¹ cf. Hamlyn, (Aristotle's, De Anima Books 2 and 3, translated with introduction and notes by Hamlyn, D. W., Oxford, 1968), 121Google Scholar: ‘This chapter is a rambling one…’.

² Kosman, L. A., ‘Perceiving that we perceive: On the Soul III. 2’, Philosophical Review 84 (1975), 499–519CrossRefGoogle Scholar.

³ Hamlyn, 121.

⁴ In this I disagree with Hamlyn (121): ‘Here he eliminates the possibility that we know that we see by means of a different sense’, and also with Kosman (500).

⁵ cf. Hamlyn, 121; Kosman, 500; Ross, (Aristotle, , De Anima, Oxford, 1961), 275Google Scholar.

⁶ This was established in book 2, chapter 7.

⁷ The inferential particle τοίνυν is usually rendered ‘therefore’ or ‘then’ implying that the comment in b20 represents a conclusion drawn from the previous aporia; but this sense of the particle is in fact relatively rare (Denniston, , The Greek Particles, 569 and 571)Google Scholar; its normal use is conversational in introducing a new point, usually from a new speaker in dialogue. Denniston, 572: ‘A rejoinder introduced by τοίνυν sometimes conveys a comment on, or criticism of, the previous speaker's words’, e.g. Plato, , Republic 358aGoogle Scholar. As a transitional particle it introduces a fresh point in the argument (Denniston, 575). We may therefore render Aristotle's words at b20 ‘well now, on the other hand, it is clear that perception is not a single thing’.

⁸ Hamlyn, 121.

⁹ In De anima 3. 15: οὐ ϒάρ οῖ⋯ τε ⋯ποϕήνασθαɩ περί τ⋯ς ὃψες ⋯τɩ ⋯ρᾷ μή ϒɩνώσκουσαν τ⋯ ⋯ρώμενον.

¹⁰ ὃψɩς is not in fact coextensive with ὃρασɩς, which represents a particular type of actualization of opsis, namely perception of colour, whereas opsis can be actualized in another [way, 425b20–22; see above, p. 402. Note that it is the sense-object (colour) which defines the actuality; cf. 425b26–426a26.

¹¹ 426a2–11. The details of this argument rely on the discussion at Physics 3, 3, 202a21 ff.

¹² 426 a 13.

¹³ 426 a 14.

¹⁴ Kosman, 501 (his emphasis).

¹⁵ There is a textual discrepancy here which enables some editors to read τ⋯ ⋯ρα⋯ν in place of τ⋯ ⋯ρ⋯ν in one or both of the occurrences at 425 b 19, but this does not provide a satisfactory solution, since if the comment at 425 b 22–3 (which must refer to τ⋯ ⋯ρ⋯ν) is to be any kind of solution to the difficulty in 425 b 19 it had better refer to at least a possible referent of the terms in b 19, and preferably to the same term. In what sense could as distinct from τ⋯ ⋯ρα⋯ν be perceptible?

¹⁶ cf. e.g. Hamlyn, 123.

¹⁷ 425b22–3. The sense in which the organ can be said to be coloured is controversial. See for example R. Sorabji, ‘Body and Soul in Aristotle’, , Philosophy 49 (1974), 72 n. 22Google Scholar.

¹⁸ op. cit. (n. 2).

¹⁹ Kosman, 511.

²⁰ In the De Somno Aristotle explicitly states that it is not by sight that we see that we see (below p. 406). Kosman has to explain this as a denial that the awareness is in sight ‘in isolation from the powers of a whole living, sentient organism’ rather than a denial that it is in sight at all.

²¹ Kosman, 517–18.

²² 426 a 13.

²³ cf. 425 b 12 αίσθανόμεθα ὃτɩ ⋯ρ⋯μεν καί ⋯κούομεν…ἤ τῄ ⋯ψεɩ…ἤ ⋯τέρα.

²⁴ οὐ ϒ⋯ρ δ⋯ τῇ ϒε ⋯ψεɩ ⋯ρᾷ ⋯τɩ ⋯ρᾷ, καί κρίνεɩ δ⋯ καί δύναταɩ κρίνεɩν ἕτερα τ⋯ ϒλυκέα τ⋯ν λευκ⋯ν… De somno 455 a 15 if.

²⁵ It is worth noting that Aristotle never says αἰσθανόμεθα ⋯τɩ αίσθανόμεθα.

²⁶ The difference between the thesis I am advocating here and that of Kosman is worth clarifying. He suggests that Aristotle's interest is in the awareness inherent in seeing: when I see white there is more than just an affection of whiteness in my organ – I am also aware of white – and this awareness is an awareness that I am perceiving white. I am suggesting, on the other hand, that the awareness in question is the awareness of a distinction between different forms of perception: how I am aware not simply that I am perceiving (as on Kosman's account), but more specifically that I am seeing. To be aware that there are different modes of perception, and which one I am currently undergoing, requires that I must have a faculty capable of recognizing more than one mode of perception; this corresponds to a recognition of different sense-objects, and it is in this way that the process of awareness is transparent.

²⁷ 426 b 15 ⋯νάϒκη δ⋯ αίσθήσεɩ αίσθητ⋯ ϒάρ ⋯στɩν.

²⁸ 455 a 12 ff.

²⁹ It is not difficult to see how the regress might be resolved, and it is perhaps unnecessary for Aristotle to specify that if there is a single common faculty which distinguishes whether one is seeing, hearing, tasting etc., we may stop the regress at that stage since there is no reason why that should not be perceptive of itself. It is only in the case of the individual special senses which cannot perceive each other's objects that judging has to be ascribed to a further faculty. On the other hand it is not clear that we do possess any faculty which discerns the common faculty, and Aristotle might stop the regress not by having a common faculty perceptive of itself but by having a faculty which is not itself perceived. Since perception of a sense is perception of the actualization of its objects, and the objects of the common sense are the objects of the various special senses, no further sense perceptive of those same objects can be distinguished.

³⁰ I am here assuming acceptance of the analysis of the potentialities and actualities involved in this passage outlined at length by Kosman, 513–14. His clarification of these details resolves most of the difficulties raised by Hamlyn, 123–5.

³¹ Some of the problems with interpretation of the passage in the past seem to be due to a shortage of technical terms for the English equivalents.

³² 426 a 13.

³³ Barker, A., ‘Aristotle on Perception and Ratios’, Phronesis 26 (1981), 248–66CrossRefGoogle Scholar.

³⁴ This traditional reading is represented primarily by Ross, W. D., Aristotle, De Anima (Oxford, 1961), 277–8Google Scholar, and Hamlyn, 125. Ross takes the argument as starting from a statement that the human voice (ϕωνή) is a συμϕωνία; Hamlyn takes συμϕωνία as a reference to any pitched sound. Neither of these is satisfactory as a bearer of the description συμϕωνία, and it is clear that a single tone of medium pitch cannot reasonably be regarded as a ratio in any sense. In addition neither reading provides Aristotle with a valid argument in order to derive the general conclusion about all ⋯κοή which Ross and Hamlyn require (cf. Barker, 250–2).

³⁵ εί δ'⋯υμϕωνία ϕωνή τίς τίς ⋯στɩν…as opposed to the version in Ross' OCT which follows Priscian and Sophonia s by printing εί δ'⋯ ϕων⋯ συμϕωνία τίς ⋯στɩν…

³⁶ Barker, 255. Thus far I am entirely in agreement with Barker's analysis. On the use to which the conclusion is put by Aristotle in the latter part of the paragraph I have some remaining doubts. Barker suggests that the argument continues with reference to a restricted range of perceptions which are proportions analogous to sound, and hence are not perceived as proportions when one or other element is excessively strong (257–9). This part of Barker's analysis, and the sense in which the OKOIJ is ‘destroyed’ seem to me somewhat strained. Alternatively we may reverse the sense and read the sentence as an argument that other sense-objects are ratios comparable to harmonies. It is because (δɩ⋯ το⋯το) the hearing of a ratio is itself a ratio that an extreme of high or low sound beyond the range of hearing (perhaps, rather than, on Barker's terms, extreme intensity of one element) destroys the actual hearing of the ratio in that the ratio which constitutes the hearing is not actualized. Aristotle argues that wherever an extreme at one end of a scale prevents the related perception from being actualized we may conclude that the perception and the object were a ratio, by analogy with hearing.

³⁷ The theory ascribed to οί πρότερον ϕυσɩολόϒοɩ at 426a20 ff. corresponds closely with that recounted by Socrates at Theaetetus 156a–c, which derives the conclusion that everything is constantly changing from the transience of the actualities of sense-objects. Aristotle objects to the fact that the theory extends the relativity to potential sense-objects such as colours when not actualized (Theaetetus 156 c 1). In the Theaetetus the theory is rather vaguely said to be a secret doctrine of Protagoras, or rather of a number of distinguished men. Aristotle's vague reference to οί πρότερον ϕυσɩολόϒοɩ could be a reference to this passage of the Theaetetus. Cf. Philoponus in De An. 475. 26, who erroneously refers to the Protagoras but must mean the Theaetetus.

³⁸ In each reference to perceiving the distinction between objects, the objects referred to are linked by καί: λευκ⋯ν μ⋯ν καί μέλαν (426 b 10–11), ϒλυκὺ δ⋯ καί πɩκρόν (426 b 11), τ⋯ λευκ⋯ν καί τ⋯ ϒλυκύ (426 b 13). This is surely parallel to the opening sentence at 425 b 12, αίσθανόμεθα ⋯τɩ όρ⋯μεν καί ⋯κούομεν. In each case the implication as that we can tell the difference between the two, or simply pick out one of the two correctly. The fact that I can identify a particular colour as black implies that I know some other colour from which I candistinguish it; the fact that I can say that I am seeing (and not just that I am perceiving) implies that I am aware of a difference between seeing and some other sort of perception.

³⁹ In this respect the details concerning touch discussed in chapter 11 of book 2 do not affect the argument. Even if it is the case that a medium intervenes in the case of touch and taste as well as the other senses, it remains true that touch does not operate at a distance and the medium functions in a different way, 423 b 12–15.

⁴⁰ Simplicius, in De An. 197. 11 ff.Google Scholar, Philoponus, in De An. 482. 30 ffGoogle Scholar. Both commentator s give an interpretation compatible with that I am suggesting for this sentence. Philoponus had referred to the Theaetetus in an earlier note on the same page, 482. 16–18.

⁴¹ Discrimination between tasting and smelling is the one region where the boundary is unclear. Aristotle does not comment on the possibility of confusion here, and was perhaps unaware how much should be assigned to smelling.

⁴² This paper originated as a contribution to a seminar in Cambridge and I am much indebted to the members of the seminar, and particularly to Geoffrey Lloyd and Myles Burnyeat who read a subsequent draft, for their constructive comments.