Creative Recollection: Bergson’s Theory of Memory

The lessons of the above objections for a theory of recognition are twofold. The first is that some other ingredient must be added to the associationist formula to explain why we can have memory-images without recognition. Second, and this will be a point familiar to readers of Merleau-Ponty (and Heidegger), that there is a level of worldly recognition below the strictly conscious or intellectual level of recognition, and it is more fundamental in explaining our engagement with the present and our successful comportment towards our environment. Merleau-Ponty calls what Bergson is calling recognition, ‘motor intentionality.’ In motor intentionality, there is no subject-object structure. Instead, “consciousness is nothing other than the dialectic of milieu and action.”11placeholder It is just the non-representational skilful interface between body and environment. In this mode of engagement, “Consciousness is in the first place not a matter of ‘I think’ but of ‘I can.’”12placeholder I believe that Bergson is saying something similar.

Bergson provides a model that accounts for these problems and explains how recognition comes about by recourse to his distinction between recollective and habit memory. It is not, as the empiricist would have you think, our recollective memories blending with perceptions that do all the heavy lifting in guiding our activity and providing recognition. Rather, the feeling of recognition is explained by a second-order consciousness of our own already habituated body immediately perceiving and navigating the world without the help of memory images. In other words, the feeling of recognition is the consciousness of successful habitual action. The process by which we obtain such a feeling is through repetition. It is impossible to understate just how much of our everyday life is constituted by this kind of habitual recognition. Bergson puts it nicely when he writes that “Our daily life is spent among objects whose very presence invites us to play a part: in this familiarity of their aspect consists.” Thus, these “Motor tendencies would, then, be enough by themselves to give us the feeling of recognition.”13placeholder It is only “because psychology has failed to separate out the motor element in memory, that we have sometimes overlooked and sometimes exaggerated what is automatic in the evocation of [recollective memory].”14placeholder

Providing this distinction avoids both pitfalls of the empiricist theory. First, it means that associative agnosia can be explained as being some defect in the parts of the body responsible for servicing our habits while presumably having perfectly functioning parts of the body that wield recollective memory. This is actually what we find in all cases of agnosia, no matter what sense modality it is: “The deficit cannot be explained by memory, attention, language problems, or unfamiliarity to the stimuli.”15placeholder In no cases have the memories themselves been destroyed. Second, it accounts for our sub-representational being-in-the-world, habit, by identifying recognition with the consciousness of precisely that activity.

However, this is not to say that we can now reduce all human activity to a bundle of habitual actions that bore novel bodily possibilities into our nervous system (though it is certainly a lot of the explanation). In most cases, we still have experience and thus recollective memory too. “[W]hile motor apparatus are built up under the influence of perceptions that are analysed with increasing precision by the body, our past psychical life is there: it survives…with all the detail of its events localized in time.” However, we seldom delve, with much real depth, into the wealth of recollective memory we theoretically have access to. This is because we are evolved creatures of habit who “follow the direction of nature” in the sense outlined above. We are geared towards successful action, and once the majority of our habits have been established, there is very little deep recollection offers us in the way of survival or navigation.

Thus, when the body is in sensorimotor equilibrium, that is, when perception is being competently connected with action through established habits, recollective memory is actively inhibited by the body as it has no use for supplementary memory-images. It is only once there is a “rift between the actual impression and its corresponding movement”16placeholder that the body calls on recollective memory as a guide in the sense hypothesised by the empiricists.17placeholder If we choose to reach for a memory-image buried deep in the past, for no particular reason, it requires a distinctive mental effort that goes against our instinctive presentism – the direction our ordinary perception inclines us.

In sum, we may return to Bergson’s initial thesis: “The recognition of a present object is effected by movements when it proceeds from the object, by representations when it issues from the subject.” As we have seen, this seems right. The feeling of recognition of a present object or environment is, for the most part, our consciousness of our own habits. On the other hand, when memory-images do need to intervene, when recognition issues from the subject, it is the work of memory-images. This section has shown that an empiricist cannot account for recognition purely with a theory of ideas, whereas a mechanist could. The following section will show that a mechanist cannot account for a wholly different feature of human experience, specifically those aspects of life where memory-images do intervene.

Gradual Passage of Recollections into Movements. Recognition and Attention.

In this extended section, Bergson ultimately wishes to argue against the idea that memories are past images of present perceptions stored in the brain or in the mind. Instead, he will argue that perception and memory are one continuous process with no distinct crossing point. Memory is the virtual (rather than actual) phase constitutive of one direction of the whole movement of thought. What this means will come a little later on. I begin where he does. He frames the initial discussion by introducing a new explanandum that will help us understand the role of recollective memory beyond mechanical recognition. This he calls attention. He defines attention as “where memory-images are regularly united with the present perception.”18placeholder In other words, the cases just discussed where your habits are insufficient for action, and your body requires regular supplementation by memory-images. Whether this does happen and how we can best explain it if it does, is the subject of this section.

There are two directions of explanation for this phenomenon. Bergson poses it as follows: “is it the perception which determines mechanically the appearance of the memories, or is it the memories which spontaneously go to meet the perception?”19placeholder The former is both the mechanist and the empiricist’s thesis. Here, someone might suppose that in every perception, there is some disturbance in the nervous system which communicates with the brain, resulting in the evocation of memory-images in the subject. If this were true, we could maintain that memory is merely a function of the brain, that these images are ‘stored in the brain’, and all behaviour can be reduced to habit mechanisms. Alternatively, one might suppose we are a unidirectional string of impressions that appear in a single file, where the present impression is perception, and memory is merely a past copy of a present perception. Here we would say that memories are stored in the mind and that behaviour can be reduced to recollective memory (or ‘habit’ in the Humean sense of an accumulation of impressions). The putative benefit of such approaches is that memories are simply copies of the present. They are the same kind of thing.

There are two problems with these views, as Bergson sees it. First, it fails to explain clinical findings in the study of memory and perception. Both theories fail to explain the existence of agnosia because it entails a separation between types of memory that neither theory has access to, which indicates the need to combine both theories. In order to explain the deficit, the empiricist needs habit memory, while the mechanist needs recollective memory. Second, it is an inadequate explanation of our experience of attention. They must explain how memory comes to bear on the present operation of perception. Before demonstrating these problems with an example, I will describe Bergson’s alternative, and then we can see how it better explains these phenomena.

Bergson wishes to affirm the latter thesis, that memory enters perception in an active and causally efficacious synthesising movement, instead of being something that merely fades away into the past or is evoked due to perception. Here, the brain does not ‘store’ determinate images copied from the past. Instead, it is the mechanism by which the body is co-ordinated into a certain attitude, configuration, or disposition, such that recollective memories can come to insert themselves, in the form of memory-images, into the body and guide its action. This process of attention is a synthesis between memory and perception. “[E]very attentive perception truly involves a reflexion…the projection, outside ourselves, of an actively created image, identical with, or similar to, the object on which it comes to mould itself…It is true we are here dealing with images photographed upon the object itself.”20placeholder Objects meet perception at the same time as memory meets the object. “Memory thus creates anew the present perception; or rather it doubles this perception by reflecting upon it either its own image or some other memory-image of the same kind.”21placeholder On this view, memory and perception are different in kind: one comes forward to meet the object, while the other goes from the object to the body

These images are selected by the body drawing on the totality of stored up recollective memory and wielding them according to similarity and utility. “For while external perception provokes on our part movements which retrace its main lines, our memory directs upon the perception received the memory-images which resemble it and which are already sketched out by the movements themselves.”22placeholder If everything required by the body is not supplied at this moment by our immediate action and our habit memory, “an appeal is made to the deeper and more distant regions of memory,” into our entire past existence, “until other details that are already known23placeholder come to project themselves upon those details that remain unperceived.” Thus, the contribution of memory undergoes a selection process wherein they become active insofar as they can be used for present action. Sometimes, in this process of recollection, the image “brought down is capable of blending so well with the present perception that we cannot say where perception ends or where memory begins.” Memory and the object create a circuit by which each contributes to the actual experience of the individual. In some cases, the memory becomes so narrowed down alongside perception that it becomes almost one with action.

A simple example of this is in reading. We do not read letter by letter, or even really word by word (unless perhaps if we are learning a language). If we are reading a language we are familiar with, “our mind notes here and there a few characteristic lines and fills all the intervals with memory-images which, projected on the paper, take the place of the real printed characters to be mistaken for them…we are constantly creating or reconstructing.”24placeholder Our constant and repeated visual acquaintance with our own language has meant that an ocean of memory stands-reserve ready to be called on in order to smooth over the reading process. This is why a book written in a peculiar dialect or a tome of philosophy full of novel neologisms is so hard to read: our regular reading process skates over the actual writing, filling in both the spelling and the expected succession of words with memory images.25placeholder However, this synthesis goes virtually unnoticed because our memory is so contracted as to be one with perception. Indeed, this state wherein the present perception is impregnated by memory is the ordinary state of affairs in our lived experience: “Perception is never a mere contact of the mind with the object present; it is impregnated with memory-images which complete it as they interpret it.”26placeholder

Since perception is an entirely extended process that unfolds in the physical becoming of the world, when memory is grafted onto perception in this way, it too tends to become an extended thing as it is eventually sublimated in action. This is attention on Bergson’s account. (It is crucial to remember what I mean by this in the following sections as I will often write “attentive experience” or “attention” and mean it in this technical sense, not in the colloquial sense.)

So, does this theory better explain our experience and the supposed constitutive role memory plays in its construction? Let us take a simple example: communication between two speakers. On the empiricist and mechanist views, memories are stored in the mind or brain, respectively. Every word, phrase, and gesture ever encountered leaves a trace in us, perhaps only rough in outline. When someone begins a conversation with us, we hear the words, and they cause the retrieval process to begin, which is perhaps almost instantaneous. By certain relations of association and similarity, we match the sounds with recordings of past perceptions to interpret them.

There are a couple of problems with this that Bergson points out. First, there is a problem of individuation. When we hear someone speak a sentence, do we store the sentence as a finite series of auditory syllables, or each word, or each sentence? Do we also store the sentence according to its pitch, volume, and the particular qualities of the voice? If memories are copies stored, it must be so. In this case, we would have thousands of the same words stored away. Thus, “If the brain chooses one of them, whence comes its preference?”27placeholder Even if we could give an account of which memory is recalled in some particular case, how exactly are the mechanisms of association relating them? Bergson describes this problem as follows:

“how is it that this same word, uttered by a new person, gives a sound which, although different, is still able to rejoin the same memory? For you must bear in mind that this memory is supposed to be an inert and passive thing and consequently incapable of discovering, beneath external differences, an internal similitude. But, for a brain that is supposed – nay, is bound – to record only the materiality of the sounds perceived, there must be, of one and the same word, thousands of distinct images. Uttered by a new voice, it will constitute a new image, which will simply be added to the others.”28placeholder

The key sentence for our purposes is the second. What does Bergson mean when he writes that materialist memory traces cannot have ‘internal similitude’ with one another?

The basic idea is that since memories are copies of the past imprinted in the brain, they are planted there as they are, for as long as the brain maintains them. Thus, how, in this case, do we recollect and select them as similar? Contiguity alone does not logically entail a relationship between such atoms of memory. Appealing to a ‘retrieval process’ to explain memory only pushes the problem back because now you have to explain why the retrieval process can remember similarity. We could likely explain the ‘why’ by appealing to evolution, but that is not an explanation, how. Each memory is ontologically individual and, despite their objective similarity, would leave our brain processes dry. Bergson makes the point as follows:

“But there is something still more perplexing. A word has an individuality for us only from the moment that we have been taught to abstract it. What we first hear are short phrases, not words. A word is always continuous with the other words which accompany it, and takes different aspects according to the cadence and movement of the sentence in which it is set: just as each note of a melody vaguely reflects the whole musical phrase. Suppose, then, that there are indeed model auditory memories, consisting in certain intra-cerebral arrangements, and lying in wait for analogous impressions of sound: these impressions may come, but they will pass unrecognized. How could there be a common measure, how could there be a point of contact, between the dry, inert, isolated image and the living reality of the word organized with the rest of the phrase? I understand clearly enough that beginning of automatic recognition which would consist, as I have said above, in emphasizing inwardly the principal divisions of the sentence that is heard, and so in adopting its movement. But, unless we are to suppose in all men identical voices pronouncing in the same tone the same stereotyped phrases, I fail to see how the words we hear are able to rejoin their images in the brain.”29placeholder

This criticism amounts to the fact that if we treat what is essentially an interpenetrating movement as if it were made up of discrete things, we come out with an underdetermined and paradoxical theory. This may not be convincing now, but it will be pursued at length in the next section regarding the metaphysics of time.

Second, it tells a worse story about our actual experience of communication. Bergson views communication as a paradigmatic case where we can recognise the process of attention, the constitutive movement of memory towards perception. For Bergson, communication is an unbroken and fluid process whereby we begin by “adopting a certain disposition which varies with our interlocutor, with the language he speaks, with the nature of the ideas which he expresses, – and varies, above all, with the general movement of his phrase”. This is the body’s way of preparing itself for the insertion of memory-images to intellectually interpret the meaning of another’s utterances out of a fluid process. At this point, we have already begun to summon auditory memory-images to stand reserve for subsequent action. Following this, we receive an unbroken stream of auditory impressions whose exact character we have never heard before. We experience an utterly new combination of words, syllables, pitch, and volume. Furthermore, at the same time, due to our disposition that covaries with the object like a circuit, our entire past life is “condensed into distinct auditory images, which, still fluid, will be finally solidified as they coalesce with the sounds materially perceived.”30placeholder

In laying distinct memory-images over an unbroken movement, we intellectually reconstruct the meaning of the utterance by reference to our past experiences that have elements in common with the current one. This is another example of attention, of how memory comes to constitute the present. Take another example of someone who has no training in music whatsoever, alongside someone who is a professional classical musician. While they are listening to a symphony, both are enjoying the same stream of stimulus in material reality, but one fuses it with previous experience so that they can intellectually reconstruct it into particular notes and its key, not just into successive phases of sound. Communication runs on the same principle, and Bergson thinks attentive perception is necessary for explaining how we interpret each other out of our flow of experience.

On this account, we can explain why certain memory-images are recalled (and not others). Furthermore, we have an account of how the process of recollection can be simultaneous with perception, rather than one that supposes that memory is only reactive to stimulus, where a process of retrieval must work tirelessly in response to an infinitely varied stimulus. This seems better to describe our experience. We have these explanations on hand because there is a common element to both perception and memory that conducts the relationship between the two. This is the body, which perceives the material world with a view to action and calls on memory to supplement its navigation. Memory-images are not the inert summoning of specific images but rather the intellectual distillation of the past into experience to effect and understand the present.

It is worth noting here that Bergson did not have access to the sophisticated and cutting-edge theories of memory, perception, and the brain we have now. (Though he was engaging heavily with the cutting-edge theories of his time.) So, no doubt, this discussion will not do justice to current sophisticated materialist theories of memory, hence why I keep the discussion short. While I think some of the critiques are still strong, they will not be convincing to the modern reader unless they properly engage with up-to-date science.  Luckily, as we will see in the next section, we have philosophical reason to accept Bergson’s theories, and it is just a bonus that it also happens to offer an elegant explanation for empirical and phenomenological phenomena. Thus far, I have described Bergson’s views and how they might be appealing explanations. Now I intend to marshal a stronger defence of them. Doing so properly requires an excursion away from the explanations of memory given in the text so far and onto more purely metaphysical matters.  Thus, before moving on to the next chapter, I do just this.

The Grounding of Memory in Metaphysics

Let us return again to the example of communication. Suppose you are listening to someone else speak. You must constantly be interpreting what they are saying. Somehow the unbroken stream of sounds produced must be cognitively individuated and understood as meaningful. The empiricist and the mechanist would have you believe that, as you perceive the distinct phrases, memories containing those words (or words similar) are called upon to supply the understanding with their meanings, based on past context, so that you can piece together and derive what is communicated. But is this really what experience tells us? The problem with these interpretations, as already mentioned, is that they make what is essentially an unbroken movement, into static and motionless indivisible occurrences, things. We have the utterance, followed by the present perception of that utterance, followed by the recollection of the idea it evokes. Each of these processes is supposedly a determinate and complete event with solid outlines. There is a serious problem with views such as this.

This is a case of what Bergson calls spatialisation, “the invincible tendency”31placeholder to treat some process in time as if it were space, as if it were a geometrical line segment and thus infinitely divisible. It is the intellect, our capacity for conceptualisation, mistaking real movement for the elements it is decomposed into for practical or scientific reasons. He writes: “The capital error of associationism is that it substitutes for this continuity of becoming, which is the living reality, a discontinuous multiplicity of elements, inert and juxtaposed.” Bergson repeats this criticism a few times in this section of the book, but he does not say exactly what is going wrong, other than insisting that ‘psychological analysis’ insists otherwise. I want to draw out here, more than he does, exactly why this is an error.

Treating our thought as though it were infinitely divisible, like space, rather than as a continuous process, in time, in order to maintain a distinction between past and present, leads to absurd conclusions. I can show this with an argument. In order to individuate the present from other moments, it must be indivisible. This is because the notion of the ‘present’ must pick out something unique and separate from the past and future. If it is divisible, this cannot be done, as it will have past and future parts. But, as Augustine has pointed out, if we take the present to be indivisible, it must also have no duration whatsoever, for the very same reason. If it were temporally extended, it would be further divisible into past and future. Parts would be temporally ahead of the present, and parts would be temporally behind.32placeholder Thus, the present has no length. It is a temporal atom, infinitesimally extended. But if the present has no length, and the empiricist and mechanist believe memories to be copies of the present, stored in the mind or brain, how could there ever be anything beyond the present? No matter how many extensionless ‘presents’ you add together, you never get anything with extension. You cannot derive duration from the duration-less.33placeholder Therefore, if you distinguish between the past and the present, you must think the present is all that exists, which is what Augustine ultimately concludes.34placeholder

However, we do not inhabit some extensionless present. We inhabit a real duration. We can see this for three reasons. First, through introspection. It is impossible to say precisely when we pass from present to past.  When we listen to a pleasant melody accompanying a song, our experience is not of individual notes sounding in some atomistic present. Nor is it of a perpetually perishing infinitesimal present, clearly demarcating itself from the past. These would be vulgar abstractions. Rather, we experience a deep and continuous interpenetration of sounds, some of which, despite their continued presence in this duration, take on a character of ‘being past’ as novel sensation, the following notes, ‘push’ them ‘backwards’ or out of relevance. It is impossible to map out this movement as it is in itself by drawing a sharp line between past and present. All we can say is that there is a succession, and this succession is in duration. Despite this, our tendency to intellectualise and demarcate kicks in. In this experience, we eventually come to label those phases which have the character of being ‘pushed out’, the past. And we label those phases which have the character of novelty, the present. There is no past-present pair in experience, just movement and novelty.

Of course, it could be held that the experience we have of this passing is merely due to some cognitive faculty that can be shown to be in the present. This is Augustine’s answer. He writes:

“It is quite clear that neither future nor past actually exists. Nor is it right to say there are three times, past, present, and future. Perhaps it would be more correct to say: there are three times, a present of things past, a present of things present, a present of things future. For those three exist in the mind, and I find them nowhere else: the present of things past is memory, the present of things present is sight, the present of things future is expectation.”35placeholder

Here both memory and expectation are durationless aspects of the present. Husserl seems to say something similar with his theory of ‘retentions’ (the immediate past), ‘original impressions’ (the present), and ‘protentions’ (the anticipation of the future), which jointly constitute our experience of time-consciousness. This view also seems compatible with presentism about time (though I do not know whether he thinks it is). While these theories are great in their own right, they are also theoretically costly. They abandon any existence beyond the present and I think this is an unnecessary move to make at this point. If our experience is temporally extended and thus an experience of succession, we ought to take it on face value and suppose that this succession is something real. We ought to see if we can save an intelligible notion of this succession, without collapsing into presentism, such that we need not sacrifice its reality. This leads me to my next argument.

Second, for abductive reasons. We would surely like to say that it is not just the present that exists, but also the universe’s history until now. It must exist because it best explains the state of the present and the efficacy of our predictive capacities in charting the direction of the future. If we cut reality off from its past by supposing it has no real existence, its progression into the future becomes random, and the present becomes inscrutable. There can be no explanation for the present (or the future) because there is no past! But as we discovered earlier, if we draw a hard distinction between the past and the present and attempt to affirm both existences by saying one is like the other, we are forced into a presentism where no accumulation of presents would ever amount to anything temporally extended. Thus, the attempt to save the construct by which the present is explicable, the past, exterminates the very possibility of ever explaining it.

Even if we could save the past and the present without collapsing into presentism, we would still get absurd abductive results. As Hume has shown us, there would be no logical or ontological connection between the past and the present once we accept a picture like this. Every event would be “loose and separate.”36placeholder On this view, we can offer no reason whatsoever from the past that something absurd will not happen. For example, we cannot say why apples (and nothing else) will not stop being affected by gravity tomorrow morning. But surely this is wrong. We must affirm that the past explains the present.37placeholder This problem corresponds to Bergson’s accusation that the materialist version of memory cannot actually appeal to relations of association, because the memories are atomistic and thus ontologically separate. To avoid this charge and maintain the explanatory efficacy of our temporally extended models, we must affirm that there is an indivisible duration or movement that passes through real and qualitatively diverse phases. It is these phases that our models are tracking. Further, we must maintain that this movement cannot be rationally demarcated into past and present and that our labels of its phases are grafted onto it after the fact. “Either, then, you must suppose that this universe dies and is born again miraculously at each moment of duration, or you must attribute to it that continuity you deny consciousness, and make of its past a reality which endures and is prolonged into the present.”38placeholder

Third, by a transcendental argument. If we attend to our experience properly, we cannot deny one thing. Namely, that: there is change. When Rowan crosses the road, we can be mistaken, even fundamentally, about whether any thing in our experience maps to things in themselves, but we cannot be mistaken that there is change. And for there to be a change in experience, there must be some corresponding change in reality. This is true even if the things we think are changing are entirely illusory; something is still changing. Even if it merely amounts to a change in our psychological constitution, since we are beings embedded in and amongst the world, there is still change. If we were not embedded in the world, there would be no experience to begin with. In both cases, there still must be change. Further, it must occur over some duration. This is because there must be real phases (that may correspond to something entirely different to what we perceive39placeholder) that bear some qualitative mark of differentiation40placeholder in order to have revealed the change to us. Again, even if these phases are ‘in the head’, there are still phases. If reality is temporally extensionless, this would be impossible because there is no temporal extent to exemplify this change. Thus, both presentism and the ontological distinction between past and present are untenable (as it collapses into presentism) because they cannot explain change.

This transcendental argument also blocks any appeal to eternalism, the view that the past, present, and future all exist simultaneously. This is because, on this view, reality is static and exists all at once. If reality never changes and is given all at once, there can be no qualitative differentiation of phase that corresponds to the change perceived. There is nothing that explains our flow of experience. If the eternalist says that it is just an illusion of the mind, they can only consistently maintain this with their denial of change by supposing that the mind is not in this world because otherwise there would still be a change occurring in the world. Such a view seems to be taken by Kant (depending on whether your reading of Kant countenances two worlds or one), where the ideal world is static, and the empirical world is changing. Since Bergson rejects all views that sunder reality in two, we must suppose that the experience of change in the mind is in this world and that there is change in this world. But for there to be qualitative phases of differentiation to occur reality cannot be given all at once, there must be the constant emergence of novelty. This indivisible duration must be constantly succeeding itself and unrolling into an indeterminate future.

The upshot of this discussion is as follows. First, since there can be no distinction between the past and the present, we must inhabit an indivisible but qualitatively diversified movement, duration. Second, since we are in and amongst the world, and since our experience of change entails the experience of duration, our thought must take this form too. Third, as present perceptions and immediately past perceptions drift into the ‘past’, there is never a threshold they cross that makes them memories rather than immediately past perceptions. There is no difference in kind, only in degree, between a long distant memory and a just perished one. Thus, “thought is a movement”,41placeholder composed of the entire immensity of our past lives all the way up to the present perception. No breaking down of it into parts can ever have a metaphysical warrant, only a practical one. To call some aspect of experience the past, the present, or even an individual memory is merely to imagine some break in the continuity of experience. Finally, this means that memories can never be destroyed because they are merely another aspect of the present, and the present can never be destroyed, since, if anything exists, it is the present. “Memory therefore has no need of explanation. Or, rather, there is no special faculty whose role is to retain quantities of the past in order to pour it into the present. The past preserves itself automatically.”42placeholder Bergson does not, but we might call this view temporal monism (in contrast to temporal atomism). So, how does this view bear on the issues raised above?

The consequences for the theories of memory Bergson has responded to in the book so far is as follows. The mechanistic theory, which reduces all memory to physiological memory and supposes that memories are stored in the brain, and the empiricist theory, which reduces all memories to copies of present impressions, and supposes that memories are stored in the mind, are both false. They are false because they presuppose the present to be a determinate thing that is copied, stored, and then retrieved. As we have seen, this cannot be true as thought is a movement, not a fusion of infinitesimal temporal atoms. Further, if one denies this, we have seen that they must accept presentism, which seems to be false for a few reasons. Therefore, it is impossible that memory could be stored in the brain or the mind, as thought is not made up of things or entities with determinate boundaries. It is a movement; it must be a movement. Memories, or more accurately, memory, has independent existence and is entirely immanent within the present. This is what Bergson calls pure memory. Pure memory is ‘virtual’, which just means that it is accessible to cognition but not active until called upon for representing possible or eventual action. Pure memory makes possible recollective memory. Thus, “our entire personality, with the totality of our recollections, is present, undivided within our actual perception.”43placeholder This is what Bergson means when he writes that “duration is the continuous life of a memory which prolongs the past into the present.”44placeholder

However, this raises a question. If the totality of our life is immanently present in pure memory, in the present, how is it that we cannot choose to experience all of it in the now? This can be answered by suggestions already made. As I said earlier, in attention, memory is called on in a utilitarian way by the body to supplement present perception for action. The reason we cannot access our entire virtual past at once is the same reason we conceptualise and carve up the world perceptually in ways that suppress certain aspects of objects in favour of others. Only some aspects of reality are useful to us.

For example, we do not see the same aspects of a tree that a bird does. Certain features that are unimportant or affectively neutral to us will be constituted as meaningful or with valence for the bird corresponding to that which is useful to it and vice versa. This is true even if there is no doubt we are inhabiting and responding to the same region of reality. We can chalk this up to differences in our biological constitution. Bergson sometimes calls this “attention to life.” The important thing about this process is that it functions to limit experience to those aspects of the entirety of the immanently present duration that are useful to us. (Keep this in mind when, in the rest of the essay, I use the word “contract” to evoke the kind of limiting process undergone in recollection.) The explanatory task of memory is precisely the opposite of what it is ordinarily imagined to be. He sums up this point as follows:

“if we take into consideration the continuity of the inner life and consequently of its indivisibility, we no longer have to explain the preservation of the past, but rather its apparent abolition. We shall no longer have to account for remembering, but for forgetting. The explanation moreover will be found in the structure of the brain. Nature has invented a mechanism for canalizing our attention in the direction of the future, in order to turn it away from the past – I mean of that part of our history which does not concern our present actions, – in order to bring to it at most, in the form of “memories,” one simplification or another of anterior experience, destined to complete the experience of the moment; it is in this that the function of the brain consists.”45placeholder

We are evolved organisms geared in large part for exploiting our changing environment. And our self, which navigates this duration, does so in a dual movement. First, perception “outlines…divisions in the continuity of the extended, simply following the suggestions of our requirement and the needs of practical life.”46placeholder It contracts our experience of the novel aspect of reality, the present, into objects to be skilfully manipulated. However, for Bergson, these objects are ultimately only cognitive abstractions from metaphysical continuity. They have practical but not metaphysical significance. This process is effected not only by our biological constitution but also by our habit memory. Second, memory. “The process by which the virtual image realizes itself is nothing else than the series of stages by which this image gradually obtains from the body useful actions or useful attitudes.”47placeholder Our self distils and synthesises phases of reality previously experienced into images relevant for immediate action. The form recollection distils these phases into is memory-images, which are also cognitive abstractions of the intellect. This process is what I have been calling recollective memory and attention. Once again, Bergson does not, but we can call this a biological theory of knowledge since our experience is synthesised according to biological needs.

Thus, Bergson’s dual biological theory of knowledge explains the two phenomena he set out to explain in this chapter. The first is recognition. This is explained through our direct perception of real objects because for us to begin habituating ourselves to certain elements of our environment (that is, recognise it), we must cognitively individuate reality into manageable parts first. Once we have cultivated our habits that deal with such objects, we need this cognitive individuation less and less. Our action stops being the stringing together of discrete steps and becomes the skilful and unbroken motion we are familiar with in our everyday activity. The second is attention. This is explained by our direct utilisation of the immanently present past through the process of selection and contraction we call recollection.

While these movements technically constitute a difference in kind for Bergson, their function is analogous. Just as we know there are always objects “out there” to perceive, given we are in the right circumstances to do so, memories are also “out there” to be utilised in attentive experience, given we are in the right circumstances to do so. Indeed, to assume the former is obvious while the latter is scandalous is mere prejudice for Bergson. He asks us: “How comes it then that an existence outside [the present perception] appears clear to us in the case of objects, but obscure when we are speaking of the subject?”48placeholder On this picture, the “body, taken at a single moment, is but a conductor interposed between the objects which influence it”49placeholder, whether they be virtual (memory) or real (perception). To speak a little too metaphorically, the latter comes in through the ‘front’ of cognition, in perception, while the former comes in through the ‘back’, in memory. We might represent it as follows:

Pure Memory → Memory-images → Experience ← Perception ← Material World50placeholder

In truth, this only represents our everyday experience as constituted by the useful selection of images from perception and memory. In actuality, the whole movement ought to be represented with bi-directional arrows. I do this for two reasons. First, there is much that dwells within the extremes of pure memory and the material world that is never experienced or utilised but is still ‘out there’, accessible in principle by an effort of mind or a movement. Second, this process is in a constant state of flux. What is perceived and contracted in one moment drifts into bodily inadequacy in the next. Thought is a movement whose content is never determined in any vulgar sense that is incompatible with Bergson’s temporal monism. It is instead a constantly interpenetrating heterogeneity. It would look as follows:

Pure Memory ↔ Memory-images ↔ Perception ↔ Material World

What constitutes experience in any given situation is determined by the bodies’ relative placement in duration and its particular biological ability to orchestrate the synthesis of memory and objects into memory-images and perception. Independent of selection, memory and the material world remain existent. As this diagram shows, the difference in kind comes not from a division between substances but rather a diversity of phases that become a difference in kind at the extremes. This will become important later on. For now, it is only important to note that this structure has polarity but not division.

Thus, our experience combines recognition, through habit memory, and attention, through recollective memory. Both movements are mediated by biological constraints that explain our particular and limited comportment towards the present. Together with a background theory of duration, temporal atomism, these movements jointly explain our experience as Bergson sees it. Thus, with some metaphysical issues cleared up and filled in, we can return to our close reading, starting with the next chapter.

Chapter III: Of the Survival of Images. Memory and Mind.

To sum up what we have established so far, Bergson has described an unbroken movement of thought containing three elements. They are (1) what he calls ‘pure memory’, the pole of thought least useful, in the limit, for present action. (2) memory-images, which are contracted into perception and (3) pure perception, which is the pole of thought that, at its limit, apprehends pure novelty. Taken together, these three elements are not really elements at all, in the sense that they are divided from one another. Instead, they must be thought of as phases of one continuous movement. Thus, Bergson writes:

“We have distinguished three processes, pure memory, memory-image, and perception, of which no one, in fact, occurs apart from the others. Perception is never a mere contact of the mind with the object present; it is impregnated with memory-images which complete it as they interpret it. The memory-image, in its turn, partakes of the ‘pure memory,’ which it begins to materialize, and of the perception in which it tends to embody itself: regarded from the latter point of view, it might be defined as a nascent perception. Lastly, pure memory, though independent in theory, manifests itself as a rule only in the coloured and living image which reveals it. Symbolizing these three terms by the consecutive segments AB, BC, CD, of the same straight line AD, we may say that our thought describes this line in a single movement which goes from A to D, and that it is impossible to say precisely where one of the terms ends and another begins.”51placeholder

A large part of this chapter is spent elucidating and fleshing out the picture given of how experience is constituted and the metaphysical consequences of such a picture. This section is where a couple of Bergson’s famous diagrams make their appearance; most importantly, the “memory cone”, which I will attempt to explain how it is helpful to us in understanding the picture Bergson has been painting.

What the Present Is

This section will describe Bergson’s account of the present, now not merely for one perceiver, but generalised to all reality. Thus far, we have developed a theory of how memory grafts itself out of pure memory and onto the present to supplement action. But, we have yet to say exactly what our present is and how we relate to it (point D on the above diagram). Bergson begins the chapter by answering precisely this question. We will now be moving from the structure of our comportment towards the world, given by recognition and attention, towards the structure of the world that supports such comportment.

Bergson thinks that what we call our present is ‘sensation’ and ‘movement.’ It is a movement because the perpetually perishing or ever-evolving now is always constituted by my action, what I am doing now. The present, for me, is sensorimotor, extended, and unfailingly active. The present is in a state of becoming. This explains the constant emergence of the new in experience and the experiential direction of time as continually moving forwards. It is this flowing mass that our bodies and minds are jointly oriented towards navigating in action. This movement of the now becomes a sensation as it traverses thought towards the immediate past. Indeed, we must experience movement only after it has occurred. Even if the duration elapsed is only minuscule, it takes some finite amount of time to be transformed and processed qua sensation.

This has some immediate plausibility. Our actions at any given moment are on the knifes edge of the present and are pure movement. My typing this sentence right now, for example, is not some abstract impression but is a pure and efficacious single, extended, movement. Immediately following this movement, it drifts into my immediate past and becomes a sensation perceived that guides my following action. I have a sensation of writing the previous sentence that leads me to write this sentence. Sensation prepares me for the future. Since all actions take some time, they all have this temporal structure from movement to sensation. Think about even the most basic action, such as getting up to turn on the light. The moment you decide to get up from your chair to turn on the light must be ‘prolonged into’ the act of flicking the light switch because you must retain, as a sensation, your initial intention so that it can be transformed into the act. In this way, the “psychical state…that I call ‘my present,’ must be both a perception of the immediate past and a determination of the immediate future.”

Further, since the duration of thought must be an undivided whole, as noted in the last section, the sensation is also not an abstract impression but is ontologically continuous with the action. According to temporal monism, there cannot be an ontological gap between temporal stages. Thus the shift from movement to sensation is not an absolute leap from matter to mind but a continuous passage that admits of degrees. By ‘prolonged into’, Bergson merely means ‘used for’ action. The word ‘prolonged’ is supposed to evoke the temporal sense of the act. Since the past is still relevant to present action, it is prolonged into it. In the case of me typing, the sensation of the first sentence was prolonged into the act of writing the second. Or even more simply, the beginning of you reading this sentence is prolonged into you reading the end of it. This is how I can seamlessly pass from typing one word to the next. My sensation of the immediately preceding word allows me to type the next one. The same is true for you reading these very same sentences. However, it must be remembered that these are not themselves individual temporal entities, only phases of an eternal present.

This present takes place in the ‘object’ pole of the structure of thought outlined above – the realm of the material world and perception. This is where our body presides, in an extended world, where our sensations and movements are localised at essentially determined points in our body. From here, the body conducts and calls on the past and its environment to intelligently construct potential paths of action, eventually to be transformed (or prolonged) into movement. However, this can be generalised. Since this body must be a thing among other things, that is, among other bodies, every other extended body is itself in this present. And since the present for me is a constant becoming, it must be for everything else. Our body is merely the part of this world that enables us to feel this flux. This section, the present, “is precisely what we call the material world.”52placeholder Both habit memory and recognition, once developed, can be accounted for merely by the body being in this material world, consisting only of movement and sensation. (However, it must be remembered that both habit memory and recognition presuppose recollective memory to get started.) The best way to imagine this is to imagine an undivided mathematical plane whose insides are in a constant state of flux, of which our body is amongst. This is what we are attuned towards in experience and where our life takes place, where each phase unrolls into the next. This is the plane of action.

The important thing to note, metaphysically speaking, is that the perception phase of experience has no ontological significance in the plane of action. Only the material world phase does. This is because, according to Bergson’s biological theory of knowledge, perception has no speculative interest, only practical. It is the process by which our body carves up this material becoming into fictional loci of possible action, not nature’s (non-existent) joints. True, we did derive movement from perception, but we are not taking any particular percept or object as metaphysically meaningful because that would be an illegitimate division in an undivided duration. Perception only has a metaphysical impact in the sense that such cognitive functionality does actually change the character of the becoming of the world, because it influences action, which is actual.

What The Past Is (and Is Not)

This section will describe Bergson’s account of the past, again, not merely now for one of us, but for all. At the beginning of this essay, we talked about recollective memory and have transitioned into conceiving of ‘pure memory’ as the virtual existence of an individual’s entire past life. But as we just saw, an aspect of perception, its underlying movement, was able to be generalised to the entirety of being, owing to our embeddedness in the world. The same can be said for memory. However, we must be careful here to get Bergson right. Deleuze takes this line of thought as follows:

“It is the same as with perception: just as we perceive things where they are present, in space, we remember where they have passed, in time, and we go out of ourselves just as much in each case. Memory is not in us; it is we who move in a Being-memory, a world-memory.”53placeholder

He is saying that we must give memory the treatment we just gave perception. Just as the personal perception of the present turned out to be just one aspect of an impersonal duration of pure becoming, so must personal memory become the impersonal and publicly accessible duration of “world-memory”, where every past life co-exists, as every present body does. However, I do not think this is exactly right.

This interpretation at once erroneously deflates and reifies personal memory. As I noted earlier, duration endures as the unrolling of the present as a kind of primordial succession. At no point in this process can any phase be rationally demarcated as ‘the past.’ Thus, strictly speaking, there is no past. There are only prior movements, succession. Therefore, when our bodies contract this movement into memory-images to supplement perception, we are in some sense inventing the past. Memory-images are marked with a ‘pastness’ that is absent from duration itself, which is only added once it is called upon for interpreting material becoming. There is no going out of ourselves because we are always already there, in the present. Thus, the Deleuzian interpretation illegitimately makes two movements of one: the present perception, the world-perception, and the past recollection, the world-memory. This is not to say that there is not a grain of truth in the notion of a world-memory, but it must be made precise. The generalisation of memory to all of being is better captured by Bergson’s notion of the unconscious because it explicitly uses his biological theory of knowledge. It shows how memory is contracted and thus how the past is invented out of the present.

For Bergson, consciousness is the images, both real and virtual, actually present in experience. Thus, the unconscious is the memory that remains powerless and unperceived as it is without utility in the material present. It is unconscious (and not non-existent) because even if it were never contracted into a memory-image by a particular being, it always retains the possibility of being recollected due to its enduring existence. The same is true for the tree that falls, which nobody hears. This is pure memory. This memory is “pure from all admixture of sensation, is without attachment to the present, and is consequently unextended.” Memory-images and the act of remembering differ profoundly from pure memory in that their status as images means that they have been brought before consciousness to represent, if not some actual movement undertaken by the body, a possible one. “From the movement that it becomes image, the past leaves the state of pure memory and coincides with a certain part of my present…The image is a present state, and its sole share in the past is the memory whence it arose.”54placeholder Taking this distinction into account, we can get at the true analogy between memory and perception, with regards to generalising the phenomena to all of being.

Before we do that, if Bergson or I ever use ‘the past’, we are referring to the unextended phases of duration that exist without being thematized, pure memory or the unconscious. The past, in this sense, is the history of duration. For each individual, their pure memory is their past. In other words, their place in the history of duration. That is, the place one’s attention to life restricts one to. Thus, recollective memory is the utilisation of this particular past for present action. We do not have access to it as if it were a ‘world-memory’ because we had to have had some actual connection, through the body, with some aspect of this history, as limited by our attention to life, in order to recollect and contract it into memory-images.55placeholder

Thus, the past is at the ‘subject’ pole of the schema given above, the realm of pure memory and memory-images. Further, memory-images are functionally analogous to perception. Both have a purely cognitive meaning as they persist with a view to action. Their objects are carefully selected and picked out of an indivisible duration to organise action and navigate its contours. On the other hand, the unconscious is analogous to the material world we discovered through perception. Just as the material world is a global, extended, becoming that all bodies share in, the unconscious is the global, unextended history of that becoming that everyone shares in, relative to their place in that becoming. The persistence of this history makes recollection possible. Just as the extended material world is, in reality, unthematized by cognition once we peel back the work of perception, so too is all of unconscious existence, once you peel back the memory-images. Thus, the unconscious transcends what is even possible for an individual to recollect, just as the entire material world transcends any individual’s simultaneous comprehension of its entirety. This past can be imagined as an ever-growing cone unrolling above the plane of action. This is the plane of dreams.

Most of the history of becoming is not even accessible in principle with our biological apparatus and relative smallness. For example, the particular thought processes of Tolstoy, while no doubt still existent virtually, are practically inaccessible to any of us finite beings.56placeholder Further, just as the material world was the grounds for the possibility of perception, pure memory is the grounds for the possibility of memory-images. Both of which are processes necessary to explain our actual experience of recognition and attention. In other words, the material world and pure memory, in their unthematized state, are the primitive reality that ground the possibility of, and are presupposed by, both directions of thought discussed in the last section. This brings us to the relationship between these two poles and Bergson’s famous memory cone.

Relation of Past & Present

Now that I have discussed Bergson’s notion of the present and what he means by the past, we can outline the relationship between the two. We have the present perception at one pole, which is a “section of the universal becoming”57placeholder, and we have the past at another pole, which is memory and our entire past lives accumulated behind us. This can be represented by Bergson’s famous ‘memory cone’ diagram:58placeholder

The idea is simple enough, and I think Bergson’s descriptions are as clear as they come:

“If I represent by a cone SAB the totality of the recollections accumulated in my memory, the base AB, situated in the past, remains motionless, while the summit S, which indicates at all times my present, moves forward unceasingly, and unceasingly also touches the moving plane P of my actual representation of the universe. At S the image of the body is concentrated; and, since it belongs to the plane P, this image does but receive and restore actions emanating from all the images of which the plane is composed.”59placeholder

Thus, in this depiction, P represents our own coincidence with the material world, which is in a perpetual state of becoming. The cone AB depicts our unique past connection to the history of becoming. This history is accessible in principle via the body’s call to insert memory images into experience, which represent a virtual array of possible actions. If they are chosen to be acted on, these memory-images are eventually turned into sensation and then action, at point S. Note that only the actual present point S is in contact with material reality: it is our Augustinian present constituted by pure action. However, even this point and our most immediately present past is already part of and continuous with our memory cone, AB.

Now that we can see what the diagram represents, it can also elucidate the connection between habit memory (here ‘bodily memory’) and recollective memory (here ‘true memory’). I quote at length:

“The bodily memory, made up of the sum of the sensori-motor systems organized by habit, is then a quasi-instantaneous memory to which the true memory of the past serves as base. Since they are not two separate things, since the first is only, as we have said, the pointed end, ever moving, inserted by the second in the shifting plane of experience, it is natural that the two functions should lend each other a mutual support. So, on the one hand, the memory of the past offers to the sensori-motor mechanisms all the recollections capable of guiding them in their task and of giving to the motor reaction the direction suggested by the lessons of experience…But, on the other hand, the sensori-motor apparatus furnish to ineffective, that is unconscious, memories, the means of taking on a body, of materializing themselves, in short of becoming present. For, that a recollection should reappear in consciousness, it is necessary that it should descend from the heights of pure memory down to the precise point where action is taking place. In other words, it is from the present that comes the appeal to which memory responds, and it is from the sensori-motor elements of present action that a memory borrows the warmth which gives it life.”60placeholder

With this, we have come to a complete explanation of Bergson’s theory of human memory. He touches on a few other topics in this chapter. However, I will limit myself to a brief discussion of the explanatory work the theory can do.

Memory & Clinical Explanation: Dreams & Near-Death Experiences

An interesting feature of Bergson’s theory advanced so far is that it has a robust explanation of dreams. Dreams are something that are notoriously opaque and tough to explain. Popular theories of dreams include those that posit them as the symbolic expression of unconscious desires, though I find this somewhat unconvincing. We also see functional theories that posit them as some kind of useful process the body undergoes, for example, to improve memory or general cognitive functioning. I find these somewhat more convincing. Even with these explanations, they still seem somewhat mysterious. Whether or not you agree with what Bergson has presented so far, his theory does have a perfect explanation of dreams.

As discussed earlier, everyday experience operates somewhere between the immediate present and pure memory, due to the synthesising movement of images orchestrated by the body. There is something of an ideal dwelling between these poles in any given situation that Bergson calls ‘good sense.’ Here you get a little philosophy of life:

“To live only in the present, to respond to a stimulus by the immediate reaction which prolongs it, is the mark of the lower animals: the man who proceeds in this way is a man of impulse. But he who lives in the past for the mere pleasure of living there, and in whom recollections emerge into the light of consciousness without any advantage for the present situation, is hardly better fitted for action: here we have no man of impulse, but a dreamer. Between these two extremes lies the happy disposition of a memory docile enough to follow with precision all the outlines of the present situation, but energetic enough to resist all other appeal. Good sense, or practical sense, is probably nothing but this.”

Of course, due to our body and its comportment towards the present, there is only so far one could go towards being the dreamer. Indeed, while it exists, almost all of our past is hidden from us because it is inhibited by the needs of present action. Furthermore, in our waking life where our body is geared onto the world and towards skilful engagement, some memory is rightly inaccessible.

Not so when we are asleep. When we sleep, our body forgoes much of its ordinary functioning and thus has no need to call on any specific memory images for present action. This means consciousness, in sleep, can find the strength to cross the inhibiting function of the body, and we can “replace ourselves, so to speak, in the life of dreams.”61placeholder This means that memory and the process of attention, the contraction of memory images, is not constrained by the relative uniformity the novelty of becoming typically offers but is instead free to associate wantonly across one’s entire past life. This explains the strange and contradictory process of non-sequiturs that emerge in dreams, bizarrely constructed out of several disparate elements, derived according to no practical norms of association. In sleep, the body no longer conducts consciousness according to an intelligible order, so it runs amok. So why do we have dreams? In sleep, the body does not limit attentive experience to memory that has practical use.

Bergson also supposes a similar explanation can be given for those who report that their “life flashes before their eyes.” People often think this experience is merely an illusion, a kind of false memory (in the same way they might think dreams are), or something to be outsourced to future science. They do not try to explain why this seems to occur. Before I explain this according to Bergson’s theory, I want to deflate this occult sounding phenomenon in three ways.

First, someone’s ‘whole’ life flashing before their eyes is probably an exaggeration. It is more likely a confused concatenation of disparate memory fragments fusing into a stream that touches on but does not exhaust many different periods of one’s life. Presumably, this explains the ‘whole life’ angle. Second, in light of the first point, the experience does not seem to me to be much stranger than dreaming, other than that we are awake and thus probably experience it more vividly (which also explains the reverence some people hold for these experiences). In both cases, it is plausible that the experience is constructed out of memories and elements of those memories. Third, if one’s scepticism comes from it not occurring in one’s own near-death experience, it is probably because it failed to fulfil the conditions required by Bergson’s theory. One cannot merely think that they could die in some situation. In these cases, one has an unambiguous directive for action, avoiding death, and thus would be quite lucid and squarely placed in the present moment. Instead, they must think that their death is inevitable, that there is no way their body, nor some other force, could save them, for this phenomenon to occur. This is because, in these cases, the body has ceased to curate useful images for eventual action. It has given up.

Thus, those that see their life flash before their eyes have bodies that no longer saw use in the present and let their attentive mind drift into the plane of dreams, rather than constraining it, as it usually would, to the plane of action. Only once they are brought back do they report such a strange phenomenon. And because of our inexhaustible prejudice towards the present, we treat it as inexplicable or false, when in reality, it is not so puzzling. Why does life flash before one’s eyes when they think they will die? The body stops tethering attentive experience to memory that has practical use in the present, so it is released into the depths of one’s own past life with no practical norms of association.

Both phenomena are not only elegantly explained by Bergson’s theory but also count as evidence for one of his essential theses of this chapter: none of our past can ever be destroyed.

Conclusion

I began this essay with an analysis of two kinds of memory pointed out to us by Bergson: habit and recollective. He then showed us of the need to posit two simple processes that explain our worldly comportment corresponding to these types of memory: recognition and attention. This simple structure led us down a rabbit hole that demanded an entire conceptual upheaval, not only of our notion of memory but of time, the history of duration, and how its virtual existence comes to support experience and action. I discussed how the temporal structure necessary to support such processes is a kind temporal monism combined with a biological theory of knowledge. Finally, I discussed some of the conceptual fallout for our ordinary notions of time, and how the theory comes to explain certain strange phenomena. As we can see, Bergson’s analysis of memory results not only in an empirical theory of how a particularly human capacity functions, but demonstrates how this function is implicated in a wider metaphysical picture. By centralizing memory as a metaphysical category, Bergson has begun to unravel some of the problems that have dogged philosophy, but it remains to be shown to what extent he achieves this. In the next issue I’ll show how Bergson’s prioritization of memory in metaphysics makes soluble the thorny problems of dualism, free will, personal identity, and fundamental ontology.

· · ·

Continued in issue #51…