2.1 Generic Likeness

An. 2.5 opens and closes by referring to a view that holds perception to be a case of ‘like being affected by like’ (LAL). When Aristotle introduces this view, at 416b35–417a2, it is already clear that he takes the position to be importantly correct, but also somehow mistaken. He then substantiates this claim throughout the chapter, spelling out the aspects of truth as well as those of falsity on two occasions – namely, at the climax of the first provisional part of the chapter (417a17–21) and in the chapter’s closing lines (418a3–6). I shall argue that, when correctly understood, these passages provide an important indication of what Aristotle intends to achieve in his first general account of perception.

Let us begin by considering how the chapter as a whole is framed. Immediately after announcing his goal of accounting for all perception – where perception is said to ‘come about in being affected’ – Aristotle introduces the view of anonymous thinkers who are reported to have believed that (in perception, apparently) like is affected by like. Aristotle then refers to his general inquiry into acting and being affected (in GC 1.7), suggesting that his treatment there should help us understand how far the anonymous thinkers were right, and how far they were wrong:

At the very end of the chapter, Aristotle sums up his response to the view put forward by these anonymous thinkers as follows:

In this closing passage, Aristotle is explicitly applying to the case of perception the scheme that has been laid out in general terms at 417a17–21. Presumably, he thinks that, with the notion of preservative πάσχειν in hand (as worked out at 417b2–29), we are finally in a position to understand how perception can fit successfully within the assimilation model. Aristotle’s response to the view that like is affected by like was first formulated as follows, with no explicit mention of perception:

The general idea of Aristotle’s response seems clear enough in both passages. Those who say that like is affected by like are right insofar as the result of being affected (expressed by the perfect tense πεπονθέναι) is considered, but they are wrong insofar as we consider the πάσχειν itself (in the present tense). That which can perceive is like the perceptual object insofar as it has (already) been affected by it, and so assimilated to it, but it is unlike it insofar as it is (still) being affected by it.

The first thing to notice about this formulation of Aristotle’s response is that it is not exactly the one that most readers expect Aristotle to have in mind in R&W₁. Standardly, Aristotle’s reference to Generation and Corruption in R&W₁ is interpreted as pointing to his resolution of the initial puzzle of GC 1.7 along the following lines.Footnote ⁴ On the one side, most of the thinkers (οἱ πλεῖστοι) believe that like cannot be affected by like and that acting is possible only between things that are unlike and different from one another (323b3–10). On the other side, Democritus claims that what acts and what is affected must be the same and like (323b10–15). After presenting arguments in favour of and against each side (323b15–29), Aristotle turns to offer his own account, which he believes can incorporate the true elements of each party’s view (323b29–324a9):

When we start reading An. 2.5 with this passage in mind, it may appear that Aristotle employs R&W₁ to refer to it.Footnote ⁵ Yet, as we have seen, the way in which he fleshes out his response in R&W_2&3 is different from the scheme presented in the quoted passage. Indeed, there are at least four reasons for not identifying Aristotle’s reference in R&W₁ with this passage.

(1) The key distinction from Genus/Species between the two interpretations of likeness is never employed in An. 2.5. The state of the patient before the acting is completed is never described here in terms of generic likeness, but always in terms of unlikeness. Instead, Aristotle’s response in R&W_2&3 hinges on a distinction between the present and the perfect tense that is not even alluded to in Genus/Species.

(2) These differences seem well motivated. Later in GC 1.7, when Aristotle turns to consider the first principles of acting (at 324a24ff.), he, effectively, throws into question his earlier adoption of the ‘Democritean’ view according to which the patient and the agent are always generically alike, and so capable of mutually acting upon each other. Instead, there are, he insists, agents that can act while being impassive in the sense that they themselves are never affected by that upon which they act. The example that Aristotle provides here is the art of medicine.Footnote ⁶ As he puts it: ‘those agents that do not have the form in matter are impassive, whereas those that [have the form] in matter, can be affected’.Footnote ⁷ The reciprocity in the latter case, Aristotle adds, is explained exactly by the fact that the agent and what it acts upon have the same matter and are of the same genus.Footnote ⁸ Fire, for instance, is of this kind, for its heat is ‘in matter’.Footnote ⁹ Medicine, in contrast, is emphatically not a form ‘in matter’, and so the ‘Democritean’ generic likeness (implying reciprocity) cannot be applied to it, at least not straightforwardly.Footnote ¹⁰ This point will hold for all productive arts, and it also has significant theoretical implications for how the agency of artisans should be understood: a medical doctor is not, qua doctor, affected by her patients because the form (i.e. the art of medicine), as the primary agent, is not ‘in matter’ (rather, it is in the doctor’s soul); she may be exhausted by her practice or become infected with an illness while in the office, but that in no way diminishes or otherwise affects her expertise.Footnote ¹¹

The reason why becoming clear on this is important for the question of how exactly An. 2.5 draws on GC 1.7 is that, as we shall see, Aristotle is eager to emphasize a similarity between the perceptive capacity and capacities like arts.Footnote ¹² One crucial aspect of this comparison is that perceptual objects are to be conceived, according to Aristotle, as impassive agents like doctors.Footnote ¹³ So, according to De Anima, the way in which the perceiver is related to perceptual objects is analogous to the way in which a patient relates to a medical doctor.Footnote ¹⁴ It would be strange then if Aristotle opened An. 2.5 by accepting a demand of symmetrical generic likeness in the case of perceiver and perceptual object, in the way that it had been introduced in Genus/Species, because that schema does not apply, in GC 1.7 itself, to the cases most relevant to An. 1.5.Footnote ¹⁵

The other pair of reasons for not identifying Aristotle’s reference in R&W₁ with the quoted passage from GC 1.7 come from a consideration of what Aristotle says about his predecessors in GC and An., respectively. (3) First, in GC 1.7 Aristotle goes out of his way to stress that nobody shared Democritus’ view according to which like is affected by like. Democritus, we read, talked about acting ‘unlike all others, alone, in a peculiar way’.Footnote ¹⁶ The view introduced at the outset of An. 2.5, in contrast, seems to be one that is widely shared.Footnote ¹⁷

(4) Moreover, we shall see reasons to believe that the thinkers Aristotle has in mind at the outset of An. 2.5 all stand, without exception, on the other side of the GC 1.7 puzzle from Democritus. The context of An. 2.5 itself suggests, as noted above, that the LAL claim was made specifically as a part of the anonymous thinkers’ reflections on perception. This seems to be confirmed by a passage from An. 1.5 which implies, among other things, that Aristotle has adherents of LKL in mind:Footnote ¹⁸ it is in the context of LKL that, according to Aristotle, these thinkers arrived at the conclusion that, in perception, the corollary LAL must hold. The immediate upshot of this observation for our question is the following. Whatever the exact motivation of their acceptance of LAL was, it seems to have been very different from Democritus’ motivation for insisting on LAL as a universal thesis about being affected. Indeed, in their broader understanding of acting and being affected, the thinkers in question seem to have roundly disagreed with Democritus and insisted that only unlike can be affected by unlike. Perception is an exception for them, which produces a tension with their general view about acting rather than being an application of it. Accordingly, if Aristotle wants to explain in An. 2.5 the extent to which their claim that in perception like is affected by like is correct and the extent to which it is not, he cannot simply rehearse the argument from GC 1.7 that explained the truth and the falsity of Democritus’ claim about being affected in general.

2.2 Aristotle’s Engagement with LKL

Thus, if Aristotle is not thinking in R&W₁ of the ‘Democritean’ generic likeness, (a) how is his response (here and in R&W_2&3) based on GC 1.7, to which he clearly refers? And (b) how did the interlocutors that he has in mind in R&W₁ arrive at LAL?

2.3 LKL and the Passivity of Perception

Aristotle’s claim that the adherents of LKL are committed to LAL for knowing (which leads to a crude contradiction) is surprising. This is so, among other reasons, because Theophrastus never says that the adherents of LKL believed perception to be a kind of being affected.Footnote ²⁶ On the contrary, he represents the view that perception is an alteration or an affection as the main motivation for those who opposed LKL and who, instead, defended CKC (Sens. 2, 27, 31).Footnote ²⁷ As we have noted, Aristotle’s De Anima remains silent about this competing view and contrasts LKL, instead, with Anaxagoras’ impassivity view.Footnote ²⁸

We shall see that Theophrastus is well aware of this latter contrast.Footnote ²⁹ The fact that his discussion is framed differently than Aristotle’s can perhaps be explained by a difference in the respective purposes of their projects. The contrast between LKL and the impassivity view turns out to be more helpful if our aim is to motivate the true account of perception (and thought) by bringing out an apparent conflict between two basic requirements on knowing.Footnote ³⁰ However, doing this would not be conducive to building a rich account of past views, given how little can be said about the impassivity view in comparison with CKC.Footnote ³¹

If this is right, the question remains why Aristotle believes himself to be justified in ascribing the claim that perceiving is a kind of being affected to the adherents of LKL and, by so doing, implicating them in the puzzle formulated succinctly in Self-Contradiction. Aristotle makes it clear here that the adherents of LKL in general believe that only unlike can be affected by unlike – that is, they disagree with Democritus when it comes to general intuitions about being affected. This immediately makes us wonder why they should accept LAL in the case of perception, and so commit themselves to such an obvious inconsistency.Footnote ³²

I think the most likely answer is that, in Self-Contradiction (and apparently as early as An. 1.2, 405b10–23), and then again in R&W₁, Aristotle is building on certain considerations and arguments that, by his lights, should have led the adherents of LKL to assume that perceiving is a kind of being affected – irrespective of whether they, themselves, actually stated this assumption or not. Regrettably, there seems to be nothing in the De Anima itself that would help us understand how exactly this argumentative move might have been employed. This is where Theophrastus’ critical discussion of LKL in his De Sensibus becomes genuinely helpful. Although it does not allow us to identify the exact considerations that Aristotle is presumably relying on in Self-Contradiction, it does provide an insight into the kind of arguments that can bring out the commitment to a passivity of perception on the side of the allegedly traditional LKL.Footnote ³³

In Theophrastus’ presentation, LKL is closely connected with the notion of effluences,Footnote ³⁴ a view wherein the perceptual object emits something that is then brought to the perceiver and meets with what is like in them. This might have already been enough for Aristotle to conclude that the perceiver, according to LKL, must in fact be affected in some sense by this effluence as ‘the ultimate and proper agent’.Footnote ³⁵ In a similarly straightforward way, Theophrastus analyses Empedocles’ concept of ‘fitting’ (ἁρμόττειν), which describes the successful meeting of the effluence with the corresponding element in the perceiver, as being referred back to (1) the concept of likeness and (2) the concept of touching (so that like is touched by like), where the latter can be argued to imply already some kind of acting and being affected: ‘For he explains knowing by both these notions, namely the like (τῷ ὁμοίῳ) and the touching (τῇ ἁφῇ). That is why he used the term “fitting” (ἁρμόττειν).’Footnote ³⁶

A more sophisticated strategy of eliciting the assumption on the part of LKL that perceiving is a kind of being affected could have been based on another kind of challenge that we also encounter in An. 1.5 (though this, too, occurs in an extremely compressed form):

Theophrastus develops a similar symmetricity argument in his polemic against Empedocles in Sens. 12: if there is nothing more about the likeness that characterizes perceivers with respect to their perceptual objects than what can be said about any two similar objects in nature,Footnote ³⁷ then LKL will have the absurd consequence that everything can perceive something – namely all the things that are in some respect similar to it.Footnote ³⁸

After this rather straightforward criticism, Theophrastus goes on to formulate two closely related objections that turn out to be directly relevant for our investigation:

The problem with the concept of likeness as allegedly employed by Empedocles is that it seems to be perfectly symmetrical.Footnote ⁴⁰ However, if this is the case and the effluence is, in fact, completely and in every respect like that which it is supposed to ‘fit’ into on the side of the perceiver, it is not clear at all (i) why we should say that the perceiver perceives the object that ‘sent’ this effluence, rather than saying that it is the effluence (or the external object by means of the effluence) that perceives the perceiver;Footnote ⁴¹ and (ii) why the element present in the perceiver does not also perceive itself in the absence of any effluence.Footnote ⁴²

We can imagine that once an adherent of LKL has followed his Peripatetic challenger this far, she may be willing to accept Theophrastus’ offer and say that, according to her view, the perceiver and the perceived object are completely like one another (ὅμοιον πάντως), but are not like one another in all respects (ὅμοιον πάντῃ). Clearly, the perceiver is the one who receives the effluence while the object sends it and the effluence is received, and this makes their relation asymmetrical, explaining why one perceives whereas the other is perceived, or is that by means of which the source of the effluence is perceived. Once this much is agreed on, it is just a small step to also agreeing that, according to LKL, the object acts, while the perceiver is affected.

This appears to be at least one possible way of showing how the adherents of LKL are, in fact, committed to a passivity of perception, and so to LAL, although this conflicts with their general view of acting, according to which only unlike can be affected by unlike. There may also be other ways of securing this result, but these need not concern us now. What matters for our purposes is that, if this result can be brought about in some way, the thinkers in question (i.e. almost all of Aristotle’s predecessors according to him) face the serious problem described succinctly in Self-Contradiction.

However, Aristotle’s aim should not be thought of as primarily destructive. Rather, in line with his view on the constructive role of puzzles, he seems to have believed that it is already a significant achievement to have successfully formulated the problem. That is because, once this is accomplished, the correct account of perceiving, which would successfully separate the truth from the error in LKL, is already in sight. It is exactly this problem, I suggest, from which An. 2.5 starts.

2.4 Another Challenge for LKL

Immediately after R&W₁, Aristotle formulates the following puzzle about perception:

When the puzzle in (i) is read as a difficulty that Aristotle raises for himself, the answer in (ii) seems quite unsatisfactory.Footnote ⁴³ However, reading the puzzle in this manner does not seem warranted, because Aristotle has not yet started formulating his own view. Rather, immediately before, he introduced the anonymous thinkers committed to LAL, while flagging that their view is only partly right. It is thus natural to read the puzzle within the context of Aristotle’s engagement with LKL, as a way of showing the limits of the reported view and thereby preparing a foil for his own account. This seems to be exactly what Aristotle is doing.Footnote ⁴⁴ More specifically, I want to suggest that the difficulty presented in (i) closely corresponds to one of the objections formulated by Theophrastus against Empedocles in Sens. 13, which has already been quoted above:

This objection is independent from the identification of perception as a kind of being affected.Footnote ⁴⁵ Rather, as we have seen, this kind of argument can serve as a way of eliciting the assumption on the part of LKL that perception is indeed a kind of being affected by perceptual objects. Aristotle’s point in (ii) is closely connected here, or so I shall argue.

He states: ‘that which can perceive is not in activity, but only in capacity’ (417a6–7), and so it requires an external perceptual object that is in fulfilment (417a9) and that can, thereby, bring it into activity.Footnote ⁴⁶ This cannot be a simple statement of the obvious fact that perceivers sometimes do not perceive X and sometimes do, depending on whether X is present to them in the external world or not. No one could disagree with this; but it would be just another way of formulating the difficulty, not proposing a solution to it. Nor does Aristotle’s point seem to draw on the technical meaning of τὸ αἰσθητικόν as being the perceptive capacity of the soul.Footnote ⁴⁷ This would in no way contribute to an explanation of why the fire and the earth in the organs should be any less capable of activating the perceptive capacity than the fire and earth outside.Footnote ⁴⁸

What Aristotle seems to have in mind in (ii) is rather that our explanation why something is perceptive must successfully account for the obvious fact that whether perceivers perceive X or not is determined by whether X is present to them in the external world or not. It is exactly at this point where the adherents to the traditional LKL view, as construed by Aristotle, fail. Their view will require them to explain why something is perceptive of, say, fire (or white colour) by an appeal to the actual presence of fire in it. Yet, this seems to imply unacceptably that the fire present in us should be perceptive of itself no less than it is of the external fire, which would make perception independent of whether or not an external object was present to the perceiver. As in Theophrastus’ Sens. 13–14, the argument seems to be aimed at putting pressure on the symmetricity of likeness presupposed by the traditional account.

One might then wonder whether Aristotle’s lesson here at least hints at an alternative account. What he offers is an analogy with burning, which first of all reaffirms the characterization of perception as a case of being affected (being burned is Aristotle’s favourite example of being affected).Footnote ⁴⁹ The point of this analogy seems to be exactly an asymmetry between that which can perceive (compared with that which can burn) on the one hand and the perceptual object (compared with fire in fulfilment) on the other. This asymmetry is developed in terms of capacity (δυνάμει) and fulfilment (ἐντελεχείᾳ). Just as combustible material requires an actual fire acting on it to burn (or at least to start burning), that which can perceive needs, in order to be activated, an external perceptual object that is already in fulfilment. This puts obvious constraints on the sense in which LKL can be true – constraints that are parallel to those put on it by arguments eliciting the assumption that perception is a kind of being affected.

As soon as this much is agreed on, we immediately see important disanalogies in (ii) that are unlikely to have been missed by Aristotle.Footnote ⁵⁰ Once the combustible material has been set on fire, (a) it continues burning on its own even after the initial source of it being set on fire has departed and (b) its capacity for burning is sooner or later exhausted. That which can perceive, in contrast, (a´) ceases to perceive immediately upon the perceptual object being removed (i.e. upon the object ceasing to act on it), but (b´) the capacity seems to be, in principle, inexhaustible. In order to account for these disanalogies, one would need to specify the kind of capacities that pertain, on the one hand, to that which can perceive, and, on the other, to that which can burn. Otherwise, the puzzle threatens to reappear, as it will seem that once the perceiver has been activated by the perceptual object, she will be always actively perceiving, or at least she will be perceiving until its capacity is exhausted, irrespective of whether the external object is still present or not. In other words, as long as we do not successfully distinguish the kind of capacity pertaining to perceivers from the kind of capacity pertaining to combustible materials, we shall not be in a position to understand why there cannot be continued perceiving without external perceptual objects being continuously present, or, indeed, how a case of continued perceiving differs from having an after-image. In this way Aristotle’s lesson in (ii) starts building a case to show that distinctions need to be drawn between different kinds of capacities and different kinds of ‘being activated’ and ‘being affected’, as is done later in An. 2.5.

Yet, already at 417a2–9, Aristotle’s puzzle – as a part of a larger Peripatetic engagement with LKL – brings out the limits of his predecessors’ theory. It thereby circumscribes the conditions that any successful account of perception will need to fulfil.

2.5 The Truth of LAL

Let us now return to Aristotle’s merciless analysis of LKL in Self-Contradiction: in perception like must be affected by like, but in general only unlike can be affected by unlike. This sounds like saying that LKL is simply wrong and beyond repair. And yet, Aristotle makes this view, and indeed its most problematic element (i.e. LAL), central to his argument in An. 2.5, returning at the two climaxes (R&W_2&3) to his explanation of the ways in which the claim is true and those in which it is false.

I have argued that the truth of LAL articulated here cannot be captured in terms of the ‘Democritean’ generic likeness. But what, then, can this truth amount to? Aristotle’s concession to LAL in R&W_2&3 can easily sound like a mere courtesy, for he appears to be effectively saying that this view is plainly false. However, against this initial impression, I shall argue that there is a robust sense in which Aristotle accepts LAL (in line with the principle of τὰ καλῶς εἰρημένα λαβεῖν).

The difficulty with determining what is right about LAL will become clearer when we consider R&W₂ in some detail:

The problem is that, if we understand the difference between πάσχει and πεπονθός in the second sentence here (and then again in R&W₃) in a strictly temporal sense, Aristotle’s acknowledgement that LAL is partly right becomes an empty gesture.Footnote ⁵¹ The passage here says that the perceiver is affected (πάσχει) by like, and explains the truth of this fact, not in terms of any qualified sense of likeness, but in terms of the perfect tense: LAL is right insofar as the perceiver has been affected (πεπονθός) and so has come to be like the perceptual object. This concession is combined with a reservation formulated in terms of the present tense: LAL is wrong insofar as we consider not the result but what is leading up to it, ‘for it is the unlike that is being affected (πάσχει)’.Footnote ⁵²

Now, if the present and the perfect tenses in the second sentence are understood as temporal qualifications in terms of Physics 6.5–6 – that is, as distinguishing between an interval t₁ of changing from A to B and an interval t₂ of having changed to B, then the initial concession that there is a sense in which the perceiver is affected (πάσχει) by like turns out to be just empty words. This is the case because the result of Aristotle’s analysis is that t₁ and t₂ meet only at a single moment T, as their common boundary, in which it is already true that A has changed to B and it is no longer true that A is changing to B. In terms of GC 1.7, this means that, starting with T, it is true that the perceiver is (in the full-fledged sense) like the perceptual object and so, starting with T, it is false that one can be affected by the other, for they now lack the requisite unlikeness. Thus, under a temporal interpretation of R&W_2&3, it is exactly at the moment in which the claim that the perceiver is like the perceptual object becomes true that it becomes impossible for the former to be affected by the latter. If Aristotle had this temporal model in mind in R&W_{2 &3}, then the sense in which LAL is right would stand or fall due merely to an improper use of the present tense in R&W₁ and in the first sentence of R&W₂, referring indiscriminately to both the result and the process leading up to it. Once this inaccurate expression comes to be duly analysed into, on the one hand, the perfect tense signifying the result and, on the other, the present tense signifying what leads up to it, the initial formulation of LAL simply becomes false and should then be abandoned altogether. If this were the case, we would be left to wonder why Aristotle decided to frame An. 2.5 by claiming that there is something right about LAL.

I suggest that, to avoid this unpleasant consequence and to understand Aristotle’s point adequately, we must not take the distinction between the present and the perfect tense in R&W_2&3 in a straightforwardly temporal sense. Instead, we should understand the perfect and present tenses to signify the result and what leads up to it – or, more abstractly, the completion and the ongoing activity. The fact that in the case of changes (κινήσεις), as analysed in the Physics and GC, the completion and the ongoing activity exclude each other (because an ongoing change has, by definition, not yet been completed) does not mean that they must exclude each other in the case of every effect caused by an agent. It may be that there are complete passive activities, which are being caused by external agents, and which at the same time pass Aristotle’s so-called tense test from Metaph. Θ.6 and EN 10.4. These activities, while being genuinely passive (as they are produced by an agent assimilating the patient to itself) would not be kinetic. Rather, their completion (πεπονθέναι) would be no hindrance for the continuation of the agent’s action and the patient’s affection (πάσχειν): the perfect tense would not exclude the present tense. I suggest that such a notion of complete passive activities is indeed what Aristotle intends to develop in An. 2.5 and that it is this notion that, ultimately, expresses the truth of LAL, purified of its elements of falsity.Footnote ⁵³ If this is right, the argument appears to run as follows: if (1) perceiving is (a kind of) being affected and having perceived is (a kind of) having been affected, if (2) everything that has been affected by something has (in some sense) been assimilated to that agent and is (in some sense) like that agent, and if (3) everything that is perceiving has perceived, then (4) everything that is perceiving X is both (in some sense) like X and (in a way) being affected by X. Hence, (5) in all perception like is affected by like.

Here we are, as was noted in Section 1.3, touching on particularly controversial matters. There appears to be a prima facie tension between two major characteristics that Aristotle ascribes to perception in An. 2.5: on the one hand, as stated in (1) and (2), perception consists in being affected and being changed/altered by the perceived object (416b33–4, 417b19–21, 418a3–6); however, on the other hand, as stated in (3), it is an activity based in a complete capacity (417a9–14, b16–18, b29–32)Footnote ⁵⁴ and is itself complete (passing the tense test), unlike motions which are essentially incomplete (Phys. 3.2).Footnote ⁵⁵ Myles Burnyeat famously discouraged interpreters from looking for any connection between An. 2.5 and the completeness expressed by the tense test.Footnote ⁵⁶ According to him, Aristotle insists at 417a14–17 that natural philosophy must proceed by supposing ‘that there is no such thing as complete or unqualified actuality’.Footnote ⁵⁷ Yet, as has been pointedly noted by Andreas Anagnostopoulos, this move seems to make Aristotle’s procedure ‘premised on a fundamental falsehood, of whose falsity the prepared reader is well aware’.Footnote ⁵⁸ As an alternative, Anagnostopoulos suggests that we treat An. 2.5 as confronting the two perspectives on perception – perception as a kind of being affected and altered and perception as a complete activity – in a sort of antinomy. At 417a14–17, Aristotle signals the provisional nature of the former perspective (affection and assimilation), but without aiming, in An. 2.5, to provide a more satisfactory understanding of it, nor to show how it can be reconciled with the latter perspective.

However, if the proposed analysis of Aristotle’s involvement with LKL is apt, then An. 2.5 attempts to do significantly more than dismiss the latter perspective, expressed by (3), or merely confront it with the former view, expressed by (1) and (2), in an antinomy. Instead, it seems that Aristotle is interested here in investigating what the latter perspective tells us about the former as well as the way in which the former needs to be transformed if it is to be reconciled with the latter. In other words, Aristotle is asking what kind of being affected perception must be according to (1), and what sort of likeness must be established between the perceiver and the perceived object according to (2), in order for (1) and (2) to be jointly compatible with (3), and so in order for the truth elements of LAL, brought out in (4) and (5), to be preserved. This is a task hinted at elsewhere, but only taken up in An. 2.5.Footnote ⁵⁹ The matter at issue here (as suggested in Section 1.3) is nothing less than offering an account of continued perceiving, which turns out to be a necessary condition of successfully accounting for perceiving in general.

In light of this, Aristotle’s identification of being affected and being active at 417a14–17 appears to be more constructive than has hitherto been appreciated. If being perceptually affected is perceiving and if perceiving passes the tense test, then it appears that, while perceiving X, the perceiver both has been affected by – and so is like – X and continues being affected by X.Footnote ⁶⁰ In this way, the passage provides us with a clue, albeit a provisional one, to the correct understanding of the assimilation model that follows immediately in R&W₂: we should learn to appreciate the connection that unlikeness has to the present tense and that likeness has to the perfect tense without assuming a straightforwardly temporal interpretation of these tenses. Without this, we shall never come to any better understanding of exactly what Aristotle thinks is true about LAL.Footnote ⁶¹

But even if we accept the tenets of this lesson, much work remains to be done. Clearly, the project will stand or fall on Aristotle’s success in accounting for the kind of being affected and the sort of likeness involved in perception. Read in the proposed way, the assimilation model implies that the perceiver is both unlike (because she is being affected by) and like (because she has been affected by) the perceived object. Understanding how this is possible will turn out to be the same task as understanding how it comes to be that an external perceptual object makes itself perceived by a perceiver rather than merely producing a perceptible replica in her sensory organs. The key idea, I submit, will be that the quality present in the perceiver as a result of her being affected by the perceived object remains a quality of the object out there rather than becoming a proper quality of the perceiver herself.Footnote ⁶²

We have yet to work our way to Aristotle’s full general account of perception. One of the remaining questions that we shall need to address in doing so is how the notion of complete passive activities differs from what happens in perceptual media, such as the air, which mediates colours, sounds, and odours.Footnote ⁶³ For now, it ought to be stressed that, as we come to R&W₂, our access to the element of truth in LAL is, indeed, only provisional. Furthermore, one might have serious doubts whether LAL can contain any coherent truth at all: under the proposed interpretation, it may easily seem that Aristotle just wants to have his cake and eat it, too. The idea of something having been assimilated to (and so being like) something else and still being affected by it (and so being unlike it) may sound paradoxical or flatly impossible. Yet this, I reply, is exactly the result of the provisional nature of our understanding at this point that was flagged right before R&W₂: as long as we assume that being affected means being changed (κινεῖσθαι), and so being active only in the incomplete sense of changing, we can never properly understand the truth of LAL. However, without taking this truth on board somehow, the key phenomenon of continued perceiving cannot be accounted for, nor can the specific object-directedness of perception. This, I suggest, is one of the reasons why ‘distinctions need to be drawn’ and discussed in what follows R&W₂.

2.6 The Tense Test Weakened

Before moving on, let me address one issue concerning Aristotle’s commitment to the truth of LAL. There is a weaker and a stronger version of it, and the weaker version transpires to be sufficient for Aristotle’s purposes.

In EN 10.4 and Metaph. Θ.6, Aristotle clearly states, at least about seeing, that it is a complete activity in the sense that there is no moment in which what performs it is still coming to its goal. Its goal has always already been reached, as Aristotle repeatedly makes clear: ‘he is seeing and at the same time he has seen’ (Metaph. Θ.6, 1048b23); ‘he has seen and is seeing at the same time the same thing’ (1048b33); ‘seeing appears to be complete at any moment of its duration’ (EN 10.4, 1174a14–15);Footnote ⁶⁴ ‘for there is no generation of seeing …, neither is [it] … a change or a generation’ (1174b12–14).Footnote ⁶⁵

One might wonder, however, whether the same can be said about the other sense modalities as they are analysed in An. 2.8–11. In the case of hearing, for instance, this is at least a questionable assumption, as Aristotle seems to follow Plato’s account – albeit not without reservations – of the difference between the high and the low being one that should be understood in terms of duration:Footnote ⁶⁶

Even without a proper analysis, this passage serves as a caveat, suggesting that the non-visual modalities may not be complete in the way developed for sight at Metaph. Θ.6 and EN 10.4.Footnote ⁶⁷ It may be that – unlike seeing – hearing and possibly also other kinds of perceiving turn out to be temporal phenomena in the sense that, for instance, discriminating a low tone necessarily takes some time (unlike discriminating something as red), as it can be recognized only on the basis of its relative slowness over a given interval of time.Footnote ⁶⁸

Fortunately, for our purposes it is not necessary to decide this question. The point is that even if this is so, LAL is no less true. For, even if hearing is not complete in the sense that there is a time when the perceiver is being affected by a sound without yet having been affected by it and assimilated to it, it will still differ from changes in that, once the perceiver has been affected and assimilated to a tone (whatever that means exactly), she can continue perceiving the very same tone – that is, continue being affected by it – which would be impossible in the case of, say, one object heating another.Footnote ⁶⁹

To this limited extent, I take it that LAL will hold, according to Aristotle, for all sense modalities, including touch. Unlike the capacity for becoming cold, the capacity to feel that something is cold allows the living being endowed with this capacity to have been assimilated to the agent and yet still be affected by it. I have perceived that this snowball is cold, and in this sense I have received its coldness, and yet it continues to act on me as something cold (which is why I continue perceiving it rather than merely being cold – in a non-sensory qualitative sense – due to itFootnote ⁷⁰). As in other cases of perception, I am being affected by the snowball as like by like. This might involve a part of my hand becoming literally like the snowball, but, crucially, this cannot be true about me as a whole: as a whole I remain unlike the snowball. This simple example manifests, I think, both the strength and the main weakness of the traditional LKL view: it shows that there is a non-trivial sense in which like can indeed only be perceived by like (i.e. by a perceiver assimilated to a perceptual object), but it also shows that standard notions of acting and likeness are not sufficient to properly articulate this fact since they offer no means of establishing the crucial distinction between ‘having perceived’ some quality and ‘having (literally) acquired’ it.

2.7 Aristotle’s Way Out

The proposed explanation of why Aristotle finds LAL to be, in a non-trivial sense, true raises a difficulty. It is not immediately clear, on this view, what he can mean by saying that it is only partly true. And, even more importantly, it is not clear how he can solve the puzzle he raised against LKL in Self-Contradiction. This question brings us back to consider just how Aristotle set the stage at the end of An. 1.2 by confronting LKL not with CKC (as Theophrastus does), but rather with Anaxagoras’ solitary demand that the knower must be impassive (405b12–23, quoted in Section 2.2). The reason for this, I suggest, is that, for Aristotle’s purposes, CKC is all too similar to LKL (at least as he construes it) – it shares the same limitations – whereas Anaxagoras’ view draws our attention exactly to the fundamental flaw of LKL (without, however, providing an independently viable alternative of its own).

I have argued that, in Aristotle’s eyes, LKL articulates – albeit imperfectly – the right intuition according to which the perceiver is like the object she is perceiving and yet continues being affected by it; this is the Acquaintance/Contact Requirement (from Section 1.4) expanded by a reflection on continued perceiving. However, the adherents of LKL have no conceptual means for explaining how in perception like can be affected by like without falling into self-contradiction. Moreover, CKC does not provide any helpful corrective.Footnote ⁷¹ The Anaxagoras of An. 1.2, 405b19–23, in contrast, puts his finger exactly on the sore spot: the perceiver cannot be assimilated to the perceptual object in the sense in which LKL envisages the likeness between the two to be realized because, if this were the case, then, rather than perceiving X, she would herself become a perceptual object similar to X, which would make her effectively unable to perceive X (or anything else). Anaxagoras’ famous claim about νοῦς knowing everything on account of its not being mixed with anything is interpreted by Aristotle as pointing towards an alternative account of knowing, broadly speaking, which can meaningfully be applied to perception. Indeed, Aristotle seems to have been preparing the ground throughout An. 1.2 for introducing Anaxagoras’ account of νοῦς as an alternative strategy of ‘defining the soul by knowing’.Footnote ⁷² This generalized account suggests that, whatever the subject of perception is, it needs to be neutral with respect to its objects even when actively perceiving them (this is the Unbiasedness/Externality Requirement). Both Aristotle and Theophrastus seem to have thought that any serious proponent of LKL would have to address this challenge.

However, as Aristotle points out at An. 1.2, 405b20–3, Anaxagoras did not develop any full-fledged alternative to LKL. Rather, in just the way that the adherents of LKL failed because they did not have control over the precise sense of likeness (and being affected) relevant for perception, Anaxagoras fails because he does not have control over the precise sense of impassivity (and unlikeness). He applies it to the subject of knowing without any qualification. But, in doing so, he ends up at an impasse, too, unable to offer any causal account (and so any proper explanation) of knowing. This failure is particularly conspicuous in the case of perception: if that which can perceive could in no way be affected by any perceptual object and could have no likeness of it in itself, then it could obviously never come to perceive anything.

In this way, LKL and the Anaxagorean view seem each to contain exactly what the other misses. I suggest that An. 2.5 can be read as a way out of this key puzzle developed throughout An. 1.2 and 1.5. We can, then, understand An. 2.5 effectively as a way of showing how the truths of both views need to be and can be combined into a coherent account of perception that is able to solve the difficulties raised for each side.

On the one hand, we must admit, against LKL and with Anaxagoras, that if the perceiver is not to lose her perceptivity, she can never come to be like the external perceptual object in the sense of herself becoming a perceptual object similar to it. If something is to feel the coldness of something else, it must be able to retain its own unlikeness with respect to it: it must remain, so to speak, equally distant from all its potential objects (i.e. unbiased) and the perceived object must remain external to it. On the other hand, if the perceptive capacity were nothing but this power of resistance and the perceiver always remained entirely neutral, there could be no explanation as to how the activity of perceiving this or that object comes about. Accordingly, against Anaxagoras and with LKL, there must be a sense in which the perceiver is affected by, and assimilated to, the perceived object, but in such a way that having been assimilated to it is compatible with being further affected by it – that is, with like being affected by like (LAL).

Although, as we have seen, the main account of perception that Theophrastus ascribes to Anaxagoras has nothing to do with the demand of impassivity, at one point Theophrastus formulates a dilemma that is almost identical to the one that we have just seen emerging from An. 1. He does so in the course of discussing Empedocles’ adherence to LKL. Theophrastus suggests here that, contra Empedocles, the senses must be impassive and must be understood as resisting becoming like their objects – with the caveat that a properly specified kind of likeness may, after all, apply to them:

The problem Theophrastus describes here is the one that is shared by both sides in An. 1.2 – namely, the lack of a sufficiently nuanced understanding of what ‘to be affected by’, ‘to be assimilated to’, and ‘to be like’ something else means. Both sides share, first of all, an all-too-narrow understanding of what ‘to act’ and ‘to be affected’ means. What is needed here are distinctions, a διορισμός – presumably like the one offered in An. 2.5 (417a21ff.).Footnote ⁷⁵ Aristotle will succeed in reconciling the two side, only if he can show that each, effectively, speaks about ‘being affected’ (and ‘being like’) in a different way and that these two ways are in fact compatible. Only then shall we see that Anaxagoras was right to imply that the perceiver must be impassive and remain unlike the perceptual object in one sense, and that LKL was right that she must be affected by, and assimilated to, the perceptual object in another sense, such that she is able to be further affected by it.

Distinguishing these two senses and showing how the two insights can be combined in a consistent account, I contend, is exactly the task undertaken by Aristotle in An. 2.5. Central to this endeavour, as we shall next see, is the notion of preservative πάσχειν.