Premonitions: Memory, Prediction, and Psychological Explanations
Author: Angela Bogdanova · Published: October 11, 2026 · Editorial Policy
Research cluster: Extraordinary Experiences & Anomalistic Psychology · Written by Artificial Sapiens
An evidence-based guide to how apparent premonitions arise, how memory and predictive judgments interact, what the research on precognition can and cannot establish, and when distress deserves professional assessment.
For a related perspective, see Meaningful Coincidences: Why Random Events Can Feel Connected.
What Is a Premonition?
A premonition is a subjective impression, intuition, dream, bodily feeling, or mental image that seems to anticipate an event before that event occurs. A person may sense that someone will call, worry that an accident is imminent, dream about an unexpected encounter, or feel unusually certain that a particular outcome is approaching. Some impressions match subsequent events; many others never become memorable because nothing comparable happens.
Psychology can study when an impression occurred, what the person actually anticipated, how specific it was, what information was available, what happened afterward, and how the experience was remembered. These are separable questions. A sincerely reported experience establishes that someone experienced and interpreted an impression; it does not independently establish that information traveled backward from an unpredictable future event. This distinction is the starting point for empirical research on premonitions.
In everyday speech, premonition may refer to an uneasy feeling, a vivid dream, a strong hunch, or a belief that one has foreseen the future. Researchers discussing claimed anomalous information sometimes distinguish precognition, the supposed acquisition of future information, from premonition, an affective apprehension of a future outcome. The research literature has not established an independently reproducible mechanism for the latter claim. Bem (2011)
What People Mean When They Say They Had a Premonition
A feeling of impending danger
An individual may suddenly become uneasy before a flight, a family conversation, or a medical appointment. If something adverse occurs, the earlier uneasiness can acquire enormous personal importance. If the feared event does not occur, the emotion may be forgotten or attributed to ordinary stress. A full account therefore records both the feeling and the many comparable episodes that did not lead to an event.
A dream that resembles a later event
Dreams combine autobiographical material, emotional concerns, familiar places, and novel associations. A dream about a hospital, a storm, or a relative can appear remarkably specific after a similar waking event. To evaluate a claimed correspondence, the dream's wording must be recorded before the outcome and compared using the same criteria applied to nonmatching dreams. Research on dream content documents incorporation of waking-life material, while studies of reported precognitive dreams have examined coincidence, ambiguity, and belief as alternative contributors. Nielsen and Powell (2004) Houran and Lange (1998)
An intuition about another person
People notice small changes in routine: a friend stops replying, a colleague sounds distracted, an elderly relative has been unwell, or weather conditions make travel risky. These cues may remain outside the center of conscious attention yet legitimately inform an expectation. If a resulting prediction proves correct, its subjective suddenness can make it seem inexplicable. Retrospective investigation should ask whether a plausible ordinary information channel existed, without assuming that every individual case can be completely reconstructed.
A sudden sense of knowing what happens next
Some people describe an almost cinematic sense that a scene will unfold in a particular way. Laboratory research on déjà vu provides a useful comparison: participants can experience a strong feeling of knowing what is coming next without predicting the next event more accurately than chance. This experimentally documented dissociation makes subjective certainty an important phenomenon in its own right. Cleary and Claxton (2018) Huebert, Myers, and Cleary (2025)
How the Brain Predicts Ordinary Future Events
Everyday prediction is indispensable to perception, movement, planning, and social life. We forecast traffic, anticipate a person's response, estimate what might go wrong, and prepare for events we have never encountered in precisely the same form. A prediction can be remarkably accurate when it is grounded in patterns, prior knowledge, and available clues; unusual accuracy is not by itself evidence of an anomalous information source.
The cognitive neuroscience of episodic future simulation describes how memory supports the imagination of possible futures. Experiences and knowledge can be recombined into prospective scenes, allowing individuals to rehearse decisions and risks. This explains why a vivid imagined future may feel personally convincing, although the research does not imply that each premonition is produced by the same mechanism. Schacter, Addis, and Buckner (2008)
A successful forecast also depends on its base rate. Predicting rain when dark clouds are approaching differs from correctly naming a randomly determined symbol in a tightly controlled experiment. In the first case the environment supplies predictive information. In the second case, if randomization and blinding work as intended, ordinary cues should be unavailable. Treating these tasks as equivalent obscures the evidential question.
Why Accurate-Seeming Premonitions Can Occur
Many predictions, few memorable matches
People generate large numbers of spontaneous thoughts about the near future, particularly about events that matter to them. Some are specific, others are broad, and many are never explicitly put into words. When only striking successes are retold, the observed record becomes enriched with hits and depleted of misses. This is a selection problem, not a claim that the original experience was insincere.
Suppose a person thinks about several relatives throughout the week and one of them phones later that day. The event may feel extraordinary, although assessing its likelihood requires the number of people considered, the length of the prediction window, the usual frequency of calls, and how many nonmatching thoughts occurred. The relevant probability is defined by the full collection of opportunities for a match rather than by one selected story.
Hindsight bias and knowledge of the outcome
Hindsight bias describes the tendency for a known outcome to influence judgments about what was foreseeable earlier. It encompasses changes in remembered judgments, impressions of foreseeability, and impressions that an outcome was bound to happen. These components can vary independently. A premonition report formed after a major event is therefore less informative about the exact earlier prediction than a time-stamped record created beforehand. Blank et al. (2008)
Experiments demonstrate that learning an outcome can even change how people retrospectively describe the certainty and quality of predictions they actually made and recorded earlier. This finding bears directly on personal stories that include phrases such as 'I knew all along.' The feeling may be genuine in retrospect while the previous degree of certainty remains a distinct empirical question. Bradfield and Wells (2005)
Reconstructive memory and source monitoring
Human memory is reconstructive. Recollection relies on stored details, current knowledge, associations, and judgments about where an idea or image originated. The source-monitoring framework explains why people sometimes confuse a dream, a thought, a conversation, an inference, and an actual observation when remembering an event. These mechanisms are well studied in cognitive psychology; applying them to any particular autobiographical report requires case-specific evidence. Johnson, Hashtroudi, and Lindsay (1993)
After hearing about an accident, for example, a person might remember having worried about a dangerous journey. That worry may indeed have occurred. The later event can nonetheless alter whether it is recalled as a vague anxiety, a strong prediction, or a sharply detailed mental image. Misattribution, suggestibility, and bias are documented possibilities, not grounds for declaring a particular witness dishonest. Schacter (1999)
Vague predictions and flexible interpretation
An impression such as 'something important will happen soon' can fit countless outcomes. A forecast that names a person, event, location, and deadline in advance is much more constrained. Similarity can also be partly semantic: a dream of a falling bridge may later be matched to a financial collapse, a breakup, or a transportation problem. Researchers should establish matching rules before learning the result, use independent raters, and report ambiguous or unsuccessful cases alongside successes.
Pattern detection and perceived meaning
People routinely search for structure in sequences and ambiguous information. In experiments, illusory pattern perception has been associated with supernatural beliefs and other interpretations of apparently connected events. This does not establish that every paranormal belief or perceived premonition has the same origin, nor does it justify assigning a cognitive deficit to an individual. van Prooijen, Douglas, and De Inocencio (2018)
A systematic review spanning four decades found heterogeneous associations between paranormal belief measures and cognitive variables, with relatively consistent patterns involving intuitive thinking, confirmatory bias, and randomness perception. The review also identified limitations in samples, study design, and generalizability. These correlations describe research populations; they cannot diagnose or explain an individual person's unusual experience. Dean et al. (2022)
The Difference Between Intuition, Anxiety, and a Premonition
Intuition refers broadly to a judgment experienced without access to a deliberate, step-by-step explanation. It may incorporate learned cues or tacit familiarity. Anxiety often involves anticipation of uncertain or threatening outcomes, sometimes accompanied by bodily arousal and heightened monitoring. A premonition is the person's interpretation that an impression points toward a future event. One experience can carry all three descriptions.
Someone who has noticed a loved one's deteriorating health may experience a clear fear about an upcoming crisis. If the crisis occurs, the anticipation could have been a reasonable inference from subtle evidence, a distress-driven forecast, coincidence, or a mixture of these. Scientific appraisal examines the timing, available clues, specificity, and alternative explanations rather than choosing a single cause solely from how intensely the experience felt.
Persistent anxious expectations deserve attention when they interfere with sleep, concentration, relationships, or decisions. A recurring thought about catastrophe may be emotionally exhausting regardless of whether the person calls it intuition, a warning, or worry. Psychological support can address distress while allowing the person to discuss the meaning of the experience respectfully.
Premonitions in Dreams: What Can Be Studied?
Dream reports pose a distinctive measurement problem because dream memories can disappear or change rapidly after awakening. A compelling resemblance reported after an event may be based on an accurate earlier dream record, an imprecise reconstruction, an interpretation of symbolic imagery, or chance. These alternatives cannot be separated reliably by confidence alone.
Studies of dreaming show that material from waking experiences enters dreams in complex and selective forms; reported day-residue and dream-lag relationships concern incorporation of past experiences into dreams rather than verified future information. Small studies of allegedly precognitive dreams have modeled reported matches as coincidences interpreted through memory and belief, but they do not establish a universal explanation for every dream report. Nielsen and Powell (2004) Houran and Lange (1998)
The strongest prospective test begins with an unedited, independently time-stamped dream account. Before any outcome is known, the investigator specifies the event categories, matching criteria, comparison periods, and way to count misses. Independent coders should be blinded to whether a record is designated a candidate hit, and analysis should account for the large number of possible future events and meanings.
Déjà Vu and the Feeling of Knowing the Future
Déjà vu is a feeling of familiarity in a situation that the person recognizes as new. In experimentally constructed virtual scenes, some participants reported both déjà vu and a sense of knowing an upcoming turn, but their actual predictions did not improve. The finding provides an especially clear example of how the feeling of foreknowledge can be psychologically real without showing foresight. Cleary and Claxton (2018)
Follow-up experiments distinguished the judgment made before an event from the impression formed after it. Participants could retrospectively feel that a scene unfolded as expected, especially when familiarity was strong. More recent experimental manipulations of familiarity increased feelings of prediction without demonstrating future knowledge. This emerging line of work identifies one established experimental pathway to premonition-like feelings, while its applicability to complex real-world stories remains an empirical question. Cleary and Claxton (2018) Huebert, Myers, and Cleary (2025)
What Does Research on Precognition Actually Show?
The original experimental claims
Parapsychological research has investigated purported anticipation of future random outcomes under controlled conditions. Daryl Bem's 2011 paper reported nine experiments interpreting certain behavioral differences as possible influences of future stimuli on earlier responses. These findings attracted sustained scrutiny because such an effect would require unusually strong methodological and independent support. The existence of a published peer-reviewed report documents the claim; it does not by itself settle whether the claim is correct. Bem (2011)
Positive synthesis and methodological dispute
A later meta-analysis by Bem and colleagues pooled 90 experiments and reported a small overall effect interpreted by the authors as favoring anomalous anticipation. This is an important part of the published debate. Its significance depends on such questions as the independence of studies, study selection, analytic flexibility, publication and reporting biases, differences among experimental tasks, and how the distribution of observed effects should be modeled. A pooled statistically significant result is not equivalent to a robust demonstration of a new causal mechanism. Bem et al. (2016, revised version)
Independent replications and preregistration
In three preregistered attempts to replicate the retroactive-recall component of Bem's work, Ritchie, Wiseman, and French did not reproduce the reported effect. A later multi-laboratory Transparent Psi Project adopted open data, preregistration-related controls, documentation, and cross-perspective protocol development. It observed 49.89% successful guesses, compared with the 53.07% originally reported by Bem in the target experiment, against a 50% chance benchmark. The target effect was not reproduced. Ritchie, Wiseman, and French (2012) Kekecs et al. (2023)
A 2025 metascientific project extended this approach across three replication studies involving 26,483 participants and 420,472 critical trials. Its results were inconsistent across the sequence: the first study did not replicate the proposed effect, a follow-up detected a small effect in the opposite direction, and a third did not replicate that reversal. This pattern does not establish a stable anticipatory mechanism. It demonstrates why independent reproducibility and control of weak-effect research are central to evaluating extraordinary claims. Walleczek et al. (2025)
What the evidence allows us to conclude
The published record contains claims of statistically anomalous results, syntheses arguing in their favor, and failures of independent replication. Controlled experiments have not established a consistently reproducible capacity to access otherwise unknowable future events. This assessment concerns the reliability of current evidence, not the sincerity of people who describe premonitions, and no individual study can settle every imaginable formulation of the hypothesis.
Methods matter as much as outcomes. When an alleged effect is small, many unnoticed analytical and procedural differences can change the result. A convincing future demonstration would need sharply specified predictions, randomization that rules out ordinary leakage, masking of researchers and participants as appropriate, secure preregistration, complete reporting, adequate power, publicly auditable records, and independent repeated success with the same operational hypothesis.
How to Evaluate a Claimed Premonition Fairly
Record the forecast before the result
A prospective record should preserve the original wording, timestamp, date range, specificity, available information, and level of confidence. Editing should be visibly tracked. A later retelling can be meaningful for autobiography but cannot replace the prior record when the question is whether a particular future event was predicted.
Define what counts as a match
Before examining results, decide which outcomes qualify. 'A person I know will contact me tomorrow' differs greatly from 'My cousin will call between 2:00 and 3:00 p.m. about a specific topic.' Scoring should account for partial matches and clear misses. Ambiguous language should not be sharpened retrospectively without marking that reinterpretation.
Count nonmatches and the opportunity space
A credible analysis includes every prospective prediction made within the agreed period, not merely the most impressive anecdotes. If a person has numerous dreams, thoughts, conversations, and intuitions, it is predictable that some will resemble later events. Base rates, multiple comparisons, ordinary cues, and chance matching must be evaluated together.
Use blinding and independent rating
Independent assessors should compare forecasts against possible outcomes without knowing which outcome is considered the apparent hit whenever the design permits. Prespecified coding prevents a moving definition of success. Analysis should include uncertainty intervals and ideally a comparison with appropriate control records, rather than relying on a count of striking examples.
Respect the limits of anecdotal evidence
Personal accounts are valuable data about phenomenology, meaning, emotion, and memory. A single narrative seldom supplies the denominator needed for estimating predictive performance. Absence of a complete record is a reason to treat causal attribution as unresolved, rather than a reason to dismiss the person or to accept a supernatural explanation.
How Individual Differences and Cultural Interpretations Matter
An identical sequence of events can be described as coincidence, intuition, a warning, a dream match, a lucky prediction, or a spiritual sign depending on language, family tradition, and personal worldview. These interpretations shape which details are remembered, which stories are retold, and how much significance is assigned to a match. The psychological question is how a meaning is formed and expressed; it is separate from evaluating the truth of a religious tradition.
Belief-related research reports group-level links with styles of reasoning and experiences of apparent patterns, with considerable variation in measures and samples. Such findings do not imply that people who report premonitions share one personality type or clinical profile. A review of cognitive correlates of paranormal belief explicitly cautioned about methodological heterogeneity and generalizability. Dean et al. (2022)
Life circumstances also affect salience. A premonition about a seriously ill family member may be entwined with careful monitoring and anticipatory concern; an impression occurring during a period of intense stress may feel stronger because its consequences seem urgent. The meaning of the experience can be acknowledged without assuming either an extraordinary external cause or a psychiatric explanation.
Are Premonitions a Sign of a Mental Health Disorder?
Experiencing an intuition, striking coincidence, unusual dream, or sense of premonition does not by itself constitute a psychiatric disorder. Clinical assessment is based on the full context, associated symptoms, distress, duration, functional impairment, and relevant medical and cultural information. Paranormal belief is not, on its own, a diagnosis.
A person who recalls a frightening prediction but continues to function well, sleeps adequately, and can consider uncertainty may need nothing more than a respectful conversation. Professional assessment becomes appropriate when the experience is persistent and distressing, seriously disrupts daily functioning or sleep, causes compulsive checking or avoidance, accompanies disorientation or other concerning changes, or prompts dangerous actions. The assessment should address possible sleep-related, psychological, neurological, or medical contributors according to the presenting symptoms.
It is especially important to distinguish a thought about a possible future event from a fixed belief with major functional consequences, and from intrusive thoughts that drive repetitive reassurance seeking. These may call for different kinds of clinical inquiry. The content of one report, a questionnaire score, or an isolated unusual experience is insufficient for diagnosis. If immediate danger is involved, seek local emergency assistance.
How to Respond When a Premonition Feels Frightening
Begin by describing the experience as accurately as possible: what was felt, when it began, what outcome was feared, and what information was available. The statement 'I had a strong feeling that something bad would happen' conveys a real experience without requiring an immediate theory of its cause. People can also recognize emotional meaning while remaining uncertain about external prediction.
Where appropriate, ordinary reality-based checks can help: verify current information, assess practical risk using evidence available now, and make proportionate decisions. Repeatedly searching for confirmation, rearranging unrelated memories, or using suggestive interviews to prove a hidden event can reinforce distress and distort the record. For a distressing recurring experience, a licensed professional can help assess anxiety, sleep, functioning, and other relevant symptoms without treating the narrative as either proof of paranormal causation or evidence of illness.
A time-stamped journal may be useful for personal reflection or a genuinely prospective scientific inquiry if it is voluntary and does not intensify obsessive checking. It should preserve all predictions and missed outcomes. People who find that recording every worry makes their anxiety worse are better served by discussing this pattern with a qualified clinician than by conducting continuous prediction tests.
Frequently Asked Questions
Are premonitions real?
The experience of having a premonition is real as a subjective psychological event. The separate proposition that it demonstrates access to an otherwise unknowable future is not established by consistent, independently replicated scientific evidence. A particular person's story may remain personally significant while its external cause is unresolved.
Why do I feel like I knew something before it happened?
Possible contributors include ordinary inference from subtle clues, familiarity, hindsight bias, selective recall of accurate-seeming predictions, source-monitoring errors, or coincidence. These are established or plausible mechanisms at different levels, and no single explanation should be imposed on a particular report without evidence. Blank et al. (2008) Johnson, Hashtroudi, and Lindsay (1993)
Can dreams predict the future?
Dreams can resemble later events, sometimes strikingly. The scientific challenge is to show that a dream predicts a future outcome more accurately than chance after accounting for its original wording, common events, flexible matches, remembered misses, and the number of opportunities to find a correspondence. That general capacity has not been established.
What is the difference between a premonition and intuition?
Intuition is a broad category of judgments experienced without conscious step-by-step reasoning. A premonition is an apparent forewarning of a future event. Some premonitions may be ordinary intuitive predictions; others may involve dream imagery, anxiety, or later interpretation.
Does a strong sense of certainty prove that a prediction is accurate?
No. Experimental studies show that a feeling of knowing what happens next can increase even without better predictive accuracy. Subjective confidence and performance are separate measurements. Cleary and Claxton (2018) Huebert, Myers, and Cleary (2025)
Why do premonitions often concern danger or loved ones?
People monitor emotionally important possibilities closely and remember consequential events vividly. Concern for loved ones, existing warning cues, and attention to threatening outcomes can make certain thoughts especially salient. An individual's recurring fear should be assessed in its own context, rather than treated as a general law about premonitions.
Can science test a specific premonition?
Yes, when the prediction and scoring rules are stated before the result, the relevant outcome can be independently verified, and adequate comparison data are collected. A prospective, specific prediction is easier to assess than a retrospective narrative or a symbolic dream interpreted after a major event.
Does a premonition mean I have a psychotic disorder?
No. An isolated unusual intuition or interpretation does not establish psychosis or any other diagnosis. Evaluation is appropriate if the experiences are accompanied by serious distress, other concerning symptoms, or impaired functioning.
Have experiments proved ESP or precognition?
Researchers have published positive claims, but high-profile findings have also failed independent replication. A large preregistration-oriented multi-laboratory test and further investigations reported in 2025 have not established a consistently reproducible effect. Kekecs et al. (2023) Walleczek et al. (2025)
How should I talk to someone who reports a premonition?
Listen to what they experienced and what it meant to them. Ask open questions about timing and specificity if investigation is wanted, and avoid suggestive questioning, ridicule, or premature diagnosis. The distinction between respecting an experience and establishing an external explanation makes a careful conversation possible.
Scientific Summary
Premonitions bring together future-oriented imagination, the interpretation of uncertainty, autobiographical memory, emotion, and the human search for meaningful connections. Research establishes that people can genuinely feel they know what comes next and that memory and outcome knowledge can change retrospective judgments. Studies of purported precognition remain contested, with consequential failures of independent replication. A responsible explanation keeps the experienced event, a plausible psychological mechanism, and an externally demonstrated cause conceptually distinct.