¹ ‘This chapter is a collection of fragments’, Hamlyn, D. W., Aristotle's De Anima Books 2 and 3 (Oxford, 1968), p. 145. See also Torstrik's verdict, quoted by Ross in the apparatus on 431al.Google Scholar

² De anima 3.2 was the subject of my earlier paper (‘Aristotle De anima 3.2: how do we perceive that we see and hear?’, CQ 33 [1983], 401–11). Here, as there, I seek to resolve some difficulties of interpretation in the light of a coherent reading that unifies the topic of the whole chapter. In so far as this forms a companion piece to the earlier paper, I intend that the two should be mutually supporting.

³ That is, even in perception the actualization of the sense is not passive, instigated by an object that simply affects it, but is active, an activity of the sense itself

⁴ This interpretation has been presented by Michael Frede in discussion, but not to my knowledge in print. His notion that there is a contrast depends upon supposing that the mind is not related to its phantasmata as a sense is related to its aisthemata, despite the similarities Aristotle suggests at 431b2–10.

⁵ This might be a use of language such as naming that does not make any assertion because it is not a proposition, or it might be mentioning a proposition without asserting it. In either case the relevant point is that, by contrast, a judgement is equivalent to assertion that such and such is so.

⁶ It is possible that this point is at least alluded to in the note at the end of chapter 6 (430b26–9), which is unfortunately somewhat obscure

⁷ There are some difficulties, of course, with the link here between evaluative judgement and pursuit or avoidance, which emerge in connection with the passage in chapter 9,432b26ff.

⁸ As far as I can see Aristotle assumes in this chapter that thoughts are initiated by phantasmata, which provide their objects, in much the same way as perceptions are initiated by aisthemata, which are their objects. There is no reference to the agent-intellec

⁹ The example seems to allude to our ability to think of future consequences and compare them with present perceptions, and choose which of the two to pursue, even in the absence of the actual items or events that are in the future. The example makes good sense if we read (431b9) as meaning ‘there in the future’, and as ‘here and now’, such that we judge the good and choice-worthy event to be future (i.e. out of sight) but act immediately to attain it.

¹⁰ This is of course the reverse of Aristotle's procedure, since he intends us to find the work on perception enlightening in understanding the intellect, and I am suggesting that we, now, might come to understand the comments on perception in the light of his account of intellect.

¹¹ The reference to exercise with respect to the perceptual mean, 431 a 11, seems to imply that being pleased or pained is a kind of perception that involves the same perceptual faculty, the same mean, as the rest of perception treated as a whole

¹² Since Aristotle sometimes elsewhere refers to the ‘appetite’ () as though it were an explanatory factor in behaviour, this passage should probably be taken as an elucidation of what that language means: in other words appetite is not another faculty but just a way in which perception operates when it perceives pleasant or painful things as such.

¹³ There is a dissimilarity in that white and sweet both belong exclusively to separate modalities, while white and pleasant are respectively a proper sensible and one that belongs on a spectrum available to any sense modality. But this does not alter the fact that two distinct judgements are made in the course of perceiving that some one thing is both white and pleasant or both sweet and pleasant, just as in perceiving that it is both white and sweet.

¹⁴ I am grateful to David Sedley for pointing out this way of taking the sentence. This seems to me to be the easiest of a number of possible ways of resolving the problem without supposing that the text is seriously incomplete. An alternative would be to take as running on from the preceding sentence with no break and explicating what is said there.

¹⁵ The single ‘mean’ refers to the idea that each proper sensible is judged relative to a mean between extremes on a single scale, the sense-faculty itself being characteristically in the mean condition and hence judging its objects by their divergence from its own condition. The natural assumption is that there will be a characteristic scale and hence a distinctive mean for every distinctive sense-modality: e.g. one mean for the range between sweet and bitter and another for the range between black and white; but Aristotle's claim is that ultimately all senses are referred to a single faculty, which judges on one scale by comparison with one mean

¹⁶ (a29). These phrases seem to indicate something that falls under two distinct descriptions or definitions, but is nevertheless one and the same entity. It correlates with the notion that some one thing X that is both F and G can have certain capacities qua F and others qua G. In some cases the individual X might seem to be individuated by matter, but I think this need not be so, and X might be a single form.

¹⁷ . I have avoided translating this ‘eye-jelly’ since it is not clear what eye-jelly is. ‘Vitreous humour’ might be a better translation but I suspect that ‘retina’ would more closely capture the modern theoretical equivalent to what Aristotle has in mind.

¹⁸ This interpretation takes as referring to the single faculty to which the data is ultimately transmitted; an alternative is that it says that the external object (to which both the colour and the sound belong) is one, and is registered by one mean, within the soul, to correspond to its objective unity.

¹⁹ I have already argued for the connection between these two questions raised in 3.2: see n. 2. The present chapter suggests that these questions cause Aristotle some recurrent difficulties.

²⁰ It would be natural for to mean ‘the latter’ and to mean ‘the former’, but there are not, as the text stands, enough items mentioned in the vicinity for this to make sense. Hence the suggestion that some were mentioned in the lacuna. An alternative that would allow for the reading I develop here is that Aristotle alludes to a chart or diagram with the correlates in parallel columns, in which he points to ‘these items’ and ‘those items’ and indicates the relations of correlation and identity by analogy. This idea is supported by the allusive references later in this passage to items labelled A, B, C, and D (431a25-bl) which evidently must have been identified on a diagram in which their reference was explicit

²¹ One possibility is to amend and , so that the passage explains what is true of the one ‘last thing’ which relates to two things at once. This suggestion was put in discussion by Michael Frede, but there is no manuscript warranty and it is hard to see why such an easy reading should have been corrupted. Ross adds an extra which is not required on Frede's reading, but is not strictly necessary even retaining the manuscript's on the interpretation I shall be offering.

²² Cf. at line 24 which indicates what is significant about these things

²³ Alternatively may be not opposites but similar items within a single spectrum, e.g. two cases of bitter, for the point that is being made is that sweet and hot are both at the same end of the common spectrum, and hence are analogously one (see further below) and thus related in the same way as two similar items at the same end of one of their respective spectra. Or they might be the opposite ends of the two spectra involved in ‘these things’, i.e. bitter and cold.

²⁴ There is a discrepancy in the manuscripts here, so that we may prefer to read ‘or numerically’ instead of ‘and numerically’. It seems that Aristotle is thinking primarily of occasions on which we perceive an object that is numerically one but has a plurality of heterogeneous sensible features. Neither its numerical unity nor any analogous unity of its sensible features prevent us from judging it as having distinct kinds of features. The stronger the unity of the hypothetical object, the stronger will be Aristotle's case, and the more apt his suggestion that the faculty has a unity correlated with the unity of the numerically single object. Hence it is more appropriate to read the text with ‘and’, though the point is little damaged if ‘or’ is read instead. With ‘and’ the reasoning would constitute an a fortiori argument

²⁵ Other alternatives have been offered by commentators for what C and D stand for, including the related perceptions correlated with white and black (Hamlyn). Sweet and hot were the last examples under consideration, but we cannot suppose that C and D pick up those, since then Aristotle would posit the claim that they stood in the same relation as contraries as his initial premise, whereas that appears to be what he has to prove. Rather he needs here a premise that is self-evident, and that seems to be satisfied by the claim that both AB and CD are opposites on their respective spectra. I am grateful to Reviel Netz for pointing out the implications of this.

²⁶ This scheme matches the discussion of correlates in the De sensu passage cited above

²⁷ Ross in the OCT alters this sentence to read C and A, D and B, respectively, but the amendment is not necessary on this interpretation. It also involves reading both pairs in reverse alphabetical order, which seems improbable.

²⁸ I am grateful to David Charles for detailed comments on an earlier draft and to him and Vasilis Karasmanis for inviting me to contribute to a workshop on this topic. Many of the points I make reflect the contributions of participants in the discussion on that occasion, not all of which I have been able to acknowledge individually. Further improvements owe their origin to the audience at the B Club in Cambridge and to an anonymous referee.