Barzykowski and Moulin present a compelling case for why déjà vu could be considered a special case of involuntary autobiographical memory, more specifically one in which retrieval was initiated but never completed. While I find the arguments and evidence convincing, I cannot let go of the “why.” Why take a conceptually and phenomenologically discrete concept such as déjà vu and redefine it as failed retrieval? I believe the authors run the risk of introducing problems that did not exist before, and it is not immediately obvious to me what the gains are.
Considering déjà vu and autobiographical memory on a continuum assumes that one can become the other with some additional components or higher intensity of activation. In line with this, the authors argue that additional activation of traces (bottom–up) or additional elaboration of details (top–down) can transform déjà vu into autobiographical memory proper. This premise begs the very intriguing question of when a memory begins and when it is completed, and what types of phenomena should be included at each end of this theoretical line.
The sensation of déjà vu may at first feel close enough to a memory that the individual initiates a deliberate search, only to decide there was no memory after all – it was a trick of the mind. An auditory hallucination could be considered analogous: An individual hears their own name and searches their apartment for the source, only to discover that no one is around. However, no amount of additional search or neurophysiological activation will lead to a source, because there is none. I believe this is also true for déjà vu at a conceptual and phenomenological level. It might also be true at a neurophysiological level, but that is an empirical question that might never be solved.
An autobiographical memory is the conscious recollection of a past event. This is the generally agreed upon definition referred to by the authors on page 6. Activation, or sensation, or any process that does not lead to recollection of a past event is not an autobiographical memory by this definition. Electronic stimulation of the temporal lobe is not autobiographical memory, and a focal temporal-lobe epileptic seizure is not an autobiographical memory. Why then award this special status to déjà vu? On the one hand, the authors base their argument on the shared neurophysiological correlates of both déjà vu and involuntary autobiographical memory, but on the other hand, they explicitly state that déjà vu and autobiographical memory branch away from each other as a result of a conscious evaluation of plausibility (target article, sect. 3, para. 1). On the one hand, déjà vu can be considered part of autobiographical memory because of the shared neurophysiological correlates, but at the same time, the conscious evaluation of plausibility is essential to the phenomenon – and the evaluation is that déjà vu is emphatically not a memory. In this way, déjà vu comes to present a problem for the (conventional) definition of autobiographical memory, if we want to include it under this definition.
Phenomenologically speaking, déjà vu is a satisfactory explanation for a distinct experience, analogous to auditory hallucination being a satisfactory explanation for hearing your cell phone “ding,” when there is in fact no message. When the individual says, “oh, it wasn't a memory, it was just déjà vu,” they are correct in an ontological sense. To say they actually experienced failed retrieval of an autobiographical memory appears to introduce an unnecessary extra step to this experience and to be at odds with the phenomenology of the experience.
Déjà vu, similarly to other tricks of the mind, could arise from a Bayesian principle in consciousness (Ramachandran, Reference Ramachandran1998): If the current situation shares a critical amount of features with a stored configuration of features, the most likely explanation is that the situation is familiar rather than occurring completely by chance. However, the stored configuration of features could refer to one or more specific past events, some fragments of multiple events, repeated events, or potentially nothing relevant at all. Importantly, there is no way of knowing (the hard problem of consciousness) and besides, the whole thing is satisfactorily resolved as the error it most likely is: A déjà vu.
I would argue that the relationship between déjà vu and autobiographical memory is not continuous, but more akin to a path diagram, where the starting points might be overlapping (pattern matching in the hippocampus), the next step might be overlapping in some situations (a sense of familiarity), but when this activation reaches consciousness, the paths diverge. At this point, one path leads to not-memory (déjà vu), while another path leads to memory (autobiographical memory). Redefining déjà vu as failed retrieval conjures up the ghost of a hypothetical subconscious memory, which introduces the hard problem of consciousness. What problem does it solve?
You have
Access
Barzykowski and Moulin present a compelling case for why déjà vu could be considered a special case of involuntary autobiographical memory, more specifically one in which retrieval was initiated but never completed. While I find the arguments and evidence convincing, I cannot let go of the “why.” Why take a conceptually and phenomenologically discrete concept such as déjà vu and redefine it as failed retrieval? I believe the authors run the risk of introducing problems that did not exist before, and it is not immediately obvious to me what the gains are.
Considering déjà vu and autobiographical memory on a continuum assumes that one can become the other with some additional components or higher intensity of activation. In line with this, the authors argue that additional activation of traces (bottom–up) or additional elaboration of details (top–down) can transform déjà vu into autobiographical memory proper. This premise begs the very intriguing question of when a memory begins and when it is completed, and what types of phenomena should be included at each end of this theoretical line.
The sensation of déjà vu may at first feel close enough to a memory that the individual initiates a deliberate search, only to decide there was no memory after all – it was a trick of the mind. An auditory hallucination could be considered analogous: An individual hears their own name and searches their apartment for the source, only to discover that no one is around. However, no amount of additional search or neurophysiological activation will lead to a source, because there is none. I believe this is also true for déjà vu at a conceptual and phenomenological level. It might also be true at a neurophysiological level, but that is an empirical question that might never be solved.
An autobiographical memory is the conscious recollection of a past event. This is the generally agreed upon definition referred to by the authors on page 6. Activation, or sensation, or any process that does not lead to recollection of a past event is not an autobiographical memory by this definition. Electronic stimulation of the temporal lobe is not autobiographical memory, and a focal temporal-lobe epileptic seizure is not an autobiographical memory. Why then award this special status to déjà vu? On the one hand, the authors base their argument on the shared neurophysiological correlates of both déjà vu and involuntary autobiographical memory, but on the other hand, they explicitly state that déjà vu and autobiographical memory branch away from each other as a result of a conscious evaluation of plausibility (target article, sect. 3, para. 1). On the one hand, déjà vu can be considered part of autobiographical memory because of the shared neurophysiological correlates, but at the same time, the conscious evaluation of plausibility is essential to the phenomenon – and the evaluation is that déjà vu is emphatically not a memory. In this way, déjà vu comes to present a problem for the (conventional) definition of autobiographical memory, if we want to include it under this definition.
Phenomenologically speaking, déjà vu is a satisfactory explanation for a distinct experience, analogous to auditory hallucination being a satisfactory explanation for hearing your cell phone “ding,” when there is in fact no message. When the individual says, “oh, it wasn't a memory, it was just déjà vu,” they are correct in an ontological sense. To say they actually experienced failed retrieval of an autobiographical memory appears to introduce an unnecessary extra step to this experience and to be at odds with the phenomenology of the experience.
Déjà vu, similarly to other tricks of the mind, could arise from a Bayesian principle in consciousness (Ramachandran, Reference Ramachandran1998): If the current situation shares a critical amount of features with a stored configuration of features, the most likely explanation is that the situation is familiar rather than occurring completely by chance. However, the stored configuration of features could refer to one or more specific past events, some fragments of multiple events, repeated events, or potentially nothing relevant at all. Importantly, there is no way of knowing (the hard problem of consciousness) and besides, the whole thing is satisfactorily resolved as the error it most likely is: A déjà vu.
I would argue that the relationship between déjà vu and autobiographical memory is not continuous, but more akin to a path diagram, where the starting points might be overlapping (pattern matching in the hippocampus), the next step might be overlapping in some situations (a sense of familiarity), but when this activation reaches consciousness, the paths diverge. At this point, one path leads to not-memory (déjà vu), while another path leads to memory (autobiographical memory). Redefining déjà vu as failed retrieval conjures up the ghost of a hypothetical subconscious memory, which introduces the hard problem of consciousness. What problem does it solve?
Financial support
This commentary received no specific grant from any funding agency, commercial, or not-for-profit sectors.
Competing interest
None.