Jump to ContentJump to Main Navigation
Perceptual Imagination and Perceptual Memory$

Fiona Macpherson and Fabian Dorsch

Print publication date: 2018

Print ISBN-13: 9780198717881

Published to Oxford Scholarship Online: June 2018

DOI: 10.1093/oso/9780198717881.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. date: 28 November 2020

Aristotle on Distinguishing Phantasia and Memory

Aristotle on Distinguishing Phantasia and Memory

(p.9) 2 Aristotle on Distinguishing Phantasia and Memory
Perceptual Imagination and Perceptual Memory

R. A. H. King

Oxford University Press

Abstract and Keywords

Aristotle is the first philosopher to give an account of memory based on phantasia, a very broad, non-cognitive power of the soul. But how he distinguishes phantasia when not used in memory, from memory, is not easy to say. Two fundamental strategies for making this contrast may be crudely distinguished—we may call their proponents ‘the Activist’ and ‘the Phenomenalist’. An Activist will say that Aristotle will say: imagining is doing something different from remembering. A Phenomenalist will say that memories and imaginings appear different to their subject. The aim of this chapter is to show that neither Phenomenalist nor Activist can stand alone in an account of memory such as Aristotle’s. For, to put it in a slogan, remembering is an activity involving appearances.

Keywords:   Aristotle, memory, phantasia, imagination, appearances, misremembering, soul, Coriscus

1. Accounting for Memory Using Imagination

Like many ancient philosophers Coriscus worked for a potentate: he left Athens where he had been a member of Plato’s Academy for Atarneus near his home city of Scepsis in Asia Minor, where Hermias held power. And Aristotle remembers Coriscus.1 What does he do, when he does this?

Aristotle is the first theorist to use ‘imagination’ or phantasia to account for memory. But just how phantasia forms part of the explanation of memory, and just how Aristotle distinguishes phantasia when not used in memory, from memory, is not easy to say. This is the first of two contrasts we shall be pursuing in Aristotle—phantasia in memory and phantasia apart from memory. Two fundamental strategies for making this contrast may be crudely distinguished—let us call their proponents ‘the Activist’ and ‘the Phenomenalist’. An Activist will say, or say that Aristotle will say: imagining is doing something different from remembering. A Phenomenalist will say that memories and imaginings appear different to their subject. The aim of this chapter is to show that neither Phenomenalist nor Activist can stand alone in an account of (p.10) memory such as Aristotle’s. For, to put it in a slogan, remembering is an activity involving appearances.

So one contrast is that between imagining and remembering. Another contrast is that between remembering rightly and wrongly.2

To see how this may look, let us see how Aristotle approaches memory through what I call the Canonical Formula:

For always whenever someone is active with respect to remembering, then he says in this way in the soul that he heard this or perceived it or thought it before.

(On memory 1 449b22–23)

Here we are given a canon or standard for deciding when memory is present, when someone is remembering. Actively remembering is saying something, and relates to a past perception or thought. Of course, the person remembering need say nothing out loud; it can be in the soul: Aristotle says to himself, I saw Coriscus. If an act of memory is saying something, one important implication is that it can be assessed as to truth and falsity. “I saw Coriscus”, said by particular person at a particular time, may be true or false. Depending on that, the memory claim of which it is the content is successful or not. One may be deceived or not, as to whether one is remembering or not, or remembering rightly or not. This is the second important contrast for a theory of memory.

Apparently, there is no ‘imagination’ here at all. We will see that phantasia is called in to explain memory; the occurrence of memory, as described here, requires no occurrent “image”. And it is good that we are neither asked to consider remembering an image, nor that all memory requires occurrent images. But while the “image” (phantasia or phantasma) is not the object of memory, it may form part of the explanation of memory. The theorist appeals to it in her theory. The theory is then that the fact that we are able to remember is based on the fact that we have phantasia. Phantasia forms part of the capacity to remember.

This approach to memory through the Canonical Formula is an application of an insight which relates generally to capacities: to understand a capacity, you need to look at the related activity, and to look at the activity you have to consider the object (De anima II 4 415a16–22). This points to the first step Aristotle takes in his account of memory, and which the Canonical Formula forms part of: to explain memory we need to understand what it is we remember. A crucial element in the answer to this question is the past (449b15), that is, I take it, past perceptions and so what we perceived in past perceptions. When someone is active with memory, the Canonical Formula says, we say I perceived this or that earlier. And this or that may be Coriscus, for example.

Aristotle is adamant that animals apart from humans possess the capacity to remember.3 The account in On Memory and Recollection applies to animals besides humans; as has often been pointed out, the Canonical Formula presents a problem here in that animals do not say anything. So, humans are the main witnesses, but (p.11) Aristotle remains committed to explaining living behaviour in general, and one determinant of his view of memory is the conviction that it is not an activity of reason, since then only humans would have it. The one cognitive faculty all animals possess is perception, and perception is, ultimately, responsible for imagination. This is an important result of the treatment of phantasia in De anima, which Aristotle refers expressly back to in the account of memory (1 449b30). The memory involved is perceptual in that it relies on an act of perception, and has as its object the object of that perception. Thus, in remembering, Aristotle says, I saw Coriscus. (For the purposes of this chapter, I ignore remembering relating to thought.)

He explains memory by relying on his account of the faculties of the soul given in De anima—above all, perception and phantasia. We will have to be selective in using this general theory of the soul here; but here are some theses, in descending order of generality which help support the theory of memory. Soul is, by definition, embodied: it is the primary activity of an organic body. Ends are involved in living behaviour. Perception gives rise to the ability to have and activate phantasia. In turn memory depends on phantasia. Thus memory is not a fundamental faculty of the soul. So we have in our case a capacity to remember, and its exercise, being active with memory.4 This is the patch where the Activist will pitch her camp.

We may distinguish two perspectives from which one should be able to distinguish between phantasia and memory. It is firstly something that you and I, Aristotle and Coriscus need to be able to do on a day-to-day basis. If we cannot distinguish between the two, we would be in trouble, practically, and, presumably, psychologically. But besides this everyday perspective, there is secondly the philosophical or theoretical question.5 Posing the puzzle of distinguishing between imagination and memory forces the question on us, at least for some ways of thinking, what each of them is. In everyday use, we do not have such an account. Nor, since we make this distinction as a matter of course, do we need to have such an account. You’re imagining things, you say. No, I remember it clearly, I reply, I heard her say that. Clearly, there are cases and cases to be distinguished in everyday talk here.

Aristotle is committed to a methodology which consists in making precise what is held to be the case, by the wise or the many, about the explanandum. In an important sense, his philosophy is a refinement of what we anyway know. He does not construct a conception of memory, and of imagination, to then rely exclusively on technical definitions to distinguish them. This would anyway be a problematic procedure in the case of memory and imagination. What right would we then have to say that what we have defined is memory? As we have seen, Aristotle in fact starts his investigation from (p.12) the way we talk about memory: we say (as co-conversationalists) that we are active with memory when we say we have perceived or thought something before.

Aristotle uses the way we attribute memory to someone, in saying something, but also what the animal or human remembering does, to investigate what is contained in an act of memory. The Aristotelian scientist relies on a basic recognition of the distinction to arrive at the definition of the explanandum.

Thus he is committed to saying that we do distinguish between memory and phantasia. But this is not a phenomenal distinction, the Activist will say: it is not “vivacity” or “intimacy and warmth”, to mention characteristics which have been appealed to more recently in the history of philosophy in this context,6 that allows us to distinguish between the two. These terms, phenomenal terms, are not the ones he uses to make the distinction. Rather, it is a matter of what we are doing. Part of what this means is that there is non-transitional awareness of what we are doing when we remember something. The activity of imagining something (‘putting it before the eye of the soul’) is only one activity of phantasia (De anima III 3 427b18–20). Here too one could argue from the Activist’s perspective that the work of distinguishing is not phenomenal, but drawn by awareness of what you are doing. So, the argument would go, when you imagine something it may well be phenomenally identical to a memory, but talking about what you are doing, you would not say: “I can remember such and such”,7 but “I am imagining such and such”, or “he appears to me such”. Since memory is an activity of phantasia, and phantasia an activity of perception, this line of thought is rooted in Aristotle’s view that perception itself is responsible for our awareness of perception (De anima III 2 425b12–25).

How can the Phenomenalist reply? He or she must admit that there is no talk of vividness or familiarity. But her arsenal is not exhausted. Firstly, she may appeal to a further aspect of Aristotle’s account: time. Aristotle links memory closely to knowing about the past, as we have seen. And one might argue that perceiving time in Aristotle’s view—“I saw her yesterday”—changes the phenomenon. Here it is not just the simple image (assuming for a moment that phantasmata are images) that passes through my mind, but this image, plus, in some sense, yesterday, or at least, past. Since time is perceived, and phantasmata arise from perception, the perception of time will not be just a symbolic temporal index. A second limb of the Phenomenalist’s reply will be expanded on presently: phantasmata may be pictures, and pictures have phenomenal characteristics.

(p.13) Time is the hero of Aristotle’s account; for memory is of the past (449b15).8 This is part of the way he distinguishes between phantasia and memory: clearly it is possible to have a phantasia without any index of time. Aristotle thinks one perceives time; and a full account would have to tackle the question of the perception of time, whether measured (yesterday, in the 50th Olympiad), or indefinite (past). The tricky question dividing Activist and Phenomenalist here is whether time has a phenomenal quality. The Activist can insist that time, strictly, needs counting; the Phenomenalist will counter that Aristotle (realistically) does not insist that every memory comes date stamped: they need merely the odour of pastness, in some sense. Memory is restricted to animals with a sense of time (On memory 1 449b24–30).

How does the Activist react to this? Aristotle uses the way we attribute memory to someone, in saying something, but also what the person remembering does, to investigate what is contained in an act of memory. There are thus two perspectives on memory: what we say about someone remembering, and what the human or animal remembering does. Thus one question concerns the change in perspective from inside to outside—for how else do we know what occurs when remembering occurs, if not because we remember? And there is no word of justification of the move from us to others, from our remembering to anyone or anything remembering. No doubt, there are things to be said that explain why Aristotle feels no need here. At this stage of the investigation into the behaviour of living things, it is already clear that soul always occurs with body, since it is the primary activity of body. Thus there is no problem here about appealing to other living things, and indeed to what they say, insofar as they speak at all, in appraising the deliverances of memory. Humans serve here as a model, and it is not clear what the brute analogue of speaking might be. This change in perspective is also important for the move from the way we talk to the scientific account offered by the Aristotelian definitions. This question is related to two others—one is about generality: are there general accounts of memory? Now, memory has been subjected to a variety of taxonomies; and Aristotle’s is very economical; one may well wonder if he can cope with all the ways we talk about memory, let alone theoretical uses in modern psychology. He argues that memory has to do with perception, and then accidentally with intellectual activities. Both thinking and perception are included in the Canonical Formula. Here, I will concentrate entirely on perceptual memory. (Thought here is the systematic thought of the scientist.)

2. Phantasia—appearances and apparitions

Now, we have already seen that Aristotle refers back to his definition of phantasia in De anima towards the beginning of his account of memory. And for the Aristotelian (p.14) scientist it is straightforward to distinguish between phantasia and memory, given their definitions. Phantasia and memory are defined differently, hence they are different. Phantasia is, according to the official definition, “a kind of change remaining from an actual perception” (De anima III 3 428b30–429a2). The final definition of memory is equally quickly stated, “the possession of a phantasma, possessed as a likeness (eikôn) of that of which it is the phantasma” (On memory 1 451b15–16). A formal and fairly trivial point may be made before we embark on the interpretation of these hard sayings. The definition of memory makes use of the notion of a phantasma, the product of the capacity phantasia. So not only are memory and imagination not identical, memory requires phantasia, and requires more than phantasia.

Let us begin by unpacking the definition of phantasia. Phantasia refers both to the capacity and its product, whereas a phantasma is only the latter (De anima III 3 428a1). Phantasia is attributed to or explained by the capacity to perceive (for phantasma cf. On memory 1 450a10–11). In a loose way of speaking, phantasia is a capacity of living things; but it is not one of the primitive capacities of living things since it is derived from perception.9

Aristotle is ambitious: phantasia is meant to explain a wide range of phenomena; and a question mark must hang over his success in this enterprise.10 There are varieties of phantasia, apparently, and it is not clear how they all fall under the general account. There are various etymological connections of the word phantasia which are relevant in Aristotle’s account. He himself relates it to phôs, light, and says that this is because there is no sight without light, and sight is the primary form of perception: etymologies of this kind are of course a feature of Greek philosophy—Plato fills a book with them (the Cratylus). But there are two other connections, reflected in the language Aristotle himself uses when discussing phantasia. One is with the verb “to appear” phainesthai, the other is with the verb phantazesthai (cf. 433b12), and hence with phantasma, “apparition, phantom”.11 These three associations may be seen as pulling the account of imagination in different directions—the connection with light and perception may suggest that phantasia reveals the way things are; whereas the connection with appearance suggests we should be careful about it; and as to apparitions, well, can they give us any guidance at all as to the way things are? Presumably not. The fact that Aristotle himself underscores a presumed connection with light and perception, suggests that he sees phantasia in some sense as a useful capacity. And the use he puts it to explaining not only dreams (which serve no purpose in his book), but also action and memory confirms its ability to guide us undeceptively, as well as to mislead us.

Let us look briefly at appearances and apparitions. Both aspects of phantasia are relevant to the treatment of memory: we can be deceived by apparent memory, but need not be. And as to phantasmata, well, it is not just phantoms that are meant in Aristotle’s book, but also what we might call images—persisting delusions. But care (p.15) is needed here in that phantasia includes more than images—appearances are also captured by phantasia.12

Aristotle remarks that we do not say that something appears such and such when we perceive it clearly (De anima III 3 428a12–15). From this he concludes that phantasia is not involved in all perception—a case of my seeing something is not merely a case of things appearing to me, it is seeing the way things are. After all, perception, when successful, is cognition. The question this raises is how mere appearance is distinguished from perception proper.

In contrast, phantasia are ‘appearances’, for example, the sun appears a foot across, but it is believed to be bigger than the inhabited world (De anima 428b2–4). Here a crucial contrast between perceptual appearance and belief is marked. Care is needed; for it is possible to be taken in by appearances, in other words to believe them. And believing them can be deceptive or not. The main point is that appearances can deceive or not; but an appearance does not allow one to decide what is the case and what is not.

In contrast, perhaps, to this aspect of phantasia, we have a range of phenomena, which relate to images—above all dreams and imagination, in the sense of an ability to call up images before the mind’s eye. This is the realm of phantasms. “This affection (pathos) [i.e. phantasia] is in our power, when we want (for it is to put something before one’s eyes, like those do using mnemonic techniques, making an image for themselves)” (De anima III 3 427b17–20).13

A key text for this strand of phantasia is from On memory and recollection. A representation, e.g. the phantasma is like the imprint of a seal, or like a portrait (On memory 1 450a29–32, 450b15–451a2). It is very common indeed to translate phantasma or phantasia (in the relevant use) by “image”, and so burden Aristotle with a view of the imagination that is squarely to do with images merely by dint of translation. Thus many readers of Aristotle regard phantasmata as images, indeed as mental images.14 But it deserves to be said that Aristotle nowhere says that phantasmata are images.

What we have been doing is unpacking some of what Aristotle’s definition of imagination contains, as part of the project of showing how easy it is for the Aristotelian natural philosopher to hold imagination and memory apart. Now what we have to do is to reconstruct the basis in everyday talk for the view of memory and imagination presented in their respective definitions.

On one reading, a phantasma is merely something which may float through your mind, a face in a daydream; remembering something requires doing something, namely taking this as a likeness (eikôn) of the original perception, and hence of that perception’s object. Aristotle remembers Coriscus, and says, I saw Coriscus.

(p.16) Aristotle thinks that, in remembering, a change remaining from perception, a phantasia, an appearance or representation, is taken as a likeness (eikôn) of the thing remembered. We have two contrasts we wish to draw, firstly that between imagination and memory, and secondly that between being right or wrong when one remembers. We will have to look more closely at this distinction between appearance and likeness later.

The background theory of the soul allows for living things to engage in activities. There are many different activities which living things engage in, and different activities relate to one another in a variety of ways; for example, one activity may use the products of another. This is a promising course for the theorist of memory to take. Memory makes use of phantasia. Then one can distinguish between a phantasma when embedded in an act of memory and when it is not so embedded. The appearance can occur to one as just that, a face floating through your mind, or else you say: I saw Coriscus.

At this point, the Activist and the Phenomenalist part company. For there are two ways of interpreting what is happening here. Either we think that the person remembering takes the phantasma in a certain way, or the phantasma appears to one in a certain way, say, as a portrait of someone.

Is the decisive point how we take the phantasma, or how it appears to one? These two possibilities mark the distinction between the Activist and the Phenomenalist. The Activist points to our activity in taking the phantasma one way or the other. The Phenomenalist points to the fact that the face, say, floating through your mind, may be the way Coriscus appeared, or someone else. The time has now come to look at the puzzle which Aristotle uses to approach his account of memory:

One may be puzzled how one remembers what is not present, when the affection (pathos) is present and the thing is absent.

(On memory 1 450a25–27)

Now, all that is present is the affection, the state produced by the past perception. But what was perceived then is no longer present, so how can one have knowledge of it? Clearly, there is an assumption at work here about only knowing things that are there or present in some form. In its wider applications, this would need extensive discussion. The relevance to memory seems clear enough, once we have the view that memory is of the past, namely how something past can be (made) present so as to be the subject of cognition. It is gone, and so not available for inspection. It becomes clear later (450b19–20, translated below) that perception is serving as the model for cognition here. You can only taste something present; how then can you remember something not present?

Aristotle also makes an “evident assumption”, as I would like to call it, about what happens in the act of perception:

It is evident that one must conceive of what happens because of perception in the soul and the part of the body holding it to be as follows: the affection (pathos), the possession of which we say is memory, is like a portrait. For the change that takes place imprints as it were a cast (tupos) of the percept (aisthêma), like people who use a seal on a ring.


(p.17) “Tupos” here translated by “cast”15 is often rendered “impression” or “imprint”; and that is fine as long as one understands that as the impression or imprint of a seal. But of course talk of “impressions” overemphasizes Aristotle’s proximity to later theories of perception. But the comparison with a seal (“as it were”) here does not use a technical term. Aristotle is adopting and adapting images of Plato’s (the wax block in the soul Theaetetus 191D–E, and for the picture, the painter, Philebus 39A–D); but they remain images. Now it may be anything but evident that we have to conceive of perception in this way. No doubt the reason that Aristotle finds the assumption evident here is that he is relying on his theory of perception, where he first uses the image of the seal (De anima II 12). At this point, we have done a lot of Aristotelian psychology. But, still, for the outsider, there are attractive aspects to this assumption—that the body is involved as well as the soul, that body and soul in fact are affected, acted on by something being perceived. The change mentioned here is the actual perception taking place. The process appears to be as follows. An actual perception contains a percept—what is perceived; and this is what we are aware of in perception. As a by-product the process produces a “cast” of the percept in body and soul. Now Aristotle appeals to our understanding of what happens when we use a signet ring, as he had done also, to rather different effect in his treatment of the perceptual capacity in De anima II 12. And to conceive of an imprint or cast made by the world on a living body as a picture, along the lines of sealing wax with a signet ring, leaves much unexplained.

Aristotle has little conception of the furious activity animals engage in when they perceive. And the notion of a mental impression is not available to him either: body and soul are affected (On Perception and Perceptibles I 436a6).

Not mentioned here but important is the likeness of the signet ring to its imprint. It is an attractive aspect of signet rings that they were used to authenticate documents, samples, and to guarantee the sealing of store-houses. To work, the imprint as an imprint of this seal or that one, not another one—that is the point of the process; and one reason some seals were so elaborate. So too with the picture or portrait mentioned here. The importance of similarity emerges in a passage in On dreams:

Each of these [phantasmata] is, as has been said, the remains of the actual percept. And it remains when the real (alêthes) thing has gone: it is true to say that the phantasma is like Coriscus, but not Coriscus. When the controlling part, i.e. the one that makes distinctions, is actually perceiving, it does not say [of the phantasma]: Coriscus, but because of the [phantasma] it says Coriscus of the true Coriscus. When it perceives the [phantasma], unless it is entirely obstructed by blood, it says this, because it is moved in perceiving this by the movements in the sense organs, and the similar thing is held to be the true (alêthes) thing. Such is the power of sleep that it makes this go unnoticed.

(On dreams 3 461b25–30)16

(p.18) What is being explained here17 is the mistake Aristotle sees us making regularly in sleep: we mistake the dream, the phantasma of Coriscus, for Coriscus because of the power of sleep. Aristotle has an elaborate story to tell about the way blood moves in sleep to inhibit the workings of perception, and hence perceptual awareness. The heart is where perception is achieved; and here he mentions a critical faculty, that is, one that can make distinctions. In sleep this critical faculty is obstructed by blood concentrating in the heart; but if it is sufficiently unobstructed to form a judgement, to make distinctions it takes the phantasma of Coriscus, remaining from actually seeing (or hearing or touching or smelling) Coriscus, for the real thing. The phantasma is a change or motion remaining in the sense organs from their activity. Here Aristotle contrasts our confusion in dreams with an explanation of why we can make true judgements when waking. In actual perception, our true judgement about the real Coriscus is based on (“because of this”) the phantasma. The phantasma is an explanatory factor in us, which enables us to make the judgement. We are able to make this judgement because of the similarity between percept and phantasma.

How does this compare to memory? Is the mistake in sleep like that in incorrect memory? It appears to be a different one. In sleep, we mistake the phantasma for the real thing. In memory, we make a mistake about what happened in the past, but we do not say that the phantasm is something real. In sleep, the mistake is one due to sleep, and the adverse effects it has on our perceptual judgement. We return, with Aristotle, to mistakes in memory later.

Let us return to memory, with the assurance that similarity is part of the explanation that phantasma play a role in memory.

After the perception, an affection, or state, remains, and possessing this is memory. The work seems to be done by the fact that the state is like a portrait, as it were, a miniature one carries with one. So Aristotle as it were carries a miniature of Coriscus with him. Having the miniature is remembering. So far we would seem to be on the side of the Phenomenalist. After all, you can claim that being a portrait is a phenomenal character of something: it may not be vivacity or familiarity, but it is a question of something appearing to me. What is less clear is how the Phenomenalist replies to the Activist’s point that all the time after seeing Coriscus, you in some sense remember Coriscus.18 Recognizing Coriscus implies memory in some sense. Aristotle knew Coriscus well, and we may ask what about his memory of Coriscus when he is doing something entirely else. Two people are talking about Aristotle in his absence, and one asks the other: does Aristotle remember Coriscus? The answer “yes” is no doubt true; then surely there is no phenomenal appearance of a portrait at all. Aristotle has the miniature in his possession, he remembers, but he does not look at it. So what is phenomenal here? Nothing. Only actually remembering is phenomenal, if at all, surely. Insofar as Aristotle is not always thinking of Coriscus, the possession of the miniature (p.19) is the potential. This point counts in favour of the Activist, against the Phenomenalist. Against our natural inclination to think that any theory which relies on imprints and pictures must be referring to phenomenal characteristics of these, if they make up the ability to remember, and this is, by definition, not an appearance at all, then the theory of memory is not pictorial at all.

One reply available to the Phenomenalist is to say that the Activist is in as bad a state. Aristotle in our example is not doing any remembering at all, but it is still true to say of him that he remembers Coriscus. Perhaps both parties to the debate would do well to agree to confine their dispute to actual remembering. This would be a great weakness. For of course, the ability to remember is just as much a part of an account of memory as actual memory. True, it is very hard to get hold of. But the Activist need not give in so easily. She can say that Aristotle is in a certain state, which is characterized, no doubt vaguely enough, as the presence of a change (kinêsis) resulting from the actual perception.

Although Aristotle has little to say about what happens in a body when we remember, he does think that certain things can be said of the body, if it is to remember at all. A striking feature of memory is that it varies through the human life cycle, as one notices, alas, as one ages. Now, Aristotle has a theory of ageing which relates to his view of the vegetative processes living things initiate and undergo as long as they are alive.19 The young grow, and so are very active in the processing of nutriments; the old shrink, and so are much less active. These vegetative processes may be seen as occupying a position analogous to modern notions of metabolism. Now, Aristotle thinks a mean state of the metabolism is necessary for memory to occur—the very young, the emotional, the very old do not remember: they either undergo too much or too little change (450a32–450b1, cf. Ch. 2 453b4–7), due to the growth or decay they are subject to because of the stage in the life cycle they are at. This is a precondition for, not a class of changes to be identified with, or concurrent with acts of memory. Bodily changes disturb the capacity to remember.

We are still at the beginning of the puzzle: at present we have the affection or state of body and soul being like the cast or imprint of a seal. This comparison is the evident assumption, from the theory of perception. The puzzle needs more development before it can be solved:

So if this kind of thing is what happens in the case of memory, do we remember the affection (pathos) or do we remember the thing from which the affection came about? If we remember the latter, then we would not remember anything absent, if the former [i.e. the affection] how would we, in perceiving this, remember the absent thing which we do not perceive?


The point being made here is that we have this state in our possession, like a miniature, and the question arises how this is going to help us to latch onto the past, in other (p.20) words, the object of the perception, in the way that it was perceived. For if we perceive just the miniature, as it were, then we just perceive the miniature. The question is how does this state here allow us to tell the truth about the past?

If it is like a cast or a drawing in us, why would the perception of this be the memory of another thing, and not of this itself? For someone who is active with their memory considers and perceives this [i.e. the ‘cast’, the ‘picture’]. So how does he remember the thing that is not present? For that would be seeing or hearing what is not present.


Here we see that in remembering we are, in a sense, perceiving. For the whole problem arises because one can only perceive something present (cf. 449b15, an ambiguity is in the Greek as well: present to me, but also present in time). Auditory changes are also considered—‘hearing what is not present’. So too presumably for other modi of perception. And here the idea of a picture or cast loses its immediate appeal. How is one to think of the “cast” or picture of a song?20 We may think of records, or even more modern forms of storage. But without these no doubt very helpful comparisons, how is one to think of the picture of a song in one? Changes that remain in the living thing, maintained by the metabolic changes in the living thing. These changes make up the potential to remember; and are ascertained only through the actualization.

3. The Solution

Aristotle’s solution appeals to a special case of perception, namely perceiving pictures:

In fact, it is possible for this to happen. For just as the animal drawn on a board is both an animal and a likeness: one and the same thing is both, but the being for each is not the same; and it is possible to consider it both as an animal and as a likeness: so too that is how we should understand the phantasma in ourselves—it is both something in itself and of something else.


So we begin from the puzzle of how to perceive something absent. And Aristotle asserts that we do this as a matter of course, namely when we look at a drawing. This familiar occurrence is then explained, and this explanation is applied to the phantasma in us. Thus we arrive at an explanation of memory. Because a phantasma is like a drawing, it too can explain how we can remember something absent, that is, past.

These lines raise a series of questions for the reader. The fundamental question is what is a phantasma? Two answers: (a) an image, (b) a change in body and soul (cf. 450a27–32, above). And the point is that a change in body and soul may be or may serve as an image. And it is this contrast which is pointed to as an explanation for the way memory basically works. The same item has two ways of occurring. Aristotle uses a piece of technical apparatus to make this point. He says that ‘The being for each is different’. (In the Physics, III 3 202a18–21, this way of speaking is exemplified by the (p.21) road from Athens to Thebes. This is the same as the road from Thebes to Athens.) One way of taking this is to say that we are invited to ask: what is it? And you get two answers: an animal, an image. Now in the context of a living thing, one attractive way of enabling some one item to be two things is to allow that it may stand in different relations, more precisely, in functional relations. Thus you look at a drawing either as an animal: then the drawing serves the function of making something present that is not present. Or else it is just a collection of marks on a board, and is only itself present. The contrast is then whether the item is serving a function or not. And what about memory? Let us assume that the person for whom it is possible to consider (theorein) the image is not the theorist, us as philosophers trying to decide what happens when one remembers quite generally, but for example, Aristotle trying to remember if he saw Coriscus (notice that Aristotle says: the phantasma in us—humans, not Aristotelians). Then he may take the phantasma either as serving a function, here: of memory, or else he does not take it this way. It is then merely a phantasma, passing idly past the mind’s eye.

But what we want to know, when confronted by these two ways of taking the remains of the perception, is: what guides this taking? Or more pointedly, what explains taking something in one way and not the other? An obvious answer seems to be the way things appear to you. This is where the Phenomenalist will take her stand. A face appears, and you say: I saw Coriscus, and the face plays, as it were grammatically, the role of “Coriscus”. And what explains you doing this is the appearance. I saw this, Aristotle says to himself, where this refers to the phantasma, taken now as representing something absent. After all, if you rely on something in forming a judgement about the way things were, prima facie, it must be the appearance. For that is what is in front of you, what you have. So we rely on appearances when saying I saw Coriscus.

But that is not the end of the story. For the way things were must be allowed to play a role. As a matter of fact, did Aristotle or did Aristotle not see Coriscus, as, for the moment, Aristotle claims to have done? This question is not decided by the way things appear. By being guided by the way things appear we approach the way things are or were. But what I am doing also depends on the way things were. For if I did not then see Coriscus, then my claim now that I did is not memory. I am being deceived. Now, of course, what is required to decide this question finally, in any given case, are more appearances or perceptions; and of course the appeal to other witnesses is not precluded, nor are other clues in the world. The soul is not an isolated mind, but necessarily in a body. In any given case, I may necessarily be remembering or not, but there is no certainty, as of the time that I am remembering, that I am remembering. We may have to assume that a memory is a memory, and do this non-inferentially, immediately on the basis of an appearance, but this assumption may be right or wrong in any given case. The fact that any case of memory is in this way defeasible, does not mean that, globally, memory is doubtful.

What about the Phenomenalist and the Activist? How does their dispute look now? It seems both have to make concessions, while insisting on their positions. The way the things appear—the phantasma can be misleading you: but you do have to go on (p.22) appearances. But if it is right to say that the phantasma is bound or not into what you are doing, then the Activist is right too. What we do in remembering depends on an appearance, but it is crucial that we are doing something as well.

4. An Explanation of the Solution

As something in itself it is a sight (theorêma) or phantasma, as something of another thing, something like a likeness (eikôn) or a memorial. The result is that whenever its change is active, if it is in itself, and the soul perceives it thus, it appears to occur like a thought or a phantasma. But if [its change] relates to another thing, [the soul] considers it like an image in a picture, without being in the state of seeing Coriscus, as being of Coriscus. Here the affection (pathos) of this considering is different, and when he [? the soul] considers it like a drawn animal: the one occurs in the soul like a thought merely, the other like a memorial, [since it is considered] as in the case of the picture as an image.


We have now almost arrived at the Final Definition of memory; and the Activist and the Phenomenalist have moved close together. In this text we have an explicit contrast, not between items in the soul, but the way items can occur or be taken. On the one hand we have a theorema or a phantasma: something which relates to nothing further. And on the other we have a mnemoneuma, a memorial. So we have Coriscus passing through my heart, or head if you prefer, and then we have Coriscus passing through my head and putting me in mind of Coriscus, say, by saying, I heard Coriscus. And this Coriscus is either taken for or occurs as an appearance or a likeness. If it is a mere appearance then that is phantasia, or its product a phantasma; and if I take it as a likeness, then it serves as a memorial.

Coriscus serves Aristotle, as we remarked at the outset, as an example for an individual. But is identifying an individual something we do merely with appearances? To what extent do all claims about the way things were depend on the correct identification of subjects of predication? These questions go deep into Aristotle’s general philosophy. For perception is closely linked to the perception of individuals,22 and while for some purposes Aristotle restricts perception strictly to perceptible qualities, in fact his full theory of perception allows for a very rich capture of the object of perception. But clearly there is no unique characteristic for an individual available in perception. (Given that on some readings of the Metaphysics he believes in individual (p.23) essences this is a controversial statement. But it should not be controversial that any essence is not simply available to perception.) For us the purpose of this is that memory depends on what perception has captured. Thus if there is no unique bundle of attributes which will necessarily pick out Coriscus, then the resulting memory will always be fallible. Of course, the fact that any particular memory is fallible is not a reason to suspect memory as a whole, as we have already remarked.

If the phantasma serves as a memorial, it may look as though there is no work to be done: we remember Coriscus using a memorial, and that is all. The Activist can reply to this by looking more closely at the final definition. To repeat, memory is defined as ‘the possession of a phantasma, possessed as a likeness (eikôn) of that of which it is the phantasma’ (On memory 1 451b15–16). But what does it mean to take a phantasma as a likeness of that which it is the phantasma of? The problem Aristotle starts from is how something that is absent, the object of the past perception, can be present: all that is present is the phantasma. Most of the explanatory work is done by a comparison with a painting, and with the different ways a painting can be taken. A painting can be taken either as a collection of lines or else as the thing represented. (Say, the ways a picture restorer and a lover might, as such, regard one and the same expanse of coloured canvas.) This suggests that what is doing the work is the phenomenal nature of the painting, but that this phenomenal nature is not entirely passive, but depends on the subject of the memory taking the phantasia in a certain way. This is an active achievement of the person remembering. Here the Activist can rest her case.

Aristotle thinks that, in remembering, a change remaining from perception, an appearance or representation, is taken as a likeness image (eikôn) of the thing remembered. The appearance can occur to one as just that, a face floating through your mind, or else you can take it, or it appears to one, as a portrait of someone. These alternatives—how it appears to you, or the way you take it—are left open by Aristotle. They may be two ways of remembering, or two views of memory. This is the final battleground of the Activist and the Phenomenalist. But as we have seen they can come together in an act of memory. The appearance can serve the work of memory.

Aristotle does not suggest that the phantasma has to fulfil certain conditions, phenomenal conditions, e.g. of likeness, or even vividness, to be able to be taken as an eikôn.

Where the Phenomenalist will say: Aristotle appeals to the ways we can take a portrait, therefore he is thinking of a visual or generally perceptual image, the Activist will say: appealing to a comparison (the phantasma can be used like a portrait) does not imply that the psychic occurrence is, sans phrase, a portrait. No, she will insist, it is like a portrait, namely in a very precise sense: not merely in being like the original (a very weak condition indeed), but in being candidate for two different kinds of attention, one fixing on the material, non-representative aspects, and the other fixing precisely on the picture or phantasma as an eikôn of the original. If, the Activist will insist, the phantasma is anyway pictorial, and being pictorial is a property things can (p.24) have whatever viewers do or do not do or think,23 then Aristotle would not put taking the phantasma in a certain way at the centre of his account.

5. Doubtful Memories

This is the reason we sometimes do not know, when changes of this kind occur in the soul from previous perceptions, whether they occur in accordance with the perceiving, and we are in doubt if it is a memory or not. At times it happens that we realize, i.e. recollect that we heard something previously or saw it. This happens when, while considering it [the phantasma] itself, one changes and considers it in relation to another thing.


The reason for the possibility of doubt: it is possible to take what occurs in the soul in two ways—either to say: that is how it was or to say nothing. This implies that the psychic occurrence can fit the world or not. Remembering is then affirming this relation to the object of memory. When making a memory claim, we assert the way things appear to one. But we can be in doubt because it is unclear whether this is the way it was or not. As he says here, we are unclear if the phantasma accords with the perceiving or not. So one considers a phantasma, first of all just in itself, and then we move to considering it as relating the way things were. What Aristotle does not say is what moves us, or justifies the move, to asserting the way it was. If all that there is to go on is the phantasma, then it must be something about the phantasma which justifies or explains us making the assertion. Supposing that the remembering here is unaided, then the justification of, and what moves us to the assertion, is the phantasma. This is a strong card in the Phenomenalist’s hand.

The Activist can reply that the phantasma is not the only thing we have to go on; for we do something with it, and this activity is one we are aware of. But surely, in itself, being aware of what one is doing does not making this doing track the truth.

The base line that Aristotle starts from are memories where there is no doubt: this is how it was. What he does not discuss is whether we can be in no doubt, and yet nonetheless be wrong. He does think that we can think that we are remembering, and be wrong about it. In contrast if one is (as a matter of fact) actively remembering, then this fact cannot escape one’s notice; the reason for this is just the definition of memory, he says (On memory 2 452b23–28).24 This is an important argument for the Activist: the question (p.25) is what makes a case of memory a case of memory, and the answer is that memory just is (as a matter of fact) bringing together, in an assertion, a phantasm and its original.

The case Aristotle now considers is one where conviction does not arise immediately, but where one comes to realize, that is how it was. This passage also mentions a move from memory to recollection; so a few words may be said on how these two activities are related. Here we move from those cases where memory is immediate, in the sense that no search is necessary, to a search for the right phantasma. In the one case, Aristotle says, I saw Coriscus; in the other case, he searches for the man he saw, because of his uncertainty. If successful, the result of recollection is memory: he says, truly, I saw Coriscus. In a sense Aristotle moves here beyond perceptual memory. For he thinks that recollection is like a kind of calculation (On memory 2 453a9–14). But he thinks that we can recollect perceptions among other things; and of course in recollecting what we are searching through are phantasmata, which are the product of perceptions. So the Activist and the Phenomenalist can argue here too about successful memory.

Aristotle denies animals the ability to recollect; either they remember or they do not. He has been criticized for depriving animals of opinions.25 But he does think that they are guided by knowledge of the world; for perception is a form of knowledge. But they have not the resources of rationality to sift appearances, subject them to criticism and so decide which to follow and which to avoid. For humans, in Aristotle’s view, things are different: we can be in doubt as to whether the present appearance corresponds to the previous perception or not. And then we can search for the right phantasma, and make the true assertion: I saw Coriscus. The path to this true assertion is recollection.

6. Hallucinatory Memories

The opposite also occurs, as it did to Antipheron from Oreus, and to other people who were mad. For they speak of the phantasmata as things that occurred, as though they were remembering. This occurs when someone considers what is not an image (eikôn) as an image.


We know nothing about the unfortunate Antipheron, nor about the kind of “madness” he and the others referred to here suffered from. Nonetheless, there are lessons to be drawn from this passage. Firstly, there was something wrong with these people: their memories were not working properly. Then as now, investigating how things work requires looking at the way things go wrong. But just what was he doing? He “spoke of the phantasmata as things that occurred, as though they were remembering”. This (p.26) suggests the following picture. A phantasma comes into his head (well, for Aristotle the heart is where these things happen), and he says: I saw Coriscus. Well clearly, that thing, the phantasma, is not what happened. No, it is what the phantasma represents that he says happened. “Coriscus was in the agora”, that happened. That a phantasma here is something representing is confirmed by Aristotle’s account of what happens in such cases. For he says that then one is taking what is not a likeness (eikôn) as a picture. So on the one hand, there is a mistake in doing something, but the mistake is connected to what is being dealt with. For, as a matter of fact, the likeness is not a likeness. So the mistake has two components: what Antipheron does, and what the phantasma is.

And Aristotle links this story to our own inability to tell, sometimes, if we are remembering or not. For the mistake Antipheron makes is to take something which is not an image as a picture. Where we have the ability to be right or wrong, Antipheron seems to have suffered from a kind of fixity. He was not able to move from considering a phantasma as a copy to not doing so. He is stuck in his delusion; he has, as it were, an idée fixe. Where we vacillate between taking a phantasma for a likeness and not doing so, he is on the contrary subject to a persistent delusion.

The moral of the story, it seems to me, is that the Activist and the Phenomenalist would do well to join forces. Neither a change in the soul alone, nor my activity alone can make up an act of memory. Something has to be there which I may take as an eikôn of my past perception. In so doing, I may be deceived or not, successfully remember or not.


Bibliography references:

Bloch, David (2007). Aristotle on Memory and Recollection. Text, Translation, and Reception in Western Scholasticism. Leiden: Brill.

Castagnoli, Luca (2018). ‘Is memory of the past?’ In L. Castagnoli and P. Ceccarelli (eds.), Greek Memories: Theories and Practices. Oxford: Oxford University Press.

Caston, Victor (1996). ‘Why Aristotle needs imagination’, Phronesis 41: 20–55.

Eijk, Philip van der (1994). Aristoteles. Werke in deutscher Übersetzung. Begründet von Ernst Grumach. Herausgegeben von Hellmut Flashar. Band 14 Teil III: De insomniis. De divinatione per somnum. Berlin: Akademie Verlag.

Hume, David (1738, 1740). Treatise of Human Nature: Being an Attempt to Introduce the Experimental Method of Reasoning into Moral Subjects. London: John Noon.

James, William (1890). The Principles of Psychology. New York: Henry Holt.

Johansen, T. K. (2012). The Powers of Aristotle’s Soul. Oxford : Oxford University Press.

King, R. A. H. (2001). Aristotle on Life & Death. London: Duckworth.

King, R. A. H. (2004). Aristoteles. Werke in deutscher Übersetzung. Begründet von Ernst Grumach. Herausgegeben von Hellmut Flashar. Band 14 Teil II: De memoria et reminiscentia. Berlin: Akademie Verlag.

King, R. A. H. (2009). Aristotle and Plotinus on Memory. Berlin: De Gruyter.

Lasserre, F. (1987). De Léodamas de Thasos à Philippe d’Oponte. Témoignages et fragments. Naples: Bibliopolis.

(p.27) Pears, David (1991). Hume’s System: An Examination of the First Book of his Treatise. Oxford: Oxford University Press.

Schofield, Malcolm (1992). ‘Aristotle on imagination’. In Martha C. Nussbaum and Amélie O. Rorty (eds.), Essays on Aristotle’s De Anima. Oxford: Oxford University Press, 249–77.

Sorabji, Richard (1995). Animal Minds and Human Morals: The Origins of the Western Debate. Cornell Studies in Classical Philology. Ithaca: Cornell University Press.

Sorabji, Richard (2004). Aristotle on Memory, 2nd edn. with new introduction. London: Duckworth.


My thanks to Fabian Dorsch and Fiona Macpherson for organizing a bracing conference, and to the participants for a spiky discussion. I dedicate this chapter to the memory of Fabian—he is sorely missed, very warmly remembered. For a (German) translation of and commentary on On Memory and Recollection, see King (2004); and for an account of the theory in English and a comparison with Plotinus’ work on memory, King (2009).

(1) Coriscus plays the role of the example for an individual in a variety of Aristotle’s works: On the Generation of Animals 767b25, 768a1; Posterior Analytics 85a24; Soph. Ref. 22 178b39ff; Phys. 219b20; Eudemian Ethics 1240b45. His name has survived largely because of this use; and also because the sixth Platonic epistle is addressed to him, Hermias, and Erastos. The final reason he is known is that his son Neleus was left Theophrastus’ library which contained Aristotle’s works and library as well (Strabo 13.1.54). Aristotle may be remembering Coriscus in the latter’s absence either before Aristotle leaves Athens for Assos and Lesbos around the time of Plato’s death in 348 BC, or else when Aristotle has returned to Athens in 336, leaving Coriscus behind, as far as we know. It is not known what happened to Coriscus after the death of Hermias at the hands of the Persians in 341. See Lasserre (1987).

(2) On these two contrasts, cf. Pears (1991), Ch. 3, ‘Memory’.

(3) Cf. Historia Animalium I 1 488b24–25 with On memory 1 449b28–30, 2 453a7–9.

(4) David Bloch (2007: 72) claims that for Aristotle memory is a passive state which is not discussed in terms of activity and capacity. As he sees, this would make Aristotle’s view of memory very different from ours. His argument is largely based on linguistic considerations; and he has considerable trouble in reinterpreting the phrase “being active with memory” in the Canonical Formula.

(5) Sorabji (2004) rests content with the theoretical distinction between the two.

(6) Respectively, David Hume, Treatise of Human Nature, I.I.3, ‘Of the ideas of the memory and the imagination’; William James, The Principles of Psychology, Ch. 16, ‘Memory’. There is no indication in Aristotle’s ‘physical’ works of the influential account of phantasia as a ‘weak perception’ from the Rhetoric (I 11 1370a28). The Rhetoric is written to appeal to what most people usually think; that is the basis of the speaker’s ability to persuade. The importance of this omission is that there is no indication of an interest in the weakness or strength, or indeed in any qualities of the phantasia.

(7) There are interesting distinctions in English between saying, I remember and I can remember, where even the latter may be used to pick out actual memory.

(8) Sorabji (2004: 13) criticizes Aristotle for this view of memory; cf. Castagnoli (2018). Aristotle will try to accommodate things that are not past, either as accidental memory, or else he will say that when actually using what we have learnt, we are not remembering. Doing maths is not remembering.

(9) See Johansen (2012), Ch. 5.

(10) See Schofield (1992).

(11) Schofield (1992).

(12) Sorabji (2004), ‘Introduction’.

(13) This is one of the rare references in Aristotle to mnemonic techniques; for the reasons that Sorabji is wrong to attach such importance to these techniques in the context of Aristotle’s theory, see King (2009).

(14) Sorabji (2004); Bloch (2007: e.g. 64, 69): the comparison with pictures makes more sense if the images are pictorial.

(15) A suggestion made by Victor Caston (1996).

(16) For similarity as the grounds for some people being easily deceived in sleep, see the start of the whole passage 461b7–11.

(17) On possible interpretations of this tough passage, see van der Eijk (1994: 238–41).

(18) For the capacity to remember, see On memory 2 452a1011.

(19) See King (2001).

(20) On “earworms”, see On memory 2 453a28–29, and Amy Kind, Chapter 11, this volume.

(21) In what follows, I ignore the complications that the introduction of theorêma and noêma, thought causes; these refer to universals, e.g. man. “Thought” means systematic thought, not merely something in the mind in the Cartesian fashion. Because Aristotle thinks that such thoughts, to be thought, anyway require a phantasma (On memory 449b30–450a10), these cases are parasitic on the case of perception, as well as requiring new moves in the theory.

(22) See e.g. Posterior Analytics A 31.

(23) Aristotle refers to mirrors in his account of the genesis of phantasmata (cf. On dreams 459b31); and clearly there are natural images—the investigation of phantasma is a natural investigation, even if it is illuminated by a comparison with the artefact portrait.

(24) 452b23–29. Sorabji (2004: 10, fn. 1) says that 451a23–28 contravenes the rule that if one is actually remembering one must be aware of it. What it says is different: that one can think one is remembering, and be wrong. One’s awareness of remembering can be mistaken; but no actual memory goes unnoticed. What the Activist is claiming is that the reason for this is that remembering is something one does, and so is aware of. Aristotle’s epistemology is built on what is the case (the truth)—here the truths are what memory is, and that one is remembering. As to what it is about actually remembering that makes it necessarily something one is aware of, Aristotle does not say that it is the activity, but there are principles in his general philosophy which would suggest that if he gives the definition of memory as explanation for one’s awareness of genuine remembering then this lies in the fact that activity and definition are identical.

(25) Sorabji (1995).