Rachel Barney

Method and Genre in Metaphysics A.3 (983a24-4b8)

Accounts of Metaphysics A.3 often begin with a disclaimer: Aristotle’s account of his precedecessors, we are warned, should not be read as history, or as a history, of philosophy.1 Sometimes explanation is provided by a gloss or contrast: Aristotle does not do history ‘for its own sake’, ‘as an end in itself’ or ‘as such’.2 But I have always found this disclaimer somewhat puzzling. After all, discussions of contemporary work in history of philosophy take for granted that it may be done with mixed motives and, quite often, a powerful philosophical agenda, without being disqualified from the genre ‘history’. Moreover the ‘end in itself’ requirement seems impracticable. Indeed, if we factor in the possibility of hidden agendas, self-deception and false consciousness, the genre ‘history 1 A complication is that the phrase ‘history of philosophy’ can be used in three quite different ways. The history of philosophy picks out a subject-matter, the one treated in a history of philosophy. When neither article is present, the phrase refers, I take it, to a broader family of literary genres which deal with some part of that subject-matter, but not necessarily in the mode of historical narrative: for instance one might claim that Diadochai, Peri Haireseôn and biography are all ancient ways of ‘doing history of philosophy’. I take the disclaimer to relate to both the second and third senses: though he is writing about the history of philosophy, Aristotle is not providing a history of philosophy, nor is he ‘doing history of philosophy’ in the broader sense. It is really this last claim which is of interest to me here. My concern will be with the term ‘history’ rather than philosophy: how far Aristotle’s predecessors were in fact engaged in a project rightly called ‘philosophy’ is another question altogether (see Mansfeld 1985c and Lloyd 1997 and 2002 for different kinds of doubts, with the comments of Laks 2005). It seems clear that Aristotle sees his story as one of a coherent enterprise to which his own project of ‘first philosophy’ is heir (cf. Frede 2004).2 “Aristotle is not, in any of the works we have, attempting to give a historical account of earlier philosophy. He is using these theories as interlocutors in the artificial debates which he sets up to lead ‘inevitably’ to his own solutions” (Cherniss 1935, xii cf. 349-50, 356-7, cf. Ross 1958, vol. 1, 128, McDiarmid 1970, 180). Reale cites Schwegler with qualified agreement: “il moderno concetto di storia della filosofia è totalmente estraneo ad Aristotele” (1968 vol. 1, 151). Collobert, exceptionally, treats the question as an open one, but in the end her answer too seems to be a qualified no (2002, 294). Guthrie 1953 defends Aristotle as a historian of philosophy; but this is not the same as defending the claim that he is a historian of philosophy. In general, the question of the genre of Metaphysics A.3 has been badly muddied by evaluative controversies, regarding how far Aristotle is fair and reliable as a source for earlier thought: I will be concerned here with the former question only.

1

of philosophy’ will threaten to evanesce in the familiar way of the Kantian ideal which the phrase ‘end in itself’ recalls: like the truly moral action, it may never have been instantiated at all. An alternative gloss on the disclaimer is in terms of objectivity.3 But this too becomes puzzling on reflection. What precisely should objectivity mean in relation to history of philosophy? Not, surely, philosophical neutrality, a refusal of all questionable interpretations or judgements of value -- no intelligible work at all could issue from that. On the other hand, if by objectivity we mean simply the desire to get one’s interpretations and judgements right, then it is no longer so obvious that A.3 fails to qualify.

More specific grounds for the disclaimer tend to run along two lines. First, A.3ff. does not meet the norms applicable to histories of philosophy: Aristotle’s account is self-serving, schematic, gappy and tendentious. Puzzling again, since these are the failings of which historians of philosophy accuse other historians of philosophy when criticising their works as history of philosophy. They cannot entail that a work falls out of the genre altogether. The second basis for the disclaimer is of more interest. This is that A.3 belongs to a different, non-historical genre: it is a dialectical survey of the endoxa, the reputable views which Aristotle uses as a starting-point for his own theorizing.4

I will discuss this aspect of A.3 in sections I and VI; but it cannot in itself settle the question of genre. To say that A.3 cannot be history because it serves a dialectical purpose is like saying that a piece of music cannot be a fugue because it is a movement of a sonata. Once we have waived the procrustean

3 E.g., “Jamais il n’adopte à leur égard l’attitude de l’historien de la philosophie, qui se veut simplement objectif” (Mansion 1961, 75). One could also infer this verdict from a much more sweeping disclaimer: “Objective history of philosophy is a modern idea” (Mansfeld 1997, 107); but this too should not be assumed without more defense and explanation than is usually given. Of course, one might for some purposes want to assume or stipulate a historicist understanding of literary form, so that by definition nothing earlier than Robinson Crusoe could count as a novel and nothing earlier than Brucker’s Historia Critica Philosophiae, say, could count as history of philosophy. I am assuming for present purposes that we can legitimately apply generic concepts in a more temporally open way.4 Cf. Cherniss 1951, Mansfeld 1986.

2

‘end in itself’ requirement, the possibility is open that a historical form may be used to execute a philosophical project, including a dialectical one.

Finally, one might hold that Aristotle could not be intending to write ‘history of philosophy’ in the absence of a pre-existent genre, for which we have no evidence. Yet it seems an intelligible possibility, not to be excluded a priori, that Aristotle is here prôtos heuretês: that in Metaphysics A.3 he is inventing something we might reasonably recognise as ‘history of philosophy’. And that is one of the possibilities I mean to explore in this paper. I will first give a step-by-step account of Aristotle’s presentation of his predecessors in A.3, 983a24-4b8, which I divide into five moves (sections I-V below). The discussion will focus on his treatment of Thales and the other ‘first’ philosophers [prôtoi]: I will have little to say about Parmenides, Empedocles and Anaxagoras in 983a24-4b8, not least because it is impossible to disentangle the issues here from Aristotle’s more extended discussions later on. I will then turn to reflections on the dialectical and historical dimensions of Aristotle’s method, in sections VI and VII respectively.

I. The Introduction of the Project (983a24-3b6):

I. Introduction of the project:a. Résumé of the four causes, which have been sufficiently studied in the Physics;b. Let us also consider what our predecessors have to say, since those who have examined ta onta have stated certain archai and aitiai;c. Either another kind of cause will be revealed or the fourfold theory will be confirmed.

Aristotle begins by announcing the project that will occupy him from A.3 through A.7 -- or, taken more inclusively, A.3-A.10, with chapters 8 and 9 offering criticisms of some of the views discussed and chapter 10 as a summation. (One might also read Metaphysics α as at least in part a series of reflections on this project: I will note its relevance on occasion below.) Aristotle

3

begins with a brief résumé of the four causes: formal, material, moving, and final. He notes:

“We have studied [τεθεώρηται] these causes sufficiently in our work on nature, but yet let us call to our aid those who have attacked the investigation of being and philosophized about reality before us. For obviously they too speak of certain principles and causes; to go over their views, then, will be profit to the present inquiry, for we shall either find another kind of cause, or be more convinced of the correctness of those which we now maintain.” (983a33-b6)5

The term τεθεώρηται (‘studied’) is carefully chosen: with it Aristotle avoids any claim to have proved the system of causes in the Physics.6 What argumentative support can be provided for his doctrine is at this point an open question; and as Mansion notes, the survey which follows is offered to some extent in lieu of the argument we might have expected in Physics ΙΙ (1961, 40).

In the course of A.3-10, Aristotle will periodically reaffirm this project in quite consistent terms.7 At the start of A.7 he announces its completion and its result: “of those who speak about principle and cause no one has mentioned any principle except those which have been distinguished in our work on nature, but all evidently have some inkling of them, though only vaguely” (988a18-23). At the end of that chapter, as Aristotle segues to critical discussion, that result is restated: “All these thinkers, then, as they cannot pitch on another cause, seem to testify that we have determined rightly both how many and of what sort the causes are” (988b16-18). After two chapters of

5 Translations are from Ross 1928, at times with revisions; the text used is Ross 1958, with qualms as noted.6 The term is most often used by Aristotle for observations of animals (e.g. Gen. An. 721a15, Hist. An. 501b21); at A.2, 983a17, we might translate as ‘discern’. Cf. Nightingale 2004 on the connotations of theôria. 7 Cf. also 986a13-15, b2-12, 987a27-8, and the particularly striking reference at 987a2-3 to “the wise men who have now sat in council with us”.

4

critique, it is reaffirmed once more, with the same diagnosis of his predecessors as philosophically immature (993a11-17). So there is a coherent project here, with a clear result -- or so, at any rate, Aristotle is determined to claim.

In a general way, this project is similar to the predecessor-reviews we find near the start of other Aristotelian works, including De Anima I.2-5, Physics I and Gen. et Cor. I.1 (and in some ways Pol. II). But these predecessor-reviews are a heterogenous group; and there are at least two respects in which the others are all like each other and unlike A.3. First, they are diairetic, proceeding by a more or less systematic classification of positions by content.8 Physics I divides thinkers according to whether they propounded one principle or many; if one, whether motionless or moving; if more than one, whether finite or infinitely many, and so on. Gen. et Cor. I.1 classes philosophers as to whether they identified generation with alteration or distinguished the two. In De Anima, Aristotle classifies views according to the ‘mark’ taken as definitive of soul: movement, thought or sensation, or both. Since accounts of soul usually derive from views on the archai, De Anima also sketches another possible diairetic treatment of accounts of the archai, dividing them into corporeal, incorporeal, or both, and by number (one or many) (404b30-5a4). So one question for the interpretation of Metaph. A.3 will be why it does not look more like this.

Second, these other discussions are all concerned to provide a starting-point. For instance, Physics I shows (among other things) that “our principles must be contraries” (189a10). The use of De Anima I is less obvious since II.1 announces a fresh start; but Aristotle’s two ‘marks’ of soul are deployed later as useful (e.g., at the start of III.3 and III.9). By contrast, in A.3 Aristotle goes to seek confirmation of a theory to which he is already at least provisionally committed and with which he assumes us to be familiar. The resulting argument is neither a mode of discovery nor merely didactic. As Mansion puts it, Aristotle “est engagé dans la recherche au moment où il interroge ses

8 Cf. also the division of positions in Theophrastus De Sensibus and the Menôneia portion of the Anon. Lond.

5

devanciers. Ce sont ses problèmes qu’il leur pose, mais il attend vraiment d’eux une réponse. Cette réponse, dans la Métaphysique, prend la forme d’une confirmation à des vues qu’il n’a pu fonder théoriquement” (1961, 56).

The form of the argument Aristotle here undertakes is thus a distinctive one: it is a search for alternatives or additions to an account already offered. Through its failure, the search will confirm the completeness and adequacy of that account, roughly as follows: ‘P. One might propose instead Q, R, S, or T; but Q, R, S and T can all be reduced to P. Therefore P (and P alone).’ Formally speaking, this is not any familiar mode of Aristotelian dialectical argument -- and it is certainly not demonstrative. So far as I know, Aristotle does not use anything quite like it elsewhere in his surviving works. The closest analogue is perhaps the argument of Metaphysics Γ. Here Aristotle’s procedure is (roughly) to explore the various avenues by which some thinkers have purported (or been alleged) to deny the principle of non-contradiction, and to show that none of them is genuinely viable, so that the (indisputability of the) principle is confirmed. The project embarked on in A.3 differs in that the operation here is presented as one of subsumption rather than refutation: still, it too seeks to confirm a claim about first principles by exploring the putative alternatives and showing that none really is an alternative.

Thus Aristotle’s project is to determine how far his own explanatory framework can effectively subsume the apparently incompatible causal accounts advanced by his predecessors. Criticisms of A.3 for the anachronistic application of Aristotle’s own conceptual scheme are thus somewhat obtuse: the whole point of the exercise, about which Aristotle is perfectly explicit, is to see how far this can be made to work. And his project imposes a certain historical fidelity as well as thoroughgoing anachronism. For Aristotle’s corroborative test is worthless unless the rivals considered (his chosen Q, R, S and T) really do represent the most promising alternatives, and are considered in such a way that anything valuable which could not be subsumed by his own system would become visible as such. Given the terms of his own project, Aristotle should be concerned to present the views he discusses with sufficient

6

clarity, accuracy and detail to convince an informed reader (one better informed than we are in a position to be) that his interpretive subsumptions are successful -- for instance, that what Empedocles in some sense ‘really meant’ was that philia is the cause of good things, and that this is best understood as a partial discovery of one of the Aristotelian causes (A.4, 985a3-10). This subsumptive strategy raises fascinating questions about interpretive method which I cannot here address.9 And whether Aristotle actually meets its constraints is of course another question. My impression is that the argument of A.3-10 often strikes contemporary readers as somehow rigged to the point of worthlessness. However, this must be at least in part an artifact of the context (or rather the lack of context) in which we encounter it. If we imagine reading A.3-10 at a time when Empedocles’ own text was easily available, and practising Platonists might object to any misrepresentations, it has a rather different air -- even more controversial and parti pris, perhaps, but less easily accused of any sleight of hand. At any rate this kind of project, in which one philosophical stance asserts its superiority by showing that it can subsume and explain its rivals, remains a central mode of historically informed philosophizing.10

II. The Reasoning of the Prôtoi (983b6-20):

2. The reasoning of the prôtoi:a. Of the earliest philosophers [prôtoi], most recognized only the material cause as archê;b. They said that the archê was preserved as substrate through all changes, as a substance underlies its changing pathê;c. They disagreed on the number and identity of this archê/ai.

9 Cf. the methodological reflections in Rorty 1984 and Brandom 2002. 10 Cf. e.g., the works of Alasdair MacIntyre, including his 1984, 1988, and 1990, as well as Korsgaard 1996 and Brandom 2002.

7

Aristotle’s survey begins with the ‘first’, materialist philosophers introduced at 983b6-8: “Of the first philosophers, most [tôn prôton philosophêsantôn hoi pleistoi] thought the principles which were of the nature of matter were the only principles of all things”. His treatment of these prôtoi falls into four phases: a general account of their reasoning (983b8-20) (discussed in this section); a discussion of Thales’ views and the reasoning behind his choice of water as archê (983b20-4a3), together with an excursus on the possibility that the earlier poets should count as sharing his doxa (983b27-4a2) (both discussed in section III); and a listing of philosophers by their archai, including Empedocles and Anaxagoras (984a2-16) (section IV). This procedure leaves the membership of the prôtoi somewhat unclear. Presumably the group strictly speaking includes only the (pre-Parmenidean) philosophers listed prior to Empedocles and Anaxagoras, since the latter group do recognise an efficient cause.11 Exactly what Aristotle means to attribute to the prôtoi is also a delicate question. According to Ross, “Aristotle does not say that the earlier thinkers recognized the material cause” (1958, 128); for the ultimate material cause is prime matter, and their analysis went only as far as the four elements. But even assuming that Aristotle believes in ‘prime matter’, this does less to fix the reference of the term hulê than the functional role of matter as the substrate of substances. And this role is just what Aristotle emphasises here by using the circumlocutory phrase, ‘in the class of <or of the kind, type, etc. of> matter’. This means something like ‘properly classed as material’; and a cause is properly classed as material if it persists as a substrate through change. 12 For Aristotle, a cause which does the work of the material cause is a material cause, however lispingly explained or incompletely recognised. 11 Later, at 987a3-9, the παρὰ μὲν τῶν πρώτων... παρὰ δέ τινων τιθέντων seems to confirm that the prôtoi are those who recognised only a material cause. At the same time, as in A.3 itself, it seems to be suggested that this group includes never-named dualists and pluralists as well as monists; cf. n. 41* below.12 That this is what Aristotle uses the phrase ἐν ὕλης εἴδει for is clear from Metaph. A.5, 986b4-7: ἐοίκασι δ’ ὡς ἐν ὕλης εἴδει τὰ στοιχεῖα τάττειν· ἐκ τούτων γὰρ ὡς ἐνυπαρχόντων συνεστάναι καὶ πεπλάσθαι φασὶ τὴν οὐσίαν (note the γὰρ). At A.5, 987a7, that the prôtoi regarded the archê as somatikê is a distinct point from their having set down those archai ἐν ὕλης εἴδει.

8

If this is right, then what qualifies an earlier thinker’s choice of cause as material has nothing to do with whether it resembles the material ingredients of Aristotle’s own cosmos. One might object that in that case apeiron should be just as much a material cause as water, and yet Anaximander is missing from Aristotle’s list of materialists. On the other hand, perhaps Anaximander is here omitted because Aristotle did not take his apeiron to function as an enduring substrate: for the opposites once segregated off are precisely not apeira. In that case, the apeiron would not for Aristotle count as belonging ‘in the class of matter’. (A) This would also explain why Aristotle says that of the first philosophers οἱ πλεῖστοι -- not πάντες -- thought there were only material causes – though why he should cope with the exceptional case of Anaximander by simply omitting him, rather than treating him as a pioneer, is an open question.13

The central interpretive claim here is thus the infamous ‘material monism’:14

(1) There is a single archê from which all things emerge and into which they are destroyed, and which persists through all change as their substrate.

Now Aristotle’s attribution of (1) to the prôtoi is clear and unequivocal, and indeed forms the principal theme of his discussion. Yet as scholars have often pointed out, we have remarkably little external evidence to support this understanding of their archê as a persisting substrate: in particular, there is reason to doubt that Thales intended anything more than a more traditional 13 Aristotle’s explicit references to Anaximander are rather few and too sketchy to decide this question (cf. Physics 187a21, 203b14, Gen. et Cor. 332a19-25). Aristotle’s reading of Anaximander is in general a vexed question, much complicated by the question whether we should see Anaximander in his references to those who postulated an intermediate substrate; I here assume Physics A.4, 187a12ff. to be decisive evidence against that identification. Admittedly Simplicius takes it for granted that the apeiron is a stoicheion in the manner of the other monistic archai, Phys. 24.13. But Graham seems to me too quick to attribute this reading to Aristotle himself (2006, Ch. 2, 3.2, and 20 n. 55). 14 Cf. Graham, who however uses ‘material monism’ to include counterparts to (2) and (3) below (2006, 49).

9

kind of cosmogonic claim, viz that all things come from water.15 There is evidence to support Aristotle’s claim only for Anaximenes, Xenophanes (who seems not to be in view here anyway) and Heraclitus;16 moreover, Daniel Graham has argued that this evidence is far from unequivocal, and that the ‘material monism’ here ascribed to the prôtoi was really the innovation of Diogenes of Apollonia (Graham 2006). I cannot here engage properly with Graham’s arguments, but will note some points of immediate relevance. First, (1) remains a possible interpretation of Anaximenes and Heraclitus, two central figures among the prôtoi; and Aristotle is highly unlikely to have had any evidence to the contrary in Thales’ case. So we should perhaps understand him as standardizing the unclear and gappy views attributed to the prôtoi rather than retrojecting post-Parmenidean ideas wholesale. Moreover, we should distinguish (1) from two further, stronger claims Aristotle might mean to impute to the prôtoi:

(2) The persisting archê is what all things really are.

(3) There is no real generation or destruction; apparent generation and destruction are really just qualititative changes in the persisting archê.

We might call (2) and (3) together monistic reductionism: everything is ‘really just’ the archê, and so all change is ‘really just’ qualitative change rather than generation or destruction. This is obviously a much stronger position than (1), and it is harder to see why Aristotle would attribute it to the prôtoi. It does not really follow from (1): one could perfectly well hold that all things are made out of water without inferring either that they are ontologically reducible to water or that their formation constitutes something less than substantial generation. Moreover, even more than with (1), all our external evidence tends to suggest that, if reductionism is rightly attributed to any of the Presocratics, it is to the

15 See Algra for a succinct statement of the reasons for doubt (1999, 49-54).16 For Anaximenes, see DKA7; for Xenophanes, B29; for Heraclitus, B30.

10

post-Parmenidean pluralists, concerned as they are to give a deflationary account of substantial generation and destruction.

Aristotle is often taken to attribute (2) and (3) to the prôtoi on the basis of the following presentation of their reasoning:

“that of which all things that are consist, and from which they first come to be, and into which they are finally resolved (the substance [ousia] remaining, but changing in its modifications), this they say is the element and the principle of things, and because of this they think nothing is either generated or destroyed, since this sort of nature [phusis] is always conserved, as we say Socrates neither comes to be absolutely when he comes to be beautiful or musical, nor ceases to be when he loses these characteristics, because the substratum, Socrates himself, remains. So they say nothing else comes to be or ceases to be; for there must be some nature -- either one or more than one -- from which all other things come to be, it being conserved.”

Here the underlined passages seem to settle the question: each amounts to (3), attributed to the prôtoi collectively. The Greek, however, is not so unequivocal. In the first, καὶ διὰ τοῦτο οὔτε γίγνεσθαι οὐθὲν οἴονται οὔτε ἀπόλλυσθαι, the outhen could perhaps be adverbial, with the touto at b10 understood as subject, i.e.: “and for this reason they think it <i.e., the archê as stoicheion and substrate> is in no way generated or destroyed” (cf. the adverbial outhen at 984a29). At b16-7, Ross is filling out what is much vaguer in the text: οὕτως οὐδὲ τῶν ἄλλων οὐδέν (more literally, “so too neither does any of the others”). This clause is correlative with the long preceding hôsper clause on Socrates and his pathê; and the point made there is that for Socrates to ‘become’ [gignetai] in a particular way (e.g., become musical) is not for him to be generated [gignetai haplôs], since he persists through such changes as their subject. Aristotle has just likened the relation of archê to ta onta to that of

11

ousia to pathê (983b9-10): the point of the analogy seems to be to illuminate this relation, showing that it explains change as something less than generation of the archê . (A later allusion in A.3 confirms that it is this claim about the archê which is definitive of the prôtoi: for the Eleatics, “the one and nature as a whole is unchangeable not only in respect of generation and destruction (for this is an ancient belief, and all agreed in it)...” (984a32-3)). So, though it is somewhat mysterious what tôn allôn in the correlative clause might be, Aristotle can hardly mean simply to extend to ordinary substances his claim about the archê, viz that it escapes generation and destruction by eternally persisting. At most, he might mean that ‘generation’ and ‘destruction’ are misleading terms in all cases of change, since the prôtoi reduce the generation and destruction of substances to qualititative change in the archê as substrate -- just as they are accused of doing in Gen. et Cor. I.1. 17

Though this amounts to (3), it would be somewhat misleading to say that Aristotle attributes monistic reductionism to the prôtoi. He does not attribute the more overtly reductionist (2) to them, and he is happy, throughout A.3-10, to speak of generation and destruction in describing their views; his criticisms in A.8 even describe them as seeking to explain these phenomena (988b26-7). (And at 986b14-15 he presumably has them in mind in contrasting the Eleatics with phusiologoi who set down being [to on] as one but nonetheless ‘generate’ from it as from matter.) Moreover, the evidence of Gen. et Corr. I.1 cuts both ways. For Aristotle opens by claiming that some earlier philosophers say that generation and alteration are the same (314a6-8); but when he actually turns to expound the views of the monists, the point is twice put in terms of what it is necessary [anankê, anankaion] for them to say (314a9-10, b1-5). This suggests that the conflation of generation and alteration is an inference on Aristotle’s part, a position to which he takes them to be committed given their other views 17 At Physics 191a24-31, Aristotle speaks of ζητοῦντες γὰρ οἱ κατὰ φιλοσοφίαν πρῶτοι τὴν ἀλήθειαν καὶ τὴν φύσιν τῶν ὄντων and says that, in order to deny generation ex nihilo, they φασιν οὔτε γίγνεσθαι τῶν ὄντων οὐδὲν οὔτε φθείρεσθαι. This looks like an equivalent to b12 as usually read, i.e., (3). On the other hand, Aristotle goes on to add that this drove them to monism, which makes it seem that he really has here in view the radicalization of this view by the Eleatics.

12

-- not an explicit doxa.18 This does not really amount to presenting the prôtoi as sharing the post-Parmenidean agenda of giving deflationary accounts of substance, generation and destruction. Aristotle’s hints towards (3) thus bring out a significant methodological point. In discussing Empedocles and Anaxagoras, he will be explicitly concerned with the implications and commitments of their views as much as with their explicit doxa: the same is true, though less visibly so, in his treatment of the prôtoi.

A final complication comes in 983b17-18, quoted above. As Ross translates, “for there must be some entity -- either one or more than one -- from which all other things come to be...”. But here our mss have (uniformly, according to both Ross and Jaeger) δεῖ γὰρ εἶναί τινα φύσιν. If δεῖ is right, then the postulation of an ongoing substrate is presented here in Aristotle’s own voice, presumably in a kind of temporary identification with the prôtoi. Rejecting this as impossible, Ross says, “It is necessary to read ἀεί with Bywater, or δεῖν with Wirth, instead of δεῖ, since the clause is still concerned with what the early philosophers thought” (1958, 129).19 Now there is rather a lot at stake here between ἀεί and δεῖ (or δεῖν). If we emend to the former, then the line merely affirms, yet again, the commitment of the prôtoi to an eternally persisting substrate. If we read δεῖ with the manuscripts, however (or δεῖν), then this position is presented as a theoretical desideratum: something must persist through all change. And it is presumably in order to meet that desideratum that the prôtoi proceed to postulate their various archai. But why would the prôtoi think that there must be an enduring substrate? An answer is suggested by some strong wording in the Gen et Cor.: “that coming-to-be proceeds out of nothing pre-existing <is>... a thesis which, more than any other, preoccupied and alarmed the earliest philosophers” (I.3, 317b28-31). Thus the transmitted reading, δεῖ, seems to preserve an important Aristotelian point about the thought of the prôtoi. Their agenda is to explain generation and destruction while denying -- and, at least in part, in order to deny -- that it ever

18 Cf. Barnes 1982, 41.19 Ross’ text follows Bywater and changes to ἀεὶ; his translation, however, has ‘must’.

13

occurs ex nihilo, particularly in the case of the cosmos as a whole. The cosmos must be formed out of a single enduring substrate, from which particular onta are in turn generated, providing ‘the all’ with unity and continuity.

We might still wonder why Aristotle attributes ‘material monism’ in the sense of (1) to the prôtoi in the first place -- for this is all the more puzzling once we take (2) and (3) off their agenda. The answer may seem obvious: after all, Aristotle’s purpose in attributing (1) to the prôtoi is clearly to show that they therefore count as having recognised the material cause. And interpreted otherwise -- if the archê of Thales, for instance, were seen by Aristotle as simply an ex hou and eis ho, without being a hupokeimenon -- it is not clear what philosophical interest they would hold for him. For if the archê is not an enduring substrate, then, as a generating cause of onta which are different in kind from itself, it must really reduce to a kind of efficient cause. But without any conception of a formal cause there is no way to explain what an efficient cause does, leaving only an arbitrary, Hesiodic succession. Whatever else it may be, Aristotle’s reading of the prôtoi as material monists is a charitable one: the terms of his project allow for no other way for them to be included among his ancestors.

But Aristotle’s commitment to (1) may also be shaped by a certain awkwardness he is faced with in constructing his investigation. Aristotle’s interest is in what the prôtoi have had to say about causes and principles; he admits at the outset that he will thus be mining their episkepsis tôn ontôn for their views about the archai. Admittedly, this need not entail conflation of onta and archai; but Aristotle does at some points speak of the archê as ousia (notably at 983b10). And in Metaphysics Z, he presents what looks like the A.3 account of the archai under the rubric of to on:

“And indeed the question which was of old and is now and always the object of inquiry, and is always puzzled over [to palai te kai nun kai aei zêtoumenon kai aei aporoumenon], viz. what being is, is just the question, what is substance? For it is this that some assert to

14

be one, others more than one, and that some assert to be limited in number, others unlimited. And so we also must consider chiefly and primarily and almost exclusively what that is which is in this sense.” (Ζ.1, 1028b2-7)

Aristotle here presents a diairesis of positions which corresponds reasonably well with the progression of archai given in Metaph. A.3, from ‘one’ (monists such as Thales) to ‘many’ (various pluralists), some of whom opt for ‘limited’ (Empedocles) and others for ‘unlimited’ (Anaxagoras). But here these are presented as accounts of ti to on, used in turn as a proxy for ousia. So Aristotle may well see his project in A.3 as involving a repurposing of earlier theories of to on, to serve as theories about the archai. But what would have made it seem natural to him to read the earlier thinkers as theorizing about to on and ousia in the first place -- to understand Thales’ water, for instance, as his theory of what all things really are? The answer is not far to seek. I will argue in sections IV and VI that A.3 is enormously influenced by the ‘doxographic’ passages of Plato’s Sophist, which, like Hippias’ work on ‘related ideas’ (see section III), serves Aristotle here as a kind of ‘reference text’.20 Now in the Sophist the Eleatic Visitor presents the early philosophical debate as a debate about ta onta, tout court. It is about ta onta that, as he puts it, his predecessors asked posa kai poia: “how many real things there are and what they are like” (Sophist 242c5-6). This catch-phrase for the framing of an inquiry is used by Aristotle for the account he takes himself to have vindicated in A.7 (περὶ τῶν αἰτίων καὶ πόσα καὶ ποῖα, 988b17); and the phrase is clearly alluded to in A.3 when he says that the prôtoi disagreed as to the plêthos and eidos of the archai (983b19).21 Moreover, that Aristotle himself accepts this Platonic reading of the

20 I do not mean this phrase to imply anything about exactly how Aristotle used either Hippias or the Sophist. But I do mean suggest a closer relation than would be conveyed by ‘source’. These texts do not just supply Aristotle with raw materials; their mode of presentation has detectable effects on his own.21 For this formula cf. also De Caelo 277b25, NE 1115a5 and 1135a14, Gen. et Cor. 329b3, Hist. An. 505b23, Meteor. 338a23, Part. An. 660a7, Phys. 194b17, Pol. 1299a31, Rhet. 1368b32, 1369b29.

15

early debate is clear, not only from his admission in A.3 that his predecessors were engaged in an episkepsis tôn ontôn but from his construal of the same material in terms of onta and ousia in Z.1.

Be this genealogical suggestion as it may, if Aristotle takes the default reading of the prôtoi to be as giving accounts of ta onta or ti to on, we can see why he would tend to fall into ascribing to them the implausibly reductionist (2) and (3). After all, if water is Thales' conception of ti to on, it does indeed follow that water is what things really are. And if, as I have argued, Aristotle actually holds back from fully ascribing these claims to the prôtoi, it is presumably because his somewhat sneaky repurposing of their views as accounts of archai leads away from this reading.

The suggestion, for which I will argue more fully in the following sections, is that in A.3 Aristotle encounters the prôtoi as already the inheritor of one or more interpretive traditions. And the most important of these, the Platonic interpretation given in the Sophist, seems to have presented the theories of the prôtoi as accounts of what things really are, thus entailing all of (1) (2) and (3) above. This is a reading which Aristotle in other places is willing to take as given; but in A.3 he reconfigures it by treating their theories as accounts not of ta onta but, more restrictively, of their archai. This change is an obvious improvement by our interpretive lights; yet it seems to be made not as a direct response to the evidence, so far as one can tell, but as a repurposing to fit the local demands of Aristotle’s project.

III. Thales, the Theologians and Hippo (983b20-4a5):

3. Thales, with an excursus on his putative predecessors:a. Thales, archêgos of this philosophy, said that [a] the archê was water, and thus that [b] the earth floated on water; his reasons were perhaps that [c] the nurture of things is moist, and [d] heat itself comes from moisture, and lives by it; and [e] seeds have a moist nature, and water is the archê of moist things. b. Excursus on the ‘theologians’ (3b27-4a2): some hold that ancient thinkers about the divine held the same position;

16

c. The evidence: (i) they made Ocean and Tethys parents of generation; (ii) they made water (‘Styx’) what the gods swear by, and the oldest thing is most honoured, and what one swears by is the most honoured; d. But whether this is really a doxa peri phuseôs is unclear.e. So we will start with Thales, and his thesis that water was the archê -- and we will be omitting Hippo.

Aristotle’s emphasis on rational reconstruction continues in his treatment of the archêgos Thales:

“Thales, the founder of this school of philosophy, says [a] the principle is water ([b] for which reason he declared that the earth rests on water), getting the notion perhaps from seeing that [c] the nutriment of all things is moist, and that [d] heat itself is generated from the moist and kept alive by it (and that from which they come to be is a principle of all things). He gets his notion from this fact, and from the fact that [e] the seeds of all things have a moist nature, and that water is the origin of the nature of moist things.” (983b20-27)

This treatment is much fuller than that given to the other prôtoi. Thales’ archê-selection [a] is followed by [b] as a corollary. This is perhaps to give a taste of the sort of use which the prôtoi made of their archai, though it has been plausibly argued that Thales’ reasoning is more likely to have been the other way around.22 Aristotle then turns to assign reasons -- empirical reasons [ὁρᾶν, b23] [c], [d], and [e] -- for the particular choice of archê made by Thales. Nothing comparable is offered for the later prôtoi listed; presumably the grounds offered here are supposed to stand in for the kind of reasons relied on by the others mutatis mutandis.

The reasoning attributed to Thales is avowedly speculative.23 The first reason mentioned, [c], is qualified with an isôs, and though this is not repeated 22 E.g. Cherniss 1935, 4. The doxa is repeated at De Caelo 294a29-30. 23 As Ross notes, Aristotle always speaks of Thales’ views with caution: cf. de Caelo 294a29, de Anima 405a19, 411a8, Pol. 1259a6, 18.

17

there is no obvious reason to give more credence to the others. Later legetai is applied even to Thales’ opinion about the archê (984a2). The reasons unfold in a stuttering, spontaneous-looking way: the διά τοῦτο at b25 suggests that [d] is an elaboration of [c] (Ross takes the two as a single reason), and [e] seems to be presented as an afterthought or correction. Even if [d] is merely a special case of [c], it suggests some significant lines of supporting argument, both cosmological and biological. Moisture and warmth are both signs of life; 24 and it seems to have been a common Presocratic view that fire as we experience it is ‘nourished’ by moisture, so that by extrapolation, the fire of the heavenly bodies must depend on exhalations from the sea (Meteorology II.2, 354b33-5a32).25

Now at least some of the reasoning here attributed to Thales has been drawn by Aristotle from Hippo. In De Anima, Aristotle ascribes [e] to Hippo – again with a slight -- as a basis for holding that the soul is water: “Of more superficial writers, some, e.g., Hippo, have pronounced it to be water; they seem to have argued from the fact that the seed of all animals is fluid, for Hippo tries to refute those who say that the soul is blood, on the ground that the seed, which is the primordial soul, is not blood”.26 And as Burnet observed, it is plausible that the arguments of the first Milesians were for the most part cosmological and meteorological, whereas reasoning from the nature of seed is more plausibly the work of a fifth-century medical writer like Hippo (1930, 48). Moreover, [d] (and thus, perhaps, the more general [c]) should be attributed to 24 “From this largely biological point of view, hot and wet are together opposed to cold and dry” (Lloyd 1964, 272). 25 At 354b32-5, Aristotle indicates that his target is much wider than the Heracliteans, and he himself makes some use of the ‘observation’ they depend on: “the fire we are familiar with lives as long as it is fed, and the only food for fire is moisture” (355a3-5). Cf. the discussions of Cherniss 1935, 133-6 and Graham 2006, 58-62. It is striking that Aristotle’s discussion of the topic emerges as a digression from the question of whether the sea is the source of fresh water. Thus the only two non-medical phusikai doxai which we arguably possess for Hippo, that the sea is the origin of fresh water and that the wet nourishes the hot in cosmological contexts, are discussed in conjunction by Aristotle. Might he be following the order of some earlier proto-doxography in which Hippo figured regularly?26 405b2, trans. Smith (ROT). The moistness of seed is also noted in the brief account of Hippo at Hippolytus Ref. I.16.

18

Hippo as well.27 According to Hippolytus, Hippo claimed that water (or the cold) and fire (or the hot) are both archai, but that water begets fire, which in turn overmasters it to form the cosmos. In other words, that the hot comes from the wet is the most basic principle of Hippo’s cosmogeny. Moreover, returning for the moment to [b], our one fragment of Hippo’s work, from a scholion to Homer, is a lumbering syllogism proving that the sea, since it goes deeper than wells, springs etc., must be the source of fresh water:

“All drinking waters come from the sea. For the wells from which we drink are surely not deeper than the sea is. If they were, the water would come not from the sea but from somewhere else. But in fact the sea is deeper than the waters. Now all waters that are higher than the sea come from the sea.”28

Here we have the purported Thalean doxa that the sea runs deeper than the earth and underlies it; so it seems plausible that Hippo would have proposed this argument in the context of argument for [b]. If we take these points together, it turns out that all of Aristotle’s report on Thales can be read as deriving from Hippo.

This report needs to be put into context with the digression (or retrogression) into the poets which immmediately follows Aristotle’s account of Thales’ reasoning. As Bruno Snell argued in an important paper, this passage seems to be (together with certain proto-doxographic passages of Plato’s Theaetetus and Cratylus) derived from Hippias’ anthology or proto-doxography of ‘related ideas’.29 Having raised the question, Aristotle dismisses the poets with little explanation, saying only “It may perhaps be uncertain whether this opinion about nature is primitive and ancient” (983b33-4a2). The doubtful point is whether the doxa that water is the archê can legitimately be attributed to

27 Simplicius Phys. 23.21-9 attributes [c] to Thales and Hippo alike.28 Schol. Homer. Genev.197, 19 Nicole, on Iliad ΧΧΙ.195 = DK38B1, trans. Barnes 1987.29 See Snell 1944, Mansfeld 1983, 1986, Patzer 1986.

19

the poets;30 since the reasoning cited in favour of doing so is valid,31 Aristotle’s scepticism must be about the very idea of reading poetic claims about Ocean, Styx et al. as expressing a phusikê doxa about water. The issue arises again at the start of A.4, when Aristotle notes, probably prompted by Hippias once more, that Hesiod might be credited with discovery of the final cause; and here too the question is basically punted (984b31-2). 32 In B.4, Aristotle for a third time invokes the poets only to dismiss them. He complains, “the school of Hesiod and all the mythologists thought only of what was plausible to themselves, and had no regard to us” (1000a9-11), in treating ambrosia and nectar as somehow, mysteriously, causes of immortality. He brushes them aside as not worth serious investigation; “those, however, who use the language of proof we must cross-examine and ask why, after all, things which consist of the same elements are, some of them, eternal in nature, while others perish” (1000a18-22). Here we seem to have what was lacking in A.3-4: a programmatic statement of demarcation, with the mythologists excluded on the grounds of unclarity and lack of argument. Still, even here matters are not so simple. It is striking that the phrasing of Aristotle’s initial complaint is taken from his other ‘reference text’ for A.3, Plato’s proto-doxography in the Sophist; there it is complained that the earlier philosophers were unclear, with the same sarcastic suggestion that this shows a self-involved contempt for their

30 See Mansfeld 1985a, 132-4 and Palmer 2000, 184-5. Note that this does not mean or entail that the poets expressed themselves unclearly: Aristotle’s remarks about the unclear state of our knowledge about certain views (cf. 984b18) should not be conflated with his complaints that various authors have expressed themselves unclearly. The latter criticism actually seems to be used exclusively for philosophers, notably Xenophanes and Empedocles (986b21-5, 993a22-4, cf. 985a4-6 etc.) (contra Mansfeld 1986, 41-44). On the ‘criterion of clarity’ and Aristotle’s use elsewhere of the theologoi, see Palmer 2000. 31 Though awkwardly expressed: Water=Oath, Oldest=Most Honoured, Oath=Most Honoured; therefore Water=Most Honoured and therefore Water=Oldest, ie water is the archê. Not a valid syllogism for class inclusions, of course, but these are identities.32 Thus Frede 2004, for instance, overestates the extent to which Aristotle is concerned in Metaph. A.3 or elsewhere to ‘make a clear cut’ between the philosophers and their predecessors. Palmer 2000 and Sassi 2002, 66-70, offer more nuanced views. Palmer indeed makes a strong case for the claim that “it is not the case that for Aristotle the history of philosophy begins with Thales” (2000, 182).

20

audience.33 Moreover, Aristotle’s main complaint here is that the mythologists -- and philosophers like Empedocles -- are arbitrary, using the same causes for opposite effects. And this criticism presupposes that we can legitimately attribute to the mythologists the same kind of doxai as the philosophers. This suggests that the question left hanging in A.3 -- is it legitimate to interpret the mythologists in such a way that their works yield doxai equivalent to those of the philosophers? -- is in fact to be answered in the affirmative (as also seems to be assumed at Metaph. Λ.6 and N.4). What leads to their exclusion from the present discussion is a different and subtler point: there is no point in trying to investigate what they say, whereas the philosophers may fruitfully be questioned. I will discuss Aristotle’s method of dialectical cross-examination in section VI.34

It is in turning back from the poets to Thales that Aristotle finally mentions Hippo, with an abrupt sideswipe: “Hippo no one would think fit to include among these thinkers [μετὰ τούτων], because of the paltriness of his thought” (983b4a3-5). Air and fire will each get a pair of advocates: Aristotle here considers the possibility of adding Hippo to round out the hydromonist team, and rejects it. This thinking-aloud may seem less odd if we consider again Aristotle’s use of his reference texts. We already have good reason to believe that Aristotle’s excursus into the poets is prompted by a counterpart passage in Hippias’ work on related ideas.35 Now we have no direct evidence that Hippo was included in Hippias’ work. However, the fragment of Hippo discussed above is from a scholion to Iliad XXI.195, where it is introduced as

33 Both passages use ὠλιγώρησαν and deny that the earlier thinkers φροντίζειν about our understanding (Sophist 243a6-7, Metaph. 1000a10-1). Frede notes the parallels (2004, 43).34 Cf. also 988b17 on the use of earlier philosophers as ‘witnesses’, and cf. Eudemian Ethics 1216 b26-35, discussed below. The basic criticism here goes back at least as far as Plato’s Protagoras, where Socrates retroactively rejects Protagoras’ proposal that they discuss ethical issues via the poets. Poets cannot be interrogated about what they say; people interpret them differently and these conflicts can never be decided. So we should set the poets aside and test our own logoi in discussion (Prot. 347e-8a).35 Patzer thinks it likely that Aristotle’s reference to Hippo is prompted by his inclusion in Hippias, and notes that DK38B1 would have fit Hippias’ purposes in relation to the locus discussed by Snell (1986, 40-1).

21

coming from the third book of Crates’ Homeric Studies: “Then in the third book [of his Homeric Studies] he says that the later natural scientists also agreed that the water which surrounds the earth for most of its extent is Ocean, and that fresh water comes from this.”36 The quotation from Hippo follows, and is followed by a standardized-looking claim that Homer said the same thing.

Now we might well wonder how Crates, a scholar and (probably Stoic) philosopher of the second century BC, came by Hippo’s ipsissima verba on this particular point. Content and context here point to the same answer. For that Homer is in agreement with the phusikoi as regards Ocean is, after all, just the point which Snell traced back to Hippias’ work on ‘related ideas’. Mansfeld has noted that in doing so Hippias must have practised something very like allegorical interpretation, and that the Stoics may have drawn on his works for their own allegorical extravaganzas.37

Given this consonance of content and function, and the apparently very restricted circulation of Hippo’s (non-medical) ideas in the Hellenistic period (see Appendix), I would suggest that Crates’ source is likely to have been Hippias or some closely derivative text. And this is worth factoring back in to our reading of A.3. I have already argued that Aristotle’s report on Thales is probably in its entirety a retrojection of Hippo; and given Aristotle’s reliance on Hippias here, we should probably see his understanding of Hippo as deriving from Hippias rather than from Hippo’s own works. In that case the engagement with Hippias here is much more extensive than Snell noted: it can be traced through the entirety of Aristotle’s account of Thales, with the excursus, and even continues as Aristotle launches in to his list of the material monists.

However, this does not explain why Aristotle thinks it worth mentioning that Hippo is not worth mentioning. The rhetorical trope here is paralipsis or

36 Trans. Barnes 1987.37 “It is therefore entirely possible that at least some of Hippias’ quotations were taken over by the Stoics, and that they lived on in the professional literature explaining the poets, or the Poet, as well as in the doxographical accounts” (1985a, 143). We cannot be sure that Hippias cited Iliad XXI, 195-6, but it would have been extremely pertinent (cf. Guthrie vol. 1, 60, Mansfeld 1983, 88 n. 17, 90, 93 n. 35). Xenophanes 21B30, also on the ocean as source, follows almost at once in the Geneva scholia ad XXI, 196.

22

praeteritio, but the standard agonistic uses of this -- ‘Far be it from me to speculate about the Honourable Member’s drinking problem’ -- seem unlikely to be called for here. Aristotle’s phrasing is striking but unhelpful. Hippo is deprecated for euteleia dianoias -- for the cheapness, shoddiness, low quality of his thought.38 Alexander suggests that it is significant that Aristotle’s criticism is of Hippo’s dianoia, rather than the doxa which he shares with Thales (In Metaph. 26.24-27.3). This is interesting but does not get us very far. For one thing, Aristotle has apparently borrowed much of Hippo’s dianoia as worthy of attribution to Thales. For another, the term euteleia is so far from being a normal term of philosophical criticism that it is hard to guess what Aristotle might have in mind by it.39 Likewise phortikoteros, used of Hippo in De Anima (though an argument of Melissus is said to be μᾶλλον φορτικός at Physics 185a9-12). Both terms suggest a kind of aesthetic, moral or even class distaste, and thus are odd ones to use in a purely theoretical context. Perhaps Hippo’s reasoning was somehow uniquely vulgar and tawdry by the standards of Aristotle’s day; but it is hard to imagine what specific features of his thought would prompt criticism of that flavour.40 We might speculate that Hippo’s resolutely materialist account of the soul, scorned in De Anima (and not shared by Aristotle’s Thales) is the real source of Aristotle’s hostility in both texts. Or perhaps Hippo irritates Aristotle simply as a throwback: a late monist without a clue about the moving cause, who thus has no role to play in the

38 Dianoia in Aristotle has a very wide range of meanings, and is commonly used both for a rather broad mental faculty (e.g. De An. 427b15, De Motu An. 700b17, Metaph. 1012a2), and for its exercise (Eud. Eth. 1214a29, 1217a6, Nic. Eth. 1166a27, 1175a14, De Interp. 16b20, 23a33, Metaph. 995a30, 1009a4): we might translate ‘thought’ and ‘thinking’ respectively. The closest uses to 984a5 are in the Metaph. itself, at 985a5 (what Empedocles had in mind, contrasted with his lisping expression), 986b10, 1009a16, etc. Cf. also perhaps De An. 404a17, De Caelo 280b3, Div. per Som. 464a29.39 According to the TLG, Aristotle’s other uses of euteleia are: De div. per Som. 463b15, Part. An. 683a25, Poet. 1448b26 and 1458b22, Pol. 1267b26, 1272b41, 1336b30, and Rhet. 1390b24, 1408a13. All are apparently used to express aesthetic or ethical contempt -- as indeed are all his uses of phortikos other than De An. 405b2.40 In another paper currently in preparation (originally presented at the Symposium Aristotelicum as an Appendix to this one), ‘On Hippo of (possibly) Rhegium’, I survey what can be pieced together about Hippo’s thought, and find no grounds for contempt visible at this distance.

23

developmentalist story told in A.3 about the serial unfolding of causal explanation. But on either count it remains hard to see why he should be singled out for scorn when Diogenes of Apollonia, for instance, is not. I will come back to this question at the end.

IV. Aristotle’s Survey of the Prôtoi (984a5-16):

4. Survey of the prôtoi (and their monistic successors):a. Anaximenes and Diogenes said that the archê was air;b. Hippasus and Heraclitus said that the archê was fire;c. Empedocles said that all four elements were archai;d. Anaxagoras said that the archai were unlimited: for all the homoiomeries are generated and destroyed in the same way that the elements are, by combination and separation.

Aristotle here proceeds by interweaving systematic and chronological ordering principles. The overarching cause-by-cause presentation which structures A.3-7 obviously represents a kind of diairetic principle; but the material cause appeared first because it was first to emerge chronologically. Aristotle’s listing of the prôtoi begins with another gesture towards diairesis: -- “Yet they do not all agree as to the number and nature of these principles”41 -- and the prôtoi are grouped systematically, by their choice of archai. But at least two of the three pairs of single-archê theorists (counting Thales-Hippo as an abortive pair) are ordered internally by chronology,42 and the pairs are in turn set out in chronological order. The four-archai theory of Empedocles and the infinitely-many archai theory of Anaxagoras are most easily read as diairetically ordered, 41 An awkwardness is being papered over here, more successfully in Ross’ translation than in the Greek. The previous sentence specified an underlying nature ‘whether one or many’; and the introduction to the prôtoi at b6-8, to which Aristotle is now reverting after the interlude on their reasoning, also spoke of archas in the plural. Later on as well, Aristotle seems to allude vaguely to early pluralist materialists (984a20); yet none are ever named. 42 This is clear in the cases of Thales and Hippo and Anaximenes and Diogenes. If we accept the traditional floruit of c. 500 B.C. for Heraclitus, Hippasus was probably a generation later (cf. von Fritz 1945 at 383-6). But Aristotle never mentions Hippasus elsewhere.

24

in the sequence one-many-infinitely many. Aristotle seems to be mildly carried away here by this diairetic progression, since the references to both Empedocles and Anaxagoras are proleptic, and we must soon backtrack to Parmenides and the discovery of the moving cause. (This prolepsis cannot represent an impulse to comprehensiveness in listing material archai, since the atomists remain offstage until the end of A.4.) The progression here is complicated by an ambiguity: having covered Empedocles, Aristotle continues: “Anaxagoras of Clazomenae, who, though older than Empedocles, was later in his philosophical activity, says...” Ἀναξαγόρας δὲ ὁ Κλαζομένιος τῇ μὲν ἡλικίᾳ πρότερος ὢν τούτου τοῖς δ’ ἔργοις ὕστερος (984a11-13). The interpretation of the proteros-husteros contrast here is enormously controversial: the phrase translated “but later in his philosophical activity” might equally well be read as expressing philosophical inferiority.43 But for our purposes the important point is fortunately the unambiguous one: Aristotle is warning that he has broken away from the birth-order of the philosophers in his order of presentation. That he bothers to alert us to the breach confirms the importance of the chronological principle even in this primarily diairetic survey of the materialists. This narrative twisting and turning may be easier to make sense of in light of Aristotle’s ‘reference text’ here, the proto-doxography of Plato’s Sophist. At 242c, the Eleatic Visitor launches into a somewhat jokey sketch of earlier thought about ta onta, posa te kai poia estin:

“It strikes me that Parmenides and everyone else who has set out to determine how many real things there are and what they are like, have discoursed to us in rather an off-hand fashion.... They each and all seem to treat us as children to whom they are telling a

43 In principle, that is: of the two, Aristotle does seem to have the higher opinion of Empedocles (see O’Brien 1968). On the other hand, 984b15-20 does not entail that Anaxagoras’ work cannot have been later than Empedocles’; and it is hard to see why Anaxagoras’ inferiority would be relevant here, least of all as an excuse for discussing him after Empedocles. If anything, given the progressivist orientation of A.3, Aristotle’s presentation should be in order of ascending merit. For up-to-date discussion and references, see Sider 2005, 3-11 and Curd 2007, 130-7, and more fully O’Brien 1968 and Mansfeld 1979-80.

25

story. According to one there are three real things, some of which now carry on a sort of warfare with one another, and then make friends and set about marrying and begetting and bringing up their children. Another tells us that there are two -- Moist and Dry, or Hot and Cold -- whom he marries off, and makes them set up house together. In our part of the world the Eleatic set, who hark back to Xenophanes or even earlier, unfold their tale on the assumption that what we call ‘all things’ are only one thing. Later, certain Muses in Ionia and Sicily perceived that safety lay rather in combining both accounts and saying that the real is both many and one and is held together by enmity and friendship.”44

The specifics of Plato’s story here are admittedly very different from the narrative of A.3, and in some respects rather baffling. Here monism seems to be an Eleatic invention, and preceding it are unattributed three- and two-archai cosmogenies of a vaguely mythological, Hesiodic character. 45 What is strikingly similar is Plato’s more careful and skillful combination of chronological and diairetic organization. We are given a ‘countdown’ diairesis from three archai to two to one (cf. Isocrates, Antidosis 268), followed by synthetic views; but this is also presented as a chronological development from traditional cosmogeny through Eleaticism (‘starting from Xenophanes and even before’) to (‘later on’) post-Eleatic synthesis.

Despite the very different content of his story, Aristotle in A.3 seems to be looking back to the methods and strategies of the Visitor’s account. First, the question is still poia kai posa, as pursued through an episkepsis tôn ontôn: as I noted in II, though his topic is the archai, Aristotle is concerned to keep his account in contact with a Platonic understanding of the same tradition as one about einai and ta onta. Second, Aristotle follows Plato in positing a division 44 Trans. Cornford 1934.45 Cornford ad loc. is silent on the identities of the two- and three-archai theorists. Pherecydes is often mentioned as a candidate for the latter: but see the cautious remarks of Schibli 1990, 195-6.

26

into three stages -- the three stages, indeed, that we still find in standard histories of Presocratic philosophy, viz pre-Parmenidean, Eleatic, and a post-Parmenidean synthesis (with the Pythagoreans, a poor narrative fit, apparently excluded from both the Sophist and A.3). Third, Aristotle follows the Sophist in framing this story as one of an inquiry into problems about change, with real progress towards a solution coming only towards the end. Finally, there is the structural similarity I just noted: in both texts, the chronological and diairetic principles are both used and, insofar as possible, made to converge. I will say more about the Sophist as model in section VI.

V. The Progress of the Inquiry and the Discovery of the Moving Cause (984a16-4b8):

5. Progress and the discovery of the moving cause:a. As inquiry progressed, the subject-matter itself made a way and compelled inquiry:b. For the material cause alone cannot explain generation and destruction: a substrate does not change itself.c. The earliest monists didn’t trouble themselves about this; but other monists, ‘as if defeated by the inquiry’, said that the one was entirely unmoved.d. None grasped the moving cause except Parmenides, and he only insofar as he said that the aitiai were in a way two. e. It is much easier for pluralists to speak of the moving cause: for instance, they can use fire as having a nature which moves, and water and earth and the rest as being the opposite.

This section of the chapter is shaped by a pair of fascinating but somewhat enigmatic statements at 984a18-9 and -- running beyond the brief of this paper, but indispensible to understand the former -- 984b8-11. First:

“From these one might think that the only cause is the so-called material cause; but as they thus advanced [προϊόντων], the very facts [αὐτὸ τὸ πρᾶγμα] showed them the way and joined in forcing them to investigate.”

27

And second, after the account of Parmenides and the unnamed pluralists:

“When these men and the principles of this kind had had their day, as the latter were found inadequate to generate the nature of things, men were again forced by the truth itself, as we said, to inquire into the next kind of cause.”46

I will call these together the internal logic claim. This striking and hugely important claim -- that philosophical progress in uncovering the causes was somehow caused by the subject-matter itself, and so was presumably necessary or inevitable -- seems to occur nowhere else in Metaph. A,47 but variants appear in other works of Aristotle. One is in a closely parallel context in the Physics, after Aristotle’s review of his predecessors on the question of the ‘number and nature’ of the principles. The moral drawn from this survey is that the principles must be contraries: for all identify their elements and principles with the contraries, “giving no reason indeed for the theory, but constrained as it were by the truth itself” (188b27-30). Then in Part. An. both Empedocles and Democritus are said to have been led to recognise the formal cause or essence in spite of themselves. Empedocles, led by the truth itself [ἀγόμενος ὑπ’ αὐτῆς τῆς ἀληθείας], was compelled [ἀναγκάζεται] to identify the nature of a thing with the ratio (642a18-21). And Democritus, because he was ‘carried away by the matter itself’ [ἐκφερόμενος ὑπ’ αὐτοῦ τοῦ πράγματος], was the first to touch on the essence and definition of substance (642a24-8). These references are a curiously close match to the pair of claims 46 μετὰ δὲ τούτους καὶ τὰς τοιαύτας ἀρχάς, ὡς οὐχ ἱκανῶν οὐσῶν γεννῆσαι τὴν τῶν ὄντων φύσιν, πάλιν ὑπ’ αὐτῆς τῆς ἀληθείας, ὥσπερ εἴπομεν, ἀναγκαζόμενοι τὴν ἐχομένην ἐζήτησαν ἀρχήν. It is controversial whether this second allusion describes a second intervention, prompting inquiry into the final cause, or is just a reiteration of the first. Ross argues for the latter reading (1958, 135-6), which is supported by 985a11 and b21. But this requires awkwardly taking the πάλιν to go with ὥσπερ εἴπομεν; and it is hard to see how it can make any sense of either the μετὰ ...τὰς τοιαύτας ἀρχάς or the ἐχομένην... ἀρχήν.47 Unless we add Parmenides’ having been compelled by the phainomena themselves, A.5, 986b31 -- which would perhaps be a gloss on the A.3 version of the claim.

28

in A.3, of which one, likewise, refers to the truth as forceful agent and the other to the pragma itself. But what is compelled in Part. An. is recognition of the formal cause, to which the internal logic claim is never applied in Metaph. A. Indeed, strictly speaking, in A.3 what is compelled is not recognition of any cause, but further inquiry or search [συνηνάγκασε ζητεῖν, ἀναγκαζόμενοι ἐζήτησαν]. In the internal logic claim as it appears in A.3, Aristotle’s claim is not that particular aspects of reality force us to acknowledge them, but that a partial grasp of reality makes further explanatory puzzles irresistable. 48 The question that is forced on the materialists as they progress would seem to be: given the continuity and stability provided by the persisting archê or archai, what happens when a natural substance comes into existence or goes out of it -- and why does it happen just when it does?

As scholars have noted, the internal logic claim must be an expression of two more general Aristotelian views: first, his understanding of the development of the arts and sciences as following a teleological (and even cyclical) progression;49 and second, more generally still, his epistemological optimism.50 Aristotle holds that all the arts and sciences have been developed and perfected many times, with the epochs of human culture divided and destroyed by recurrent natural disasters. Fragments of wisdom somehow survive in mythic form to provide the starting-point for the next civilization (Metaph. Λ.8). Aristotle also seems to think of the development of a science

48 Unfortunately Aristotle does not spell out how it does so. The clearest account I know is Irwin’s: “<the material monists> have to explain the variation of non-essential properties at different times; and they cannot simply mention the permanent subject and its composition, since these alone do not explain why the changes happen at some times and places rather than others.... Even though they do not initially recognize the efficient cause, their own questions, not some questions raised by another theorist from within quite a different theory, require the recognition of an efficient cause” (1988, 158). This is an attractive picture of how philosophical investigation might progress as a series of self-propelling problems and solutions. The difficulty is that it presupposes that Aristotle’s monists were all along trying to explain ‘the variation of non-essential properties’, which seems unlikely. Aristotle’s critique in A.8 describes them as “trying to state the causes of generation and destruction, and doing natural science with regard to all things” (988b26-7). 49 Mansfeld 1985c, 3-10, Palmer 2000, 196-202.50 Cf. Denyer 1991, 183-213.

29

within each epoch as naturally exhibiting the kind of teleological progress to perfection characteristic of natural organisms. This is presumably because -- and this is the second view -- we are naturally adapted to opt for the true over the false: humans “have a sufficient natural instinct for what is true, and usually do arrive at the truth” (Rhetoric 1355a15-8, cf. E.E. 1.6, 1216b28-31, N.E. 1.8, 1098b27-9, Pol. 1264a1-4). This general orientation to the truth implies that (barring special countervailing factors of some sort) good explanations will tend to drive out bad ones, the false parts of theories will tend to be corrected, and the sciences, philosophy included, will tend naturally towards completion. The link between this epistemological optimism and Aristotle’s progressivist conception of philosophy is made explicit in Metaph. α.1, where it is perhaps meant to serve as a kind of retrospective gloss on the project of A.3-10. As it explains, the investigation of the truth is in a way hard, but in another way easy. While no one gets the entirety of it, “no one fails entirely, but every one says something true about the nature of things, and while individually they contribute little or nothing to the truth, by the union of all a considerable amount is amassed” (993b1-5). Contributions to the arts and sciences build on each other and all are of value (993b11-19); so it is not surprising that Aristotle would tend to view progress in thinking about the archai as virtually inevitable.

VI. Notes on Genre and Method in Metaphysics A.3, 983a24-4b8: Aristotelian Dialectic

It is natural to take Aristotle’s predecessor-surveys, and Metaph. A.3 in particular, as dealing with the endoxa -- the reputable opinions which he identifies as the starting-point for philosophical inquiry in a famous passage of NE. VII.1 (1145b3-7).51 I think that this is basically right, but needs to be made

51 On Peripatetic doxography as dialectic more generally, see Mansfeld 1992b and Baltussen 1992.

30

more precise. For how to fit together Aristotle’s practice in A.3 with his account of dialectic in the Topics and his programmatic remarks elsewhere is a vexed question. The dialectic of the Topics is essentially elenctic, a matter of refutation through question-and-answer,52 and it is difficult to read Aristotle’s project of subsumption in A.3 that way. Moreover, in thinking about A.3 as dialectic about the archai, we need to be careful not to assume that the referent of archai is fixed. The term seems to be used for three different things in Metaph. A: (1) in A.1-2, for the archai in an honorific sense, the knowledge of which deserves the name ‘wisdom’; (2) in A.3-10, for the archai of the cosmos and ‘the all’, as identified in various ways by the early philosophers; (3) for Aristotle’s own system of the four causes -- but these are referred to as archai only somewhat indirectly and, it seems, in order to preserve some kind of continuity with (1) and (2) (e.g., 983b3, 4a25-6, 988a21-3, b18-19). Aristotle habitually collocates archai and aitiai, but clearly prefers the latter in relation to his own theory from the Physics (983a24-7, b54a17, 984a25, 988b16-21, 993a11). And it is not clear that the fourfold causal scheme should count as archai of physics (Physics I at any rate investigates the archai of physics in a very different sense); and even if they do, that will not be sufficient to make them the archai of first philosophy as well. Now A.1-2 are concerned to identify certain marks of the science which constitutes wisdom (it is universal, theoretical, an end in itself, divine, etc.) which in fact are met quite well by the philosophical inquiry narrated in A.3-7. So one might think that the argumentative structure of Metaph. A as a whole runs as follows: (1) wisdom is argued to be knowledge of the ‘honorific’ archai (i.e., the archai to be studied by first philosophy) (A.1-2); (2) the archai of the earlier philosophers are discussed as likely candidates for these honorific archai (A.3-7); (3) the archai of the earlier philosophers are shown to reduce to the Aristotelian four causes (A.3-10).53 Thus the four causes turn out to be the leading candidates for the 52 Cf. Brunschwig 1967: “La tâche du questionneur est d’élaborer une argumentation tendant à établir la proposition contradictoire de celle que soutient le répondant (xxix-xxx).53 On the overall project of Metaph. A see Frede 2004.

31

honorific archai. And perhaps this is the structure of Aristotle’s argument; but it is important to bear in mind that Aristotle’s next move (in Metaph. B) is to set out aporiai, an operation which still belongs to dialectic, rather than to proceed to any systematic science of metaphysics. 54 Moreover, it is doubtful that any later part of the Metaphysics treats the four causes as a set of ‘first principles’ in the Post. An. sense, as opposed to providing a handy framework and starting-point for further dialectical inquiry into what those first principles might be. So if A.3-10 is dialectic, it is of a kind which is preliminary even in relation to the dialectical investigations of the later books of the Metaphysics.

How A.3 operates as dialectic can be made more visible by further comparison with Plato’s Sophist. I have already noted the parallels between Sophist 242cff. and A.3 as surveys of views poia kai posa ta onta/hai archai. But in terms of method, A.3 is still more endebted to the second phase of Plato’s discussion, Sophist 246a-9d. In this famous passage, ‘the battle of the gods and the giants’, two earlier views are expounded and scrutinized at much greater length: a kind of materialism and what sounds very much like the middle-period theory of Forms. The Visitor launches into a critical dialogue with each faction in turn. This involves positing ‘improved’ materialists, who are willing to answer ‘in a more orderly way’ than their real-life originals. In defense of this idealization, the Visitor notes that agreement from these ‘better’ interlocutors is worth more anyway; and that our concern is not with them but to seek the truth. In the interrogation which follows, he begins from the fact that the materialists set down and even define being as body (246c7, b1). But they must admit that there is such a thing as an animal; that animals are ensouled bodies; that souls therefore exist; that souls can be qualified as just and intelligent or the opposite; that they gain these qualities by the presence of 54 So the Metaphysics itself shows that a predecessor-review and a working out of aporia are, or at least may be, distinct dialectical operations (whereas in Physics I, say, they seem to be combined: nb I.8, 191a23-5). Aristotle concludes in A.10: “but let us return to enumerate the difficulties that might be raised on these same points; for perhaps we may get some help towards our later difficulties” (993a24-7). The identity of these two sets of difficulties is controversial, but at any rate those of A.8-9 are clearly ad hominem and a completion of the predecessor-review, whereas the aporiai of B are forward-looking.

32

justice and intelligence and their opposites; that justice and intelligence are not visible and... at this point the answers fall into disarray, but the conclusion is that the giants must admit that there are non-perceptible and by extension non-bodily existents. The Visitor at this point proposes to them a new criterion of being, one coherent with this newly expanded ontology: being is power [dunamis], the ability to affect and be affected. (Note that this new criterion might plausibly be claimed to be what the giants ‘really’ had in mind all along: it would explain why they readily accepted that the qualities which affect bodies must exist just as much as those bodies.) As for the friends of the Forms, they are presented as initially denying the new criterion: ‘Being’ properly speaking does not have such capacities to act and be acted upon, since all ‘becoming’ is excluded from it. However, they in turn must admit that knowing and being known are a form of acting and being acted on; that what really is has intelligence; that it therefore has life; that it therefore has soul and is alive; and that it is therefore subject to motion and change (though not complete motion and change, i.e. flux). In sum, each faction is forced to concede, on the basis of its own thesis in conjunction with further premises acceptable from its own viewpoint, the contradictory of that initial thesis, which in turn supports a new constructive account of being.

The form of each argument is thus that of a Socratic elenchus with a positive upshot affirmed. And this upshot -- that both change (or motion) and rest really exist, and that Being must be understood as not the same as but different from both of them -- constitutes the basic framework of the account of the ‘greatest kinds’ which the Visitor goes on to produce. Each side can thus be seen to have a positive contribution to make to that framework, once its one-sidedness is corrected. The materialists were right that what exists must have powers to act and be acted upon (a point to be pressed again at 252c-e); but they were wrong to assume that such powers could only be realized by bodies. The friends of the Forms were right that what really is must be knowable and stable, but wrong to assume that this entailed complete stasis. As each is corrected, they converge on the correct account; and Plato presents this as

33

something that they (or somewhat idealized counterparts) could be brought to recognise themselves through dialectical question-and-answer. This question-and-answer is at once refutative and constructive; though superficially elenctic, it is also oriented towards sorting the true from the false, and amassing convergent grains of truth from superficially conflicting sources.

In the Eudemian Ethics, Aristotle describes a dialectical procedure along very much the lines displayed in the Sophist:

“We must try, by argument, to reach a convincing conclusion on all these questions, using, as testimony [μαρτυρίοις] and by way of example, what appears to be the case. For it would be best if everyone should turn out to agree with what we are going to say; if not that, that they should all agree in a way and will agree after a change of mind [μεταβιβαζόμενοι]; for each man has something of his own to contribute to the finding of the truth, and it is from such <starting-points> that we must in a way demonstrate: beginning with things that are correctly said, but not clearly, as we proceed we shall come to express them clearly, with what is more perspicuous at each stage superseding what is customarily expressed in a confused fashion.” (1216b26-35, trans. Woods 1992)

If we read μεταβιβαζόμενοι here as simply ‘after [or, better, ‘with’] a change of mind’, it sounds like a sort of fantasy by stipulation: after they change their minds, everyone will agree! But this would make nonsense of the crucial claim that everyone has something of his own to contribute. So Aristotle cannot have in view here the simple replacement of other people’s views with his own. Instead, we should see here a reference to the kind of dialectical clarification and correction we find in the Sophist, as the following lines of the passage confirm; it is as we move from confusion to clarity that everyone can be shown

34

to (‘in a way’) agree, just as the gods and giants come to do.55 Several points of the Eudemian passage also recall Aristotle’s procedure in Metaphysics A.3-10. A progression from the confused to the clear and corrected is precisely what he here aims to offer, not only in the narrative progression from more primitive views to more articulate ones but, even more, in the dialectical clarification of those views from the vantage point of Aristotle’s own theory. “But while he would necessarily have agreed if another had said this, he has not said it clearly”, as he says of Empedocles’ approach to the formal cause in A.10, 993a22-4. That everyone, once their views have been fully clarified and adjusted, has a distinctive contribution to make to the truth is then thematized in α.1.

In sum, Aristotle’s project in A.3-10 is one of theoretical subsumption through dialectical engagement. More precisely put, in A.3-10 Aristotle is working with a species of originally Platonic dialectic exemplified in the Sophist (and perhaps also the interrogation of Protagoras in the Theaetetus), one which he himself describes in the Eudemian Ethics. To distinguish this from other forms of argument for which ‘dialectic’ is sometimes used, I will refer to it as ‘Eudemian dialectic’. Eudemian dialectic consists in a critical, dialogic examination of earlier views, in the course of which they are clarified so that what is true in them may be sorted from what is false. The end result is that superficially conflicting views are shown to converge on a framework which can be used as a starting-point for constructive theorizing about first principles. Though Eudemian dialectic is critical, the extent to which it is refutative (rather than being seen as a constructive exercise in the clarification, correction and

55 Cf. also the use of metabibazô at Topics 101a33: dialectic is useful in ‘encounters’, “because if we have correctly reckoned up the opinions of the many, we will speak to them not from foreign opinions but from their own, changing their minds [μεταβιβάζοντες] about anything they do not seem to us have well said”. Again, whatever exactly is envisaged here, again it can hardly be wholesale belief-replacement. What is ‘not well said’ must be not well said for the purposes at hand, e.g., by mixing up several topoi. Aristotle is describing the correction of views not in order to make them identical to his own, but simply to make them usable in discussion: as Smith says, “replacing our audience’s clumsy formulations of their own views with better ones we have worked out in advance” (1993, 351).

35

harmonization of views) is subject to variation.56 In A.3-10 (and likewise De Anima I), the emphasis is on finding the distinctive, though incomplete, contribution to the truth made by the views under examination. In the Sophist, the emphasis is on their correction, and the discussion is elenctic in form; but there is still a profoundly important positive upshot. In the Theaetetus’ refutation of Protagoras and in Metaph. Γ, virtually nothing survives the examination, and the path to first principles is pointed by negation. But these differing outcomes represent not so much differences in method as in the value of the raw materials to hand.

This mode of argument deserves the name of dialectic for several reasons. It is clearly a close descendant of the Socratic elenchus. As per Aristotle’s own account of dialectic, it deals with endoxa (specifically the views of the wise), and takes a dialogic form; it is a preliminary stage of inquiry, and a useful non-demonstrative way of getting clear about prospective archai (cf. Topics I.2). As we can see from A.3-7 itself, Eudemian dialectic is the kind of enterprise for which a collection of doxai would be a necessary resource, and Mansfeld has noted that such collections are proposed for dialectic in Topics I.14.57 At the same time, Eudemian dialectic is not quite the same thing as the dialectic of the Topics; for it need not be syllogistic in form, nor refutative in its fundamental aim.58 Nor does Aristotle’s much-discussed sketch of method in NE VII.1 map properly on to this kind of dialectic. For though A.3 is preparatory to the consideration of aporiai, and might thus be classed as an elaborate way of tithenai ta phainomena, it is hardly the case that the rest of the Metaphysics is simply concerned to vindicate as many of those phainomena as possible.59 56 It is unclear to me how far all Aristotelian dialectic is supposed to be refutative in form, or how the refutative dialectic of Topics VIII is related to any broader genre: cf. Topics VIII, S.E. 2, 165b1-4.57 Mansfeld 1986, 25, Mansfeld 1992b.58 Strictly speaking, according to Topics Ι.12, a dialectial argument is either an induction [epagôgê] or a deduction [sullogismos]: I see no good way to read A.3 as either.59 I cannot here engage with the rich and complex reading of Aristotelian dialectic offered in Irwin 1988. But if I am right to connect Metaphysics A and Γ with the Eudemian Ethics and the Sophist, the gap between the Topics and dialectic as used in the Metaphysics is not primarily to be explained in developmental terms.

36

Eudemian dialectic remains a canonical mode of philosophical argument. In her Tanner lectures, The Sources of Normativity, Christine Korsgaard constructs an account of normativity in part by way of engagement with such predecessors as Grotius, Hume and Kant. In the final lecture, Korsgaard concludes: “I hope by now it is clear that all of the accounts of normativity which I have discussed in these lectures are true” (1996, 164). As she explains, voluntarists like Pufendorf are right about the necessity for a moral legislator -- but wrong about who that legislator is. Realists like Nagel are right that normativity involves objectivity -- but objectivity does not entail what Nagel thinks. Meanwhile Kant is right sans phrase that autonomy is the source of obligation; and Aristotle is (after some corrections) right that normativity gets its authority from human nature. These critical subsumptions are quite similar to Aristotle’s procedures in Metaphysics A; the underlying assumptions about philosophical progress cannot be far from those articulated in Metaph. α.1.

VII. Conclusions on A.3 and ‘History of Philosophy’

The Sources of Normativity is catalogued as ethics, not history of ethics: Eudemian dialectic today lies on the borderlands of ‘history of philosophy’.60

But Eudemian dialectic is clearly not the only genre in play in A.3; for it cannot explain Aristotle’s use of chronology as a central structural principle. This principle governs his decision to begin with the material cause and with Thales in particular. It helps to explain why, having decided to begin with Thales, Aristotle immediately wonders whether he should move back to still earlier thinkers; once he returns to forward gear, it determines the order in which the various candidate archai are discussed. When the introduction of Empedocles before Anaxagoras seems to break with chronology, he is careful to explain the apparent deviation; having jumped ahead by mentioning them both before Parmenides, Aristotle reverts and fills in the missing steps. And all this can be

60 Cf. also Rorty 1984 on posthumous dialectical ‘re-education’ as a historical mode, and Brandom 2002.

37

seen as just continuing a perspective adopted even before A.3: for in A.1 and A.2, Aristotle muses repeatedly on the origins of philosophy in order to identify wisdom with non-utilitarian knowledge (A.1, 981b14-24) and non-productive episteme (A.2, 982b11).

This chronological emphasis is unique in Aristotle’s predecessor-surveys, and calls for explanation. Two complementary lines of explanation seem to me to be available. First, we have plenty of evidence, both internal and external, that Aristotle has a serious ongoing interest in certain kinds of historical question about the sciences.61 We know that Aristotle’s students and successors researched and wrote studies of the history of philosophy, mathematics, medicine and theology: it is hard not to infer that Aristotle himself inspired this ‘historical turn’. And this was just part of a quite general interest in the development of civilization and the arts and sciences, commonly put in terms of the prôtos heuretês -- the subject, apparently, of a whole Peripatetic genre.62 This research in turn belongs to a more general Peripatetic engagement with history of all kinds, which also led to collections of constitutions, political histories and chronologies.63 It would be perverse to insist that, despite this powerful concern with history of all kinds, Aristotle can have had no interest in the history of philosophy as such, or that such an interest can have played no role in shaping Metaph. A.64

Still, this does little to explain why Aristotle would adopt the chronological mode here in particular, when it is dispensible for his purposes in (say) De Anima. The obvious explanation, it seems to me, is that A.3 is special because of the availability of the ‘internal logic’ claim. Aristotle here presents earlier 61 See Zhmud 2006 for a full account, esp. Ch. 4.62 ibid., 149-52; cf. his Ch. 1 on heurematography in general.63 ibid., 136-40. 64 Cf. the sensible Jaeger 1937, contra Cherniss 1935: “We must not separate Aristotle’s interest in the history of philosophy from his historical research in all these other fields of civilization” (p. 354). Zhmud 2006 points to Aristotle’s treatise De Inundatione Nili (preserved only in a Latin abridgement), which gives a doxographic survey of earlier opinions on a problem which Aristotle takes to have been conclusively solved. Zhmud goes too far in inferring the primacy of historical motivations in Peripatetic doxography; but clearly such motivations had some independent force (2006, 143-4).

38

thought in narrative form because in this case he has a story to tell, one in which the progression of inquiries and discoveries takes on a life of its own.65

(Admittedly this is not yet much of an explanation: what we would like to know is why thought about the archai would have been uniquely self-propelling in this way. But so far as I can see A.3-10 offers no answer to that question.) Aristotle’s announced demonstrandum, the negative claim that no cause has been detected beyond his own four, does not require the narrative mode; but in the execution, this purpose is overtaken by a more ambitious and constructive one. In fact, it is doubly overtaken, by a ‘Eudemian’ argument for dialectical convergence and by the historical claim that his own four causes have at least in part been forced upon philosophical inquiry. And the latter is a claim that can only be made good in the narrative mode.

Like ‘Eudemian’ dialectic, narratives driven by the ‘internal logic’ claim are alive and well in contemporary history of philosophy. So, to return to our opening question, it is worth noting how far other features of Aristotle’s method in A.3-7 also fit our understanding of the historian’s undertaking. A number of his moves look very familiar.66 Aristotle worries about scanty evidence and calibrates his claims to reflect the quality of his information. He also refers us to the texts, occasionally with a slight air of bluff (985b2-3). He raises but fudges the problem of the demarcation of philosophy from its precursors and kindred works. He reconstructs reasoning which he finds murky, to tell us what Empedocles or Anaxagoras was really getting at; he offers charitably anachronistic interpretations, usually flagged as such, in the interests of philosophical clarification. Overall, he is at pains to tell a story of philosophical progress, i.e. of increased clarity in the posing and solution of central philosophical problems. If Aristotle is not a historian of philosophy, it is remarkable how many of the historian’s problems, worries, desiderata, and strategms he has managed to accumulate. 65 Cf. Frede 2004, 13-4.66 In the terms of Rorty 1984, still the most useful typology of ‘history of philosophy’, these features place A.3 in the category of both rational reconstruction and (‘Whiggish’) Geistesgechichte.

39

Another marker of genre deserves special mention. This is Aristotle’s praeteritio, noted in section III, of poor old Hippo. For this too is best understood as part of the construction of a specifically historical genre. So far as we can tell, Aristotle has no good reason to dismiss Hippo so scornfully: but he does have good reason to dismiss somebody. To narrate is to select; and to select among philosophers is to offer up a canon. Aristotle’s explicit exclusion of Hippo serves to signal this undertaking of canon-formation. Moreover, in avowing that the exclusion is based on the inferiority of Hippo’s thought, rather than the falsity of his doctrine (or the crudeness of his expression, or any other consideration), Aristotle avows that his canon will be governed by considerations of philosophical merit. To return to a question mooted at the outset, this is surely a large part of what we mean by ‘objectivity’ in the history of philosophy. We are objective as historians when, respecting this distinction between doxa and dianoia, we give a fair hearing to philosophical reasoning with which we disagree. And the historical enterprise depends on a certain optimism about the prospects for convergence between the demands of narrative and of canon -- between considerations of historical importance and philosophical quality. When he drop-kicks Hippo out of canon and story alike, Aristotle affirms that optimistic presumption. He thereby, I suggest, announces himself as something fully recognisable only in retrospect: a historian of philosophy.

So Metaphysics A.3 is a palimpsest of methods and genres. There are vestiges of a diairetic structure; there are twists and turns inspired by Aristotle’s reference texts; there is an exercise in Eudemian dialectic; and finally there is a historical narrative -- a story -- and with that story a canon. The dialectical exercise and the historical narrative are jointly the core of the enterprise; and while the former is in some ways a Platonic inheritance, the latter seems to be all Aristotle’s own.

40