③ when the images brought to mind by memory are so real as to displace the reality this tells us images are not stored or processed in the brain
because in the last entry what I was trying to say was lost in its elaboration, I will try to keep this brief.
2 statements, 1 which requires explanation, 2 which reflects on the other: they are both Bergson's. (As before I will cite them from the online version of Matter and Memory.) He has just slammed the treatment of memory-images and ideas that assigns them to 'centres' in the anatomy of the brain, which is still how memory, consciousness, perception and thinking are treated. (51) He says it's nothing but the association theory of mind, that it comes from the metaphysical presupposition that all progress can be cut up into phases, and that afterwards they can be made concrete things. I would only add, we tend to stop short of the things, to stop at the phases. Bergson:
... it has neither the advantage of following the movement of consciousness nor that of simplifying the explanation of the facts. (Ibid.)
He advises that we follow this illusion up to the point where its contradiction becomes obvious. To which:
We have said that ideas—pure recollections summoned from the depths of memory—develop into memory-images more and more capable of inserting themselves into the motor diagram. (Ibid.)
Although suggesting a kind of editing function, like for example the flashback, I take him to mean that these images more and more resemble perceptions. They become, he says, more complete, more concrete, more like conscious representations. To the degree they do we can confuse them with the particular image which we perceive. "Therefore," he starts,—
... there is not, there cannot be in the brain a region in which memories congeal and accumulate. (Ibid.)
The emphasis is mine, because the answering question is wherefore? How does this statement reach the point of self-contradiction he alerts us to at the start? Why is it that when memory-images, ideas solidify for us to the degree they can be confused with outward perceptions, the metaphysical presupposition we can cut them up into phases and separate them into processes, assigning them to centres in the brain, falls apart?
He continues by returning to the topic of brain-injuries, stating that any alleged destruction of memories resulting therefrom is but a break in the continuous process by which they actualise themselves. This presents itself as a promising lead: we might recall from Deleuze the distinction between virtual and actual and that between real and the possible, for which, in his book Bergsonism, 1988, he credits Bergson. However this is a philosophical lead. It adds nothing here but more to explain.
Another lead suggests itself if we split our focus between Deleuze's juxtaposition of Bergson and cinema and the terms Bergson uses, images in conjunction with memory, but also the term continuity, which has a particular technical sense in view of cinema. The proof Bergson gives us of language-processing not being centred in the brain, or in areas of it, for example Broca's area and Wernicke's area (both of which 'found' well before Bergson was writing, by Broca in 1861 and by Wernicke in 1874), is then that memory-images can take on the appearance of lived experience as if they were films. Deleuze himself supports this notion. His third commentary on Bergson, "From Recollection to Dreams," in Cinema 2, 1985, is all about memory-images, to the point of including in a note Bergson's diagramme from Matter and Memory showing the circuits of perception:
The perception invoked by Deleuze is however cinema's, as are the memory-images, and if the circuits belong to cinema that does nothing to dispel the illusion of language-centres in the brain.
In fact the diagramme looks like a brain and the flattened curve at 0, representing the degree zero of perception, like a screen. From it we may infer that it is not in the centres of the brain we find images, which include memories and perceptions, since both are forms of images, but that the brain itself is like a cinema. The simile is so pervasive as to have gone from being a commonplace, to become a shorthand, despite cases of aphantasia, for saying what the brain is doing.
Why is it decisive for Bergson when the images brought to mind by memory are so real for us they displace the reality that this tells us images are not stored or processed in the brain? This seems so totally wrong to us. We would rather have it be that our brains show us an illusion than have those images, which for Bergson are primarily auditory, put on the outside.
We might compare this to the attitude to those with aphantasia, who don't 'see' images in their heads. Bergson notes that neither does such psychic blindness have anything to do with actual blindness nor does psychic deafness, resulting from aphasia, have anything to do with actual deafness. It's easier for us to believe the latter, despite the primacy of language in cognitive function, than the former: psychic blindness raises many more questions, often to do with cognitive function, for example, how do you recognise images if you do not see them internally?
For Bergson this is easily explained. Both the memory-image and the perceptual image are outside the brain and one is the outgoing answer to the other's question. The memory-image is the issue of an action of the nervous system. We can note here that any perceived image can elicit any number of respondents. The brain's role is to choose amongst them. Now this we know but it is somehow subsumed under the illusion that goes out to meet that illusion with which Bergson is concerned.
Bergson's diagramme that Deleuze uses for cinema shows us successive circuits of memory-images in their response to an image. It shows as well a deepening understanding of the perceived image, the loops below the 'screen.' The nerves act in response to it and it is their 'actions' which are successively chosen in the interval between perception and recognition. We know it from children learning how to read.
When we might say we see their brains working it is more like a sifting process. Each image is first placed in contention for being the correct reading or interpretation of the perceived image and when a word is recognised its meaning is further and further refined as a deeper understanding is gained. The memory overlay in this case doesn't falsify the image but is in reciprocal relation with it, so that it's less for whether the response chosen fits the case given than to reach out from the point of view first established, with each circuit of attentive reading widening the compass of meaning, and also refining it, to, as we might call it, the direction it indicates.
If the image were somehow locally stored—which is the illusion Bergson says is pushed to self-contradiction— . . . and I should thank Jon Hickman. Partway through the paragraphs above I received a cannot connect with the server message. To no avail, I checked the connection and tried to restore the window. I went into drafts on Ghost, the platform hosting this website, where I have made a practice of writing, relying on its autosave function. The most recent draft was from yesterday. Starting to panic, I went through the browser history and looked for anywhere the autosave might have misdirected this entry, without luck. Finally, in a panic, I wrote to Ghost, Jon answered and suggested a restore url, which, he said, would only work if the material was stored locally, on the machine I am using. There it was, as good as current—isn't this also how the brain does it? No, but it supports the illusion that it is. . . each new image would displace the last. And each time the reader confronted the same word, it would not be a memory-image as such, but the referent image, held up for its match, against the afferent image, would increase in number exponentially, through successive circuits of suggested responses, and lead to an equal increase in processing time. If the processing time were to be kept down, the resources, including energy expenditure, would undergo a proportionate increase. And this is in fact what we are seeing with LLMs, an energy expenditure disproportionate to the difficulty of the exercise.
Bergson's reflection, statement 2, comes in a footnote, where he states that no trace of an image can remain in the substance of the brain . . .
Instead, there are merely, in that substance, organs of virtual perception, influenced by the intention of the memory, as there are at the periphery organs of real perception, influenced by the action of the object. (52n77)
. . .
