The text of Book 3 of Aristotle’s De anima is notoriously troubled, particularly the seventh chapter, which Torstrik in his 1862 edition claimed to be so internally inconsistent and ungrammatical that ‘no writer, not even the worst, could have put such incongruous things together’.Footnote ¹ This judgement, although perhaps not its tenor, has been largely accepted by subsequent editors and scholars.Footnote ² This note focusses on the first and more significant of Torstrik’s two grammatical objections, the second of which has been satisfactorily dealt with by Corcilius.Footnote ³

According to Torstrik,Footnote ⁴ the transmitted text contains at 431a4–7 a μέν without a corresponding δέ (or another adversative particle). Ηere is the passage, bolding the μέν and lightly adapting Shields’s translation:

It is evident that the object of perception is what turns the perceptual faculty into something actual from being in potentiality; for it is neither affected nor altered. Consequently, this kind is different from motion, since motion is the actuality of something incomplete, whereas unqualified actuality, that of something complete, is something else.

Defending Aristotle’s grammar, Corcilius argues that the μέν is solitarium, following Denniston’s comment that μέν can, like γε, mark a contrast to something the author ‘does not, even in the first instance, intend to express in words, or even (sometimes) define precisely in thought’.Footnote ⁵ Aristotle, Corcilius suggests, is implicitly contrasting perception with all the mental episodes subsequently discussed in the chapter, which he denies are caused by complete and actualized items. These include simple desires (for example pleasure and pain), basic motor behaviours of flight and pursuit, and thinking in images.

This is odd, however, since the things discussed in the rest of the chapter are ipso facto expressed in words and so the contrast could have been explicitly marked in any number of ways. There seems no reason for Aristotle to have enlisted the exotic μέν solitarium (which, as Denniston’s examples illustrate, occurs more frequently in verse and dialogue, and typically in specific contexts not matching the one here).

Let us consider a better option for preserving the grammar. The text, as standardly punctuated, suggests an argumentative structure that can be marked with indentation indicating subordination:

(3) is clearly subordinate to (2), and the period after ἀλλοιοῦται in (3) indicates that (4) returns to the main line of argument. (5) and (6) are coordinates, jointly subordinate to (4), with Aristotle explicitly contrasting the actuality of something incomplete (as we find in a typical κίνησις) with the unqualified actuality of something complete (as we find in perception).

Yet, while the earlier discussion in De an. 2.5 makes clear that a κίνησις is defined by an end external to it, and hence is incomplete, in what sense is there unqualified actuality in perception? On Corcilius’s reading,Footnote ⁶ what is unqualifiedly actual is the object of perception, τὸ αἰσθητόν, identified in (2), so that (6) ultimately restates and further illuminates (2).

Aristotle, however, insists that unqualified actuality is of something complete (τοῦ τετελεσμένου), while the object of perception is a sensory quality, for example whiteness or sweetness. How can a sensory quality, depending for its existence on the object that ‘has’ it, be something complete? Even if, as Corcilius maintains, the αἰσθητόν is the subject of the verbs in (3), neither changing nor being altered in perception, that still falls some distance short of being complete.Footnote ⁷

But we can read the passage with a different argumentative structure, making modest changes in punctuation at (3), (4) and (5):

Thus, (2) through (5) yield a continuous characterization of τὸ αἰσθητόν, all in the scope of the initial μέν. (4) is coordinate with (3) and explained further by (5). The only period comes at the end at (5), and (2) and (6) now form the two balancing parts of a μὲν … δὲ … clause, pace Torstrik.

What contrast, then, is Aristotle drawing? The key is the immediately preceding claim that knowledge and the object of knowledge are one and the same (a claim made verbatim at De an. 3.5 430a19–21). Knowledge, as Aristotle construes it, is a better candidate than, say, whiteness for something that is complete and unqualifiedly actual.

This brings us back to Torstrik, who suggests that the discussion of knowledge in the first part of the chapter is completely discontinuous with what follows, because ‘ad metaphysicam, non ad psychologiam pertinet’.Footnote ⁹ The reading proposed here suggests, rather, that it plays the role of a touchstone or foil. One indication of this is that the δέ in (6) responds to the μέν in (2), which is not solitarium but ordinarium.