Asking About Déjà Vu: Separating Experience from Report

Answering yes to a question about déjà vu at the end of an experiment is not the same measurement as spontaneously marking an experience while the experiment proceeds. The first asks a person to look back across a task. The second asks them to identify a particular moment. The difference may involve not only memory, but also what repeated questions suggest should have happened. When researchers count a subjective experience, the way they ask can contribute to the count.

Radka Jersakova, Chris J. A. Moulin and Akira R. O’Connor examined this problem in a 2016 study of déjà vu and tip-of-the-tongue states. This article focuses on the first experiment’s baseline procedure and the reporting change introduced in Experiment 3. The study was not a test of precognition or extraordinary memory ability. It asked how often participants reported these experiences during a recognition task not expected to generate them, and how assessment procedures affected those reports. [1]

Familiarity is not sufficient

The researchers defined déjà vu as familiarity with a situation accompanied by awareness that the familiarity is inappropriate because the situation is new. Merely knowing what a familiar word means does not satisfy that definition. Neither does correctly recognizing an item shown earlier in an experiment. Déjà vu, as operationalized here, includes a mismatch between a feeling and the person’s appraisal of the situation.

A tip-of-the-tongue state involves failure to retrieve information together with awareness that it should be accessible. The study’s explanation also included partial recall of characteristics and a feeling that retrieval was approaching. Both experiences concern an appraisal of one’s own retrieval process, but the resemblance does not make them identical. A procedure that changes reporting of one need not change reporting of the other in the same way. [1]

A task about earlier presentations

Experiment 1 collected online data from 224 participants divided across four conditions. Volunteers arrived through the laboratory website, online experiment listings and social networks. Recruitment did not produce a random population sample. Participants judged a sequence of 120 word presentations using a mouse to indicate whether each word had already appeared in the task. There were eighty distinct words: forty appeared once and forty appeared twice. [1]

The relevant question was whether a word had appeared in this task, not whether the participant had ever encountered it in life. A common noun could be generally familiar and nevertheless be new within the experimental sequence. This separates knowledge of a word from memory for its presentation. The researchers had not designed a critical manipulation to produce déjà vu or tip-of-the-tongue states. That expectation does not prove that no individual could have any genuine subjective experience during the task.

One group was periodically asked about déjà vu, another about tip-of-the-tongue states, and a third about seeing words printed in yellow. No word was actually yellow. A fourth group completed the task without these interruptions. In the first three groups, questions appeared after every twelve trials, making nine checks during the task. At the end, everyone received the same questionnaire asking about all three kinds of event. The yellow-word question supplied a comparison concerning an experimentally known feature.

High reporting without repeated prompts

Even in the uninterrupted control condition, 41.1 percent later reported déjà vu and 33.9 percent reported tip-of-the-tongue states. The proportions endorsing déjà vu at the end did not differ significantly between groups. Tip-of-the-tongue reporting did differ: the group repeatedly questioned about that state reported it more often than the control and colour groups. The two subjective experiences did not respond identically to repeated assessment. [1]

These percentages are not estimates of everyday population prevalence. They describe answers given after a specific word-recognition task, with definitions supplied. Yellow-word reports were much lower, which complicates the idea that people simply endorsed every possible event indiscriminately. Yet this comparison does not independently verify that each endorsed déjà vu experience matched the investigators’ definition. Establishing that no yellow word appeared is a different evidential problem from establishing whether someone experienced a particular subjective state.

Replacing nine questions with an available button

Experiment 3 included 123 participants, sixty-four in the déjà vu condition and fifty-nine in the tip-of-the-tongue condition. The recognition task remained the same. Instead of interrupting it with recurring retrospective questions, the researchers left a reporting button available throughout. Participants could press it whenever the defined experience occurred. They still completed the common questionnaire at the end. [1]

Free report here did not mean writing an unrestricted account. It meant initiating the report rather than responding at one of nine predetermined checks. Because the button was continuously available, the procedure did not simply reduce the number of opportunities to report. It did, however, introduce its own demands: participants had to monitor their state and remember to use the button while completing recognition judgments. The method cannot be regarded as a completely transparent window into experience.

Fewer participants endorsed déjà vu in the final questionnaire than in the corresponding group from Experiment 1. The proportion endorsing tip-of-the-tongue states did not differ significantly. The article’s Experiment 3 summary and general discussion also describe fewer within-task occurrences reported for both experiences among participants who endorsed them at the end. This latter outcome counts reports within a selected subgroup. It differs from the proportion of all participants who say they had at least one experience. [1]

Fewer reports do not identify a single mechanism

The authors propose that recurring questions may encourage participants to reinterpret uncertain sensations in terms of the category being queried. Someone unsure how to describe a feeling might infer that the experimenter expects déjà vu or a tip-of-the-tongue state. Self-initiated reporting could reduce pressure toward agreement. This explanation does not show that participants intentionally supplied false information, nor does it measure the motive behind every individual response.

It would also go beyond the evidence to conclude that the new procedure made the brain produce fewer episodes of déjà vu. Several aspects could change: the experience itself, the threshold for identifying it, later reinterpretation or the act of reporting. There was no independent indicator determining whether each subjective episode was genuine. Experiments 1 and 3 also used different samples, rather than measuring every experience of the same participants under both methods. The findings make assessment method important without eliminating all alternative explanations.

The absence of an experience-generating manipulation remains a reason for scrutiny, not proof that every report was false. The authors acknowledge that some participants could have genuine experiences. The issue is whether the unexpectedly high rates can be interpreted confidently as successful induction of the phenomena, when ordinary recognition and reporting practices may account for part of the pattern. Subjective evidence remains necessary, but the conditions under which it is elicited require attention.

The measurement belongs in the conclusion

The supported research principle is to treat question format as part of the data, rather than reject subjective accounts outright. A self-initiated report during a task, an end-of-task questionnaire and a control condition contribute different information. Agreement between them can be useful, but does not make them interchangeable. Likewise, a stable response pattern does not by itself establish that the intended experience was measured accurately.

This study does not provide a disease diagnosis or a personal assessment of memory health. Its contribution concerns how researchers turn first-person descriptions into evidence. Before asking how many people experienced déjà vu, it helps to ask what definition was supplied, when the question was posed and which response was counted. Those details determine whether a number describes a momentary report, an overall endorsement or an inference about the underlying experience.

Sources

[1] Jersakova, R., Moulin, C. J. A., and O’Connor, A. R. (2016). “Investigating the Role of Assessment Method on Reports of Déjà Vu and Tip-of-the-Tongue States during Standard Recognition Tests.” PLOS ONE, 11(4), e0154334. Open access, Creative Commons Attribution; used with independently written exposition and interpretation. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0154334