Aiming to provide a means of integrating research on the two phenomena, Barzykowski and Moulin (B&M) propose an approach to involuntary autobiographical memory and déjà vu as natural products of memory retrieval. Their proposal must, if it is to achieve this aim, have a level of detail sufficient to account for the core functional and phenomenological properties of the target phenomena. Focusing on déjà vu, this commentary shows that the proposal does not have the requisite level of detail. We argue, after identifying three core properties of déjà vu experiences, that each of these leads to a problem that the proposed account, in its current form, is unable to resolve.
The first property that we will consider is strangeness. Déjà vu experiences are regularly characterized as strange, weird, or even eerie by both subjects and theorists (e.g., Brázdil et al., Reference Brázdil, Mareček, Urbánek, Kašpárek, Mikl, Rektor and Zeman2012). In the kind of experience on which we will focus here – sometimes referred to as “déjà vécu” (O'Connor, Lever, & Moulin, Reference O'Connor, Lever and Moulin2010) – the strangeness of the experience is due to the fact that it concerns a singular event and not merely a repeatable item, “as if time had slipped a cog and were now repeating itself” (Woodworth, Reference Woodworth1940: 357). The second property is infrequency. While most people experience déjà vu, they do so very rarely – on the order of few times a year (Brown, Reference Brown2003). The final property is suddenness. Triggered by a variety of situational factors, déjà vu experiences typically begin abruptly, last no longer than a few seconds, and end just as abruptly as they began (Brown, Porter, & Nix, Reference Brown, Porter and Nix1994).
Consider strangeness first. According to B&M's proposal, the distinctive phenomenology of déjà vu results from the combination of incomplete memory retrieval, which produces a sense of familiarity, with metacognitive appraisal of that familiarity as implausible. The strangeness of déjà vu, however, is due not merely to the implausibility of familiarity with a repeatable item – that is, with a type – but also, in cases in which déjà vu amounts to déjà vécu, to the felt impossibility of familiarity with the currently experienced token event. In order to account for the strangeness of déjà vu, then, the proposal needs to explain not only why incomplete retrieval produces a sense of familiarity with an item (a feeling that the current event resembles something experienced in the past) but also why, at least in some paradigmatic cases of déjà vu, unexpected familiarity is not simply brushed off as arising from limited access but is instead taken to indicate event repetition (a feeling that this very event has already been experienced). If the distinctive phenomenology of déjà vu results from a metacognitive assessment of plausibility, in short, then what should be assessed is not the plausibility of event resemblance but of event repetition.
For a similar reason, B&M's proposal cannot account for the infrequency of déjà vu. According to the proposal, incomplete retrieval – producing familiarity but not specific memory content – should be relatively frequent. This is required to explain the relative frequency of tip-of-the-tongue experiences (Brown, Reference Brown1991). Now, if our argument above is on the right track, the metacognitive assessment of plausibility responsible for generating déjà vu pertains to event repetition, not to mere event resemblance. Since event repetition is impossible, however, familiarity unaccompanied by specific memory content should typically, if not invariably, be assessed as impossible or at least highly implausible. But this entails that déjà vu experiences should be relatively frequent, which strongly contradicts the available data. Thus, if it is to account for both the strangeness and the infrequency of déjà vu, the proposal will have to invoke a wider range of (meta)cognitive processes. We will not attempt to determine here with it is feasible for a modified form of the proposal to do so.
The suddenness of déjà vu also presents a problem for the proposal. B&M contrast the rapidity of content retrieval in involuntary memory with the gradual intensification of familiarity in déjà vu. The onset of familiarity is sudden, but the epistemic feeling characteristic of déjà vu is a result of the conflict between familiarity and metacognitive expectation, which triggers additional search. Hence, as the authors suggest, the feeling of unexpected familiarity in déjà vu should gradually intensify until it is resolved or explained. (Compare to the gradual intensification, and persistence, of familiarity in tip-of-the-tongue experiences.) But déjà vu experiences have a characteristically sudden onset: Not only do they begin abruptly, but often the strangeness that characterizes them is strongest in the first moments of the experience. The proposal, at least in its current form, does not account for this property of déjà vu.
Before concluding, we note that the strangeness, infrequency, and suddenness of déjà vu have sometimes been viewed as pointing to an underlying neural or cognitive malfunction (e.g., Critchley, Reference Critchley1989). Their characterization of déjà vu as a natural product of memory retrieval suggests that B&M intend to argue against such views. If, on the one hand, their intention is to argue that déjà vu results from properly functioning retrieval processes, much more evidence is required to support their claim. If, on the other hand, they do not mean to argue for a claim about function, then their characterization of déjà vu as “natural” adds little to their proposal. Greater clarity about this issue would be welcome.
B&M might respond by arguing that our focus on déjà vu for token events (i.e., on déjà vécu) is unjustified, suggesting that future work, and the development of more detailed models, will shed more light on this particular phenomenon. While we sympathize with this sentiment, we worry that operationalizing déjà vu as any form of inappropriate familiarity (cf. Neppe, Reference Neppe1983) obscures much of what is puzzling, and indeed difficult to explain, about the phenomenon. We thus maintain that integrating déjà vu into autobiographical memory research will require tackling the problems that we have highlighted head-on.
You have
Access
Aiming to provide a means of integrating research on the two phenomena, Barzykowski and Moulin (B&M) propose an approach to involuntary autobiographical memory and déjà vu as natural products of memory retrieval. Their proposal must, if it is to achieve this aim, have a level of detail sufficient to account for the core functional and phenomenological properties of the target phenomena. Focusing on déjà vu, this commentary shows that the proposal does not have the requisite level of detail. We argue, after identifying three core properties of déjà vu experiences, that each of these leads to a problem that the proposed account, in its current form, is unable to resolve.
The first property that we will consider is strangeness. Déjà vu experiences are regularly characterized as strange, weird, or even eerie by both subjects and theorists (e.g., Brázdil et al., Reference Brázdil, Mareček, Urbánek, Kašpárek, Mikl, Rektor and Zeman2012). In the kind of experience on which we will focus here – sometimes referred to as “déjà vécu” (O'Connor, Lever, & Moulin, Reference O'Connor, Lever and Moulin2010) – the strangeness of the experience is due to the fact that it concerns a singular event and not merely a repeatable item, “as if time had slipped a cog and were now repeating itself” (Woodworth, Reference Woodworth1940: 357). The second property is infrequency. While most people experience déjà vu, they do so very rarely – on the order of few times a year (Brown, Reference Brown2003). The final property is suddenness. Triggered by a variety of situational factors, déjà vu experiences typically begin abruptly, last no longer than a few seconds, and end just as abruptly as they began (Brown, Porter, & Nix, Reference Brown, Porter and Nix1994).
Consider strangeness first. According to B&M's proposal, the distinctive phenomenology of déjà vu results from the combination of incomplete memory retrieval, which produces a sense of familiarity, with metacognitive appraisal of that familiarity as implausible. The strangeness of déjà vu, however, is due not merely to the implausibility of familiarity with a repeatable item – that is, with a type – but also, in cases in which déjà vu amounts to déjà vécu, to the felt impossibility of familiarity with the currently experienced token event. In order to account for the strangeness of déjà vu, then, the proposal needs to explain not only why incomplete retrieval produces a sense of familiarity with an item (a feeling that the current event resembles something experienced in the past) but also why, at least in some paradigmatic cases of déjà vu, unexpected familiarity is not simply brushed off as arising from limited access but is instead taken to indicate event repetition (a feeling that this very event has already been experienced). If the distinctive phenomenology of déjà vu results from a metacognitive assessment of plausibility, in short, then what should be assessed is not the plausibility of event resemblance but of event repetition.
For a similar reason, B&M's proposal cannot account for the infrequency of déjà vu. According to the proposal, incomplete retrieval – producing familiarity but not specific memory content – should be relatively frequent. This is required to explain the relative frequency of tip-of-the-tongue experiences (Brown, Reference Brown1991). Now, if our argument above is on the right track, the metacognitive assessment of plausibility responsible for generating déjà vu pertains to event repetition, not to mere event resemblance. Since event repetition is impossible, however, familiarity unaccompanied by specific memory content should typically, if not invariably, be assessed as impossible or at least highly implausible. But this entails that déjà vu experiences should be relatively frequent, which strongly contradicts the available data. Thus, if it is to account for both the strangeness and the infrequency of déjà vu, the proposal will have to invoke a wider range of (meta)cognitive processes. We will not attempt to determine here with it is feasible for a modified form of the proposal to do so.
The suddenness of déjà vu also presents a problem for the proposal. B&M contrast the rapidity of content retrieval in involuntary memory with the gradual intensification of familiarity in déjà vu. The onset of familiarity is sudden, but the epistemic feeling characteristic of déjà vu is a result of the conflict between familiarity and metacognitive expectation, which triggers additional search. Hence, as the authors suggest, the feeling of unexpected familiarity in déjà vu should gradually intensify until it is resolved or explained. (Compare to the gradual intensification, and persistence, of familiarity in tip-of-the-tongue experiences.) But déjà vu experiences have a characteristically sudden onset: Not only do they begin abruptly, but often the strangeness that characterizes them is strongest in the first moments of the experience. The proposal, at least in its current form, does not account for this property of déjà vu.
Before concluding, we note that the strangeness, infrequency, and suddenness of déjà vu have sometimes been viewed as pointing to an underlying neural or cognitive malfunction (e.g., Critchley, Reference Critchley1989). Their characterization of déjà vu as a natural product of memory retrieval suggests that B&M intend to argue against such views. If, on the one hand, their intention is to argue that déjà vu results from properly functioning retrieval processes, much more evidence is required to support their claim. If, on the other hand, they do not mean to argue for a claim about function, then their characterization of déjà vu as “natural” adds little to their proposal. Greater clarity about this issue would be welcome.
B&M might respond by arguing that our focus on déjà vu for token events (i.e., on déjà vécu) is unjustified, suggesting that future work, and the development of more detailed models, will shed more light on this particular phenomenon. While we sympathize with this sentiment, we worry that operationalizing déjà vu as any form of inappropriate familiarity (cf. Neppe, Reference Neppe1983) obscures much of what is puzzling, and indeed difficult to explain, about the phenomenon. We thus maintain that integrating déjà vu into autobiographical memory research will require tackling the problems that we have highlighted head-on.
Acknowledgements
The authors thank Denis Perrin for helpful discussion.
Financial support
N. A.'s research is funded by the European Union's Horizon research and innovation program under the Marie Sklodowska-Curie grant agreement 101062754. K. M.'s research is supported by the Institut universitaire de France.
Competing interest
None.