 704a3-b3; IA 18, 714b20-
24). It is less likely that the former references were added by a later editor, more likely with the latter,
especially in light of occasional inconsistencies about the order of inquiry in the latter references.
Nevertheless, some of the references forward and backward at the beginnings and endings of texts
explicitly point to relations of conceptual or explanatory dependency (e.g., those at the beginning of
GA i and PA ii) and for that reason are potentially valuable for the current project. With these texts,
an interesting test of their usefulness as evidence of Aristotle’s views is considering whether the
beginning of the text would make sense in their absence.
34
She cites Owen’s British Academy lecture, “The Platonism of Aristotle” (1966/1986), as support for
this claim, but Owen’s only explicit reference to the issue in that lecture is this: “[Aristotle] allows
that sometimes one science may take over and apply the arguments of another; but these are the
exceptions” (Owen 1986, 213).
35
An assumption that is explicitly acknowledged: “[MA] will not only use one science to get
knowledge about another, but also claim that the second provides a necessary part of any valid
justification of the principles of the first” (Nussbaum 1978, 112). The issue of interest to us is not so
much her concern that the MA seems to constitute an unwarranted mixture of sciences, but the
assumption lying behind that concern, that the different natural investigations will necessarily be
contributions to separate sciences.
36
Kung 1982, 65–76. Barnes’ report of this dispute (Barnes 1993, 131) is somewhat misleading. Cf.
Wilson 2000, 47–52, who endorses Kung’s response.
37
Ibid., 68. In fact, MA, after noting that the different movements found in different kinds of animals
and their causes have been studied in other places (an apparent reference to the investigations
reported in IA and perhaps HA), refers to itself as investigating “in general the common cause of being
moved with any motion whatsoever” (698a4-5).
38
This is not an uncommon assumption; for example, note in the following remark the assumption
that for Aristotle, physics and biology are distinct sciences: “The same thing may be treated by more
than one science. For example, man is treated qua a thing-that-is by first philosophy, qua a substance
whose principle of motion and rest is in itself by physics, and qua animal by biology” (McKirahan
1992, 63).
39
Problems would, of course, still arise if there were a serious role played by moral psychology or
geometry in a natural treatise, but for now I will restrict my focus to the role of principles from Ph.
viii in MA, and its discussions about the movement of the whole cosmos. Interestingly, in the
sixteenth and seventeenth centuries, a number of researchers in the Aristotelian tradition built on
the various hints toward geometric models of limb motion in MA and IA to work on creating a sub-
subaltern science of biomechanics – that is, applying geometrical mechanics beyond the realm of
simple machines to animal locomotion. Examples of such works include Hieronymus Fabricius ab
Aquapendente, De motu locali animalium secundum totum and an unpublished work by William
Harvey, recently edited and published also with the title De motu locali animalium (which appears to
be the title of the first chapter rather than of the entire work). For details about this creative
extension of Aristotelianism, see Distelzweig 2014a and 2014b.
40
This is simply to be skewered by the other horn of the dilemma posed on p. 111.
41
The verb used here for ‘apply,’ ἐφαρμόττειν, often expresses the relationship between a principle of
demonstration that is more ‘abstract’ or more general and one that is more ‘concrete’ or specific (cf.
APo. i.7, 75b42-76a7; i.9, 76a22-25; i.32, 88a32, 33). Passages in which the term seems to be playing
a similar semantic role (Cael. iv.2, 308b2; Resp. 7, 474a10-11; Ph. iii.1, 201b14; v.4, 228b25; Pol. iii.1,
1275a33-35, b32; iii.2, 1276b25) suggest translating it ‘to apply.’ Especially important for fixing its
meaning is MA 1, 698a11-14: “One should grasp this [the existence and nature of the first unmoved
mover] not only by a universal account, but also in the presence of the particulars and perceptibles,
on account of which we search for the universal accounts, and to which (ἐφ’ ὧν) we think these
accounts should apply (ἐφαρμόττειν).” Though it cannot be reflected in a readable translation, the
prepositional phrase ἐφ’ ὧν and the infinitive ἐφαρμόττειν are contiguous in the Greek, indicating
that the preposition is simply emphasizing the verb’s prefix. Given its use in the passages we are
looking at, I think Nussbaum’s ‘to harmonize’ is not precise enough; Aristotle likely has in mind the
application of universal accounts via proof to the particulars, since it was in order to understand
them that the search for the universal began.
42
At one level; Aristotle does, of course, see significant differences in the eggs of birds and fish (cf. GA
ii.1, 732b1-8; iii.1, 749a10-34).
43
The thought here is followed out in much more detail in APo. i.13, 78b34-79a16.
44
Elsewhere I have argued that three passages in Aristotle (in the opening and closing chapters of the
Parva naturalia – Sens. 1, 436a18-b2, Resp. 21, 480b22-31 – and PA ii.7, 653a1-10, suggest that he did
see the relationship between medicine (a productive science) and the science of nature on the
subordinate science model (see Lennox 2005, 55–72, esp. 66–68).
Since the nature [of a natural thing] is a source of motion and change,
and our methodical inquiry is concerned with nature, [the question]
what is motion must not escape our notice; for necessarily when we
are ignorant of this we are also ignorant of nature.
(Ph. iii.1, 200b12-15)1
So then, about the being of sensible things, what it is has been stated,
in the methodical inquiry of the naturalists about the matter, and later
about the <being> in accordance with activity; but since the investi-
gation is whether there is some being besides the sensible beings that is
immoveable and eternal or not, and if there is, what it is, we must first
consider what is said by others.
(Metaph. Μ.1, 1076a8-14)2
Chapter Summary. Part i made a case for seeing the Posterior Analytics
ii in a new light. In that new light, it does not provide poor answers to
questions people bring to it about the path by which inquiry takes us from
perceptual experience to first principles of demonstration. Rather it pro-
vides very good answers to questions about what questions ought to shape
any inquiry aimed at knowledge and about how the answers to those
questions are related to one another – it provides, that is, an erotetic
framework for inquiry. Many of the norms that guide scientific inquiries,
however, are domain-specific, and are not to be found in APo., though they
are best seen as different ways of specifying the framework provided by the
Analytics. Part ii of this book tests these ideas in the context of specific
1
Ἐπεὶ δ’ ἡ φύσις μέν ἐστιν ἀρχὴ κινήσεως καὶ μεταβολῆς, ἡ δὲ μέθοδος ἡμῖν περὶ φύσεώς ἐστι, δεῖ μὴ
λανθάνειν τί ἐστι κίνησις· ἀναγκαῖον γὰρ ἀγνοουμένης αὐτῆς ἀγνοεῖσθαι καὶ τὴν φύσιν.
2
Περὶ μὲν οὖν τῆς τῶν αἰσθητῶν οὐσίας εἴρηται τίς ἐστιν, ἐν μὲν τῇ μεθόδῳ τῇ τῶν φυσικῶν περὶ τῆς
ὕλης, ὕστερον δὲ περὶ τῆς κατ’ ἐνέργειαν· ἐπεὶ δ’ ἡ σκέψις ἐστὶ πότερον ἔστι τις παρὰ τὰς αἰσθητὰς
οὐσίας ἀκίνητος καὶ ἀΐδιος ἢ οὐκ ἔστι, καὶ εἰ ἔστι τίς ἐστι, πρῶτον τὰ παρὰ τῶν ἄλλων λεγόμενα
θεωρητέον.
121
5.1 Introduction
As I will discuss in some detail shortly, the opening sentences of the Physics
strongly suggest that the erotetic framework for inquiry in the pursuit of
scientific knowledge presented in the second book of the Posterior Analytics
shapes Aristotle’s views about natural inquiry. Assuming that is the case,
there are a number of features one would expect to find. In light of the
framework discussed in Chapter 2, there ought to be two intimately related
goals of natural inquiry: knowledge of what natural beings are, ideally to be
formulated in definitions; and knowledge of causes used to demonstrate
why natural beings have the necessary but nonessential attributes that they
do; and fact-establishing stages on the way to each of those goals.3 The
intimate relationship between these two paths of inquiry stems from
Aristotle’s conviction that the essences of things are, in various ways,
causally responsible for their nonessential features – if one has scientific
knowledge of what something is, one will thereby be able to explain why it
has the other nonaccidental features it has.
Nevertheless, notwithstanding the generality of this framework, I have
been arguing that Aristotle sees different subject matters or domains as
governed by norms that are quite specific to them. This specificity derives,
from (at least) three sources: (i) differences in the objects being investi-
gated, (ii) differences in our epistemic access to those objects, and (iii)
differences in the perspective we take on those objects. Thus, while the
Posterior Analytics provides what I have called an erotetic framework for
scientific inquiry, it does not provide the domain-specific norms needed to
guide actual research.
3
See APo. ii.1, 2. The view is, of course, much more complicated than this brief summation suggests.
For detailed discussion of the complications, see Charles 2000, 2010; Ferejohn 2013; Lennox 2005.
4
Cf. APo. i.2, 71b33-72a5; EN i.4, 1095b2-4.
5
For the sake of consistency, I will render κίνησις as ‘motion’ and μεταβολή as ‘change,’ but ‘motion’
will thus be referring to a variety of processes besides local motion. Cf. Ph. iii.1 and v.1–3 for
Aristotle’s distinction between μεταβολή and κίνησις and discussions of the various kinds of change.
6
I will, to some extent, be assuming the results of Lennox 2008a here, which is almost entirely devoted
to an interpretation of this chapter.
7
Reading λέγεται at 193b22. Later in the chapter, at 194a12, we are told that nature is διχῶς, τό τε εἶδος
καὶ ἡ ὕλη, a result Aristotle argues for from 193a9-b5, after which he makes the case for the form being
a composite’s nature ‘more than’ the matter.
8
I make the case for this understanding of the argument in Lennox 2008a.
9
That is the precise language he uses, but I assume he means the natures of natural things (i.e., the
forms of natural things).
10
Cf. Metaph. Z.10, 1035a4-5, Z.11, 1036b21-31.
11
This passage is controversial, of course. For a review of one central controversy with a valuable
discussion of the parallels I am relying on between Metaph. Ε.1 and Ph. ii.2, see Peramatzis 2011,
Appendix 1, 312–316; cf. Bostock 1994, 162–165; Gill 1989, 130–134.
12
Cf. Mete. iv.12, 390a10-12: “Everything is defined by its function. For things able to perform their
own function are truly what they are in each case, for example, an eye if it sees; whereas one that
cannot is an eye in name only, for instance, a dead eye or one made out of stone.”
13
The interpretation of this chapter is defended more fully in Lennox 2008a; for a different reading see
Peramatzis 2011, 59–64, 130–137; our disagreements are highlighted at 62 n. 5, 65, and 130.
14
Aristotle identifies the two natures in relation to one another in various ways: at Ph. ii.1, 193a28-31,
the contrast is of the primary underlying matter with the shape and form according to the account
(cf. Ph. ii.1, 193b3-5); at PA i.1, 640b28-29 it is of the nature in respect of form with the material
nature; at PA iii.2, 663b22-24, the nature in respect of the account is contrasted with the necessary
nature; and at GA iv.4, 770b9-18, the nature in respect of the form is contrasted with the nature in
respect of the matter.
15
Here ἐπιστήμη is being used in a ‘big tent’ sense to include the crafts.
16
Some other options for translation are Waterfield (1996): “Questions remain – in what sense is
anything separable? What is it that is separable? – but it is the job of first philosophy to answer
them”; Hardie and Gaye (1930): “The mode of existence and essence of the separable it is the
business of the primary type of philosophy to define”; Pellegrin (2000): “Ce qu’il en est du separable
e ce qu’il est, c’est l’affaire de la philosophie première de le determiner”; Angioni (2006): “Mas
delimiter como se comporta e o que é o separável é tarefa da filosofia primeira.” Certainly, the nature
of the separability of the objects of mathematics, Platonic Forms, the forms of natural substances,
and a first unmoved mover is a theme that runs through the entirety of the Metaphysics.
17
This reading is defended in Lennox 2008a. It is usually translated ‘natural and perishable forms,’ but
since Aristotle consistently denies that forms (as opposed to composites) are perishable, I prefer to
follow the Λ mss. and translate as above.
18
καὶ ἐπεὶ ἡ φύσις διττή, ἡ μὲν ὡς ὕλη ἡ δ’ ὡς μορφή, τέλος δ’ αὕτη, τοῦ τέλους δὲ ἕνεκα τἆλλα, αὕτη
ἂν εἵη ἡ αἰτία, ἡ οὗ ἕνεκα.
19
We will return to this passage in Chapter 10.
20
Cf. Lennox 2009, 187–214, and Leunissen 2009, 215–237.
23
Cf. 195a29-30.
24
Cf. Metaph. H.4, 1044a32-b20. Importantly, in this passage, after citing the four causes and saying
that in the study of generated natural substances it is necessary to discuss all the causes, he states that
we should proceed by ‘another logos’ when discussing the natural and eternal substances (1044b2-6)
and that certain natural processes, such as eclipses, perhaps do not have final causes (1044b9-13). In
fact, no final cause is ever provided in the Meteorology for any of the objects and processes discussed
there. Cf. Wilson 2013, 93–107. Despite this, Scharle 2015 makes a valiant effort to argue that “there
is nonetheless an important explanatory role for teleological processes to play in the treatise [i.e.,
Meteorologica]” (79).
25
Cf. PA i.5, 644b22-24: “As many substantial beings as are constituted by nature, some are
ungenerated and imperishable for all eternity, while others participate in generation and perishing.”
See too Cael. i.1, 268a1-6, Metaph. H.4, 1044b3-6.
26
The discussion in Metaph. E.1 depends on comparing the objects of mathematics, natural science,
and a still hypothetical first philosophy along two axes, separability and changeability, whereas here
separability is not discussed. Recall that in Ph. ii.2 he refers that topic to first philosophy.
27
One central conclusion of Ph. viii is that there must be such a thing, but the investigation of its
nature is the topic of Metaph. Λ.
28
There are a number of puzzling aspects of this remark I will not be able to deal with here; but see
Ross 1936, 527–528; Philoponus (in Ph. 303.8-306.12); Simplicius (in Ph. 366.30-369.13).
29
The details of the defense of the teleological character of natural processes and their outcomes in this
chapter are hotly disputed. My primary focus here is not to enter into those disputes, but in recent
years there have been a number of publications that not only enter into these disputes but do an
excellent job of surveying the issues and the various alternative interpretations on offer: Gotthelf
2012, esp. chs. 1–4; Johnson 2005; Quarantotto 2005, esp. chs. 3–4; Leunissen 2010, esp. ch. 1.
30
Cf. de An. ii.4, 415b8-28.
31
For the Greek, see p. 129, note 18.
32
Cf. Charles 2012, 230–232; Quarantotto 2005, 178–212. 33 Cf. PA i.1, 639b12-16; 642a14-18.
34
To put the point in a less epistemic mode: the nature of the goal determines what the appropriate
materials and patterns of changes are to be, and that causal relationship needs to be reflected in our
accounts of natural things.
35
This example should be compared to PA i.1, 642a9-13: “for just as, since it is necessary for the axe to
split, it is a necessity that it be hard, and if hard, then either bronze or iron, so too, since the body is
an instrument (for each of the parts is for the sake of something, and likewise too the whole), it is
therefore a necessity that it be a certain sort and from such materials, if this [end] is to be.” The
differences are as important as the similarities: in PA i.1 the analogue to the living body, the axe, is
explicitly mentioned as the object being defined. In Ph. ii.9, however, it is unclear whether the
definiendum is the saw or the activity of sawing. The example begins with reference to defining the
work of sawing (perhaps analogous to defining respiration or sleep); however, this might be
a reference to the form of the saw, since it is that work that then determines that the saw must
have teeth and that they must be made of a specific material. And by listing two material options, the
PA example makes it clear that it is the dispositional property (hardness) of the material that is
conditionally necessary, not a specific uniform material.
36
The term methodos appears four more times in the Physics – once in iii.5, 204a34-b11, and three times
in book viii (251a5-8, 253a32-b10, 261a27-31). As one would expect if the position I am defending is
correct, all are concerned with delimiting the approach to the topic at hand so that it is appropriate
to natural inquiry, as opposed to mathematics or first philosophy.
37
These are chapters that have received a great deal of attention. For a clear review of the issues and
options, see Coope 2009. For defenses of various options, see Gill 1989, ch. 6; Graham 1988;
Heinaman 1994; Kosman 1969; Kostman 1987.
38
If they exist at all, that is. Aristotle denies the existence of void and actual infinity. This is one way in
which the erotetic framework of APo. ii structures the inquiries in Ph. iii–iv – the inquiries proceed
by first considering if the purported object of inquiry exists and then, if it does, moving on to
consider what it is. Cf. iii.4, 202b35-36 (re infinity); iv.1, 208a27-30, iv.4, 210b32 (re place); iv.6,
213a12-14 (re void); iv.10, 217b29-32 (re time).
39
I remain neutral on how the various books of our Physics came to be related to one another. Relevant
to answering that question are the following points: (i) The opening of book iii both looks back to
the earlier books, noting that the account of nature in those books assumes that there is motion in
nature, but it remains to investigate what it is; and it then previews all of the topics covered in books
iii and iv; (ii) book v can be viewed as a natural appendix to the general discussion of change; and
(iii) since all of the topics in books iii and iv arose out of the undefended assumption that motion
‘seems’ (δοκεῖ, 200b16) to be continuous, the discussion of continuity in book vi fills an obvious gap
in the earlier discussion as well.
40
I am here distancing myself from a thesis defended by David Bostock (Bostock 2006, 4–7 and
passim) that Aristotle’s Physics is largely an exercise in ‘conceptual analysis’ and that Aristotle
regularly confuses such an exercise with empirical inquiry. (See Lennox 2008b.)
41
And as is clear from the literature referenced in note 37, Aristotle’s account of what motion is is
among the most subtle and complicated discussions in the entire corpus. This is one of many cases in
the Aristotelian corpus where the distinction between an inquiry into τὸ ὅτι and τὸ διότι, a key
feature of the erotetic framework provided by APo. ii, is of fundamental importance.
1
οὗτος μὲν οὖν ὁ λόγος καθόλου λίαν καὶ κενός· οἱ γὰρ μὴ ἐκ τῶν οἰκείων ἀρχῶν λόγοι
κενοί, ἀλλὰ δοκοῦσιν εἶναι τῶν πραγμάτων οὐκ ὄντες. οἱ γὰρ ἐκ τῶν ἀρχῶν τῶν
γεωμετρικῶν γεωμετρικοί, ὁμοίως δὲ καὶ ἐπὶ τῶν ἄλλων· τὸ δὲ κενὸν δοκεῖ μὲν εἶναί τι, ἔστι δ’ οὐθέν.
141
6.1 Introduction
In Chapters 3 and 5, I surveyed the ways in which Aristotle deploys the
concept methodos generally and in the context of the general investigation
of nature. There are a number of more or less invariant features of its use,
as we saw: (i) It is nearly always invoked in the context of inquiry; (ii) it
refers specifically to how one ought to proceed toward the goal of
knowledge, or to an actual inquiry viewed from the perspective of its
investigative methods; and (iii) these ‘methodic’ norms are invariably
domain-specific. These three features of the concept’s use lead to an
interesting distinction in its reference: often it refers to the methods
that are or ought to be used in a specific domain of investigation; but
occasionally it refers to the domain of investigation itself, looked at from
a methodological perspective. In this chapter and the next, I look in detail
at two distinct introductions to two distinct inquiries: De partibus ani-
malium i, generally agreed to be a philosophical introduction to
Aristotle’s various investigations of animals; and De anima i.1, the intro-
duction to Aristotle’s investigation of the soul. The concept methodos
plays a critical role in both introductions, and attending to that role will
provide additional insight into the ways in which different epistemic/
methodological norms govern different inquiries.
knowledge of the subject, but in a way that makes explicit that a person can have the general form of
discernment without having what would count as knowledge of the subject.
6
I here flag a number of features of these passages that make it somewhat difficult to nail down precisely
what sort of expertise he is referring to. The PA i.1 passage appears to contrast a transdisciplinary
discriminative ability with one limited to a specific domain, while the Pol. iii.11 passage distinguishes
three sorts of physician, one of which apparently does not actually practice medicine and yet is
discerning enough about the practice to be called a doctor. Tentatively, I suggest that the third sort of
doctor is akin to what today might be called a philosopher of medicine; and the person referred to in
the PA i.1 passage as discerning about the inquiry into nature is akin to a philosopher of science.
Given the almost exclusive focus on animals of PA i, we might, as David Balme (1992, 69) does, think
of it as a philosophy of zoology. For a contrasting viewpoint, see Tipton 2014, 19–29.
7
Notice that it is natural inquiry (ἡ περὶ φύσιν ἱστορία) that is said to be in need of certain standards.
I cannot, therefore, agree with David Balme’s assertion that this introduction “makes it clear that he
[Aristotle] is not setting out to discuss scientific method . . . [nor] considering how to arrive at an
explanation . . . but how to judge an explanation when made” (Balme 1992, 69; for a contrasting view,
see Le Blond 1945, 52). In fact, as we shall see, many of the norms eventually established in this
discussion are explicitly identified as norms for guiding inquiry along the right lines. Cf. Kullmann
1997, 43–62 (which builds on Kullmann 1974, ch. 3), who sees PA i generally as propaedeutic to, and
as a study of the presuppositions of, the scientific study of animals.
8
Cf. Karbowski 2019, 116–117.
9
Since Aristotle holds that the same objects may be studied from a variety of distinct perspectives, and
that each perspective may be governed by different normative constraints, the subject matter will be
determined intensionally rather than extensionally. As he says in Metaph. M. 3: “It would be best to
study each thing in the following way – take separately that which is not separated, which is what the
arithmetician and geometer do. For human being, qua human being, is one and indivisible, and the
arithmetician takes him as one indivisible thing and then considers whether something is attributed
to man qua indivisible. But the geometer does not study man qua man nor qua indivisible, but qua
solid” (1078a21-26). Note that this in itself is a norm of inquiry. He had previously suggested that
standards of precision will change depending on whether or not one prescinds from considering
objects as having magnitude or undergoing change (1078a9-14).
10
And though I am less sure about this (given the range of uses to which Aristotle puts this word), it
may also stress that the study is theoretical, as opposed to productive or practical, in nature (cf.
Metaph. Ε.1, 1025b18-28 and EN vi.3, 1139a14-22 on this distinction).
11
As noted in Chapter 3, the noun ἡ ὁδός refers to a road, path, or track and is used metaphorically in
much the same way as those English expressions are. The dative form used here often has adverbial
force, conveying the idea of staying on the road to one’s destination, thus my ‘on course.’
12
Though of course the subject matter, and the fact that the investigation is theoretical rather than
practical or productive, will need to be taken into account in articulating those norms and standards.
13
That is, the participle is formed from δείκνυμι, not ἀποδείκνυμι, the verb cognate to ἀπόδειξις
(demonstration). However, later in the chapter Aristotle claims that the goal of the inquiry he has in
mind is a different mode of demonstration, grounded in a different mode of necessity (ἀλλ᾽ ὁ τρόπος
τῆς ἀποδείξεως καὶ τῆς ἀνάγκης ἕτερος; 639b30-640a1).
14
There is a legitimate worry about whether, if Aristotle ends up recommending starting, wherever
possible, with what is universal and common, this could be a norm about how to begin an
investigation, since on Aristotle’s general epistemic view one begins the quest for knowledge with
perceptual experience of particulars. This passage does not conflict with his perceptual foundation-
alism, however. The issue being raised in this passage in PA i.1 is how inquiry into a domain of
experience (say, animals) should proceed. As we will see, Aristotle’s answer (which is a view defended
by many cognitive psychologists today) is that the right way to begin, when possible, is with
categories I refer to as ‘entry-level kinds,’ categories like ‘bird’ and ‘fish’ – and as I am about to
argue, he arrives at that conclusion by reflecting on the practices that led ordinary Greeks to
successfully grasp those general categories. Aristotle is arguing that in most cases, this is the
appropriate first step in inductive inquiry. I thank Victor Caston for pushing me on this point.
As he pointed out, in a way this is the same puzzle that bothers commentators about the claim in Ph.
i.1 that inquiry starts with the undifferentiated universal, and epistemic progress is achieved by
making the appropriate differentiations.
15
There are, of course, variations in things like how long and how often animals sleep, and on whether
they dream or not, but these are not formal differences in sleep per se. On the methodological norms
of the inquiry into sleep, see Code 2015, 11–45.
16
Three notes of caution on this point are necessary: (i) The expression μέγιστον γένος is not used
here; (ii) in PA iv.8 he identifies four μέγιστα γένη of soft-shelled animals – that is, he uses the
expression μέγιστον γένος to refer to sub-kinds of one of the μέγιστα γένη mentioned in HA i.6
(490b10); and (iii) even in HA the term is used rarely, and never in a rigid and fixed way. See
Gotthelf 2012, 293–306.
17
For more detailed discussion, see Balme 1992, 114–119; Lennox 2001a, 158–165.
18
Notice that this also links the discussion of division with the earlier argument that only a functioning
body or part is the body or a part of an animal (640b30-641a6).
Winged Flyer
Sheathed Non-sheathed
Membranous- Membranous-
winged winged
Nonsheathed Nonsheathed
Diptera Tetraptera
19
I have discussed these in Lennox 2005, 87–100; a different, more ‘theory-laden’ reading of the same
text can be found in Charles 2000, 312–316.
20
Ἀπορήσειε δ’ ἄν τις (644a12).
21
I am here correcting my published translation in two important respects: first, by explicitly translating
τῶν γενῶν in a17, so it is clear that Aristotle is saying that kinds that differ more and less are brought
together under one kind; second, I now read τοῦτον τὸν τρόπον at a16 as referring forward, and thus
take the γὰρ in the next sentence as ‘explanatory’ or ‘prefatory,’ picking up on τοῦτον.
22
The aporia with which this discussion opens is stated with reference to ‘flyers’ and ‘water dwellers.’
This could be a preliminary (and rough) way of introducing ‘bird’ and ‘fish’ – rough, since Aristotle,
as we have seen, recognizes two kinds of flyers other than birds, and discusses in great detail water
dwellers (such as cephalopods, crustaceans, and cetaceans) that are not fish. Another possibility is
that he is withholding the terms ‘bird’ and ‘fish’ until he has introduced the idea of the appropriate-
ness of uniting into one kind those animals that differ only by more and less.
23
Ἐπεὶ δ’ οὐσίαι μέν εἰσι τὰ ἔσχατα εἴδη (644a23-4).
24
Charles 2000, 312, explicitly treats the καὶ as epexegetic; Gotthelf (2012, 276) follows Balme’s
translation but in his analysis of the passage takes it as “more or less epexegetical” (277). This may
well be right, but I am leaving open the possibility that Aristotle is distinguishing the way the words
‘bird’ and ‘fish’ are used in everyday Greek from the criteria that need to be explicitly identified in
order to determine when we “should speak in common according to kind.” That is, when, in
addition to people speaking of kinds in a well-defined manner, animals meet two other criteria, we
should then identify them as kinds. People may speak in a well-defined way of groups of animals
that do not meet these other two criteria.
25
Throughout I am reading the text of Bekker, not Peck, who for reasons that are not clear to me
makes a number of emendations with no manuscript support.
26
As Gotthelf notes (2012, 277, 299) Aristotle’s formulation allows for two different cases of groups
that embrace forms in the way that kinds do, one in which the kind is recognized but not designated
by a proper name (but perhaps by a ‘name-like expression’), and one in which there is a group that
meets the criteria but has not been recognized.
27
Charles 2000, 313–316, worries about this problem as well, but resolves it by reading ideas from
chapter 5 – in which Aristotle is discussing the ways in which, once we know the proper function of
a part, we can explain its presence teleologically – back into this chapter, which I take to be aimed at
answering “entry level” questions (Lennox 2005), questions about how to initiate an inquiry into
animals, and which level of generality is appropriate for different groups. I have an alternative
reading that depends only on evidence from the context and from parallel passages in HA. Cf. the
careful, comparative analysis of our views on this issue in Gotthelf 2012, 275–287, 296–298; and in
the fine dissertation of Keith Bemer (Bemer 2014).
28
Charles 2000, ch. 2.
JAÇANA
SHAG (WALKING ON
(SWIMMING) FLOATING PLANTS)
COOT
(SWIMMING)
Figure 6.2a Forms of bird feet, indicating differentiation for different ways of life.
29
Lennox 1987a (reprinted in Lennox 2001b, ch. 1), 2005. The relationship between Aristotle’s use of
the expression ‘the more and the less’ and Plato’s use of it in the Philebus is important and needs
further exploration. There is a more extensive discussion of these different sorts of sameness or
identity in the first chapter of HA.
30
Lest the reader think the hummingbird beak is an exception: it is actually structurally just like other
beaks, though it is longer, thinner, and more flexible than most. The hummingbird opens its beak
and extends a long tongue with grooves and a feathery tip to extract the nectar from flowers.
TOUCAN
HAWK
SHOVELER
PETREL CURLEW
SPOONBILL
COCKATOO
GROSBEAK
Figure 6.2b Forms of bird beaks, indicating differentiation for different ways of life.
32
Eledone cirrhosa, the ‘lesser’ or ‘curled’ octopus. See figure 6.3a.
33
Aristotle’s account and those found on contemporary websites are very similar: “Their [cuttlefish]
preferred diet is crabs or fish, and when it [sic] is close enough it opens apart its eight arms and out
shoots two deceptively long feeding tentacles. On the end of each is a pad covered in suckers that
grasp hold of the prey” (www.tonmo.com/articles/basiccuttlefish.php). One principal difference is
that Aristotle consistently mentions two functions for these tentacles, feeding and mooring during
storms, while modern texts I have consulted only mention feeding. One audience member in
attendance at a lecture in which I mentioned this discrepancy claimed to have repeatedly observed
the mooring behavior described by Aristotle while snorkeling in the Mediterranean.
Figure 6.3a Eledone cirrhosa (the ‘lesser’ or ‘curled’ octopus) with a single row of
suckers on each arm.
34
An important issue that I cannot properly address here, but which I want to flag, is the difficulty of
distinguishing between Aristotle’s actual methods of empirical investigation and his written reports
of the results of those investigations. When historians of science investigate many post-Renaissance
natural philosophers/scientists, they have access to such things as field notes, letters, laboratory
notebooks, autobiographical accounts of investigations, and so on. This allows the historian to make
a principled distinction between the published book or journal article, on the one hand, and the day-
to-day struggles of information collection, analysis, interpretation, and integration of new findings
with old, on the other – aspects of science often suppressed in the publications that result. There is
no clear demarcation of this kind in the extant Aristotelian corpus – certain works seem more like
a published book, others more like an investigator’s notes on an ongoing investigation. (The ancient
catalogues mention a number of works, no longer extant, that may well have reflected early stages of
investigation: Divisions [seventeen books], Dissections [eight books], Selections from Dissections [one
book] along with the [mostly?] post-Aristotelian Problems.) This poses an intriguing meta-level
problem for an inquiry like the one in which I am currently engaged – is it possible to distinguish, in
Aristotle’s texts, norms for presenting the results of an inquiry from norms for carrying out an
empirical inquiry? This footnote is intended to signal only an awareness of the problem, not
a solution to it.
36
We will revisit these issues in Chapter 8. For a full discussion of the passages in Cael. in which
Aristotle expresses the epistemic limitations imposed on such research, see Bolton 2009, 51–82.
Falcon 2005 does a fine job of pointing out how unusual the views defended in Cael. were, which
helps explain the widespread skepticism with which they were greeted.
37
A claim repeated at GA i.1, 715a8-9: ὅ τε γὰρ λόγος καὶ τὸ οὗ ἕνεκα ὡς τέλος ταὐτόν.
41
There is a considerable literature around the question of the nature of this priority. For a recent
survey of the issues at stake that takes a somewhat different view from that defended here, see
Leunissen 2009, 99–108.
42
Complicating the picture presented here, Aristotle explicitly puts off the causal investigation of the
parts related to generation in PA until the first book of GA. Cf. PA ii.7, 653b13-18; iv.10, 689a4-20;
GA i.1, 715a11-18.
43
Of note in this review is that ‘form’ is not mentioned – in its place is ὁ λόγος τῆς οὐσίας, the account
of the substantial being.
44
To this discussion of how the elements, uniform parts, and nonuniform parts are related, compare
PA ii.1, 646a13-b28.
45
David Balme in his translation brackets the GA i.1 introduction and in his notes says that it may be
post-Aristotelian (1992, 127). His only grounds are that it is ‘stylized’ and is ‘so different’ from the PA
i introduction. That is true: but there are striking differences between the introductions of every
work in the corpus; and PA i, as Balme himself stresses, is not the first book of the investigation of
animal parts but a philosophical discussion about how the investigation of living nature in its
entirety ought to be carried out. I will therefore treat the opening lines of GA as a genuinely
Aristotelian introduction to the study of generation, comparable to the opening pages of PA ii, to
which it clearly alludes.
46
See section 3.4.2.
48
And recall the way in which that set of questions is introduced by first reminding the reader of the
Analytics framework: “For since our study is for the sake of knowing, and we do not think we know
each thing until we grasp τὸ διὰ τί about each thing . . . ’ (Ph. ii.3, 194b17-19); see the discussion in
Section 5.2.1.
49
For reasons of length I have limited the discussion to the three initial questions in PA i.1. It is already
clear, however, from looking at the norms recommended for answering those questions, that Aristotle
also develops a series of norms for the proper use of division in a zoological context (PA i.2-4, 642b5-
644a11); for applying necessity to teleological processes (639b22-640a9; 642a1-13); for the study of the
material and formal nature of living things, especially in light of the formal nature of a living thing
being its soul (640b4-641a32); and for delimiting the parts of the soul to be investigated by the natural
scientist (641a33-641b23). Finally, in chapter 5, there are a number of complex norms about how best to
integrate the methodology of division that dominates the discussion in chapters 2–4 with the
teleological and hylomorphic perspective articulated in chapter 1.
50
I depart from the translation in Lennox 2001a in two respects: I now translate ζήτησις consistently as
inquiry, and I now think that the use of τὸ κάλον in this text has a decidedly aesthetic aspect. I owe
the latter change of heart to reflecting on conversations with Richard Kraut.
The Soul
One Subject, Two Methods?
It remains then for reason alone to enter from without and alone to be
divine; for bodily actuality does not participate at all in its actuality.
(GA ii.3, 736b27-29)1
7.1 Introduction
Let us return, yet again, to the beginning of the Meteorology. It will be
recalled that Aristotle recites an appropriate course of study for an investi-
gator of nature (338b20-339a10) in order to locate the investigation of
meteorological phenomena. For the topic of this chapter, two things are
significant about this course of study: (i) nothing corresponding to the
subject of the De anima, the investigation of the soul, is mentioned; and (ii)
the course of investigation of nature is said to culminate in a study of
animals and plants. It is sometimes assumed that the investigation of the
1
λείπεται δὴ τὸν νοῦν μόνον θύραθεν ἐπεισιέναι καὶ θεῖον εἶναι μόνον· οὐθὲν γὰρ αὐτοῦ
τῇ ἐνεργείᾳ κοινωνεῖ <ἡ> σωματικὴ ἐνέργεια.
174
8
The latter passage states that “it is necessary to suppose that they (the spheres of the stars and planets)
participate in action and life.”
9
I am going to use ‘starting point’ as a translation for ἀρχή here, since at the very least Aristotle is not
merely saying that the concept of soul is a scientific first principle. The possessors of the ἀρχή here are
living things. ‘Starting point’ is sufficiently generic that it can refer both to the principles of a science and
the sources of motion or being in natural things, and by adopting it I can allow for the possibility that
Aristotle is either thinking of soul as an ἀρχή in both senses, or intentionally suggesting that such
questions should be left open at the beginning of the inquiry. Of course, in de An. ii.4 Aristotle argues
that the soul is a cause in three ways: (i) as source of motion, (ii) as the source of being for living things,
and (iii) as that for the sake of which the instrumental body exists (415b8-27).
10
Given that, according to Metaph. Δ.1, there are perhaps six or seven distinct sorts of ἀρχή, it is
perhaps surprising that, despite all the other differences, all of the translators referred to earlier settle
on ‘principle.’
11
Metaph. Δ.4, 1015a12-13.
15
This may seem an odd question. However, as one reads through his careful critique of previous
accounts of the soul in de An. i, it becomes clear that, apart from a number of materialist theories, the
other alternatives are that the soul is a ‘self-moving number’ (404b30), a ‘ratio,’ or a ‘harmony’ of
some sort (407b28ff; i.e., in the category of quantity), or that it is movement of some sort (see note
17). In this context, determining how to categorize the soul is an important first step and in fact, as
I’ll discuss shortly, this is a specific application of a general norm of inquiry specified at APo. ii.8.
16
Aristotle comments on the relationship between these terms at Metaph. Θ.3, 1047a30-1047b2, also
discussed in Blair 1992, Menn 1994, and Graham 1995. In this same passage, Aristotle contrasts
investigating being from a categorical perspective and from the perspective of potency, fulfillment,
and function – the two perspectives reflected in Questions 2 and 3 about the ontological status of the
soul in de An. i.1.
That is, he begins by identifying the category in which the soul is found,
but then notes that there are a number of senses in which something can be
in the category of substantial being. Note the precise correspondence with
the first metaphysical question of de An. i.1, 402a24-25, about “whether it is
a this and a substantial being” (τόδε τι καὶ οὐσία) – Aristotle points to the
role of substantial being qua form as the source of ‘thisness.’ He later goes
on to argue against seeing the soul as either the matter or the compound,
and concludes:
The soul must, then, be substantial being as form (οὐσίαν ὡς εἶδος) of a
natural body that has life potentially. (de An. ii.1, 412a19-21)
That is, the very first positive step in providing his own account of the soul,
in opposition to the views he has reviewed and rejected in book i, is to
answer the first metaphysical question, about the soul’s categorial status,
presented in his methodological introduction. And, by identifying it as
belonging in the category of substantial being in the sense of form, he has
affirmed that it is substance in the sense of that which makes the ensouled
being a this, a τόδε τι.
17
Soul as a self-moving number, perhaps a view held by Xenocrates, is mentioned at 404b27-30 and
critically examined in chapter 4 (408b32-409a31); and the view that it is the οὐσία of soul to move
itself is noted at 406a17, various versions of which are critically examined in chapter 3 (406a16-b25).
At i.4, he argues that the soul, properly speaking, cannot move, which leads him to point out at
408b30-32 that one implication of this is that it cannot move itself.
This idea of the soul as the living body’s fully realized ability to perform a
kind of activity accounts for a vast range of facts about ensouled beings,
such as that they remain fully capable of the rapid mobilization of coordin-
ated actions needed to, say, successfully elude a predator, even while they
are at rest and doing no such thing. A living thing in a state of rest is neither
merely a body capable of life nor is it fully engaged in living, and Aristotle’s
‘first’ fulfillment aims to capture this important and fundamental fact
about life as it had not been captured previously, and has rarely been since.
This opening paragraph, then, is following precisely the normative
demands of de An. i.1 about which questions ought to be answered first in
an inquiry into the nature of the soul. Questions about its categorical and
modal status are among the first mentioned, but there are more to come.
Aristotle next raises questions about whether the soul has parts and, if so,
in what sense;19 about whether or not there is a general account of soul as
there is of animal;20 whether there are many souls or just many parts of
soul, and in either case, whether there is a proper order in which they are to
be studied and whether they differ in form or in kind from each other; and
whether we should investigate the parts first or their functions, or perhaps
the objects correlated with each function. Every one of these questions is
stated in the form of a norm (σκεπτέον, 402b1; εὐλαβητέον, 402b5), and
each plays an important role in structuring the inquiry to follow in de An.
ii–iii. As noted, Aristotle refers to the account of the soul with which ii.1
opens as “the most common account” and a bit later refers to “first
fulfillment of a natural body with organs (ὀργανικοῦ)” as a statement
common to every soul, noting that he has stated “what soul is generally
19
de An. i.1, 402b1-5: “We must also investigate whether the soul has parts or is without parts, and
whether every soul is alike in form or not; and if they are not alike in form, whether they differ in
form or in kind (εἴδει διαφέρουσιν ἢ γένει). For at the moment, those speaking and inquiring about
the soul seem to investigate the human soul alone.”
20
402b5-9: “And one should take care not to overlook whether there is one account of the soul (εἷς ὁ
λόγος αὐτῆς ἐστι), just as there is of animal, or a different account for each, as there is of horse, dog,
human being, god, the universal ‘animal’ being either nothing or posterior; and similarly even if any
other common thing were predicated.” Note the similarities and differences between this question
as formulated at the beginning of the de An. and the first methodological question raised in PA i.1.
Both are questions about the level of universality at which an investigation is to focus, but here the
question is about whether there is a common account of soul or a different one for each (kind or part)
of soul; and there is also apparently a question raised about the ontological status of whatever it is
that corresponds to the universal account.
22
As one pointed example of the lack of such coordination between PA i and PA ii–iv: in PA i.1 there
are repeated discussions of a special kind of necessity, conditional or hypothetical necessity, that is in
play in investigating animals and how it is distinguished from unqualified necessity; but in PA ii–iv,
though the language of necessity is a constant feature of the explanations on offer, Aristotle never
distinguishes different kinds of necessity. Similarly, while much is made about a distinction between
material and formal natures in PA i, that distinction is rarely mentioned in his actual study of animal
parts.
23
This recommendation has parallels in a number of other texts; for example, HA i.6, 491a9-14: “After
[we have grasped the differences and attributes of every animal] we must attempt to discover their
causes. For to pursue the study in this way is natural, once there is a body of systematic information
[ἱστορία] about each kind, since from these both the facts to be demonstrated and the principles
from which demonstration ought to be performed become apparent.” Here the focus is on finding
causes after a study of differences and attributes has been carried out, but from the standpoint of
APo. ii this is a matter of focus and emphasis, since scientific definitions identify causes. Compare
APr. i.30, 46a3-28.
24
This may be an example of “knowing that it is accidentally” (APo. ii.8, 93a24-25), in which case,
Aristotle tells us, “necessarily we have no fix on what it is, for we do not know that it is” (a25-26).
25
Though I am not sure, it appears that Philoponus may have seen the connection that I am
discussing; cf. in de An. 44.22–26 Hayduck.
26
Hicks 1907, 198.
29
See Miller 2012, 306–339 for a good review of the options.
30
Hicks 1907: “conjoined with”; Hamlyn 1968, Smith 1931: “involve”; Irwin and Fine 1995: “require”.
31
Reading with mss. C and E. For reasons that are opaque to me, Ross defends ἔνυλοι on the basis of
the assertion, at Metaph. Ζ.7, 1033a4, that “the bronze circle has matter in the logos.”
32
As I noted earlier, de An. ii.2 asserts that theoretical intellect seems to be a different kind of soul, one
that can be separated, “just as the eternal from the perishable” (413b24-27). Cf. Shields 2016,
100–101.
Here Aristotle first clearly distinguishes the inquiry into the soul by itself
from an inquiry into animals (and plants, presumably). The distinction
between distinctive and common activities is clarified later in the passage –
some of the activities considered are shared by all, or practically all,
animals; others, such as respiration (and probably waking and sleeping)
belong to some but not all; still others (youth and old age, life and death)
belong, not just to animals, but to everything with a share in life
(436a12-17).
But he uses ‘common’ (κοινά) in a different sense here as well, and one
we are familiar with – for the activities to be considered in PN are ‘those
that are common to soul and body.’ Consequently, as Thomas Johansen
has pointed out, it is “a point of continuity between the de An. and Sens.
that affections that are common to body and soul are to be approached as
such, in the manner of the physikos” (Johansen 2006, 142). And I agree
wholeheartedly with Johansen’s conclusion that the distinction Aristotle is
making here
33
There are especially helpful discussions of this introduction and the way it distinguishes the inquiries
in PN from the de An. in Morel 2006 and Johansen 2006 in King (ed.) 2006, 121–139, 140–164.
The first chapters of both of these inquiries, then, aim to specify conditions
that need to be met in order for attributes of the soul to be appropriate
subjects for investigation by natural science. That task involves consider-
ation of what it means to study something as a physikos; which in turn
requires a discussion of how the investigations of a physikos, and his objects,
are to be distinguished from other related investigations (and their objects)
with which it may be confused. Aristotle turns to that topic near the end of
de An. i.1.
Aristotle begins by distinguishing between a supposed account of anger
to be given by the physikos and another to be given by the dialektikos.35 The
first would give a statement about the matter, the second would give ‘the
form and the account.’ At this point, he briefly notes that this logos must be
in a certain sort of matter if it is to exist. Using a house as his example, he
now characterizes three distinct accounts, one that talks about a house by
reference to its sheltering function, a second that catalogues the materials
from which it is built, and a third that talks about the form in these
materials for the sake of certain functions (403b3-7). In the end, the account
appropriate for an Aristotelian natural science is the third, the account that
is ‘from both’ (i.e., an account that draws on knowledge of the form
present in the matter for the sake of an end).36
34
There are many other points in this perceptive discussion that support the argument I’m making in
this chapter.
35
403a29-b2. Initially it is surprising that, immediately after he has given a model definition of an
embodied pathos, he introduces such a distinction. But while movement, a source of movement, and a
goal of movement are mentioned in these accounts, and thus objects of such accounts will be
appropriate for natural investigation, a couple more steps are needed before Aristotle can conclude
that the natural scientist’s definition must involve essential reference to both form and matter. There
is an interesting parallel here with the structure of the Physics. The form of the definition given at
403a25-27 is what one might expect based on the discussion of Ph. i. But I argued in Chapter 5 that
Ph. ii enriches the discussion of the ‘principles of change’ in book i with the idea that natural things
have natures in two respects, as form and as matter, and that natural science is engaged in a search for
four different kinds of cause.
36
Translating is tricky. I take the ὁ δὲ at 403b5 to answer to the ὁ μὲν at b4. Thus, when Aristotle asks at
403b7 τίς οὖν ὁ φυσικὸς τούτων, the most natural way to take it is “Which of these, then, is the
natural <logos> (i.e., the logos of the natural investigator)?” Following through on that suggestion,
however, leads one to see why the majority of translators (e.g., Shields 2016) rather assume that ὁ
φυσικός is being used substantively in the same way as it was at 403a29, to refer to the natural
investigator. For on the alternative reading, a natural way to continue translating is: “Is it the <logos>
about the matter, ignoring the logos, or the <logos> about the logos alone? Or is it rather the <logos>
derived from both? And which of these is <the logos> of each of the other two?” In other words, we
must suppose that the meaning of logos in this passage keeps shifting between ‘account’ and ‘formal
aspect of natural thing’. There are problems in either case. In favor of reading logos all the way
through is that it is suggested by the opposition with which the passage begins, and it makes best
sense of ὁ ἐξ ἀμφοῖν (b8-9). (To see the hoops one must jump through otherwise: the normally lean
and textually precise Irwin and Fine translation translates that phrase, “Or is the <real> student of
nature more properly the one who mentions both form and matter?”. They are aware of the problem
with their choice; when they translate ὁ δὲ at 403b5 as ‘someone else,’ a footnote reads: “Or perhaps
‘another account’” (Irwin and Fine 1995, 172 n. 15). I lean toward supposing the definite articles pick
up on logos throughout. I can even imagine Aristotle smiling at the fact that there will be logoi that
ignore the logos, logoi that are of the logos alone, and logoi that derive from each of these other logoi.
37
And it must be separable in a way stronger than in account, since the forms of natural things are “not
being separable, other than in account” (Ph. ii.1, 193b4-5; cf. ii.2, 194b12-13).
Commentators have tried to limit the scope of this claim in various ways.
In “On Aristotle’s Conception of the Soul,” Michael Frede, though with-
out explicitly citing this passage, surely has it in mind when he says:
I think it is with reference to this kind of intellect [active intellect which is
immaterial and not dependent on the body in any way] that Aristotle
sometimes says that not every kind of soul falls within the province of the
study of the natural scientist. (Frede 1992, 105)
Sarah Broadie, however, and I think correctly, notes that when one reads
through this passage in its context, it is hard to imagine that Aristotle has in
mind “an active intellect [that] is not a human intellect, that is not an
39
And of course, not only in recent years. I’ve already had reason to note Philoponus’ (mis)use of the
‘correlatives’ argument, and this passage was of central concern to the Renaissance Italian commen-
tators on PA i since both Averroists and their critics saw its relevance to their debate over the unity
and separability of the intellect, on which see Perfetti 1999, 297–316; Mahoney 1982, 264–282. See
Broadie 1998, 163–176, and Caston 1998, 177–192; cf. Charlton 1987, 408–423; Lennox 1999, 1–16;
and Lennox 2002, 139–145.
40
See Frede 1992, 106; Broadie 1998, 163–166. A full consideration of this topic would need to include
careful consideration of GA ii.3, 736b5-30, a passage that puzzles over how best to understand the
temporal emergence of the different capacities of the soul and concludes with the passage I have
chosen as the epigraph for this chapter, referring to reason entering ‘from without’. Burnyeat asserts
that this may “perfectly well refer . . . to a second potential νοῦς acquired through the agency of a
teacher” (2003, 70 n. 111), a suggestion already hinted at in Charlton 1987, 411–412.
41
On which, see the opening lines of MA 1, 698a1-7.
42
See Charlton 1987, 411; Lennox 1999, 4–5. Mary Louise Gill in her comment on an early draft of
Lennox 2009 challenged me to look at this material, for which I thank her. She also argued that there
was a tension between the two passages, and offered a most ingenious reading of the PA i.1 passage
on the hypothesis that it was an aporetic consideration of a view Aristotle in the end rejects. Part of
her argument rested on Aristotle’s remarks about order being more apparent in the cosmos than in
animals, but the sentiment in that passage is exactly mirrored at De caelo ii.8, 290a29-35. In the end,
since I do not see a significant tension between the PA i.1 text and de An. iii, I do not feel the need to
press a different interpretation.
7.6 Conclusion
In the end, de An. i.1 leaves the issue of whether natural science studies all
soul or not undecided; but it is quite clear that if there is a capacity of the
soul that is not to be studied by natural science, it is reason (νοῦς). That
question, like many others raised in this preliminary, problem-setting
chapter, is not there answered. But some advances are made. First, we
learn something about the nature of soul from a study of its pathê, namely
that most, if not all, of its attributes are associated with the body and are
not attributes distinctive to the soul, meaning that neither the approach of
the traditional dialectician nor that of the traditional (materialist) natural
scientist alone can study the soul. There needs to be a way of inquiring into
the soul modeled on the form of inquiry defended in Ph. ii, one that
studies teleologically unified composites of matter and form. Whether reason
can be studied in this way remains, however, an open question in book ii,
well after Aristotle has presented his own theoretical account of the soul as
the first fulfillment of a natural, instrumental body; and however mysteri-
ous the discussion of nous in book iii.3–6 is, it gives us every reason to
think he found the question of the proper way to inquire into its essence a
profoundly difficult one.
PA i.1, on the other hand, is straightforward on the issue of whether such
an inquiry should be considered a natural one. Of the various sorts of
change that are the proper subject of natural science, the capacity for
thought is not a proper source or principle for any of them. Neither it
nor its activities are a proper subject of natural investigation. It is not a
nature (i.e., not an inherent source or principle of natural motion). As far as
one can tell from that passage, however, its activities could well be
Chapter Summary. In this chapter and the next I explore the ways in
which Aristotle uses the results of one natural inquiry as starting points for
inquiry in another area. This is a practice in which Aristotle routinely
indulges, and yet there is, at least prima facie, a problem with him doing so,
given his views about how we arrive at first principles and the propriety of
those principles to specific domains. In APo. i.7–13, he allows that geomet-
ric and arithmetic premises can be used in a range of fields of inquiry he
refers to as ‘subalternate’ branches of mathematics, which he elsewhere
describes as ‘the more natural of the mathematical sciences’ (i.e., optics,
astronomy, mechanics, and harmonics). However, whether and, if so, how
this practice might apply within the science of nature is never explicitly
addressed. In this chapter, I address this question by exploring the depend-
ence of Aristotle’s discussion of the application of the concepts ‘right’ and
‘left’ to the motions of the heavens on his discussion of directional dimen-
sions in De incessu animalium.
8.1 Introduction
In the previous two chapters we have seen that, though the inquiries into
animals and into the soul are, in whole or in part, natural inquiries
structured according to the erotetic framework of APo. ii and the causal
framework of Ph. i–ii, they were each governed by norms specific to their
1
Ἀρχὴ δὲ τῆς σκέψεως ὑποθεμένοις οἷς εἰώθαμεν χρῆσθαι πολλάκις πρὸς τὴν μέθοδον τὴν φυσικήν,
λαβόντες τὰ τοῦτον ἔχοντα τὸν τρόπον ἐν πᾶσι τοῖς τῆς φύσεως ἔργοις.
200
2
Though it is not the primary focus of this chapter, the opening lines of the De caelo raise questions
that are directly relevant to the topic of this book. There is no connecting particle in the first
sentence, and the way in which the science of nature is identified is not based on, nor is there any
reference to, anything said in the Physics about the objects of natural investigation. Here ‘body’ is
simply treated as magnitude in three dimensions, and the principles eventually identified are
principles of magnitude. And while continuity is mentioned at the beginning of Physics iii, it is
mentioned simply as a property of motion that points to the need to discuss other things that are said
to be continuous (infinity, time, space). For an excellent discussion of the difficulties posed by De
caelo’s introduction, see Falcon 2016, 423–436, esp. 428–432. On its status as natural science, see
Leunissen 2010, 156–158.
3
As in previous chapters, where ἀρχή clearly refers to the starting-points of a science, I translate this
term as ‘principle’ or ‘starting point’ (keeping in mind that Aristotle distinguishes starting points of
inquiry from principles of demonstration). However, where it appears to refer to the source or origin of
a change of any kind, I translate ‘source’ or ‘origin,’ in this case keeping in mind that a nature is
defined as an inherent archê of motion, and what is often referred to as his ‘efficient’ cause is described
as ‘whence comes the archê of motion’ (but, as noted in ch. 6, he often insists that the telos or that for
the sake of which is also an archê). As the entries in Metaphysics Δ.1 (archê), 2 (aition), and 4 (physis)
make abundantly clear, Aristotle is fully aware of this problem.
4
Timaeus argues that the Demiurge chose the form of the sphere for the cosmos on grounds that it is
most complete and like itself (Ti. 33b-c), and introduces the three pairs of directional dimensions at
43b-c, but in a manner essentially unconnected to Aristotle’s. Excellent discussions of the connection
between the two works may be found in Falcon 2005, 31–36; and Johansen 2009, especially 18–24.
5
This passage also includes (i) an odd endorsement of the Pythagorean dictum that ‘the all and
everything are defined by threes’; (ii) a claim that we take the number three from nature and make use
of it in worship of the gods; and (iii) a claim that the completeness of three is shown by the fact that
the first time we say ‘all’ (rather than ‘both’) is when we refer to three items. I will take up later the
question of why Aristotle stretches in these various ways for support of the idea that objects with these
three dimensions of magnitude are complete objects. For now, however, it is enough to note that it is
part of a set of wider epistemological questions raised by the method and style of the De caelo, on
which see Bolton 2009 and Falcon and Leunissen 2015, 217–240.
6
Interestingly, Cat. 6 distinguishes discrete and continuous quantity (πόσον) but does not mention
magnitude (μέγεθος) (4b20-22, 5a1-6); while in Metaph. Δ.13, Aristotle immediately divides quantity
into plurality (πλῆθος) and magnitude (μέγεθος) based on whether the quantity is countable or
measurable, and whether it is divisible into discontinuous or continuous parts (1020a7-11). In line
with Cael. i.1, he goes on: “Of magnitudes those continuous in one dimension are lengths; in two,
breadths; in three, depths. Of these, limited plurality is a number, [limited] length a line, breadth
a surface, depth a body” (1020a11-14).
chapters 5–7 Aristotle mounts an extended argument against the idea that the cosmos is infinite; in
chapters 8–9, against there being more than one cosmos; and in chapters 10–12, against the cosmos
being generable or destructible.
8
ἐν τοῖς περὶ τὰς τῶν ζῴων κινήσεις. The standard formula Aristotle uses when cross-referencing
another investigation (i.e., ἐν + τοῖς + περὶ + the subject of investigation) is used here. The phrase
used for the subject of investigation, τὰς τῶν ζῴων κινήσεις, is the same as that used in the first
sentence of MA to refer to a prior inquiry reported on in IA.
9
Cf. Simplicius, in Cael. 384.9-20; Leggatt 1995, 224; Stocks 1922, 284b14 n.1; Guthrie 1939, 138. While
all these commentators and editors correctly identify this as a reference to IA, none note the
identification, in IA 2, of these directional principles as among the archai frequently posited as
starting points of natural investigation, which, from the viewpoint of understanding the dependence
of Cael. on the discussion in IA, is critical.
10
Recall Cael. i.1, 268a22-24: “of magnitudes body alone [vs. line or plane] would be complete; for
body alone is defined by the three [dimensions], and this is all [the dimensions there are].” An
anonymous reader for the press has suggested translating διάστασις as ‘interval’ as a way of lessening
the tensions I discuss in this section. I have decided to follow the standard English translation
(‘dimension’ is used by both Guthrie and Stocks) in part because I think ‘interval’ simply changes
the tension without eliminating it, and because it does seem that dimensions are what are under
discussion. Nevertheless, it is important to remember that it is, along with διάστημα, a term used in
harmonics (cf. Ti. 36a6-7) for musical intervals, and less technically for gaps or separations.
11
Immanuel Kant takes the opposite stance and derives the directional distinctions from the spatial
dimensions; cf. Kant 1768/1992, 2:375–383.
12
Although on at least one occasion, he allows that “the privation is in a way form” (Ph. ii.1,
193b19-20).
13
The three meanings of ‘heaven’ (οὐρανός) distinguished in Cael. i.9 are (i) the substantial being of
the outermost circumference of everything; (ii) the body continuous with the outermost circumfer-
ence that includes the sun, moon, and planets; (iii) the entire body surrounded by the outermost
circumference (278b11-22).
14
Cael. i.1, 268a19-b10.
15
For recent discussions of Aristotle’s quite common appeal to what is εὔλογος, see Falcon and
Leunissen 2015, 217–240 in Ebrey 2015; Bolton 2009, 51–82 in Bowen and Wildberg 2009; and
Karbowski 2014, 25–38. A classic and still valuable discussion is LeBlond 1938. Both occurrences in
this passage support the view that a claim is εὔλογος if it is in agreement with a more general or more
fundamental proposition that Aristotle accepts as true.
16
The closing lines of chapter 1, to which this sentence is syntactically linked, say “For that (ὅτι μὲν)
these facts about animal locomotion happen to be so is clear from the inquiry into nature (τῆς
ἱστορίας τῆς φυσικῆς); we must now investigate the reason why (διότι δέ, νῦν σκεπτέον)” (IA 1,
704b9-10). Yet another link is that between the verbal adjective in this sentence (σκεπτέον) and the
opening words of chapter 2, Ἀρχὴ δὲ τῆς σκέψεως. Note here the clear echo of the ὅτι/διότι
characterization of the two stages of inquiry in APo. ii.1.
17
A good example of the ambiguity in the use of this term is discussed in note 2 above.
18
Referred to as the functions (τὰ ἔργα) of motion at MA 10, 703a19-20.
19
That is, these are three archai that are hupotheseis often used in natural investigations.
20
For similar language in a very similar context, see Plato, Meno 86e4-5, 87b2-5.
21
Lennox 1997, 199–214 (repr. Lennox 2001b, 205–224).
22
See APo. i.2, 72a18-24. There are two uses of the first, negative half of this presupposition in Cael. as
well, at 271a25 (or a32 if one adopts Moraux’s suggested reordering of the text) and 290a30-33. Later
in this chapter we will be discussing a use of the positive version of the principle at 288a2-12, in
defending the applicability of ‘front’ and ‘back’ to the heaven, an especially problematic move. In
this case, uniquely, the principle is ‘God and nature do nothing in vain,’ a form of the principle
adopted by Galen.
23
The first appears a bit later in Cael. ii.5: εἰ γὰρ ἡ φύσις ἀεὶ ποιεῖ τῶν ἐνδεχομένων τὸ βέλτιστον
(288a2-3). For a valuable discussion of Aristotle’s invocation of this principle in Cael., see Leunissen
2009, 152–175; a different approach is defended by Johnson 2005, 136–140, though this particular text
is not discussed.
24
For an exact parallel, see Aristotle’s discussion of the organs that function both for excretion and
generation at PA iv.10, 689a4-b1, and the note to this passage in Lennox 2001a, 323–324. (τοῦτο δὲ
νῦν μὲν ὑποκείσθω, ὕστερον δὲ δειχθήσεται περὶ αὐτοῦ, 689a10-11; διορισθήσεται δὲ καὶ περὴ
τούτων ὕστερον, νῦν δὲ ὑποκείσθω μόνον . . ., 689a12-14). That the seed and the menstrual
discharges are moist residues is for now to be assumed, and their nature is to be defined and proven
later – presumably a reference to the discussion in GA i.
25
Oddly, however, while chapter 3 is a general discussion about the need in locomotion for something
against which limbs must press, the concepts mentioned as ἀρχαί in chapter 2, ὦσις and ἕλξις, are
never mentioned again in IA, and appear only once in MA!
26
ἐπεὶ δ’ εἰσὶν. The syntax of this and the next paragraph is complex. The use of ἐπεὶ with the
indicative likely has the concessive force of ‘although.’ Aristotle wants to begin with a contrast
between most animals, which are bounded by six directional dimensions, and a wider class, which
includes plants, that is captured if we limit the directional dimensions to “the above and below part”
(τὸ μὲν ἄνω καὶ κάτω μόριον). The μὲν at 705a28 at first looks to be solitarium, adding emphasis to
the dimensions that belong to all living things. But the next paragraph begins ὅσα δὲ μὴ μόνον, and
I suspect the δὲ here loosely marks a contrast with the μὲν clause at a28. The contrast is very loose,
however, since a group of living things that has directional orientations in addition to above and
below is being contrasted not with a wider group of animals but with the above and below possessed
by that wider group.
27
This passage should be supplemented by PA ii.10, 656a11-14; iv.7, 683b21-25; iv.10, 686b31-36, and
Juv. 1, 467b13-468a12. In fact, in PA ii.10 Aristotle not only says that human beings have their
functional above and below aligned with cosmic above and below – he says that their natural parts
are according to nature (τὰ φύσει μόρια κατὰ φύσιν ἔχει), where κατὰ φύσιν pretty clearly means
according to the nature of the cosmos.
28
See PA ii.1, 647a21-24; 8, 653b19-27; de An. ii.3, 414b1-10. Aristotle was definitely onto something
here. A recent discussion of “animal architecture” notes that “[m]ost familiar animals are bilaterally
symmetrical in that they have matching right and left sides with a central axis of symmetry running
down the middle of the long axis of the body. This design also imposes a front/rear orientation to
animals and has enabled the evolution of many efficient modes of locomotion” (Carroll 2005,
33–34). The author goes on to mention other forms of symmetry, especially among echinoderms
such as sea urchins and starfish. He would no doubt appreciate Aristotle’s long discussions of sea
urchins at PA iv.5, 680a4-681a9 and HA iv.5, 530a32-531a7.
29
On the idea that the better of two related principles should be separated from the worse, recall GA
ii.1, 732a3-7: “Since the nature of the primary cause of motion, to which the logos and the form
belong, is better and more divine than the matter, it is also better that the superior be separated from
the inferior. On this account in those cases where it is possible, and to the extent it is possible, the
male has been separated from the female.”
30
Cf. Clark 1975, 28–40; Lloyd 1983, 26–43; Gregoric 2005.
31
Aristotle is explaining that, though blooded and therefore in possession of four limbs, two of them
do not qualify as ‘feet.’ The very next lines explain that by foot he means a part that has a ‘point of
contact’ (σημείον) with the ground for movement in place. An anonymous press reader correctly
noted that wings appear to initiate flight simultaneously – apparently it is only their walking that is
under consideration here.
32
A clear example where Aristotle is using ἀρχή to mean both the starting point or origin of motion
and a starting point or principle of reasoning.
33
For an extended discussion of this passage and a comparison of it with PA iii.3, 665a9-21,
Theophrastus, Metaphysics 11a8-13, and Plato, Timaeus 45a-b; cf. Lennox 1985, 143–164 (repr. in
Lennox 2001b, 266–272). I am still in substantial agreement with that discussion and review critical
reactions to its original publication in Lennox 2001b, 226–228. For two alternative readings of these
passages see Lloyd 1991, 27–53 (a reprint of his classic “Right and Left in Greek Philosophy” (Lloyd
1962) with a valuable introductory update); and Carbone 2016, 1–31.
34
In light of his criticisms of the Pythagoreans in Cael. ii.2, it is of interest that in his presentation of
the Pythagorean ‘table of opposites’ in Metaph. Α.5 (986a22-b2), the only directional dimensions that
appear in the table are right and left. For further consideration of that passage, see Schofield 2012, in
Steel (ed.) 2012, 141–166.
35
Simplicius (in Cael. 385.15–22 Heiberg) rehearses Aristotle’s general views about priority, but it is
clear that here he is relying directly on the arguments of IA.
36
Pace Leggett 1995, 225, who rules out this meaning because it would make the above claim
‘tautological.’ Certainly, Aristotle does say that to call a body complete in this sense is to say that
it is extended in width, depth, and breadth, so that to say ‘complete body’ is, since bodies are three-
dimensional, strictly redundant. But here Aristotle is reminding the reader of his grounds for calling
bodies complete.
37
Note this is the precise formula that Aristotle uses for one of his causes, the misleadingly labeled
‘efficient’ cause.
38
Here we confront a widely noted difficulty with Aristotle’s fundamental claim that all natural
substances have within themselves a source of motion and rest: by occasionally listing the four
sublunary elements as natural substances (e.g., Cael. i.2, 268b16, Ph. ii.1, 192b9-12), he appears to
contradict the claim that only ensouled beings are self-movers; while by sometimes denying that
even animals can, strictly speaking, move themselves without an external mover (Ph. viii.2, 253a11-
21; 6, 259b1-16), he seems to deny self-motion even to ensouled beings! For a variety of discussions of
this problem, see the papers by Furley, Gill, Freeland and Sauvé Meyer in Gill and Lennox 1994;
Kelsey 2003, 59–87; Graham 1996, 171–192; and Waterlow 1982, ch. 5. Simplicius (in Cael. 387.5–24
Heiberg) provides us with Alexander’s view on this issue (that the elements have only a passive but
not an active source of motion), a view with which Simplicius takes issue.
39
If this is a reference to Cael., the only plausible candidate would be i.9, 279a17-28, on which see the
discussions of Kosman and Judson, chapters 7 and 8 in Gill and Lennox 1994. But the passage
currently under consideration is the first in which Cael. explicitly claims the heaven to be ensouled,
so this reference may well be to another work.
40
οὐ δεῖ ἀπορεῖν.
41
As indeed may have Aristotle; see the discussion of this passage in Bolton 2009, 56–63. Notice the
uncharacteristic appeal to thought experiments and counterfactuals to make the case that we need
not be puzzled. For more on thought experiment in Cael., see Leunissen 2010, 165–166, and Rapp
2019, 79–88.
42
Aristotle may have in mind either that the elements do not have within them a source of initiating
motion from rest (see Ph. viii.4, 255b13-256a3) or (as a reader suggested) that there is no distinct part
of an element that can be isolated as the initiating part.
43
So that motion from the East will be ‘from the right.’ Yet another peculiarity of this discussion: as we
have seen repeatedly, above and below used with reference to the cosmos typically refer to the outer
periphery of the cosmic sphere and its center, respectively, not to the poles.
44
Cf. i.9, 279a18-30; ii.3, 286a7-12; ii.6, 288a28-b8; ii.12, 292a15-28. Admittedly, these passages are
often accompanied by those discussed in Bolton 2009, where Aristotle acknowledges the difficulties
of the questions and the relative lack of empirical support for the needed answers (e.g., ii.3, 286a3-8;
ii.5, 287b29-288a2; ii.12, 291b24-28).
45
It seems that some of the arguments in Cael. ii demand that ‘the heaven’ must be used in its ‘second’
sense (i.e., to refer not only to the outermost sphere but also to the spheres of the planets).
46
And I here find my conclusions largely confirmed by Robert Bolton’s study of the underlying
epistemology of the De caelo in Bolton 2009, 51–82. Cf. Burnyeat 2004, 15–16, Lloyd 1996, 171.
47
Cf. Leunissen 2010, 160–164.
48
On the specific force of ἀνθρωπίνως here, see Bolton 2009, 68–70.
49
See Chapter 6, 161, 171–173.
50
For a thorough account of the complex impact of the Timaeus on the De caelo, see Johansen 2009,
9–29. And the attack on the Timaeus goes beyond the discussion of the heavenly motions:
a significant portion of book iii is devoted to criticizing the geometric account of the elements in
the Timaeus; cf. iii.1, 300a1; 2, 300b17; iii.7–8, 306a1-307b18. Pythagorean attempts to generate
material elements out of numbers are also under attack there (e.g., iii.1, 300a17).
51
This is most fully spelled out in the midst of criticizing Democritus’ account of the growth and
shedding of teeth in GA v.8: “Since we posit, doing so based on what we see (ἐπεὶ . . . ὑποτιθέμεθα,
ἐξ ὧν ὁρῶμεν ὑποτιθέμενοι), that nature neither defaults nor does anything in vain given the
possibilities for each thing, then it is necessary that they have instruments for working on nutrients
so that they will be able to grasp nutrients once they stop suckling” (788b20-24).
9.1 Introduction
In Chapter 8, we focused on the dependence of Aristotle’s discussion, in
Cael. ii, of whether the concepts of ‘right and left’ have application to the
movements of the heavenly spheres, on his discussion of the three pairs of
1
Ὑπὸ δὲ τῆς ἐντὸς θερμότητος τά τε νεῦρα καὶ τὰ ὀστᾶ γίγνεται ξηραινομένης τῆς ὑγρότητος. διὸ καὶ
ἄλυτά ἐστι τὰ ὀστᾶ ὑπὸ τοῦ πυρὸς καθάπερ κέραμος· οἷον γὰρ ἐν καμίνῳ, ὠπτημένα ἐστὶν ὑπὸ τῆς
ἐν τῇ γενέσει θερμότητος.
227
2
As we saw in Chapter 4, Metaph. Ε.1, 1025b30-1026a5 distinguishes uniform and nonuniform parts of
plants as well.
3
Victor Caston (Caston 1997) argues that Galen is the first “emergentist” in the history of philosophy.
If my interpretation of Mete. iv is correct, Galen was anticipated by Aristotle. For the purposes of this
chapter, I’m following the account of emergence in Humphreys 1997, according to which the fused
elements may lose certain dispositional properties and will no longer exist as distinct elements, while
the emergent entity generated by their fusion will be characterized by novel dispositional properties
not found at the lower level.
4
The following well-known remark of Myles Burnyeat exemplifies the first extreme: “Aristotle simply
does not have our task of starting with the existence of matter as physics and chemistry describe it and
working up to the explanation of secondary qualities” (Burnyeat 1992, 22). The claim that Aristotle
“takes himself to have given material conditions sufficient to account for the generation and
continued existence of homoeomerous bodies” (Charles 1988 33) exemplifies the second extreme,
even given his claim that such an account is insufficient to fix the essence of homoeomerous parts.
A similar position is defended in Freudenthal 1995.
5
By saying they are not ‘mechanistic,’ I mean that Aristotle rarely makes explanatory appeals to
contact interactions such as pushing and pulling; by referring to the explanations as ‘thermo-
dynamic,’ I am drawing attention to the fact that the primary causal power or δύναμις involved in
these explanations is heat, either internal or environmental – that is, I am appealing to the Greek roots
of the term. From that perspective, the title of Furley 1983, “The Mechanics of Meteorology iv,” is
misleading – even though, by the criteria for ‘mechanistic explanations’ defended in, for example,
Machamer, Darden and Craver 2000 or Machamer 2004, the explanations in Mete. iv would count
as mechanistic.
6
As Aristotle points out explicitly: cf. GA ii.1, 734b31-735a26; ii.4, 740b12-741a3; ii.6, 742a10-16,
743a26-b5. This remains the central explanandum of developmental biology today, as this very clear
statement of the problem from a contemporary advanced textbook shows: “This generation of
cellular diversity is called differentiation. Since each cell of the body . . . contains the same set of
genes, how can this identical set of genetic instructions produce different types of cells? How can the
fertilized egg generate so many different cell types?” (Gilbert 2006, 4; emphasis in original).
7
παθητικά spoken of κατὰ δύναμιν καὶ ἀδυναμίαν (Mete. iv.8, 385a11-12) – affective properties spoken
of in virtue of their capacities to undergo or resist specific kinds of change (e.g., the capacity or
incapacity to harden or to melt). For the full list of eighteen pairs, see 385a12-18.
8
There are, it should be noted, a number of references to ‘completion’ (τελείωσις) and ‘end’ (τέλος)
during the discussion of concoction (πέψις) and its forms – for example, 379b18-22, b25-28, 380a13-
16, a31 – and inconcoction and its forms are characterized as incompletions (ἀτέλεια) at 380a6-9, a31-
33. This language is understandable given that Aristotle is discussing processes of cooking and
ripening and failures of such processes. What is completely absent is the language of ‘final causality,’
that is, τὸ οὗ ἕνεκα or ἕνεκα τοῦ.
9
The view that they are mixtures of preserved elemental constituents he attributes to Empedocles –
the idea being that uniform bodies consist of (say) actual quantities of earth and of water, combined
according to a certain ratio.
10
That there is a range of intermediates is clear from 334b28 (τὸ δὲ μέσον πολὺ καὶ οὐκ ἀδιαίρετον)
and from the idea that the λόγος (here, certainly ‘ratio’) of the contraries in the compound can be
2:1, 3:1, or some other (334b15-16). See Frede 2004 289–314. In John Cooper’s contribution to the
same volume (Cooper 2004, 315–326), he poses the question of whether the theory of mixture
defended in GC i.10 and slightly elaborated in ii.8 implies that every part of a mixture has
(potentially) precisely the same balance of elements as every other part. His view is that the theory
elaborated in GC does not imply this. While I cannot explore the issue here, Mete. iv must be of
considerable value in deciding the question.
11
I defend this way of understanding the notion of potential presence in the exchange between Sorabji
1989 and Lennox 1989, published in Cleary and Shartin, eds. 1989, 35–76. Cf. Popa 2009.
12
πρὸς τὸ μέσον is a very common prepositional phrase in Aristotle (used, with very different
references, in discussing the virtues (EN 1106b9, 1108b28; EE 1222a26, 1234b7), the middle term of
a syllogism (APr. 28a17, b5, 29a1, 31a19, b11), and the center of the cosmos (Cael. 269a27, 270a7,
295a13, a34, b21, b25 – and many more; Ph. 212a26, 214b15), but this is the only instance of its use with
this verb.
13
Much of the literature on this topic restricts itself to a discussion of GC. Bogaard (1979, 11–29)
provides a valuable discussion of mixis in GC, but refers to Mete. iv as a “collation of empirical data”
and more or less ignores it. So too Needham 2009, 149–164. Frede 2004 does devote a few pages to
Mete. iv.
14
Compare the discussion of contrariety in Metaph. Ι.4–7 (thanks to Devin Henry for reminding me of
this background).
15
I realize the expression ‘secondary property’ might be misleading, given its use in John Locke and
modern empiricist literature generally, but it seems the most straightforward way to denote the
contrast with the primary dispositions. Aristotle does make an interesting distinction in Mete. iv, in
introducing his list of emergent dispositions, between those that differ in relation to the senses (πρὸς
τὰς αἰσθήσεις) such as smells and tastes, and those, like solidity and flexibility, that are more
appropriately termed affections; and those he goes on to list all fall into the latter category (Mete.
iv.8, 385a1-11).
16
For example, PA ii.1, 646a16-17; Long. 5, 466a21-3; Mete. iv.1, 378b32-34, 10, 388a21-22. For two
different attempts to make philosophical sense of this idea, see Gill 1989 78–82, 243–252; Furth 1988,
76–83. Claiming that the four contraries are the matter of the composites is at odds with the idea that
they are qualities inherent in some underlying ‘prime’ matter.
17
Frede 2004 unfortunately only looks at chapters 8–12, and thus not at the causal theory of the
production of the dispositions that differentiate uniform compounds.
18
The reference need not be directly to GC, however, since the same doctrine is stated in Mete. i.2–3 as
well.
19
Or perhaps they are the matter of the uniform bodies. As we will see, in discussing the causal
processes that give rise to the uniform bodies, Aristotle’s preferred language is that of heat and cold as
agents transforming bodies on the moist/dry continuum – though he often describes these bodies as
earthen or ‘of earth’ and watery or ‘of water.’ This is apparent in 388a21-25, quoted below.
20
For valuable discussions of the range of uses for these terms, see Düring 1944, 68–74 and Lloyd 1996,
83–103. Neither comments, however, on the virtually complete absence of this taxonomy of forms of
concoction and inconcoction throughout the remainder of Mete. iv. (Inconcoction is mentioned
once in ch. 7 at 384a33 and concoction once in ch. 11 at 389b8.) In the biological works, concoction
and inconcoction and their cognates are used repeatedly to describe the way heat operates in
nutrition and generation, but again there is no systematic use of the elaborate classification found
in Mete. iv.2–3.
21
An implication of this description, tellingly supported by the way Aristotle grounds a number of his
explanations, is that he thinks you can ‘get at’ the underlying nature of these bodies by means of
experiment. Needless to say, this runs counter to a common view about Aristotle’s science: however,
see Newman 2004, especially the discussion of ‘Aristotelianism and Experiment’ that opens chapter
5, 238–242; cf. Newman 2006.
22
HA iii.6, 515b30-516a6 also discusses a kind of fiber that comes to be in blood. Fiber does not arise in
the blood of deer, antelope, roe, or hares (b33-35), but blood solidification in these kinds ranges along
a more/less continuum. Deer blood solidifies like milk without rennet; antelope’s blood solidifies
better, somewhat like that of sheep. For help in thinking about what Aristotle might have in mind
by ‘fiber’ and removing it from blood, I consulted with Dr. Resia Pretorius, director of the Applied
Morphology Research Centre in the Department of Physiology, University of Pretoria. She informs
me that if blood is allowed to sit undisturbed in a container, the components will separate in the
same way they do when centrifuged, though she wondered if it would need to be mixed with an
anticoagulant (any liquid high in citric acid would do). When this is done, the fibrin is concentrated
in a yellow layer above the red blood cells. If Aristotle extracted that layer, the blood would then not
clot – or so Dr. Pretorius speculated, since she had not tried it.
23
Thin blood due to abnormally low platelet production is a symptom of a wide and diverse class of
illnesses. PA ii.7, 652b33-653a8 explains that fluxes begin in the head because when the blood around
the brain is cooled this region has “the capacity to produce phlegm and serum.” Compare the
following, from the Hippocratic text On Fleshes, on the formation of the liver: “When you slaughter
a sacrificial animal, as long as its blood remains warm it is liquid, but when it becomes cold, it
congeals. If, however, you shake it, the blood does not solidify, for its fibers are cold and gluey”
(Carn. 8 [Potter trans.], 1995).
24
In recent philosophy, there have been a number of attempts to develop a taxonomy for the latent or
‘nonoccurrent’ attributes of things: powers, potencies, capacities, dispositions, abilities, capabilities,
and so on. I will let context determine which of these terms I use to render Aristotle’s δύναμις, which
has a generic sense (which encompasses nature (φύσις)), referring to any source of change, and
a specific sense referring only to a source of change in another thing (or in the thing itself qua other)
or to a source of being changed by another (cf. Metaph. Θ.1, 1046a9-36).
25
He speaks of these properties as universally present in uniform materials, but sometimes appears to
restrict them to things that are at least to some extent solid, as we shall see presently. For doubts
about whether all the uniform bodies on the hard/soft spectrum are compounds, see Popa 2009.
26
οἰκείῳ ὅρῳ. I take the contrast to be with fluids or piles of sand, which have a defined boundary only
by being placed in a container.
27
Nor is there in Mete. i–iii; see 134, note 24. Famously, Mete. iv.12 notes this absence; on which see
Gill 2014.
28
When considering the use of the concept of ‘form’ (εἶδος) here, it helps to keep in mind that the
entities under consideration are uniform materials (copper or clay, bone or blood) being considered
in themselves, and not as parts of organisms or artifacts. As elsewhere, ‘form’ here refers to what it is
to be something – in this case, those dispositional properties that, once acquired, differentiate
a specific uniform body from all others, qua uniform body. For the use of εἶδος to refer to bodily
affections, see Ph. i.7, 190b28-29; ii.1, 193b19-20.
29
Quoting Burnyeat 1992, 22, who denies that Aristotle sets himself this task; see note 4, above.
30
What I have in mind by this claim is that the statements typically have the following properties: (i)
they are universal in scope; (ii) they are taken by Aristotle to identify natural necessities; and (iii)
they are intentionally worded so as leave the extension of the relationships identified open. This
gives these statements the ‘projectability’ expected of scientific laws.
31
It must be remembered that these are true compounds, however – ‘earth’ and ‘water’ are present
potentially, not actually, as explained earlier.
32
ὅσα δὲ κοινὰ γῆς καὶ ὕδατος, καὶ ὑπὸ πυρὸς πήγνυται καὶ ὑπὸ ψυχροῦ, παχύνεται δ’ ὑπ’ ἀμφοῖν.
33
ὅσα μὲν οὖν ὑπὸ ψυχροῦ πήγνυται τῶν κοινῶν γῆς καὶ ὕδατος, πλέον δὲ ἐχόντων γῆς, τὰ μὲν τῷ
τὸ θερμὸν ἐξεληλυθέναι πηγνύμενα, ταῦτα τήκεται θερμῷ εἰσιόντος πάλιν τοῦ θερμοῦ.
34
ὅσα δὲ διὰ ψύξιν, καὶ τοῦ θερμοῦ συνεξατμίσαντος ἅπαντος, ταῦτα δὲ ἄλυτα μὴ
ὑπερβαλλούσῃ θερμότητι, ἀλλὰ μαλάττεται.
35
ὅσα δ’ ὑπὸ ξηροῦ θερμοῦ πήγνυται, τὰ μὲν ἄλυτα, τὰ δὲ λυτὰ ὑγρῷ.
36
These two prominent features of the presentation in Mete. iv are also prominent features of the
presentation of information about animal differentiae in HA. For the prominent role of ‘open
scope’ universals in the presentation of information in HA, see Lennox 1991 (repr. Lennox 2001b,
ch. 2). However, there is one crucial difference: the propositions in Mete. iv cited above often
identify causal relationships, while those in HA identify universal correlations without mentioning
causation – and its method is explicitly said to be pre-causal (HA i.6, 491a7-14; PA ii.1, 646a8-12).
See Popa 2014, 320–323 for a discussion of this contrast.
37
These principles are implicit in chapters 4–7, but not highlighted. Aristotle regularly notes that
compound bodies are combinations of moist and dry, or earth and water, and are called by one or
the other name depending on which one is predominant (cf. 381b24-27, 382a4-6, 383a27-28). But the
language of uniform bodies (τὰ ὁμοιομερῆ σώματα), literally ‘bodies all the parts of which are alike,’
is used for the first time in chapter 8 and is not used frequently until chapters 10–12.
38
383b24-25, 384a15-16, 385b4-5, and 388a32 all seem to say that it is a compound of air and water with
air dominant; 388a8-9 says it is ‘more of moist’ than dry, suggesting there may also be a small amount
of earth present – and the other passages are worded so that they allow for the presence of earth,
provided the moist, or water, predominates.
39
Interestingly, at GA ii.1, 735b1-37, a number of the same observational claims are cited in establish-
ing that there is pneuma in semen, and at one point (735b13-16) the behavior of olive oil is cited as
evidence.
42
The pores mentioned here, and in the cognate biological passages, are not, by the way, a non-
Aristotelian idea, as has occasionally been claimed (e.g., by Gottschalk 1961) as an argument for the
inauthenticity of Mete. iv. Aristotle’s notion of a uniform body in no way excludes that body having
a porous conformation. The compound may be porous, and if so it will have some other material in
its pores; but Aristotle is extremely careful to stress that what is inside the pores is no part of the
compound. The theory of pores he is arguing against in GC i.8 (326b8-29) is not the theory
defended here. On this topic, see the thorough discussion in Düring 1944, 74–78; cf. Furley,
73–93; Lee 1962, xiii–xxi; Lewis 1996, 3–15; Pepe 1978, 503–523; and the discussion in Popa 2014,
6–12 and note 26.
43
Cf. Gill 2014, 339–344. This passage tells against the claim of Mirus 2006 that the differential
dispositional properties discussed in Mete. iv provide true, nonbiological definitions of the uniform
bodies. Aristotle is quite clear here that we have yet to determine what each of these uniform bodies
is, though we have learned, by studying how they come to be, the material kinds to which they
belong. The language of “getting hold” (ἔχομεν) of these bodies by grasping their kinds is evocative
of APo. ii.8, where we come to know that there is, say, thunder, by coming to know that there is
a certain noise in the clouds, or that there is man by coming to know that there is a certain sort of
animal (93a20-24).
44
While certainly open to dispute, I take this reference to “whole works of nature” to be to animals
(and perhaps plants), and that this provides the context for the rest of the chapter. Mirus (2006) puts
much stress on the fact that, amidst a long discussion of how animals and their parts are what they
are in virtue of their capacities to perform living activities, Aristotle once mentions, in conjunction
with the parts of plants, inorganic metals (bronze and silver, 390a17-18). But (i) this actually
reinforces the point that the chapter is primarily about animals, since he stresses that he is extending
the point he is making about animals to these cases; they are “just as flesh and sinew,” and (ii) the
point he is making here is generalizable – “all these things are what they are by a capacity of acting or
being acted upon” (390a18-19). That does not imply that all capacities are the same, and it allows that
some things may be defined primarily by passive capacities. It is also worth noting that when
Aristotle talks about the processes of heating and cooling being able to produce the affections of
uniform bodies, but not nonuniform parts like a hand or foot, he goes on to make an analogous
point about bronze or silver – heating and cooling may be a cause of their generation, but “not of the
generation of a saw, a cup, or a box: in this case art [is the cause], whereas in the other case nature or
some other cause” (390b12-13).
45
Recall the virtually identical list of parts at Metaph. Ε.1, 1025b30-1026a5, to make the point that the
natural scientist studies both the matter and form of natural objects.
That is to say, moist, dry, hot, and cold are matter of the composite bodies,
while the other differences, for example, heavy and light, dense and rare,
rough and smooth, and the other such bodily affections, follow from
(ἀκολουθεῖν) these ones. (PA ii.1, 646a16-20)
The four primary powers are here said to be the matter of composite bodies,
and other powers are said to ‘follow’ them. Follow, in what sense, one
might ask? Mete. iv, as we saw, provides the answer: these emergent
dispositional properties are due to a variety of interactions between pro-
cesses initiated by heat and cold and material compounds that fall on the
moist/dry, watery/earthen continuum.49
After presenting this ‘bottom-up’ compositional hierarchy, Aristotle
immediately contrasts the way things are ordered in generation (ἐπὶ τῆς
γενέσεως) with the way they are ordered in substantial being (τῆς οὐσίας):
“for things posterior in generation are prior in nature, and the final stage in
generation is primary” (646a25-27). The implication for animals and their
parts is clear: the living being, which is the end product of generation, is
metaphysically primary.
Hence, as Mete. iv.12 leads us to expect, PA ii.1 also ranks dispositional
properties based on the teleological priorities inherent in the nature of
living things.
And since the actions and motions present in animals as a whole and in such
[nonuniform] parts are complex, the things from which they are composed
must have distinct potentials; softness is useful for some things, hardness for
others; certain things must be elastic, others flexible. Now in the uniform
parts such potentials are distributed part by part (one is soft, another hard,
one moist, another dry, one viscous, another brittle), while in the nonuni-
form parts they are distributed to many and in combination with one
another; for a different potential is useful to the hand for squeezing than
for grasping. That is why the instrumental parts have been constituted from
bones, sinews, flesh, and the other parts of this character, not the latter from
the former. (646b14-27)50
49
An anonymous referee commenting on Lennox 2014 noted the oddity of heaviness and lightness
being mentioned here. They are not among the list of emergent dispositions in Mete. iv.8, and Cael.
iv defines heavy and light in terms of a natural capacity to move either toward the center or the
periphery of the cosmos. The two elements that possess natural heat (fire and air) are the ones that
tend upward, so Aristotle may have in mind that the presence of heat is essential to lightness – but
I am not aware of any passage where he actually says that.
50
The Greek of the opening phrase here is critical: Πολυμόρφων δὲ τῶν πράξεων καὶ τῶν κινήσεων
ὑπαρχουσῶν τοῖς ζῴοις ὅλοις τε καὶ τοῖς μορίοις τοῖς τοιούτοις, ἀναγκαῖον ἐξ ὧν σύγκεινται, τὰς
δυνάμεις ἀνομοίας ἔχειν.
51
See Popa 2014, 306–334.
52
In modern chemistry ‘compound’ refers to the product of a chemical bonding of different elements,
which gives rise to properties not present in the elements, while ‘mixture’ refers to a physical
intermingling of different elements or compounds that retain many of the properties of the elements.
Given those options, ‘compound’ is closer to what Aristotle has in mind by μῖξις than ‘mixture’ is.
53
The account in PA iv.10, 687a3-b25 includes a good deal of discussion of how well constructed the
hand is for grasping and squeezing.
54
These were already said to be “the matter of the composite bodies” at 646a17.
55 56
Cf. 237–238, above. Cp. HA iii.6, 515b27-516a7; iii.19, 520b10ff.
57
Cf. de An. i.1, 403a24-b9; and for systematic discussion of this aspect of Aristotle’s natural
philosophy and its relevance for his ethics and politics, see Leunissen 2017.
58
Cf. Freudenthal 1995, 48–56; Laberrière 1984, 405–428; Leunissen 2012, 507–530.
59
Cf. Lennox 2001a, 187, 188, and 201–203; Leunissen 2012, 517–520 and 2017. As we will see shortly,
many of the explanations of pathêmata of organic parts in GA v have the same character.
60
Gotthelf 1987b 167–198, esp. 172–178 [reprinted in Gotthelf 2012, 153–185, esp. 158–164] makes
a compelling case that PA ii–iv is structured in such a way that the premises needed for later
explanations are established earlier in the treatise.
61
This is one of many examples of what Leunissen refers to as ‘secondary teleology’. Cf. iii.7, 670a30-
32, iii.9, 672a14, iv.12, 694a22-23, 694b5; cf. Leunissen 2010, 95–97, 140–141.
62
As Popa 2014 stresses, these material-level explanations often have a transparently syllogistic
character.
64
For other comparisons of embryogenesis to figure drawing, cf. HA iii.6, 515a35; PA ii.9, 654b29-32;
GA iv.1, 764b30-31.
65
As Peck notes (Peck 1942, 219 note e); cf. Düring 1944, 83. This must be a reference to Mete. iv.
66
Cf. Mete. iv.3, 380b8-11.
67
As a reminder, by using the term ‘law’ here I only mean to stress that these are universal propositions
identifying necessary causal relations between forms of heating and cooling, on the one hand, and
the changes they bring about, on the other – it is not a historical claim about Aristotle originating
the notion of a ‘law of nature.’ For a discussion of this issue, see Popa 2014.
68
For a detailed analysis of the explanatory project of GA v and its place in the overall project of GA,
see Gotthelf and Leunissen 2012 (in Gotthelf 2012, 117–141).
69
Cf. PA i.1, 640a33-640b4 on the distinct forms of explanation referenced here. There is no
implication that none of the affections discussed in this book are for the sake of something. The
last chapter of this book (GA v.8), for example, discusses differences in whether animals shed a first
set of teeth that are later replaced, and Aristotle strongly criticizes Democritus both for providing
too general an explanation for this phenomenon (788b9-19) and for failing to investigate what it is
for the sake of (789b2-15). On this, and GA v generally, see Gotthelf and Leunissen, in Gotthelf 2012,
117–141; on precisely what is and is not implied by the opening discussion (778a29-b19) see the
appendix to their chapter, 139–141.
70
As a reminder of the relevant conclusions at work here: “Therefore since everything definite and
constituted is either soft or hard, and these are due to solidification, no composite and definite body
can exist without solidification” (Mete. iv.5, 382a24-27). “All things are dried either by being heated
or by being cooled, but in both cases by heat – that is, by internal or external heat. For even things
that are dried by cooling . . . are dried by the internal heat that causes the moisture to evaporate
along with it, the heat being driven off by the surrounding cold” (iv.5, 382b16-22).
71
For example, Cohen 1989. The apparent inconsistency is also discussed in Gill 2014, who like me sees
no conflict, though she resolves it in a different way than I defend here.
390b2-9). The dispositions that differentiate bone from the other uniform materials are brought
about by the actions of heating and cooling and the associated motions discussed in Mete. iv.1–11,
but not in such a way as to organize a properly functioning skeleton.
75
Cf. 388a21-22.
9.4 Conclusion
In this chapter and the previous one, my goal has been to highlight the
complex interdependencies among Aristotle’s many and varied inquiries
into the natural world. These dependencies point to a preferred, if not
necessary, order of inquiry; and while this order of inquiry is likely related
to the order that Aristotle might have considered best from a pedagogical
point of view, or the order dictated by his views regarding explanatory
priority, it is distinct from either of those orderings. What is revealed in the
two case studies that make up this chapter and Chapter 8 is that progress in
certain aspects of inquiry into one domain of nature (e.g., into the
directionality of the motions of the heavenly bodies, or the material nature
of the uniform parts of animals) depends on having already reached at least
preliminary conclusions in another domain. This need not be reflected in
the pedagogical order, where questions such as what is clearer to the novice
versus the expert and what can be treated as locally unsupported assump-
tions for the purposes of teaching will play an important role. Similarly,
given Aristotle’s view that, in the study of living things, explanatory appeals
to form (living capacities and activities) and end (what various parts and
processes are for) take priority over efficient causes based in the material
natures of living things, the dependence of PA ii and GA ii and v on the
results of the explorations reported in Mete. iv will not necessarily be
reflected in the explanatory priorities of those treatises. But what I hope has
become clear in these chapters is that the norms of inquiry that we’ve been
exploring extend to norms about the order in which inquiries need to be
carried out.
76
Some of the material in this section of Chapter 9 is based on Lennox 2014d, which in turn was based
on a presentation at an HSS/HOPOS symposium in November 2012, organized by Tiberiu Popa –
which also included the papers by Popa and Mary Louise Gill on which I have relied in this
discussion. The translations of texts from Mete. iv in this chapter are based on one prepared by the
three of us for publication.
Aristotle on Respiration
Framework Norms Meet Domain-Specific Norms
1
Δεικτέον δ’ οὕτως, οἷον ὅτι ἔστι μὲν ἡ ἀναπνοὴ τουδὶ χάριν, τοῦτο δὲ γίγνεται διὰ τάδε ἐξ ἀνάγκης.
ἡ δ’ ἀνάγκη ὁτὲ μὲν σημαίνει ὅτι εἰ ἐκεῖνο ἔσται τὸ οὗ ἕνεκα, ταῦτα ἀνάγκη ἐστὶν ἔχειν, ὁτὲ δ’ ὅτι
ἔστιν οὕτως ἔχοντα καὶ πεφυκότα. τὸ θερμὸν γὰρ ἀναγκαῖον ἐξιέναι καὶ πάλιν εἰσιέναι ἀντικροῦον,
τὸν δ’ ἀέρα εἰσρεῖν. τοῦτο δ’ ἤδη ἀναγκαῖόν ἐστιν. τοῦ ἐντὸς δὲ θερμοῦ ἀντικόπτοντος, ἐν τῇ ψύξει
τοῦ θύραθεν ἀέρος ἡ εἴσοδος καὶ ἡ ἔξοδος. ὁ μὲν οὖν τρόπος οὗτος ὁ τῆς μεθόδου, καὶ περὶ ὧν δεῖ
λαβεῖν τὰς αἰτίας, ταῦτα καὶ τοιαῦτά ἐστιν.
264
2
And see PA iii.6, 669a4-5, which refers to what was said ἐν τοῖς περὶ ἀναπνοῆς; cf. de An. ii.8, 421a6;
iii.9, 432b10-11.
3
Cf. 90a14-15: “In all these cases it is apparent that ‘what it is’ and ‘why it is’ are the same.”
4
And it is this view that accounts for a number of otherwise puzzling aspects of APo., among these the
fact that he is willing to discuss ‘partial demonstrations’ (i.24); to distinguish demonstrations of the
fact from demonstrations of the reason why (i.13); and to argue that in at least some cases a causal
demonstration of the ‘reason why’ can be reconfigured as an account of what the subject of inquiry is
(ii.10). At one point in ii.8, we are told that in order to grasp that something is non-incidentally, we
need to grasp something of the thing itself (93a22-28).
5
I am using the title On Respiration simply for convenience’s sake, since it does delimit the inquiry on
which we are focused. Ross (1955, 2–3) presents the evidence against treating this as a separate work.
The opening lines of Juv. 1, 467b10-13 show that this inquiry is continuous with that into the causes of
youth and old age and the later inquiry into the causes of life and death. See too Hannon 2011, 5–14
and Morel 2000, 51. In Section 10.3, I discuss one especially telling piece of textual evidence for this
continuity.
6
He takes up the cetaceans in Juv. 18/Resp. 12, which opens: “Among the water-dwellers, one might be
puzzled about the cetaceans, that is about the dolphins, whales, and as many as have the so-called
blowhole – but in fact these are in accordance with our account” (476b13-16). The problem someone
might raise, he says, is that these creatures are footless, take in sea water, yet have lungs (b16-17). He
assures the reader that they do, nonetheless, manage to cool themselves by breathing and are thus
κατὰ λόγον (b14), in accord with the account he has been giving. Note again the open extension
characterization of the subject – the relevant group is ‘as many as have a blow hole.’ Aristotle wants to
restrict the term κητώδη to a group of sea creatures united by a system of related similarities, but he
does so in a way that allows for the discovery of other forms of cetacean that have not yet been
identified.
7
Kullmann 1974, 81 n. 8; Lennox 1987, 102 n. 26 (reprinted in Lennox 2001b, 35 n. 25); Lennox 1991,
267–270 (reprinted in Lennox 2001b, 43–45).
8
See page 208, note 16.
9
As an aside, this is another example of the methodological importance of order of inquiry and the
dependence of one inquiry on the results of another, the focus of the previous two chapters.
10
These two criticisms, as we’ll soon see, correspond to two important normative recommendations
Aristotle makes at the end of Resp. 3.
11
See Ross 1955, 2–3.
12
It should also be noted that On Length and Shortness of Life is not mentioned, though Aristotle
claims to be mentioning only ‘the most important’ (τὰ μέγιστα, 436a13) pairs of topics.
13
The Loeb translator, W. S. Hett, makes a pair of glaring missteps here, translating: “Democritus
states that respiration serves a certain purpose in animals that respire . . . but he never says that this is
why nature evolved respiration.” This comes close to stating the exact opposite to the view Aristotle
attributes to Democritus and suggests that Aristotle is looking for an evolutionary alternative!
14
An earlier comment makes the reference clear: “The circulation written about in the Timaeus”
(Juv.11/Resp. 5, 472b6).
15
Hannon 2011, 75, titles this chapter “Reiteration of the First Principles of the Theory.”
16
Note, by the way, the claim that one part is present for the sake of another part. This is a form of
causal dependence explicitly defended in PA i.5: “if some actions are in fact prior to, and the end of,
others, it will be the same way with each of the parts whose actions are of this sort” (645b30-32).
17
HA iii.2–4 quotes extensively from earlier thinkers’ accounts of the blood vessels and discusses in
detail problems in observing the vascular system via dissection.
18
“It is, then, apparent from these and like considerations that blood is present in blooded animals for
the sake of nourishment” (PA ii.3, 650b2-3).
19
See the discussion of 468b31-469a1 on the previous page.
20
“Concerning the bloodless animals, then, we have stated that they are aided in relation to living in
some cases by the surrounding air and in others by the surrounding water” (475b15-17). Once again,
we are reminded of the unreliable nature of these chapter breaks that were introduced in the
Renaissance; this reads much better as the conclusion of the previous discussion.
21
The use of the double quantifier ὅσα . . . πάντα is common in Aristotle and particularly common in
the Historia animalium, and appears to be used to identify coextensive or counter-predicated
attributes such as lungs and the taking in of air for cooling and (in the next quotation) gills and
the taking in of water for cooling. For some discussion of its use cf. Lennox 2001b, chapters 1 and 3;
Gotthelf 1988 (reprinted in Gotthelf 2012, 307–342). This passage should be compared with HA ii.
15, 505b32-506a8.
22
The point is not as clear in English as in Greek, which has distinct words for ‘fish scale’ (λεπίς) and
for the scale of a reptile (φολίς).
23
The δὲ at 476a1 is coordinate with the μέν at 475b17. The background class is the animals with blood
and a heart, and within that class he is distinguishing those that have a lung and cool themselves with
air from those that have gills and cool themselves with water; and then within those two classes he is
distinguishing various subdivisions. The use of ὅσα . . . πάντα in those two cases is to stress the
coextensive relationship between the organs of cooling and the medium used (air or water).
24
In case you are wondering about snakes or cetaceans, the background class here is animals with gills,
so it is only footless animals with gills that are under consideration.
25
The Greek for ‘wing’ is πτέρυξ and for ‘fin’ is πτερύγιον, the plural of which is πτερύγια.
26
Most salamanders have a larval stage during which they breathe with external, feathered gills and
lack lungs. In an attempt to be clearer about the sort of organism Aristotle was observing and the
data behind the claims he makes about the kordulos I have been greatly aided by Dr. J. W. Arntzen of
the Netherlands Centre for Biodiversity Naturalis in Leiden. On the basis of our email correspond-
ence he conjectures that Aristotle is referring to a so-called paedomorphic newt, regularly found in
mountain populations of Ichthyosaura alpestris, in Greece. Paedomorphism is the retention into
adulthood of features of the larval stage, in this case including the presence of gills and lack of lungs,
which occurs in some but not all members of this species. Aristotle would have had no reason not to
think that the paedomorphs were a distinct species.
27
And how is he sure about that? Another instance where dissection is clearly playing a role.
28
And he is seriously mistaken: there are numerous species of lungfish around the world that have, and
make good use of, both.
29
The same stress can be found at HA vii(viii).2, 589b26-7: ἓν δὲ μόνον νῦν ὦπται τοιοῦτον, ὁ
καλούμενος κορδύλος· οὗτος γὰρ πνεύμονα μὲν οὐκ ἔχει ἀλλὰ βράγχια.
34
For example, at HA iii.1, 511a14-15, PA iii.4, 666a9, and iv.13, 696b15-16.
35
The majority of manuscripts have ἐμπύρωσιν, but Ross defends ἐμπύρευσιν (which is found in two
manuscripts) on grounds that the spelling of the verb used by Aristotle is ἐμπύρευω. Ross notes
another piece of evidence supporting his recommendation, namely the use of the phrase ἡ φύσις
ἐμπεπύρευκεν αὐτήν earlier in Resp. 8, 474b13, which uses the verb, as opposed to the noun based on
it, to express exactly the same thought.
36
There is one very brief account of the connection between heart and gills at HA ii.17, 507a3-10, but it
more or less duplicates what is said here.
37
Cf. HA i.17, 497a32 (ὦν ἡ μὲν ὄψις θεωρείσθω ἐκ τῆς διαγραφῆς ἐν ταῖς ανατομαῖς); PA iv.5, 680a1-4
(ὅν δὲ τρόπον ἔχει τούτων ἔκαστον, ἔκ τε τῶν ἱστοριῶν τῶν περὶ τὰ ζῷα θεωρείσθω καὶ ἐκ τῶν
ἀνατομῶν· τὰ μὲν γὰρ τῷ λόγῳ τὰ δὲ πρὸς τὴν ὄψιν αὐτῶν σαφηνίζειν δεῖ μᾶλλον); and GA ii.7,
746a14-16 (Δεῖ δὲ ταῦτα θεωρεῖν ἔκ τε τῶν παραδειγμάτων τῶν ἐν ταῖς ανατομαῖς καὶ τῶν ἐν ταῖς
ἱστορίαις γεγραμμένων).
38
And Aristotle is well aware of this fact about diagrams, as the following passage makes clear: “The
same affection comes about in thinking as in diagramming (ἐν τῷ διαγράφειν); in the latter, though
it is of no use that the triangle is a definite quantity, nevertheless we draw (γράφομεν) one of
a definite quantity; and likewise one who is thinking, even though he does not think of a quantity,
places a quantity before his eyes, but thinks of it not qua a definite quantity” (Mem. 1, 450a1-6).
39
Today referred to as the conus or bulbus arteriosis.
40
It is difficult to be certain what he is referring to here; Ogle conjectures that he is generalizing from
examining a ray (1897, 131, n. 136) in which the aortic arches are idiosyncratic, but that seems
unlikely. Ogle assumes that Aristotle is only referring to the aortic arches stemming from the conus
arteriosis to the gills. But since Aristotle does not make our functional distinction between veins and
arteries, and it seems clear he is discussing vessels that are not connected to the conus arteriosis, it
occurs to me that he may be referring to the blood vessels that we think of as taking oxygenated
blood from the gills to the body.
41
This is more explicit in the discussion of gills in PA iv.13 (which refers to our discussion): “A cause of
the number of gills being larger or smaller is a larger or smaller amount of heat in the heart; for the
movement must be more rapid and stronger for those with more heat. And those with more gills and
double gills have a nature of this sort more than those with simpler and fewer gills” (696b16-20).
42
τὸ τέλος, rendered “final authority” here, as recognized by LSJ (s.v. τέλος i.2), has the sense of the
final authority over decision-making (indeed, LSJ’s recommended translation for its use here is
“decisive difference between”).
breathing and by the actions of gills. Equally puzzling, PA ii.7 makes no mention of the lung as
performing a similar function to the brain.
45
But it is worth recalling that in APo. ii.14, 16, and 17 biological examples, some quite technical, are
also used to make similar points.
46
See PA iii.6, 669a22-b7; discussed in Lennox 2001a, 265–268.
appendix
Ch. 1. Others have spoken about breathing, but not well. Among
other errors, many seem to claim all animals breathe. Clearly
all with lungs do, but there are differences even there.
Chs. 2–3. Anaxagoras and Diogenes. What is wrong with their ideas
about fish breathing. Their failure to ask teleological questions
while performing dissections.
Ch. 4. Democritus. Like others, Democritus also does not consider
what respiration is for, and there are lots of more specific
problems with his account.
Ch. 5. Plato in the Timaeus gets things all wrong anatomically and
also does not say what respiration is for.
Ch. 6. Some say it nourishes the intrinsic fire. Lots of problems with
that idea.
Ch. 7. Empedocles has much to say about breathing, much of it is
quoted, and many problems in his account are identified. He
also fails to consider the question of what breathing is for.
Ch. 8. A reminder of the views about heat, soul, and nutrition in On
Youth and Old Age. Final nutrient is blood; blood is produced
in the heart, which is the source of blood vessels as well. (On
this point, the Dissections need to be consulted.) The produc-
tion of nutritive blood requires inherent heat, which is the
instrument of the nutritive soul.
Ch. 9. Given the conclusion of chapter 8, there is a subsequent
need for moderation of heat by cooling. Bloodless animals
are sufficiently cooled by their surroundings. A few
307
314
eclipse, 33, 43, 181 Falcon, A., 77, 102, 104, 113, 114, 161, 176, 201, 202,
as object of inquiry, 43–44, 59–60, 290 203, 207
definition and demonstration of, 63 Ferejohn, M., 14, 45, 47, 50, 61, 107, 122
knowledge of, 17–18, 43–44 first philosophy, 6, 23
no final cause of, 134 and separability, 128–129, 192, 197
of moon, 17–18, 44, 59, 134 and the soul, 192–194
elephant, trunk of, 18 vs. natural/second philosophy, 75, 99, 105,
embryology, 77–89, 169, 171, 256–257, 277 126, 128
emergence, 230 form(s), 78, 84, 85, 236–237, 240, 242
during embryogenesis, 80 as nature, 118, 125, 126–128, 136, 177, 249, 260
in GA, 231, 256–260 as that for the sake of which, 129, 134–138
in GC, 232–235 as what a thing is, 128, 134
in Mete. IV, 229–232, 236–247 in relation to matter, 104, 106–107, 108, 128,
in PA, 249–256 150, 171, 263
of dispositional properties, 230, 231 of kinds, 57–58, 61, 97, 147–149, 150–159, 184,
Empedocles, 103, 127, 165, 166, 232, 273, 291, 292 209, 255
empiricism, 16, 20, 25, 33–34, 234, See also Platonic, 66–68
experience (empeiria) soul as, 43, 106, 177, 182–184, 186, 189, 230
entelecheia. See fulfillment studied by natural science, 128–131, 134–135,
epistemic access, 5, 34, 65, 80–81, 102, 122, 141, 171 191, 193, 198, 202
epistemic map, 103–104 framework. See causal framework; erotetic
epistemology, 19, 24–27, 29–30, 141 framework
of inquiry, 30, 32, 82 Frede, D., 233, 234, 235
erotetic framework, 1, 12, 18, 37–40, 95, 102, 179 Frede, M., 19, 24–27, 32, 35, 105, 107, 194–195
and causal inquiry, 44–64, 278–279 Freudenthal, G., 230
and domain-specific erotetic structure, 180, fulfillment (entelecheia), 181
185–186, 235 in relation to activity (energeia), 181
and domain-specific norms, 82, 122–128, soul as, 183–184
169–170, 264–265 vs. potentiality, 181, 183
as provided by APo. II, 40–44, 121–122, 139, Furley, D., 218, 230, 247
200, 282 Furth, M., 235
eternal natural objects, 4, 109, 135, 203
and circular motion, 220 Galileo Galilei, 66
and unconditional necessity, 161, 164 Gilbert, S., 231
vs. perishable natural objects, 4, 161, 171–172, Gill, M.-L., 107, 195, 218, 235, 240, 247, 260,
185, 224 261–262, 263
exhalations, 243, 245 gills, 267–268, 279–280
experience (empeiria) analogical relation to lung, 280–281, 282
and expertise, 88–89, 267–268 function of, 282–288, See also cooling, as
and inexperience, 79, 267–268, 271 function of lungs and gills