Hallucination as a Memory of the Present
the third post in a triptych on hallucination
The first post of this triptych discussed some philosophical approaches to hallucination, and the second post discussed some recent psychiatric approaches, especially calls by (mainly) phenomenological psychiatrists to reconsider the ‘perceptual quality’ of hallucinations (with a focus on auditory verbal hallucination).
And yet, many of us do have ‘perception-like’ hallucinatory experiences. Perhaps the lesson is that there might be some ‘essence’ of hallucination that is found in both the overtly perceptual cases and in other cases. Perhaps this ‘essence’ may be captured by the idea that hallucination is a ‘memory of the present.’ Two thoughts lie behind this proposal: (1) hallucinations essentially bear some kind of ‘sense of reality’ for the one who experiences the hallucination (and being ‘perceptual-like’ is one way of ‘seeming real’); and (2) the ‘sense of reality’ associated with hallucination is very similar to the ‘sense of truth’ typically associated with memory. (The idea is not that hallucinations are memories. A version of that idea has been proposed (Flavie et al 2006)—consideration of its merits and flaws is for another time.)
Let’s start with memory. Based on experiments performed in the 1910s and 1920s, the Cambridge psychologist Frederic Bartlett published Remembering in 1932, which proposed a theory of memory that, in outline, is still championed by some theorists (notably Schacter 2021, 2022). One of the book’s more striking experiments involves having Cambridge students read a short indigenous Canadian story, “War of the Ghosts”.[1] At various times after reading—from several minutes to several years—they were asked to recount it. The original story would have felt quite strange and disjointed to a typical western mind. In recalling the story, subjects tended to modify it in the direction of ‘making more sense’ (to a 1920s Cantab): “The net result,” writes Bartlett (1932, 86), “is that before long the story tends to be robbed of all its surprising, jerky and apparently inconsequential form, and reduced to an orderly narration.” ‘Canoes’ become ‘boats.’ The two main characters, whose relationship is unspecified in the story, become brothers. The departure of one of them—unexplained in the story—gets an explanation (in one retelling he was returning home to see family). And so on. The net effect of such modifications, says Bartlett, is “to make the whole incident ordinary and rational” (ibid., 127).
A theory of memory emerges from these observations. Bartlett founds the theory on the idea of a ‘schema’, introduced by Head and Holmes (1911, 187), for whom a schema is, roughly, a mental representation of one’s bodily configuration, in terms of which one understands present and immediately future bodily configurations, reacts to current perceptions, and predicts future experiences. The schema is updated in response to perceptual (especially proprioceptive) experience, and the updated schema further guides both action and perception, which further inform the schema, and so on. The notion was originally introduced to explain, among other things, the perception of phantom limbs and similar phenomena. In addition to its role in explaining these notable phenomena, the notion of a schema is also supposed to be explanatory in utterly mundane cases of bodily perception and control.
In Bartlett’s hands, a schema becomes more general, no longer about bodily configurations alone, but about coordinated sequences of actions and our experience of them: “…an active organization of past reactions, or of past experiences, which must always be supposed to be operating in any well-adapted organic response” (1932, 200). These experiences “operate, not simply as individual members coming one after another, but as a unitary mass” (1932, 200-201).
As a simple example, compare how an experienced versus new driver might navigate a traffic circle—the experienced driver invokes an existing schema, while the new driver reacts individually and serially to each aspect of the situation. A similar example (on my mind at the moment!): I recently was taken mountain biking, and while she was flowing easily down the trail, I was seeing every root and rock, and reacting to each one (fingers resolutely planted on the brakes). She has schemata for this activity. I lack them. (This situation will change! I am currently shopping for a mountain bike…)
Based on these ideas, Bartlett offers a constructive theory of memory: “The first notion to get rid of is that memory is primarily or literally reduplicative, or reproductive” (1932, 204). Instead, proposes Bartlett, memories are constructed under the guidance of schemata. In the simplest cases, memory follows a schema—in recalling what you did this morning, you may invoke a schema concerned with the ‘usual things that I do in the morning’. Other times, one does not simply ‘fill in’ a fixed schema with specific content (the time that I awoke, the cereal that I ate). Rather, just as a schema in Head and Holmes’s sense may both guide action and be continually updated by perceptual input, so also, as the schema in Bartlett’s sense guides the construction of memory, it may also be modified, or updated, or replaced with a ‘nearby’ schema: “I awoke at 8:00am; that would be too late for the usual cereal…”. And thus a modified or perhaps ‘nearby’ version of the schema comes into play to guide further construction of memory. Thus there may be a kind of feedback loop between the schema and the memory, in which the former guides the construction of the latter, which may further modify the schema (or evoke a new one), and so on.
This role of schemata in the construction of memory is, on Bartlett’s account, a ‘feature not a bug’:
Remembering is not the re-excitation of innumerable fixed, lifeless and fragmentary traces. It is an imaginative reconstruction, or construction, built out of the relation of our attitude towards a whole active mass of organised past reactions or experience [a schema], and to a little outstanding detail which commonly appears in image or in language form. It is thus hardly ever really exact, even in the most rudimentary cases of rote recapitulation, and it is not at all important that it should be so. (1932, 213)
So a memory is (at least in part) a construction of a meaningful representation, and it is supposed to involve a kind of meaning-making, the construction of a ‘story’ that comports with patterns of experience (schemata) and, in virtue of doing so, allows one to make sense of the world. Moreover, memories are experienced with a sense of truth (‘reality’) in part to the extent that this construction is successful, to the extent that the constructed stories really do ‘make sense of’ the world.
The speculative idea, here, is that hallucination may also be a kind of constructive meaning-making, except what is constructed in this case is a present, not a past, experience. Moreover, the sense of reality of this experience is influenced by the sense of meaningfulness in more or less the same manner as the sense of truth for memory. In other words, meaningfulness, in the sense of assimilation to a schema, may be an important component not only of the sense of truth of a memory, but also of the sense of reality of a perception-like experience (hallucinatory or not).
The idea that hallucinations are tied to meaning is old (e.g., Miller et al. 1993). The point here is to bolster this general idea by suggesting that there is a strong analogy between memory and hallucination, in two ways. First, some of the same constructive mechanisms—the kinds of mechanism suggested by Bartlett’s account of memory and its successors—may be at work in both cases. Second, the way that sense of reality is generated for hallucinations may be very similar to the way sense of truth is generated for memory—meaningfulness may be central in both cases.
The general idea is illustrated nicely (albeit partially) in an experimental protocol developed by Bernasconi et al. (2022) to induce hallucination in typically non-hallucinating subjects.[2] While they do not explicate the protocol in these terms, we may understand it as relying on something like Bartlett’s schemata. The schema in question is something like this: experiencing human-like physical contact that is not self-caused is coincident with other forms of interaction with another person. In (one version of) their protocol, the participant pokes a lever and feels a corresponding poke on the back. Subjects report the experience as poking themselves in the back (because, one might speculate, the pattern of action and sensation matches with relevant schemata). A temporal delay (between lever presses and pokes) is introduced, and gradually the experience ceases to match the schema “I am touching myself” and begins to feel as if something is wrong, and eventually as if the poke is coming from somebody else. The world is not behaving as it should. Experimenters then introduce background noise, and it turns out that participants are highly susceptible to developing auditory hallucinations of voices under these circumstances (Orepic et al. 2023). One’s schemata ‘fills in the gaps’ in perception. The auditory hallucination is a way of making sense of the overall situation.
There is much more to say about this case (even though it is itself, as lab experiments tend to be, a severe simplification of the real-world phenomenon of hallucination). Perhaps most notably, the experimental protocol also disturbs a subject’s ‘sense of self’, because of the confusing shift from ‘I am poking myself’ to ‘somebody is poking me’. Setting all of that complexity aside (for now), we are left with the idea that hallucination may be a kind of meaning-making, a kind of ‘memory of the present’. And if hallucination is meaning-making, then it makes sense that just as ‘meaningful memories’ may seem true in spite of the facts, so also hallucination may ‘seem real’. And, for some of us at least, ‘seeming real’ apparently manifests as ‘being perception-like’.
Notes
1. The story came to the awareness of western scholars via the (late 19th century) Canadian anthropologist John George Bourinot, who apparently heard and transcribed it from indigenous story tellers in eastern Canada.
2. I learned about this protocol from Olaf Blanke (who helped to develop and test it), so thanks to Olaf for bringing it to my attention.
References
Bartlett, F. 1932. Remembering. London: Cambridge University Press.
Bernasconi, F., E. Blondiaux, G. Rognini, H. Dhanis, L. Jenni, J. Potheegadoo, M. Hara, and O. Blanke 2022. Neuroscience Robotics for Controlled Induction and Real-Time Assessment of Hallucinations. Nature Protocols 17:2966–2989.
Head, H., and G. Holmes. 1911. Sensory Disturbances from Cerebral Lesions. Brain 34:102–254.
Larøi, F., T. Luhrmann, V. Bell, W. Christian, S. Deshpande, C. Fernyhough, J. Jenkins, and A. Woods.
2014. Culture and Hallucinations: Overview and Future Directions. Schizophrenia Bulletin 40(4):S213–S220.
Orepic P., F. Bernasconi, M. Faggella, N. Faivre, and O. Blanke 2023. Robotically-Induced Auditory-Verbal Hallucinations: Combining Self-Monitoring and Strong Perceptual Priors. Psychological Medicine 54:569–581.
Schacter, D. 2021. The Seven Sins of Memory: How the Mind Forgets and Remembers. Houghton Mifflin Harcourt.
Schacter, D. 2022. The Seven Sins of Memory: An Update. Memory 30:37–42.
Waters, F., J. Badcock, P. Michie and M. Maybery 2006. Auditory Hallucinations in Schizophrenia: Intrusive Thoughts and Forgotten Memories. Cognitive Neuropsychiatry 11(1):65–83.