## Hendrik Lorenz

Print publication date: 2006

Print ISBN-13: 9780199290635

Published to Oxford Scholarship Online: September 2006

DOI: 10.1093/0199290636.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 19 January 2019

# The Workings of phantasia

Chapter:
(p.148) 11 The Workings of phantasia
Source:
The Brute Within
Publisher:
Oxford University Press
DOI:10.1093/0199290636.003.0015

# Abstract and Keywords

This chapter discusses Aristotle’s version of the association of ideas. It exposes a rich conception of sense, or the perceptual system, which can account not only for acts of sense-perception, but also for associating one thing with another, remembering things, and being reminded by something of something else. So conceived of, the perceptual system can also account for the occurrence of action-guiding representations that are indeterminately complex, and at the same time relevant and suitable to the perceiving subject’s current circumstances.

Keywords:   Aristotle, association, sense, perception, memory

It may be helpful to begin by recalling the roles perception and phantasia are presented as playing in the conceptual framework that Aristotle employs in his discussions of animal locomotion, in De Anima 3.9–11 and in the De Motu Animalium. Much of this conceptual framework is on display in two rather similar passages, one from De Motu Animalium 6, the other from De Anima 3.10:

We see that the movers of the animal are thought (δι$α ´$νοια), perception, phantasia, decision, wish, spirit, and appetite. And all of these can be reduced to thought (νο$υ ˜$ς) and desire. For phantasia and perception hold the same place as thought: for all of these involve discernment, while they differ in ways that have been stated elsewhere. (De Motu Animalium 6, 700b17–22)

These two are seen to produce movement, either desire or thought (νο$υ ˜$), if one were to take phantasia to be like a kind of thinking ($ω ‘$ς ν$ο ´$ησ$ι ´$ν τινα): for many follow phantasiai against knowledge, and in the other animals there is neither thinking (ν$ο ´$ησις) nor reasoning (λογισμ$ο ´$ς), but there is phantasia. Both of these, then, can produce movement in respect of place, thought and desire—but thought which reasons for the sake of something and is practical. (De Anima 3.10, 433a9–14)

In both passages, Aristotle proposes to account for animal locomotion in terms of cognition and desire. He also makes clear, in these passages and in their respective contexts, that there are, on his view, different kinds of cognition, and different kinds of desire. The relevant kinds of cognition are thought or thinking (δι$α ´$νοια, ν$ο ´$ησις) on the one hand and phantasia and perception on the other. In the passage from De Anima 3.10, Aristotle suggests that we take phantasia to be ‘like a kind of thinking’. He nevertheless implicitly insists, in the same passage, on the distinction between phantasia and thinking: he credits all or almost all non-human animals with phantasia and at the same time denies them the capacity for thinking. In the present chapter, I shall discuss some points of contact between phantasia and thought, hoping to shed light on what Aristotle may have in mind in suggesting that phantasia can be taken to be ‘like a kind of thinking’. In the next chapter, I shall turn to the question of why Aristotle, in spite of whatever similarities there may be between the two, nevertheless insists on their distinctness.

Given that the forms of cognition that Aristotle makes available for the explanation of animal movement are thought, phantasia, and perception, the cognition (p.149) involved in the purposive movement of non-human animals must on his view be explicable in terms of phantasia and perception alone. What I intend to do in what follows is to consider some forms of non-human animal behaviour that Aristotle observes and discusses, and to reflect on the question of how it might be that the cognition involved in such forms of behaviour can be explained simply in terms of phantasia and perception, as he conceives of them. My main objective will be to bring out the remarkably powerful notion of phantasia with which Aristotle operates. For this purpose, it will not be necessary to provide a comprehensive or exhaustive survey of the forms of behaviour that he observes and discusses. Rather, I shall focus on a few cases that seem especially helpful in showing the remarkable power of phantasia, as he conceives of it.

In the present chapter, as well as in the next one, I shall focus on non-human animal motivation as providing the clearest case of non-rational motivation, as Aristotle conceives of it. What I intend to bring to light is a rich and, I think, rather attractive conception of non-rational motivation that is in principle applicable both to non-human animal behaviour and to human behaviour that fails to manifest reason. In Chapter 13, I shall turn to the question of the extent to which Aristotle takes that conception to be applicable to the behaviour of adult, ordinarily developed humans.

In a passage from Nicomachean Ethics 3.10 that we had a look at in Chapter 9, Aristotle discusses a situation in which a predatory animal notices some suitable prey somewhere in its environment. In that passage, he is interested in the pleasure that the predator takes in such circumstances. ‘What pleases the lion’, he insists, ‘is not the sight of “a stag or a wild goat”, but that he is going to get a meal.‘1 The lion's pleasure, Aristotle thinks, is a pleasure of anticipation, and so he must take it to involve apprehending the prospect of having a meal. This makes clear that he thinks non-human animals can, in some way or other, anticipate or envisage prospects. Independently of this, it seems to be an implication of his account of animal locomotion, in De Anima 3.9–11 and in the De Motu Animalium, that non-human animals can envisage prospects. He evidently thinks that they are capable of locomotion for the sake of goals, and this capacity seems to presuppose the capacity for envisaging prospects.

It is fairly easy to see at least some ways in which perception and phantasia may enter into accounts of the types of animal response and behaviour that Aristotle notes in Nicomachean Ethics 3.10. Perception supplies the predator with the information that some suitable prey is located nearby in its environment. A phantasia which, in some way or other, presents the prospect of having a meal will play a role in the explanation both of the lion's pleasure of anticipation, and of its purposive locomotion towards its prey.

If Aristotle has in mind an account along these lines, as it seems clear that he does, he must assume, not only that non-human animals can envisage prospects, (p.150) but also that there is some mechanism which brings it about that in cases of the kind described in Nicomachean Ethics 3.10, animals envisage prospects that are suitable to the circumstances they find themselves in, whatever these may be. It is plain, after all, that there is a rather tight fit between the prospect the animal apprehends by way of phantasia and its current situation, which is presented to it by way of its senses. What a lion typically anticipates on seeing (say) a stag is having a meal, rather than, for instance, copulating. This fit between prospective and present situations cannot be a mere coincidence. It must stem from an ability that lions and many other kinds of animals have as a matter of being naturally constituted the way they are, namely to envisage prospects that are, more often than not, suitable to their present circumstances.

These points may be made in another way. The suggestion so far has been that, on Aristotle's view, animals with the capacity for retaining sensory impressions are capable of envisaging prospective situations, with the latter capacity playing a crucial role in purposive locomotion. However, one deficiency of the account so far offered on Aristotle's behalf is that it fails to explain the fact that non-human animals can, in appropriate circumstances, be relied on to behave in rather specific ways. There are circumstances in which a lion, when presented with a stag, will hunt it down and sink its teeth into it.2 According to Aristotle's account, the lion's behaviour expresses and realizes a purpose. Forming that purpose, I have suggested, requires rather specific exercises of the capacity for phantasia. Thus Aristotle must, I take it, assume that there are circumstances in which lions can be relied on, when presented with some prey, to have some suitable phantasia that will in some way or other represent eating the prey, rather than having no phantasia at all, or having some quite different phantasia.

A theory which, like Aristotle's, proposes to account for the cognitive achievements involved in the purposive behaviour of non-human animals in terms of perception and phantasia should then be able to account, in these terms, not only for their ability to envisage prospective situations, but also for the fact that, given certain conditions, they can be relied on to envisage prospects that are suitable to the circumstances they find themselves in. Otherwise there would be an important gap in Aristotle's account. Now, Aristotle does not explicitly confront the question why it is that some of the brute animals can, given certain circumstances, be relied on to envisage rather specific prospects. It is nevertheless possible to make a detailed and, I hope, persuasive case for the view that perception and, in particular, phantasia, as he conceives of them, can, or anyhow are meant to be able to, account for an animal's ability to envisage prospects that are suitable to its (p.151) circumstances. Making this case will be my task in the remainder of the present chapter.

According to Aristotle's psychological theory, for any animal capable of perception and phantasia, it is the same part or aspect of its soul that accounts for its being capable both of perceiving and of having phantasiai.3 We can see this in the first chapter of the De Insomniis. Aristotle begins that treatise by asking in virtue of what part or aspect of the soul it is that we have dreams and, specifically, whether dreams are affections of the part or aspect of the soul that is concerned with thinking (τ$ο $ νοητικ$ο ´$ν), or the one concerned with perceiving (τ$ο $ α$ι ’$σθητικ$ο ´$ν) (458a33–b2).4 At the end of the chapter, the question is answered: dreaming belongs to the part or aspect of the soul that is concerned with perceiving, in so far as it is concerned with phantasia (459a21–2); for a dream appears to be a kind of phantasia$α ´$ντασμα), hence to belong to the part or aspect concerned with phantasia$ο $ ϕανταστικ$ο ´$ν), and that part or aspect is in fact the same as the part or aspect concerned with perceiving (τ$ο $ α$ι ’$σθητικ$ο ´$ν) (459a14–22). It is Aristotle's view, then, that there is a part or aspect of the soul, which may be referred to as the perceptual part, that enables certain living things both to perceive and to have phantasiai.5 It turns out that (p.152) this part or aspect is, on Aristotle's view, also responsible for dreaming, and for remembering.6

The perceptual part of the soul is meant to account for a variety of interrelated activities in which animals engage, much as the soul as a whole, according to Aristotle's psychological theory, accounts for an even wider variety of interrelated activities. And as the perceptual part is conceived of as a part or aspect of the soul as a whole, so the activities it accounts for form a subset of the set of interrelated activities that the soul as a whole accounts for. Activities that Aristotle takes to belong to the soul, but not to its perceptual part, include digestion and thought.7

Now, an ordinarily developed living thing that is equipped with a perceptual soul-part is an organism with a certain structure. This will typically involve having a variety of sense-organs and a central organ of perception; I shall refer to that configuration of organs as the animal's perceptual apparatus. For animals capable of phantasia, this apparatus will be complex enough to support, not only the reception of sensory impressions when appropriate objects are present to its senses, but also the retention of such impressions when the objects in question are no longer present. It is part of Aristotle's psychological theory, I suggested, that the ability to retain sensory impressions enables animals to envisage prospects, and to form purposes that may impel them to engage in movement from one place to another.

There is, moreover, good reason to think that, on Aristotle's view, the perceptual part of a suitable animal's soul can account for the fact that, given certain conditions, it can be relied on to envisage prospects that are suitable to the circumstances in which it finds itself. As is clear from a number of texts in the Parva (p.153) Naturalia, Aristotle takes it to be part of the functioning of the perceptual part of the soul that connections or associations between sensory impressions are formed and maintained in the perceptual apparatus of suitably constituted animals.8 As a result, he is in a position to explain an animal's ability to envisage prospects that are suitable to its present circumstances in terms of associations between sensory impressions. He may, for instance, hold that a suitably conditioned animal associates eating, presented to it by way of phantasia, with the look and the smell of animals of certain kinds, as presented to it by way of its senses.

In order to support, and give more content to, this suggestion, I shall discuss two texts from the Parva Naturalia in which Aristotle presents and employs a rather elaborate theory of ordered sequences of sensory impressions. These passages are chapter 3 of the De Insomniis and chapter 2 of the De Memoria et Reminiscentia. Both texts rely explicitly on the account of phantasia offered in De Anima 3.3.9 According to that account, a phantasia is a change (κ$ι ´$νησις) which arises from the activity of perception; it is like the perception that produced it; and it can persist beyond the activity of perception that produced it (De Anima 3.3, 429a1–5). As we shall see, both texts make clear that Aristotle takes the changes or, as I shall call them in what follows, affections10 that constitute phantasiai to be in some way or other retained or preserved in the animal's perceptual apparatus.11

Both texts, moreover, present theories according to which it is, in suitably constituted animals, part of the functioning of the perceptual part of their souls that sensory affections are preserved in their perceptual apparatus in an orderly way, with dispositions obtaining among them to the effect that specific representations tend to become active together with, or to be followed by, other specific representations. As a result, Aristotle can account for a remarkable degree of order in the mental lives of non-human animals. Perceptual experience, he is in a position to hold, can bring it about that phantasiai are activated in an animal's perceptual apparatus when and as appropriate, and that phantasiai form ordered sequences of indeterminate duration and complexity. All of this may happen, he can add, without thought being involved in any way at all. The two texts present a coherent and relatively detailed view of the affections that constitute phantasiai, and of what accounts for the order which sequences of such affections may (p.154) exhibit. For our purposes, the two texts complement each other rather nicely. In the De Insomniis, Aristotle goes into considerable detail concerning the material basis and underlying physiology of phantasia. In the De Memoria et Reminiscentia, he makes some very interesting remarks about what accounts for the order which sequences of sensory affections tend to exhibit.

I shall begin with chapter 3 of the De Insomniis. By the time we get to that chapter, Aristotle has answered the question which the De Insomniis begins by asking, namely what part of the soul it is to which dreams belong. His answer, as we have seen already, is that they belong to the perceptual part, in so far as it is responsible for phantasiai. For dreams, he holds, are phantasiai of a certain kind. He has also restated, and in fact amplified somewhat, the account of phantasia offered in De Anima 3.3. He has added to that account that the affections that constitute phantasiai are qualitative changes, caused by the qualitative changes that constitute perceptions (459b1–7). Moreover, in chapter 3 itself he adds that these affections, or at least the active ones among them, are ongoing disturbances in the animal's perceptual apparatus: ‘We must suppose that like the little eddies that form in rivers, so each of the changes [sc. sensory affections] occurs continuously (γ$ι ´$νɛσθαι συνɛχ$ω ˜$ς). Often they remain in the same way. Often they are broken down into other shapes because of collisions’ (De Insomniis 3, 461a8–11). We should note that the retention of such affections requires that disturbances created by acts of perception are in some way or other preserved in the animal's perceptual apparatus.12 These disturbances, moreover, are contentful. As they arrive at the central organ of perception—the heart, that is (De Iuventute 3, 469a5–7)—they generate sensory experiences:

In blooded animals, as the blood becomes calm and separated out, the change belonging to percepts13 from each sense-organ is preserved (σ$ω ˛$ζομ$ɛ ´$νη). This makes dreams connected (ɛ$ι ’$ρ$ο ´$μɛνα),14 makes things appear to the dreamer, and brings it about that they seem to see on account of the changes descending from sight, to hear on account of those coming from hearing, and so on with those that proceed from the other organs. For also when one is awake, it is because of the change from there arriving at the starting point [sc. the central organ of perception] that one seems to be seeing, hearing, and perceiving. (De Insomniis 3, 461a25–b1)15

(p.155) One phenomenon in which Aristotle is interested is the contrast between unconnected, disorderly dreams and dreams that are well connected and life-like. His explanation of that contrast is that the heat associated with the activity of digestion generates large-scale disturbances in the relevant parts of the body, which can interfere with the more delicate disturbances that carry the contents of dreams (461a14–25). It is in the blood of suitably constituted animals, he holds, that contentful affections originally created by acts of perception are preserved.16 What he says suggests that he takes such affections to be preserved primarily in the blood located in the peripheral sense-organs.17 In sleep, much of that blood travels to the heart, carrying with it affections that are contained in it. When the blood around the heart is agitated by the large-scale disturbances of digestion, the contentful affections travelling from the peripheral sense-organs to the heart may be altogether destroyed, or they may be thrown into disarray—for instance, by being broken up in collisions—so that disorderly and unconnected dreams ensue. By contrast, when the blood around the heart is relatively calm, the affections travelling to the heart may be preserved in their order and complexity, in which case they generate dreams that are coherent and life-like. Such dreams may present to the dreamer, not monstrosities, but people he or she knows,18 or the actions and pursuits of their waking lives.19 ‘When someone is asleep’, Aristotle adds,

as most of the blood travels down to its source, the changes present within it—some potentially, some actively—travel down with it. They are so disposed that in this change, that one will emerge from the blood, and as this one perishes, that one.20 They are disposed towards one another (κα$ι ´$ πρ$ο $ς $α ’$λλ$η ´$λας δ$η $ $ɛ ’ ´$χουσιν) like the artificial frogs that rise to the surface of water as salt is being dissolved.21 In a similar way, these changes are in us potentially, and become active when what arrests them is relaxed. And as they are released, they are active in the little blood that remains in the sense-organs, taking on a resemblance, as cloud-shapes do, which in their rapid changes we liken to humans and centaurs. Each of them is, as has been said, a remnant of a percept in activity ($υ ‘$π$ο ´$λɛιμμα το$υ ˜$ $ɛ ’$ν τ$η ˛ ˜$ $ɛ ’$νɛργɛ$ι ´$ $α ˛$ α$ι ’$σθ$η ´$ματος); (p.156) and when the real percept has departed, it persists, and it is true that it is like Coriscus, but is not Coriscus. (De Insomniis 3, 461b11–24)

The passage presents a remarkably elaborate theory of sensory affections. They are in the perceptual apparatus either potentially or actively. Active affections, I take it, are ongoing contentful disturbances. Potential affections are potentialities for such disturbances. They are arrested in some way or other, and they become active when what arrests them is removed or relaxed. Moreover, Aristotle plainly thinks that sensory affections are, or tend to be, ordered in certain ways, so that the activity of one particular affection is followed by the activity of another particular affection, which is followed by the activity of yet another one, and so forth.22 This is important for his account of dreaming, I suggest, because he wants to explain why dreams can represent, in a well-connected and life-like manner, complex events and processes that unfold over considerable periods of time, as when a builder dreams of building a house, or a sculptor of making a statue (cf. De Divinatione per Somnum 1, 463a21–30). If Aristotle's account is to be able to explain the occurrence of such dreams, he plainly needs to allow, not only that affections produced by acts of perception can be preserved and re-enacted, but also that the order in which such affections are received can be preserved and re-enacted. This, I submit, is exactly what he does allow in our passage.

Now, it is worth emphasizing that having dreams, no matter how complex and elaborate they may be, is not, according to Aristotle's theory, an exercise of the capacity for thinking. Nor does he think that dreams are limited to humans. He evidently thinks that some of the brute animals have dreams.23 In fact, his explanation of connected dreams is meant to apply, not only to humans, but to blooded animals in general,24 or anyhow to those among them which are capable of dreaming. Aristotle's account distinguishes sharply between dreams themselves and thoughts about dreams that a dreaming person may have—for example, the thought that the experience in question is a dream (De Insomniis 3, 462a28–9; 462a5–7; cf. 1, 458b15–20; b25). Such thoughts, if and when they occur, belong (p.157) to the intellectual part of the soul. Dreams themselves, by contrast, Aristotle assigns to the perceptual part of the soul, in so far as it is concerned with phantasia. What this means is that having dreams, no matter how elaborate and ‘connected’ they may be, is on Aristotle's view an activity that, in and of itself, involves no more than suitable exercises of the capacity for phantasia. That capacity, moreover, belongs to the system of capacities that he refers to as the perceptual part of the soul.

Let me recapitulate. Aristotle thinks, I take it, that the phantasiai that constitute dreams can exhibit order, in that they can represent complex events and processes in a connected and life-like manner. He wants to explain the possibility of such order by appealing to dispositions among sensory affections which are in some way or other preserved in the animal's perceptual apparatus. Sensory affections, he holds, are preserved in the perceptual apparatus either as active, contentful disturbances or as potentialities for such disturbances. He takes it that such affections can, in suitable organisms, be preserved in an orderly way, so that the activity of one particular contentful disturbance in the animal's perceptual apparatus is, or tends to be, followed by the activity of another particular disturbance, which is or tends to be followed by the activity of another particular disturbance, and so forth. It must then be part of his psychological theory that animals capable of preserving sensory affections in an orderly way are constituted so that appropriate dispositions can be formed among sensory affections that may be preserved in their perceptual apparatus. Moreover, since at least some of the brute animals are, on his view, capable of having ‘connected’ dreams, his theory must make the preservation of order among sensory affections available to suitable kinds of non-human animals as well as to human beings.

We should now attempt to get a clearer view of the dispositions which Aristotle thinks can come to obtain among sensory affections in the perceptual apparatus of suitable kinds of animals. Does he offer an account of how it is that such dispositions are formed and maintained? It seems to me that we can extract at least some crucial parts of such an account from a few passages in the second chapter of the De Memoria.

The main topic of De Memoria 2 is recollecting ($α ’$ν$α ´$μνησις, $α ’$ναμιμ ν$η ´$σκɛσθαι). This follows a discussion of remembering (μν$η ´$μη, μνημονɛ$υ ´$ɛιν, μɛμν$η ˜$σθαι) in the first chapter. It is in discussing recollecting that Aristotle makes especially prominent use of his theory of ordered sequences of sensory affections. However, he also relies on that theory in specifying what is involved in remembering something. As we shall see, this turns out to be rather important for our purposes. Now it is not immediately obvious what Aristotle means either by remembering or by recollecting. Before we turn to chapter 2 and its discussion of recollection, then, I want to make some remarks about Aristotle's conception of remembering, and to draw attention to some aspects of the discussion in chapter 1 that it will be important to bear in mind as we approach chapter 2.

(p.158) The objects of memory, Aristotle holds, are things that lie in the past.25 More precisely, what can be remembered, he takes it, are things that one perceived or thought of in the past. And remembering something, he thinks, is not just a matter of having in mind something that you perceived or thought of in the past. It also involves being aware that you perceived or thought of this thing in the past (De Memoria 1, 449b18–23; 450a19–21). As a result, he takes it that when you are remembering, say, a forest fire, this involves not just the retrieval and re-enactment of sensory affections that were actively present in your perceptual apparatus at the time. It also involves your being aware, perhaps in a certain distinctive way, that you did perceive what is now being represented to you at some more or less specific time in the past, or at the very least at some time or other in the past (De Memoria 2, 452b23–453a4).

It is worth noting that this conception of memory is cognitively more demanding than Plato's in the Philebus, even just so far as perceptual memory is concerned. In the Philebus, memory (μν$η ´$μν, μɛμν$η ˜$σθαι) is defined simply as the preservation of perception (34 A 10–11). One way in which memory, so understood, is employed is in putting a thirsty or otherwise depleted animal in cognitive contact with the appropriate type of replenishment, so as to enable the animal to form a desire. Socrates offers no indication that such exercises of memory as are required for the formation of desire must involve not only a re-enactment of a previously received sensory affection, but also some kind of awareness of having had past dealings with the thing in question. Plato, in the Philebus, seems to regard the mere re-enactment of a sensory affection preserved by the soul as an exercise of memory. Aristotle distinguishes between such mere re-enactment and re-enactment accompanied by awareness of past interaction with the thing in question. He regards only the latter as amounting to an act of remembering. The former he treats as a case of phantasia.

This distinction is made close to the end of chapter 1, where Aristotle responds to the difficulty of how it can be that what is remembered is not a sensory affection or appearance that, at the time, is actively present to the animal,26 but the absent object from which that affection or appearance derives (450a25–7; 450b11–15). A picture of, say, the Eiffel Tower is both a picture in its own right and a representation, or ‘likeness’, of the Eiffel Tower. You can observe it all by itself and (p.159) simply as the picture it is. But you can also look at it as a representation of the Eiffel Tower. Likewise, Aristotle suggests, a phantasia that is involved in an act of remembering is something all by itself (α$υ ’$τ$ο ´$ τι καθ’ α$υ ‘$τ$ο ´$), and it is at the same time a representation of the thing, now absent, from which it derives (450b20–7). Correspondingly, he distinguishes between two ways of employing a phantasia. The soul, he thinks, can attend to the appearance involved in a given phantasia all by itself and simply as the appearance it is; but it can also employ a suitable phantasia as a representation, or ‘likeness’, of the particular thing from which it derives (450b27–451a2). Aristotle regards what occurs in the former case as merely an act of phantasia, and only what occurs in the latter case as an act of remembering.

Now, it should be clear that both ways of employing a phantasia involve having experiences with representational content.27 Even to have an ordinary phantasia of, say, a forest fire is to have a forest fire represented to one in some way or other. Remembering some forest fire, as Aristotle thinks of it, goes beyond such representation. It is not just a matter of having a forest fire represented to one. It also involves being aware, perhaps in a certain distinctive way, that what is represented to one is something that one did perceive at some time in the past. Having articulated the notion of employing a phantasia as a representation of what it derives from, Aristotle is ready to say what he takes remembering to be: the having of a phantasia as a representation of the thing it derives from (ϕαντ$α ´$σματος, $ω ‘$ς ɛ$ι ’ ´$κονος ο$υ ‘ ˜$ ϕ$α ´$ντασμα, $ɛ ‘ ´$ξις).28 That is to say I take it, that remembering something is a matter of having a phantasia in a way that involves being aware, perhaps in a certain way, that what is represented to one is something that one perceived or otherwise experienced at some more or less specific time in the past, or at least at some time or other in the past.

Remembering, Aristotle holds, belongs to the perceptual part of the soul, in so far as it is responsible for phantasia.29 This answers the last one of the three questions about remembering that the De Memoria begins by asking: in virtue of what part of the soul does remembering occur (449b4–5)? What it means is that remembering is, like dreaming, an exercise of the capacity for phantasia, which, as we have seen already, is part of the system of capacities that is the perceptual part of the soul. Given that Aristotle takes remembering to be a matter of utilizing sensory impressions in a certain way, one can readily see why he assigns the activity of remembering, via the capacity for phantasia, to the perceptual part of the soul. However, although the phantasiai that Aristotle takes to be involved in remembering (p.160) can represent an enormous variety of things, they nonetheless are subject to the limitation that they are sensory representations. They cannot in themselves provide cognitive contact with intelligibles such as, for instance, essences or natures.30 At the same time, Aristotle's discussion from the start includes references to remembering, not only perceptibles, but intelligibles as well—for example, remembering some object of study (De Memoria 1, 449b15–23). However, if remembering in general is a matter of utilizing sensory impressions in a certain way, it is not clear how anyone can possibly remember, say, what it is to be a human being. Somewhat surprisingly, Aristotle does not explicitly flag this as a difficulty, but he does attempt to answer the question.

Every act of the human intellect, he holds, involves and requires representing features such as magnitude and time, features whose representation involves and requires suitable exercises of the capacity for phantasia. It is at least part of the idea that thinking anything at all, anyhow for thinkers like us, requires visualizing the objects of thought by means of the sensory imagination.31 The visualizations in question are phantasiai. Aristotle rather naturally extends this idea and claims that visualizing is required, not only for grasping an object of thought in the first place, but also for subsequent acts of remembering the thing in question: ‘memory also of intelligibles’, he says, ‘does not occur without a phantasia’ (450a12–13). This makes acts of phantasia necessary for remembering intelligibles. Aristotle seems to think, however, that it also establishes that remembering in general belongs in its own right (καθ’ α$υ ‘$τ$ο ´$) to the perceptual part of the soul, in so far as it is responsible for phantasia, and at best incidentally to the intellect (450a13–14). In any case, Aristotle plainly does hold that remembering in general belongs in its own right to the perceptual part of the soul, and incidentally to the intellect. He also holds, relatedly, that the proper objects of memory are, as he puts it, things of which there is phantasia32—by which, I take it, he means things that phantasia can represent.33 Things that cannot be grasped without phantasia, he adds, are incidental objects of memory. In the context, it is clear that the latter items are meant to be intelligibles. They cannot themselves be represented by the sensory affections that constitute phantasiai, but their grasp by the intellect requires appropriate acts of phantasia.

(p.161) What Aristotle appears to have in mind, then, is something like this. It is after all possible to remember intelligibles, such as, for instance, what it is to be a human being. Intelligibles, however, are not remembered in their own right. Remembering intelligibles is always parasitic on remembering things that are remembered in their own right, and these are things that are represented by phantasia. If this is Aristotle's view, as it seems to be, he will say that what actually happens whenever someone remembers an intelligible object is that he or she in the first place remembers something that is represented by phantasia, and that memory happens to be accompanied by an act of the intellect that is the thought of the object in question, perhaps in that this act of the intellect is prompted by the relevant exercise of phantasia. The upshot is that things that can be represented by phantasia can be remembered directly and immediately, whereas intelligibles can only be remembered indirectly, in a way that is mediated by remembering things that are represented by phantasia. If that is Aristotle's picture, this makes at least some sense of his view that intelligibles are incidental objects of memory, and that remembering belongs to the intellect incidentally. For on that picture remembering intelligibles will always accompany, and depend on, remembering things that are represented by phantasia, and such acts of the intellect as may be involved in remembering will always accompany, and depend on, appropriate acts of phantasia.

The question remains, of course, why Aristotle adopts a picture along these lines. His adoption of some such picture is motivated, I suggest, by his acceptance of the following premisses.

1. (1) The proper objects of memory are things which are capable of being represented by representational items (states, processes, or whatever) which can be preserved in the animal's organism.

2. (2) Sensory affections are the only sort of representational items that can be preserved in an animal's organism.

3. (3) Sensory affections cannot represent intelligibles.

These premisses entail the conclusion that intelligibles are not among the proper objects of memory. To accept premiss (1) is to adopt a rather natural view of the functioning of memory as a matter of storing and retrieving representational items of some sort or other. Committing something to memory, on that view, crucially involves forming and retaining some sort of representation of it, and remembering it involves retrieving that representation and employing it in a certain way. As we have seen, Aristotle does embrace a view of memory along these lines.34

(p.162) Given Aristotle's psychological theory, moreover, the representational items in question could either be thoughts or sensory affections. Now, we have seen that Aristotle takes sensory affections to be contentful modifications in the hylomorphic structure that is the animal's perceptual apparatus. There is nothing mysterious about how such modifications can be preserved indefinitely in the animal's perceptual apparatus. By contrast, Aristotle holds that there is no such thing as a bodily organ or apparatus of thought (De Anima 3.4, 429a22–7). Thoughts are not, on his view, modifications of any kind in a bodily structure, nor are they constituted by such modifications. Since he does not take them to reside in a bodily structure in the first place, he cannot make sense of their preservation in a bodily structure.

On Aristotle's view, then, sensory affections are the only sort of representational item that can be preserved in the animal's organism. However, since sensory affections cannot represent intelligibles, Aristotle is compelled to accept that intelligibles are not among the proper objects of memory. He does want to say, though, that it is in a way possible to remember intelligibles. To show how, he resorts to the rather ingenious idea that intelligibles are incidental objects of memory. When you remember, say, the proof of a geometrical theorem which you studied the day before yesterday, what actually happens, Aristotle might say, is that you remember how you visualized the items mentioned in the proof (as being extended objects of such-and-such sizes and shapes) as well as how you visualized the operations performed on them (cutting them in halves, and the like). These memories are not memories of the proof itself. But they are, or may well be, accompanied by the thought of the proof itself, perhaps in that they may prompt an intellectual act that is the thought of the proof. If so, Aristotle can say that in a way you are remembering the proof. You are remembering it incidentally, because you are remembering how you visualized it, and that memory happens to be accompanied by the thought of the proof itself.

We should now turn to chapter 2 and its discussion of recollecting (τ$ο $ $α ’$ναμιμν$η ´$σκɛσθαι). I begin with some linguistic points. Anamimnēskein is a transitive verb, meaning ‘to remind’. The present infinitive anamimnēskesthai can be construed either as middle, ‘to remind oneself, to recollect’, or as passive, ‘to be reminded’. Now, if one looks at the passages in the Meno and the Phaedo in which Plato presents and discusses his so-called theory of recollection,35 it becomes clear that he strongly tends to use the infinitive form anamimnēskesthai in contexts in which someone actively sets out to call something to mind (middle rather than (p.163) passive construal),36 rather than contexts in which it just so happens that someone is reminded of something without having tried to call the thing in question to mind.37 For the latter type of case, Plato uses expressions that are unambiguously passive, like anamnēsthēnai, whenever such expressions are available.38 At the same time, it is noteworthy that Plato uses the noun anamnēsis both in the middle sense of recollecting and in the passive sense of being reminded.39 Against that background and in light of the fact that Aristotle is echoing Plato's characterization of recollection,40 it is reasonable to expect that when Aristotle proposes to discuss to anamimnēskesthai, he has in mind deliberately recollecting something, as opposed to cases in which it just so happens that something reminds someone of something else. This expectation is in fact fully borne out by the discussion in De Memoria 2.

Here is Aristotle's statement of what he takes recollecting to be: ‘When someone recovers ($α ’$ναλαμβ$α ´$ν$η ˛$) a piece of knowledge, a perception, or that thing the having of which we said is memory, that recovery, when it occurs, is recollecting one of the things mentioned; and it turns out that this is followed by remembering and memory’41 (De Memoria 2, 451b2–6). This is a preliminary statement only, because he takes it to be true only with a qualification that he is not yet in a position to articulate fully. Not every case of recovering a piece of knowledge, a perception, or a phantasia is, he thinks, a case of recollecting. For someone can, for instance, recover a piece of knowledge, not by recollecting it, but by learning the thing in question all (p.164) over again. Recollecting, Aristotle says somewhat obscurely, requires the presence within of a principle over and above that required for learning (451b9–10).

Before he can offer his full statement of what distinguishes recollecting from relearning, he must first present his theory of ordered sequences of sensory affections. Recollecting occurs, he holds, because sensory affections form ordered sequences, so that the active occurrence of some particular contentful disturbance in one's perceptual apparatus tends to be followed by the active occurrence of another such disturbance:

Acts of recollection (α$ι ‘$ $α ’$ναμν$η ´$σɛις) happen because, naturally, this change [sc. sensory affection] occurs after that one ($ɛ ’$πɛιδ$η $ π$ɛ ´$ϕυκɛν $η ‘$ κ$ι ´$νησις $η ‘ ´$δɛ γɛν$ɛ ´$σθαι μɛτ$α $ τ$η ´$νδɛ). If this is so by necessity, then plainly whenever one undergoes the earlier one, one will undergo the later one. If it is not by necessity but by habit, one will for the most part undergo the one after the other. (It is a fact that some changes become more habitual with just one occurrence than others that have occurred many times. And this is why after seeing some things once, we remember better than we do after seeing other things many times.) In recollecting, then, we undergo some one or other of the earlier changes, until we undergo the one that is habitually followed by the change in question. It is for this reason also that we hunt for (θηρɛ$υ ´$ομɛν)42 that which follows in the sequence (τ$ο $ $ɛ ’$ϕɛξ$η ˜$ς), beginning in thought (νο$η ´$σαντɛς) with the now or with something else, and with something similar to the thing in question, something opposite to it, or something proximate to it (το$υ ˜$ σ$υ ´$νɛγγυς). Recollection occurs for this reason: for the changes that belong to these things are in some cases the same ones, in other cases they occur together, in yet other cases the one change contains part of the other, so that after the earlier one only a little remains to be undergone. It is in this way, then, that people search, but also without searching, they are reminded in this way,43 when the change in question occurs after some other one. And for the most part the change in question does occur after the occurrence of other changes of the kinds we mentioned [sc. affections belonging to items similar, opposite, or proximate to the item represented or called to mind by the affection in question]. (De Memoria 2, 451b10–25)44

(p.165) As we have seen, Aristotle's main topic in the chapter is recollecting, which is a matter of deliberately recalling something or other. In the passage just quoted, he is focusing on recollecting, but he also addresses being reminded of something without seeking to recall it. His theory of ordered sequences of sensory affections is, I take it, meant to explain both the fact that one thing frequently reminds us of another, and the fact that by means of suitable mental activity we sometimes manage to recollect things that we perceived or thought of in the past, but that do not now come to mind right away or without effort.

Aristotle begins by saying that the order that obtains among sensory affections is either necessary or habitual. In the subsequent discussion only habit recurs (at 451b28–30 and at 452b26–8). Necessity seems to drop out of consideration. It is, in any case, not easy to see how necessity might be relevant.45 Aristotle takes it, moreover, that we tend to associate things with one another on the basis of such relations as similarity, opposition, and proximity (by which he probably means both spatial and temporal proximity). He does not address the question of how habituation and such patterns of association are interrelated. He may well think that such patterns are themselves at least in part due to habituation, in that we are used to thinking of opposites together, or to hearing thunder after seeing lightning. But he may also think that relations that obtain between suitable things can facilitate, or even bring about, the formation of habits of association, as when one comes to associate toads with frogs because they are rather similar. However that may be, it is clear that Aristotle is meaning to account for recollecting and being reminded by appealing to ordered sequences of sensory affections.46 These are affections of the same kind as the ones that he mentions in De Insomniis 3. For the purposes of that text, he assumes that dispositions among sensory affections can be formed in the perceptual apparatus of suitable kinds of animals, so that active sensory affections can come to follow each other in orderly ways.47 As we saw, he relies on that assumption in explaining how their dreams can be ‘connected’. The present text adds significant detail to that picture. It says that the dispositions among sensory affections obtain either by necessity or as a result of habituation; and that sensory affections typically are so disposed that, at any rate so far as humans are concerned, things that are similar, opposite, and proximate to one another tend to be represented, or called to mind, together or in immediate succession.

(p.166) A slightly later passage seems to offer a little more detail as to how things that are proximate to each other come to be associated with one another:

It is by habit that changes follow one another, this one after that one. And so when someone wants to recollect ($α ’$ναμιμν$η ´$σκɛσθαι), he will do this: he will seek to get hold of a starting point, after which the change in question will occur. And it is for this reason that from some starting point acts of recollection occur most swiftly and finely. For just as the things in question are related to one another in terms of one thing after another ($ω ‘$ς γ$α $ρ $ɛ ’ ´$χουσι τ$α $ πρ$α ´$γματα πρ$ο $ς $α ’ $λληλα τ$ω ˛ ˜$ $ɛ ’$ρɛξ$η ˜$ς), so also are the changes. (De Memoria 2, 451b28–452a2)

Aristotle thinks that we obtain sensory affections from interacting with perceptible and intelligible objects. These objects themselves exhibit order in various ways. Thunder comes after lightning, the sea after the sandy beach, the conclusion of an argument after its premisses. It is a fact about some kinds of animals, Aristotle holds, that they are able, not only to preserve sensory affections that they obtain from interacting with perceptible or intelligible objects, but also to retain these sensory affections in an orderly way, a way that reflects the order of the objects they derive from.

He is now ready to revisit the difference between recollecting and relearning:

Recollecting differs from relearning in that the person in question will be able in a certain way to be conveyed through himself (δυν$η ´$σɛτα$ι ´$πως δι’ α$υ ‘$το$υ ˜$ κινηθ$η ˜$ναι) to what follows the starting point. When this ability is absent, and the person depends on someone or something else, he no longer remembers (ο$υ ’$κ$ɛ ´$τι μ$ɛ ´$μνηται). It often happens that one is unable to be reminded, but with some searching one is able to, and finds what one is looking for. This occurs when one initiates many changes (κινο$υ ˜$ντι πολλ$α $), until one initiates one that is such as to be followed by the thing in question. (De Memoria 2, 452a4–10)

Having presented his theory of ordered sequences of sensory affections, he can now give more content to his earlier remark that recollecting requires the presence within of a principle over and above that required for learning. What it requires, he takes it, is the presence in the person's perceptual apparatus of suitable sensory affections, and the existence of suitable dispositions among them, so that he or she will be able, by selecting an appropriate starting-point, to set off a sensory affection, or a series of such affections, so that the object in question will come to be present to his or her mind. I take it to be Aristotle's view, moreover, that where the object of recollection is intelligible rather than perceptible—say, a theorem or a definition—it will not itself be represented by the sensory affections that the person manages to excite, but those sensory affections will be accompanied by an intellectual act that is the thought of the relevant intelligible object.

It is worth pointing out that Aristotle's account contains the resources needed to distinguish recollecting, not only from relearning by being instructed, but also (p.167) from relearning by rediscovering for oneself.48 Suppose you once knew the proof of a geometrical theorem, but you subsequently forgot it. It so happens that you are unable to recollect it, but by utilizing your general knowledge of geometry you manage to work the proof out by yourself. In a way, you have recovered a piece of knowledge through yourself rather than through someone or something else. Aristotle can say, however, that you nonetheless did not recollect the proof because you were not ‘conveyed’ to it in the way that is distinctive of recollecting. For you were not conveyed all the way to it by a series of sensory affections preserved within you, so that some, or one, of these affections turned out to be accompanied by the intellectual grasp of the proof. Instead, you had to work the proof out by exercising other pieces of knowledge and hence by employing your intellect in ways other than the identification of an appropriate starting-point for recollection and the subsequent grasp of the proof itself.

Furthermore, it is important to note that, in distinguishing recollecting from relearning, Aristotle is making a fresh point about what is involved in remembering something. He says that when someone has lost the ability to be appropriately ‘conveyed’ to the active cognition of something or other, he or she no longer remembers the thing in question (452a6–7). Now, what he has in mind in saying this is plainly not that in this case the person in question is not at that time performing an act of remembering. His point is rather that in this case the person has lost the acquired ability to remember the thing in question.49 He also spells out what he takes to be involved in having the ability that we deny to someone when we say that he or she no longer remembers something or other: ‘Remembering (τ$ο $ μɛμν$η ˜$σθαι) is the presence within one of the power that conveys one [sc. to the thing in question], so that one is conveyed to it from oneself and from the changes one has within oneself, in the way described50 (De Memoria 2, 452a10–12). This characterization of what may be called dispositional memory51 applies Aristotle's theory of ordered sequences of affections to memory and remembering. It is a rather complicated characterization, and it deserves careful attention. It characterizes the acquired ability to remember something as the presence within one of a power to bring about some change, or some changes. The exercise of that power results in one's being affected so that the object of memory is represented to one or is called to one's mind.

(p.168) Now, Aristotle evidently does not think that having the acquired ability to remember, say, what Cebes looks like entails being able to perform a suitable act of remembering whenever one pleases. He thinks it happens frequently that one does not manage to activate dispositional memory. One might suppose that the case of a person who has dispositional memory but does not manage to activate it is a counterexample to Aristotle's chararacterization of dispositional memory. But this would be mistaken. His characterization of dispositional memory requires only that there is in fact some way in which it could be activated; it may be difficult for its bearer to identify that way. What he has in mind in the context is that there is some affection or other, say one that represents Cebes’ companion Simmias, such that the active occurrence of that affection would be followed, or anyhow would tend to be followed, by the active occurrence of an affection that represents what Cebes looks like. The upshot is that Aristotle takes dispositional memory not only to involve sensory affections that are retained or preserved in the organism. He also takes it to involve—in many cases and perhaps in general—the existence of dispositions that obtain among those sensory affections, such that one specific sensory affection tends to become active together with, or in succession to, the activity of another specific sensory affection.

This, I submit, is a significant addition to the account of memory and remembering offered in chapter 1 of the De Memoria. In that chapter, Aristotle concentrates on the act of remembering, having little or nothing to say about dispositional memory. That chapter, moreover, has nothing to say about the question of how it is that representations that are retained in an organism are accessed and recalled. In other words, chapter 1 has nothing to say about the transition from having dispositional memory to the act of remembering. According to Aristotle's account, the perceptual apparatus of a suitably constituted and ordinarily developed animal will retain countless sensory affections. He tells us nothing, in De Memoria 1, about how and why it is that sometimes some of these countless affections come to be active in the animal's perceptual apparatus, so that the animal is remembering this or that particular thing.

Now, one might think that he takes it to be specifically by way of recollecting, as that is discussed in De Memoria 2, that representations retained in an organism are accessed and recalled. This, however, cannot be the whole story. First, recollecting, as he thinks of it, is a matter of deliberately recalling, and there obviously are many acts of remembering that do not involve deliberately recalling whatever the thing in question may be, as when you are remembering something because it just so happens that you are reminded of it by something else. Secondly, recollecting, as Aristotle thinks of it, is a matter of deliberately recalling in a rather specific way, namely by thinking of something else that, with some luck, puts one in mind of the thing in question.52 So if you manage to call something to mind directly and (p.169) without first thinking of something else, as no doubt you sometimes do, this will not be a case of recollecting, as Aristotle thinks of it, at any rate for the purposes of De Memoria 2. In fact, his account of recollecting presupposes the ability to call something to mind directly, since the starting-points of many acts of recollection will be thoughts of things that one manages to call to mind directly.53 Presumably, calling something to mind directly is supposed to be a matter simply of thinking of it, rather than of recollecting it. Thirdly, while Aristotle evidently attributes the ability to remember to some of the brute animals, he denies the ability to recollect to all of them:

Recollecting differs from remembering not only with regard to time,54 but also in that many of the other animals, too, have a share in remembering, whereas it may be said that, apart from humans, none of the known animals has a share in recollecting.55 The reason is that recollecting is rather like a kind of reasoning (ο$ι ‘ ˜$ον συλλογισμ$ο ´$ς τις). For the person who is recollecting reasons (συλλογ$ι ´$ζɛται) that he saw or heard the thing in question before, or that he was affected by it in some other such way, and recollecting is rather like conducting a search of some kind (ο$ι ‘ ˜$ον ζ$η ´$τησ$ι ´$ς τις). To do that, however, naturally belongs only to creatures whose soul has a deliberative part as well. (And indeed deliberation, too, is a kind of reasoning.) (453a4–14)

Given how Aristotle conceives of recollecting, and how he discusses it throughout De Memoria 2, it is not difficult to see why he holds it to be limited to reasoning creatures. He seems to think that reason is involved in recollecting in at least two ways. First, anyone who sets out to recollect something or other believes that he or she did at some stage perceive or think of the thing in question, and Aristotle takes that belief to depend on some kind of grasp of reason. Thus I may believe that I went through Plato's argument for the tripartition of the soul, because I know that I studied book 4 of the Republic, and I also know that this is the text which contains that argument. Or I may believe that I heard Cebes’ name at a dinner party last week, because I know that I was introduced to him by his companion Simmias. Aristotle's thought might simply be that while you are not actually remembering the thing in question, it could only be by way of some appropriate bit of reasoning that you are aware of having perceived or thought of it at some time in the past.

Secondly, once you start recollecting, you are, according to Aristotle's theory, conducting a search, or something rather like a search, for a representation that will represent the thing in question, or call it to mind (453a15–16). This will (p.170) require finding a suitable starting-point,56 a thought that involves the occurrence of an active sensory affection, so that this affection is followed by another such affection that will represent or call to mind the object of recollection. Aristotle, naturally enough, associates this search for a starting-point with deliberation (453a12–14). Like deliberation, it is a matter of having a goal and of identifying a suitable starting-point, something that one is now in a position to do with a view to achieving one's goal.

It is plain, then, that recollection, as Aristotle thinks of it, is a rather special way in which representations retained in an organism may become active, and one that, moreover, he takes to be unavailable to the brute animals. If Aristotle's account of memory is to be anything like tolerably complete, he must at least indicate how representations can become active independently of recollection, as he characterizes it in De Memoria 2. Furthermore, it will not do simply to point to the fact that one can sometimes call something to mind directly and without first thinking of something else, as when you exercise some piece of knowledge,57 or when you think of the colour of your own eyes. For this would still not do justice to the fact that memory often becomes active without anything being deliberately called to mind, as when it just so happens that the scent of some flower reminds you of a walk you took during last year's summer vacation. It is, moreover, doubtful whether Aristotle is prepared to attribute to any non-human animal the ability deliberately to call something to mind, directly or otherwise. It seems that he regards directly calling something to mind as a case of thinking (De Memoria 2, 451b18–20), and hence as an act of the intellect. If so, it too is unavailable to the brutes. One thing that Aristotle does need to do, in any case, is to indicate a way for representations to become active which does not involve deliberately recalling the thing in question, directly or otherwise. I submit that he does precisely that when he characterizes dispositional memory as involving—in many cases and perhaps in general—the existence of dispositions among sensory affections to become active together or in succession in ways that are determined, at least in large part, by past sensory experience and habituation.

In characterizing dispositional memory in this way, Aristotle makes it clear that he takes acquiring the ability to remember, say, what Cebes looks like not simply to be a matter of retaining an appropriate sensory affection somewhere or other in one's perceptual apparatus. He also takes acquiring such an ability—in many cases and perhaps in general—to involve retaining the relevant sensory affection in a way that relates it to other such affections by way of appropriate dispositions to become active together, or in immediate succession. As a result, we can see how, (p.171) according to his account of memory, sensory affections can become active in a way that does not involve deliberately recalling the thing in question. This can happen when one type of sensory affection ‘triggers’ another type. For example, your dispositional memory of what Cebes looks like may be activated by sensory affections that actively occur in your perceptual apparatus as you see Simmias.

Moreover, Aristotle evidently holds, as we have seen, that memory and remembering belong to the perceptual part of the soul.58 On the basis of this assignment, it is, I think, reasonable to attribute to him the view that all activities and operations that form part of the ordinary functioning of memory are exercises of capacities that belong to the perceptual part of the soul, or are exercises of one such capacity. In fact, I take this to be no more than a fuller statement of his claim that memory belongs to the perceptual part of the soul. Now, we have seen that he takes it to be part of acquiring and maintaining dispositional memory that sensory affections are retained in ways which relate them to other such affections by way of appropriate dispositions to become active together or in succession. He must take it, moreover, that acquiring and maintaining dispositional memory is part of the ordinary functioning of memory. If this is along the right lines, then it is in fact clear that Aristotle is committed to the view that preserving sensory affections in a suitably structured way is a matter of exercising capacities that belong to the perceptual part of the soul, or of exercising one such capacity. The most plausible candidate for this task is, of course, the capacity for phantasia. This, after all, is the capacity that accounts for the preservation of sensory affections. Moreover, Aristotle indicates a special connection between memory and phantasia when he says that memory belongs to the part of the soul to which phantasia belongs as well (De Memoria 1, 450a22–3).

As Aristotle is quick to point out, by assigning memory to the perceptual part rather than the intellect, he is making memory available to at least some of the non-human animals (De Memoria 1, 450a15–16).59 Moreover, we have now seen that he takes it to be part of the functioning of memory, anyhow in suitably constituted animals, that sensory affections are preserved in their perceptual apparatus in a structured way, with dispositions obtaining among them to co-occur or follow one another in certain ways. By assigning memory to the perceptual part of the soul, he therefore makes the formation and maintenance of such dispositions among sensory affections available, at least in principle, to suitably constituted non-human animals. In virtue of the perceptual part of their souls, (p.172) Aristotle is in a position to hold, such animals can preserve sensory affections in suitably interrelated ways. This may enable them, for instance, to associate one thing with another, to be reminded by something of something else, and to have ongoing representations of indeterminate duration and complexity.

This position, it should be noted, is not only one for which his psychological theory fully provides the resources. It is also one that he needs to adopt if he is to be able to account for the cognitive achievements involved in forms of non-human animal behaviour that he describes in considerable detail. Consider, for instance, his report of adult deer leading their young to their lair, habituating ($ɛ ’$θ$ι ´$ξɛιν) them to the place where they should seek refuge (Historia Animalium 8.5, 611a20–1). From the point of view of Aristotle's psychological theory, such behaviour plainly needs to be accounted for in terms of the preservation of sensory affections in orderly ways, so that the habituation of juvenile deer can be seen to equip them with appropriately complex representations that are preserved in their perceptual apparatus, so as to guide their speedy return to the lair in moments of peril.

Moreover, we saw earlier in the present chapter that Aristotle's psychological theory needs to be able to account for the suitability of a non-human animal's phantasiai to its current circumstances, which Aristotle must think is manifested in anticipatory pleasure as well as in purposive locomotion. What is minimally required for explaining such phenomena is what Aristotle's account of memory in fact makes available: namely, that brute animals of many kinds can form and maintain appropriate dispositions among sensory affections retained in their perceptual apparatus, so that they may associate one thing with another, or be reminded by something of something else. Thus when a lion notices a stag in its environment, its current perceptual experience may put it in mind of what it is like to eat a stag, and that representation may both occasion anticipatory pleasure and play a crucial role in impelling the lion to go after its prey.

The interpretation that I have presented and argued for gives Aristotle no more than the bare bones of an account of non-human animal cognition in terms of connections or associations between sensory impressions. To do justice to the cognitive achievements of non-human animals, such an account would no doubt require extensive supplementation and refinement. Something would, for instance, have to be said about how it is that among all the countless possible connections or associations between impressions that might be formed, such connections as are required for the animal to survive, and to get around in the world, actually get formed. Such an account might appeal to a mechanism which privileges sequences of impressions that lead to, or involve, pleasurable experiences, e.g. ‘stag-eating’. But we should also bear in mind that Aristotle leaves open the possibility that at least some sequences of representations may be a matter, not of habit, but of necessity:

Acts of recollection happen because, naturally, this sensory affection occurs after that one. If this is so by necessity, then plainly whenever one undergoes the earlier one, one will (p.173) undergo the later one. If it is not by necessity but by habit, one will for the most part undergo the one after the other. (De Memoria 2, 451b10–14)

The underlying idea might well be that the perceptual apparatus of some kinds of animals is constituted so that they are predisposed to proceed from one specific type of representation to another, provided that the animal in question actually receives sensory affections of the relevant types. In other words, the idea might be that some kinds of animals are ‘wired up’ in such a way that their perceptual apparatus contains, as it were, ‘slots’ specifically for certain types of affections, in which affections of these types are stored as soon as they are received. Affections of one type will then be linked to affections of some other type, with the effect that the animal in question invariably proceeds from representations of one type to representations of another. In this way, Aristotle's Empiricism60 about phantasia could turn out to be a less extreme position than it may appear to be: although an animal has to acquire by experience whatever sensory affections it needs, its nature might be such as to facilitate—or even, given a suitably conducive environment, to predetermine—the formation of such connections or associations between impressions as are required for it to be able to live in the way that is characteristic of its species.

## Notes:

(1) [sc. χα$ι ´$ρɛι] $ο ’ ´$τι βορ$α $ν $ɛ ’ ´$ξɛι: Nicomachean Ethics 3.10, 1118a18–23.

(2) These circumstances include, for instance, that the lion is in reasonably good health and not completely sated, and perhaps also that it has acquired appropriate levels of relevant experience and skill of the sorts that lions naturally acquire in their habitat. The difference between a healthy, hungry lion and a sick or sated one will not lie in what prospects they can envisage, but presumably in which ones they find pleasurable and thus desirable.

(3) There are several passages in Aristotle's psychological writings in which he mentions μ$ο ´$ρια τ$η ˜$ς ψυχ$η ˜$ς, a notion which I intend to capture by writing of parts or aspects of the soul. Although Aristotle is not very specific about what he has in mind in mentioning these items, a number of points are nevertheless clear. Being a part of the soul is contrasted with being a soul (De Anima 2.2, 413b11–16); as a result, conceiving of (for instance) whatever it is that is concerned with perceiving (α$ι ’$σθητικ$ο ´$ν) as a part or aspect of the soul enables Aristotle to resist the view that an animal may have more souls than one, since it has something concerned with perceiving, something concerned with nutrition, and so forth. At the same time, he evidently finds the notion that the soul is a thing of parts—a composite object, that is—to be deeply and seriously problematic, as we saw in Chapter 3. The aporia for soul partition that is articulated at De Anima 1.5, 411b5–14, is never, in fact, resolved. Thus Aristotle may well have in mind a notion as weak as ‘aspect’. Claims about parts or aspects of the soul may simply be claims about how the various capacities which constitute the soul are related to one another, and about which ‘psychic’ capacities are needed to account for a given activity or operation which living things perform in virtue of being ensouled.

(4) I assume that the expression τ$ο $ α$ι ’$σθητικ$ο $ν at 458b2 refers to a part or aspect of the soul, just as the expression τ$ο $ νοητικ$ο ´$ν in the same sentence. Thus the second question seems to me to be a specification of the first one, narrowing down the range of candidates to two. τ$ο $ α$ι ’$σθητικ$ο ´$ν has been introduced as a part or aspect of the soul in the preceding treatise, De Somno (which, at 453b17–20, announces the De Insomniis): at 454a11–19, Aristotle mentions τ$ο $ α$ι ’$σθητικ$ο ´$ν as one of the items that are spoken of as parts or aspects of the soul (μ$ο ´$ρια τ$η ˜$ς ψυχ$η ˜$ς), and later in the same chapter he refers to it as the part or aspect concerned with perceiving (τ$ο $ α$ι ’$σθητικ$ο $ν μ$ο ´$ριον): ‘Sleep is an affection of the part or aspect concerned with perceiving, a kind of fetter and lack of movement; so that it is necessary that everything that sleeps has a part or aspect concerned with perceiving’ (454b9–12). This result is assumed in the De Insomniis: ‘Let us assume what is quite obvious, that dreaming is an affection of that which is concerned with perceiving (τ$ο $ α$ι ’$σθητικ$ο ´$ν), just as sleep is: for dreaming does not belong to another part or aspect of animals than sleep’ (459a11–14).

(5) It is with, or in virtue of, the α$ι ’$σθητικ$ο ´$ν that we have certain cognitions: see De Insomniis 1, 458b2–3. In other words, the α$ι ’$σθητικ$ο ´$ν, rather than itself doing the perceiving, enables us to perceive. This form of expression reflects Aristotle's view that it is ‘perhaps better’ to say that we pity, learn, or think with, or in virtue of, the soul, than to say that the soul pities, learns, or thinks (De Anima 1.4, 408b13–18). For more on this view, see Conclusion, pp. 203–4.

(6) ‘Memory also of intelligibles does not occur without a phantasia. Hence it would seem to belong incidentally to that which is concerned with thought, but in itself to the primary part or aspect concerned with perceiving (τ$ο $ πρ$ω ˜$τον α$ι ’$σθητικ$ο ´$ν)’ (De Memoria et Reminiscentia 1, 450a12–14). This suggestion answers one of the questions posed in the first sentence of the De Memoria et Reminiscentia, namely ‘to which of the parts or aspects of the soul does this affection [sc. remembering] occur?’

(7) I reject J. Whiting's suggestion, argued for in ‘Locomotive soul: the parts of soul in Aristotle's scientific works’, Oxford Studies in Ancient Philosophy, 22 (2002), 192–200, that thought, or at least practical thought, belongs to the α$ι ’$σθητικ$ο ´$ν. Aristotle never says or implies that a person thinks or deliberates in virtue of the part of their soul that is concerned with perceiving. Nor does he ever identify the α$ι ’$σθητικ$ο ´$ν with the νοητικ$ο ´$ν, or its practical aspect. If thought belonged to the α$ι ’$σθητικ$ο ´$ν, then ‘α$ι ’$σθητικ$ο ´$ν’ and ‘νοητικ$ο ´$ν’ would be two designations for one subject, in precisely the way Aristotle in fact takes ‘α$ι ’$σθητικ$ο ´$ν’ and ‘ϕανταστικ$ο ´$ν’ to be. That this is not his view is clear from the De Insomniis and the De Memoria. He begins the De Insomniis by asking, as we have seen, whether dreaming belongs to the part concerned with thinking (νοητικ$ο ´$ν) or to the α$ι ’$σθητικ$ο ´$ν. His answer is that it does not belong to the part responsible for belief and thought (De Insomniis 1, 459a8–9)—the νοητικ$ο ´$ν, that is—but in fact to the perceiving part, with the qualification that it belongs to it in so far as it is concerned with phantasia (459a10–11, 21–2). In De Memoria 1, he notes that he is assigning memory, specifically not to either one of the intellectual parts (450a16–17), but to the part or aspect to which phantasia belongs (450a22–5). (For the moment, I am leaving aside the complication that since objects of thought are incidental objects of memory, memory belongs incidentally or derivatively also to the intellect. I shall shortly offer some comments on the role of the intellect in Aristotle's account of memory.)

(8) As Beare saw, the texts in question contain Aristotle's version of the ‘association of ideas’: Greek Theories of Elementary Cognition from Alcmaeon to Aristotle (Oxford: Oxford University Press, 1906), 306, 318.

(9) References to De Anima 3.3 are at De Insomniis 1, 459a14–18, and at De Memoria et Reminiscentia 1, 449b30–1.

(10) Aristotle speaks both of changes and of affections (π$α ´$θος): De Insomniis 2, 459a26, b5; De Memoria 1, 450b5, 12, 18.

(11) De Insomniis 2, 459a26–7; b5–7; 3, 462a9: $ɛ ’$ν το$ι ˜$ς α$ι ’$σθητηρ$ι ´$οις (‘in the sense-organs’). Cf. De Memoria 1, 450a28–9: $ɛ ’$ν τ$η ˛ ˜$ ψυχ$η ˛ ˜$ κα$ι $ τ$ω ˛ ˜$ μορ$ι ´$ $ω ˛$ το$υ ˜$ σ$ω $ματος τ$ω ˛ ˜$ $ɛ ’ ´$χοντι α$υ ’$τ$η ´$ν (‘in the soul and in the ensouled part of the body’); 2, 453a24: πɛρ$ι $ τ$ο $ν α$ι ’$σθητικ$ο $ν τ$ο ´$πον (‘around the place concerned with perception’). Note also De Anima 1.4, 408b17–18: recollection involves changes and ‘states of rest’ (μον$α ´$ς) in the sense-organs ($ɛ ’$ν το$ι ˜$ς α$ι ’$σθητηρ$ι ´$οις).

(12) It is worth pointing out the striking closeness in conception between Aristotle's phantasiai and memory in Plato's Philebus. According to the Philebus, perceptions are contentful disturbances (σɛισμο$ι ´$) undergone jointly by body and soul (Philebus 33 D 5, E 11; note also π$α ´$θος at 34 A 3 and κ$ι ´$νησις at 34 A 4); and memory is the preservation (σωτηρ$ι ´$α) of such disturbances (34 A 10–11).

(13) τ$ω ˜$ν α$ι ’$σθημ$α ´$των $η ‘$κ$ι ´$νησις at 461a26 is, I take it, a shorter expression for α$ι ‘$ $υ ‘$π$ο ´$λοιποι κιν$η ´$σɛις α$ι ‘$ συμβα$ι ´$νουσαι $α ’$π$ο $ τ$ω ˜$ν α$ι ’$σθημ$α ´$των at 461a18–19. What Aristotle has in mind is something that can be thought of either as one complex disturbance or as any number of interrelated disturbances that jointly travel from the peripheral sense-organs to the central organ of perception.

(14) Like Ross, Beare in J. Barnes (ed.), The Complete Works of Aristotle, and Gallop, I accept Lulofs's conjecture ɛ$ι ’$ρ$ο ´$μɛν$α ´$ at 461a27.

(15) My translations from the De Insomniis are indebted to those by Beare, in Barnes (ed.), The Complete Works of Aristotle, and by Gallop, in Aristotle on Sleep and Dreams.

(16) This is not to say that on Aristotle's view the blood itself receives and preserves the affections in question. His view rather seems to be that it is pneuma contained in the blood that is the bearer of sensory affections. F. Solmsen, ‘Greek philosophy and the discovery of the nerves’, Museum Helveticum, 18 (1961), 172–8, and G. Freudenthal, Aristotle's Theory of Material Substance: Heat and Pneuma, Form and Soul (Oxford: Clarendon Press, 1995), 130–4, offer detailed discussions of this point.

(17) Note De Insomniis 2, 459a24–8; $ɛ ’$ν το$ι ˜$ς $ο ’ ´$μμασιν(‘in the eyes’), 459b10–11; 3, 461b16–21.

(18) Coriscus, for example: 462a2–8.

(19) De Divinatione per Somnum 1, 463a23–7: ‘when we are about to do something, or are in the middle of doing something, or have done something, it often happens that in dreams we find ourselves with these acts and find ourselves doing them—the reason being that the change [sc. the sensory affection that constitutes the phantasia in question] happens to have its path prepared (προωδοποιημ$ɛ ´$νη) as a result of our daytime beginnings.‘

(20) ο$υ ‘ ´$τω δ’ $ɛ ’ ´$χουσιν $ω ‘ ´$στɛ $ɛ ’$ν τ$η ˛ ˜$ κιν$η ´$σɛι τ$η ˛$δ$ι $ $η ‘ ´$δɛ $ɛ ‘$πιπολ$α ´$σɛι $ɛ ’$ξ α$υ ’$το$υ ˜$ $η ‘$ κ$ι ´$νησις, $α ’ $ν δ’ α$υ ’ $τηφθαρ$η ˛ ˜$, $η ‘ ´$δɛ.

(21) According to Sophonias, 37, 12–24, and Michael of Ephesus, 72, 8–19, Aristotle has in mind a number of wooden frogs that are buried in layers of salt one on top of the other. As water is added and the salt dissolves, one frog after another rises and, in rising, becomes visible.

(22) The sentence in which this becomes clear (461b13–15) is, I think, often under-translated. Consider, for instance, Gallop's translation in Aristotle: On Sleep and Dreams: ‘They are so disposed that in any given movement, one movement will rise from it to the surface; and if that one perishes, then another will do so.’ (Similarly Hett's Loeb translation, and Beare in The Complete Works of Aristotle.) According to this translation, Aristotle's point is that there is a steady flow of sensory affections in the perceptual apparatus, such that for any given one, there is another that follows it. However, this interpretation fails to give force to the demonstrative pronouns in Aristotle's Greek. (After all, he could have written, say, $α ’ ´$λλη τις κ$ι ´$νησις instead of $η ‘ ´$δɛ $η ‘$ κ$ι ´$νησις in line 14, and $α ’ ´$λλητις instead of $η ‘ ´$δɛ in line 15.) The present sentence should be compared with a passage from the De Memoria which, I take it, expresses the same idea of order obtaining among sensory affections: ‘Acts of recollection happen because, naturally, this change ($η ‘$ κ$ι ´$νησις $η ‘ ´$) occurs after that one (τ$η ´$νδɛ). If this is so by necessity, then plainly whenever one undergoes the earlier one, one will undergo the later one. If it is not by necessity but by habit, one will for the most part undergo the one after the other’ (De Memoria 2, 451b10–14).

(23) De Divinatione per Somnum 2, 463b12–13; note also De Insomniis 1, 459a13–15.

(24) $ɛ ’$ν το$ι ˜$ς $ɛ ’$να$ι ´$μοις (‘in blooded animals’), De Insomniis 3, 461a25–6.

(25) $η ‘$ δ$ɛ $ μν$η ´$μη το$υ ˜$ γɛνομ$ɛ ´$νου (‘memory is of the past’): 449b15, 27–8.

(26) That what one remembers might be sensory affections preserved in one's perceptual apparatus is so abstruse a thought that I hesitate to attribute it to Aristotle even for purposes of articulating an aporia. Aristotle plainly uses the word ϕ$α ´$ντασμα (‘phantasia‘) to refer both to sensory affections preserved in an animal's perceptual apparatus (e.g. at De Memoria 1, 450b10) and to appearances which he takes to be involved in the active occurrence of such affections (e.g. ibid., 449b31–450a1, 451a10). I am inclined to think, partly on the basis of the present passage, that he uses the related terminology of π$α ´$θος (‘affection’), τ$υ ´$πος (‘impression’), and the like, in the same twofold way. If so, it is open to us to interpret the difficulty discussed at 450b11–451a17 as dealing with the question of whether one remembers appearances that are present to one's mind at the time, or absent objects from which such appearances derive. This is a good question to ask, and Aristotle's subsequent discussion seems to me to offer a plausible and interesting answer to it.

(27) I agree here with S. Everson, Aristotle on Perception, 196.

(28) I assume that what Aristotle is meaning to define is the activity of remembering ($ɛ ’$νɛργɛ$ι ˜$νκατ$α $ τ$ο $ μνημονɛ$υ ´$ɛιν, $ɛ ’$νɛργɛ$ι ˜$ν τ$η ˛ ˜$ μν$η ´$μ$η ˛$), since activities are definitionally prior to capacities (De Anima 2.4, 415a18–20). In his definition of remembering, Aristotle may be using the word $ɛ ’ ´$ξις in precisely the way Plato uses the same word in the Theaetetus’ aviary simile. There, $ɛ ’ ´$ξις is contrasted with κτ$η ˜$σις (Theaetetus 197 B 1–4). The latter denotes possession; the former is illustrated by having a cloak on, and by holding a bird in one's hand. Note also the aorists σχ$η ˛ ˜$ at De Memoria 1, 449b19, and σχɛ$ι ˜$ν at Theaetetus 197 C 9; this means something like ‘to get hold of’.

(29) De Memoria 1, 450a22–3: τ$ι ´$νος μ$ɛ $ν ο$υ ’ ˜$ν τ$ω ˜$ν τ$η ˜$ς ψυχ$η ˜$ς $ɛ ’$στι μν$η ´$μη, ϕανɛρ$ο ´$ν, $ο ‘ ´$τι ο$υ ‘ ˜$πɛρ κα$ι $ $η ’$ ϕαντασ$ι ´$α (‘it is clear then which part of the soul memory belongs to: the part that phantasia belongs to as well’). Cf. 451a16–17.

(30) This is because they are in themselves simply exercises of sensory capacities. They belong to the perceptual part of the soul, after all. Only acts of the intellect can provide cognitive contact with intelligibles.

(31) Cf. De Memoria 1, 450a4–5: κα$ι $ $ο ‘$ νο$ω ˜$ν $ɛ ‘$σα$υ ´$τως, κ$α ’ $ν μ$η $ ποσ$ο $ν νο$η ˛ ˜$, ποσ$ο ´$ν (‘in the same way a person who is thinking, even if he is not thinking of something with a size, places something with a size before his eyes’).

(32) 450a23–5: $ɛ ’$στι μνημονɛυτ$α $ καθ’ α$υ ‘$τ$α $ μ$ɛ $ν $ω ‘ ˜$ν $ɛ ’$στι ϕαντασ$ι ´$α, κατ$α $ συμβɛβηκ$ο $ς δ$ɛ $ $ο ‘ ´$σα μ$η $ $α ‘ ´$νɛυ ϕαντασ$ι $ας (‘things of which there is phantasia are objects of memory in their own right; things which are not grasped without phantasia are incidental objects of memory’).

(33) I should perhaps note that in writing of phantasia being able or unable to represent something or other, I am meaning to convey the idea that it is able or unable to provide cognitive contact with the item in question, the way sight, for instance, is able to provide cognitive contact with colours but not with flavours.

(34) The view he adopts is, incidentally, indebted to Platonic antecedents. In writing, at 450a29–32, of ‘something like a painting’ (ο$ι ‘ ˜$ον ζωγρ$α ´$ϕημ$α ´$ τι) being retained in the living organism, and of ‘something like an imprint’ (ο$ι ‘ ˜$ον τ$υ ´$πον τιν$α ´$) being stamped in the organism ($ɛ ’$νσημα$ι ´$νɛται), the way seals are imprinted with signet rings (καθ$α ´$πɛρ ο$ι ‘$ σϕραγιζ$ο ´$μɛνοι το$ι ˜$ς δακτυλ$ι ´$ $ι ´$οις), Aristotle is echoing not only the Philebus‘ simile of the painter in the soul, but also the Theaetetus‘ wax block model of memory and knowledge. According to the latter, we have in our souls a block of wax, and ‘we make impressions ($α ’$ποτυπο$υ ˜$σθαι) upon this of everything we wish to remember among the things we have seen or heard or thought of ourselves; we hold the wax under our perceptions and thoughts and take a stamp from them, in the way we take the imprints of signet rings ($ω ‘ ´$σπɛρ δακτυλ$ι ´$ων σημɛ$ι ˜$α $ɛ ’$νσημαινομ$ɛ ´$νους)‘ (Theaetetus 191 D 4–8). Aristotle adopts Plato's picture with two significant modifications. First, the generation of imprints does not depend on what one wishes to remember, but occurs simply as a matter of the ordinary functioning of the animal's cognitive apparatus. Secondly, there are, for Aristotle, no imprints of thoughts (450a27–32).

(35) Meno 81 C 5–86 C 2; Phaedo 72 E 1–77 A 5.

(36) Note, for example, Meno 85 D 6–7, where Socrates asks: τ$ο $ δ$ɛ $ $α ’$ναλαμβ$α ´$νɛιν α$υ ’$τ$ο $ν $ɛ ’$ν α$υ ‘$τ$ω ˛ ˜$ $ɛ ’$πιστ$η ´$μην ο$υ ’$κ $α ’$ναμιμν$η ˛ ´$σκɛσθαι $ɛ ’$στιν; (‘Is not one's own recovery of knowledge in oneself recollection?‘) Also 86 B 4: $ɛ ’$πιχειρε$ι ˜$ν ζητε$ι ˜$ν κα$ι $ $α ’$ναμιμν$η ˛ ´$σκɛσθαι; (‘try to seek out and recollect’). Cf. Phaedo 75 E 2–7.

(37) This point is missed entirely by Sorabji, Aristotle on Memory, 40–1.

(38) For example, Phaedo 73 D 10–11: Σιμμ$ι ´$αν τις $ι ’$δ$ω $ν πολλ$α ´$κις К$ɛ ´$βητος $α ’$νɛμν$η ´$σθη (‘on seeing Simmias, one is often put in mind of Cebes’); E 6–7: κα$ι $ Σιμμ$ι ´$αν $ι ’$δ$ο ´$ντα γɛγραμμ$ɛ ´$νον К$ɛ ´$βητος $α ’$ναμνησθ$η ˜$ναι (‘on seeing a picture of Simmias, [someone may] be reminded of Cebes’). At Phaedo 73 C 6–74 A 7, Socrates discusses cases of one thing reminding someone of another thing, in order to make certain points that he takes to apply to every case of $α ’$ν$α ´$μνησις, crucially including active, deliberate recollection. At that stage of the discussion, Socrates is notably careful in using unambiguously passive forms of $α ’$ναμιμν$η ´$σκɛιν whenever they are available. The only ambiguous form is $α ’$ναμιμν$η ˛ $σκɛται at 74 A 5. This is present tense indicative, where no unambiguously passive form is available. Cf. Aristotle, Nicomachean Ethics 9.4, 1166b15 (a passage mentioned by Sorabji, Aristotle on Memory, 99), where Aristotle says that bad people avoid being alone, because while alone they are reminded of many distressing things ($α ’$ναμιμν$η ´$σκονται γ$α $ρ πολλ$ω ˜$ν κα$ι $ δυσχɛρ$ω ˜$ν).

(39) The noun is to be construed in the middle sense in the slogan that learning is recollection. Passive uses are in evidence at Phaedo 73 D 10, E 1, and 74 A 2. The noun is clearly used in the middle sense at De Memoria 2, 453a15. There is an interesting passive use at Nicomachean Ethics 3.10, 1118a12–13: self-indulgent people take pleasure in the smells of perfumes and tasty dishes, because ‘through these they are reminded of the objects of their appetites’ (δι$α $ το$υ ´$των $α ’$ν$α ´$μνσις γ$ι ´$νɛται α$υ ’$το$ι ˜$ς τ$ω ˜$ν $ɛ ’$πιθυμημ$α ´$των).

(40) To recollect is to recover knowledge ($α ’$ναλαμβ$α ´$νɛιɛ … $ɛ ’$πιστ$η ´$νην): Meno 85 D 6–7; Phaedo 75 E 4; De Memoria 2, 451b2–3. Cf. also Philebus 34 B 6–8.

(41) With all extant manuscripts, I read τ$ο $ rather than τ$ω ˛ ˜$ at 451b5. Ross and Sorabji follow Michael (c. AD 1090) and Sophonias (c. AD 1300) in reading the latter. Ross's reason for rejecting the reading of the manuscripts is that it makes μν$η ´$μην ‘a mere repetition’ of μνημονɛ$υ ´$ɛιν. This is true, but in view of the same kind of repetition at 449b4, 451a14–15, and 453b8–9, it is no good reason to abandon the reading of all manuscripts.

(42) This is another Platonic echo, this time from the Theaetetus‘ aviary model: 197 D 1, 198 A 2, A 7.

(43) ζητο$υ ˜$σι μ$ɛ $ν $υ ’ ˜$ν ο$υ ‘ ´$τω, κα$ι $ μ$η $ ζητο$υ ˜$ντɛς τɛς δ’ ο$υ ‘ ´$τως $α ’$ναμιμν$η ´$σκονται. The word order suggests strongly that Aristotle intends a contrast between ζητο $υ ˜$σι (‘people search’) and μ$η $ ζητο$υ ˜$ντɛς(‘without searching’) rather than, as Sorabji takes it, between ‘people search in this way’ and ‘without searching in this way’. (See Kühner–Gerth, Ausführliche Grammatik der Griechischen Sprache, Zweiter Teil: Satzlehre (Hannover: Hahnsche Buchandlung, 1904), §528.). Beare takes the sentence the way I do; Sorabji, 99, admits that his reading strains the Greek. Sorabji's problem is that if the text is read in the way it is most natural to read it, Aristotle seems to speak of recollecting without searching. But he repeatedly characterizes recollecting as a matter of searching (esp. 453a15–16; cf. 451b30, 452a8, and 453a12). However, $α ’$ναμιμν$η ´$σκονται need not be construed as middle; it can just as naturally be read as passive. Sophonias (10, 1–2), for what it is worth, takes the second clause to describe a case of being reminded without having searched: $ο ‘ ´$ταν δ$ɛ $ μ$η $ ζητο$υ ˜$σιν $α ’$ναμνη σθ$η ˜$ναι του γ$ɛ ´$νηται (note the passive!). The idea, I take it, is this. People search in this way: namely by thinking of something or other that is somehow related to the thing they are searching for—for example, something similar, opposite, or proximate to it. In this way, too, people, may be reminded of something without searching for it: by thinking of something or other that is somehow related to the thing in question—for example, by being similar, opposite, or proximate to it.

(44) My translations from the De Memoria are indebted to those by Beare, in Barnes (ed.), The Complete Works of Aristotle, and by Sorabji, Aristotle on Memory.

(45) I shall offer a suggestion at the end of the chapter.

(46) This claim ought to be acceptable independently of my view that Aristotle is explicitly talking about being reminded at 451b23–4. For also the series of associations employed in recollection will typically involve multiple cases of being reminded by something of something else (e.g. 452a13–16).

(47) The De Insomniis comes after the De Memoria in Bekker's edition of Aristotle's works. This may well be in line with Aristotle's view about the order in which the two texts should be studied. The De Sensu, at 436a5–17, sets out a programme of topics to be discussed which the Parva Naturalia follows loosely. In that order, memory comes right after sense-perception and precedes, among other things, sleep and waking. Dreaming, moreover, is announced as a topic to be discussed at the beginning of the De Somno (453b17–20). There is some reason for thinking, then, that the De Insomniis takes as read the De Memoria‘s rather more detailed statement of Aristotle's theory of orderly sequences of sensory affections.

(48) Sorabji, Aristotle on Memory, 38–9, claims that Aristotle goes wrong in failing to recognize that one can relearn something through one's own efforts, ‘and without depending on someone else’. On my view, that objection misfires. Aristotle does not, and need not, deny that one can relearn by oneself. He can gladly accept this, since he has the resources needed to distinguish recollecting from that kind of relearning, too.

(49) This acquired ability corresponds to the second potentiality, or first actuality, that is knowing something without contemplating it: De Anima 2.5, 417a21–417b2; 2.1, 412a22–3.

(50) τ$ο $ γ$α $ρ μɛμν$η ˜$σθα$ι ´$ $ɛ ’$στι τ$ο $ $ɛ ’$νɛ$ι ˜$ναι δ$υ ´$ναμιν τ$η $ν κινο$υ ˜$σαν*** το$υ ˜$το δ$ɛ ´$, $ω ‘ ´$στ’ $ɛ ’$ξ α$υ ‘$το$υ ˜$ κα$ι $ $ω ‘ ˜$ν $ɛ ’ ´$χɛι κιν$η ´$σɛων κινηθ$η ˜$ναι, $ω ‘ ´$σπɛρ ɛ$ι ’ ´$ρηται.

(51) The formulation is due to Sorabji, Aristotle on Memory, 1.

(52) Aristotle makes it very clear, throughout De Memoria 2, that what he thinks of as recollecting is always a matter of mentally proceeding from something else to the object of recollection: see, e.g., 451b16–18, 18–22, 29–31, 452a4–6, 8–10, 12–16, etc.

(53) De Memoria 2, 451b18–19: δι$ο $ κα$ι $ τ$ο $ $ɛ ’$ϕɛξ$η ˜$ς θηρɛ$υ ´$ομɛν νο $η ´$σαντɛς $α ’$π$ο $ το$υ ˜$ ν$υ ˜$ν $η ‘ $ $α ’ $ λλου τιν$ο ´$ς κα$ι $ $α ’$ϕ’ $ο ‘$μο$ι ´$ου $η ’ $ $ɛ ’$ναντ$ι ´$ου $η ’ $ το$υ ˜$ σ$υ ´$νɛγγυς (‘it is for this reason that we hunt for that which follows in the sequence, beginning in thought with the now or with something else, and with something similar to the thing in question, something opposite to it, or something proximate to it’).

(54) What Aristotle has in mind, I take it, is that recollecting typically occurs some time after memory has first been established. This is because recollecting requires that the item in question has, so to speak, absented itself from one's mind, for example in that it has been forgotten, or simply in that one has not thought of it in a while. Cf. De Memoria 2, 451a31–b2.

(55) Aristotle offers a more confident statement of this view at Historia Animalium 1.1, 488b26.

(56) Cf. 451b29–31: $ο ‘ ´$ταν το$ι ´$νυν $α ’$ναμιμν$η ´$σκɛσθαι βο$υ ´$ληται, το$υ ˜$το ποι$η ´$σει** ζητ$η ´$σɛι λαβɛ$ι ˜$ν $α ’$ρχ$η $ν κιν$η ´$σɛως, μɛθ’ $η ‘ $ν $ɛ ’$κɛ$ι ´$νη $ɛ ‘ ´$σται (‘when someone wants to recollect, he will do this: he will seek to get hold of a starting-point, after which the change in question will occur’).

(57) Cf. De Anima 2.5, 417b22–4: ‘Knowledge is of universals, and these in a way are in the soul itself. For this reason thinking is up to the person, and he can think whenever he wishes to (δι$ο $ νο$η ˜$σαι μ$ɛ `$ν $ε ’$π’ α$υ ’$τ$ω ˛ ˜$, $ο ‘$π$ο ´$ταν βο$υ ´$ληται).’

(58) De Memoria 1, 450a22–3; 451a16–17; this is confirmed at the end of the treatise, 453b8–10. To do justice to the complexity of Aristotle's position, we should add that he takes the simple statement that memory belongs to the perceptual part of the soul to be appropriate so far as the proper objects of memory are concerned. As we saw earlier, he takes the intellect to be involved in remembering intelligibles. For present purposes, however, I can afford to limit myself to Aristotle's views on remembering the proper objects of memory; and so I disregard the complications introduced into his theory of memory by remembering intelligibles.

(59) Other texts in which Aristotle attributes memory to non-human animals include Historia Animalium 1.1, 488b25–6, and Metaphysics A 1, 980b21–7.

(60) For a brief account of Empiricism, see Introduction, pp. 4–6.