Metamemory: Knowing What You Remember and Forget

Author: Angela Bogdanova · Published: October 16, 2026 · Editorial Policy

Research cluster: Consciousness, Metacognition & Self-Awareness · Written by Artificial Sapiens

Metamemory is the knowledge, beliefs, monitoring, and regulation people apply to their own memory. It lets someone estimate whether a fact has been learned, recognize that an answer feels familiar but remains inaccessible, judge the reliability of a recollection, and decide whether to study, search, verify, or stop trying to remember. These judgments shape behavior, but their accuracy depends on what information the person uses to make them. The scientific study of metamemory therefore asks two related questions: How do people know what they remember, and how accurately does that self-knowledge guide their decisions? Dunlosky and Thiede's review and the foundational Nelson–Narens framework60053-5) describe these processes as part of memory monitoring and control.

A student may finish a chapter and feel confident about knowing it, then struggle to explain its ideas without looking. Another student may encounter a temporarily forgotten name and correctly expect to recognize it among several options. A person may confidently recognize a photograph but misremember where it was first encountered. All three experiences involve the relationship between memory performance and a judgment about memory, rather than memory performance alone.

What Is Metamemory in Psychology?

Memory concerns the processes by which information is encoded, retained, reconstructed, and retrieved. Metamemory concerns what a person knows or believes about those processes and how that information is used. Researchers commonly distinguish general knowledge about memory, moment-to-moment judgments of learning or retrieval, and the choices made in response to those judgments. Hertzog and Curley's account of metamemory explains why a single global rating of one's memory cannot capture this multidimensional system.

The term belongs to cognitive psychology and the broader research tradition of metacognition. Metacognition includes monitoring and regulation of thinking, comprehension, decisions, and other cognitive operations. Metamemory focuses specifically on memory-related knowledge, predictions, confidence, and regulation. It is a scientific construct, not a clinical diagnosis, personality type, or score with a universally accepted cutoff.

A brief research history

Studies of what individuals know about their own remembering expanded during twentieth-century developmental and cognitive psychology. Early work investigated people's understanding of memory capacity, task demands, and strategies. Later experimental traditions examined feeling-of-knowing judgments, predictions of later recall, and study-time allocation. In 1990, Thomas Nelson and Louis Narens articulated an influential model separating monitoring from control across interacting levels of cognition. Their theoretical framework60053-5) remains an organizing reference, although later research has developed more detailed accounts of the evidence underlying specific judgments.

The central insight is that memory self-evaluation can be studied empirically. A researcher can collect a prediction of success, then compare it with observed recall or recognition. Differences between confidence and performance become measurable results, which can be tested across individuals, materials, settings, and conditions.

The Three Main Components of Metamemory

1. Metamemory knowledge and beliefs

Metamemory knowledge includes beliefs about how memory works, one's strengths and limitations, and which strategies are likely to help. Someone may know that unfamiliar vocabulary requires repeated encounters, that names are harder to recall without cues, or that an appointment should be placed on a calendar. These beliefs can be useful without being perfectly accurate. A person can also hold mistaken theories, such as believing that an easy second reading guarantees long-term retention. Dunlosky and colleagues' comprehensive review of learning techniques demonstrates why perceived ease and the tested effectiveness of a strategy need to be evaluated separately.

Knowledge about memory may concern people (what is typically difficult for me), tasks (which material or tests demand more retrieval), and strategies (how spacing, cues, rehearsal, or retrieval practice may affect future performance). The evidence does not justify assuming that a person with sophisticated general beliefs will always monitor each individual memory accurately.

2. Metamemory monitoring

Monitoring means estimating the current or future state of one's memory. It appears in predictions of later recall, the feeling that a forgotten answer would be recognizable, confidence about an answer just given, and judgments that more study is needed. Monitoring is inferred from reports or decisions gathered under defined experimental conditions; researchers compare these observations with subsequent performance. Schwartz's review shows that judgments made during learning and those made during retrieval may draw on different sources of information.

Monitoring can be prospective, when it predicts a later outcome, or retrospective, when it evaluates an answer or remembered event after retrieval. It can target a specific item or take the form of an overall assessment of one's memory. Those distinctions matter because a person may have fairly good confidence about answers already produced and much poorer insight into material that has not yet been tested.

3. Metamemory control

Control is the use of memory assessments to select actions: continue studying, change strategies, devote more time to a difficult item, attempt retrieval again, seek a hint, verify the source, or rely on an external reminder. Monitoring and control form a feedback relationship. A judgment that a definition is not yet secure may trigger practice; the outcome of practice provides new evidence for a revised judgment.

A monitoring judgment can be inaccurate and still strongly influence behavior. Conversely, someone can identify an uncertain memory without allocating sufficient time to investigate it. This is why the quality of metamemory must be evaluated at both the judgment stage and the action stage. In two studies of college students, Dunlosky and Rawson investigated how overconfidence in learning decisions could undermine later retention.

How Metamemory Appears in Everyday Life

Metamemory is visible in ordinary moments: predicting whether you can give a presentation without notes, recognizing the beginning of a forgotten password, wondering whether an account of an event is reliable, deciding which chapters require revision, and adding a reminder because you anticipate forgetting. These examples describe different functions and should not be reduced to a single scale of confidence.

At work, a person can remember the substance of a meeting while being uncertain about a specific date or figure. Good metamemory helps separate a stable summary from a detail requiring confirmation. In learning, someone may correctly see that familiar material is easy to recognize but difficult to retrieve freely. In conversation, the familiar tip-of-the-tongue experience can indicate that retrieval is partially activated, although it cannot guarantee that the sought information is available or correct.

Metamemory also shapes how people use external aids. Calendars, notes, search tools, and checklists are choices about memory management. They can compensate for expected future forgetting, but their successful use depends on whether the reminder is available, noticed, and acted upon when needed. This article focuses on the cognitive judgment behind that decision; the larger science of memory systems and forgetting belongs to the adjacent memory domain.

Judgments of Learning: Can You Predict What You Will Remember?

A judgment of learning, often abbreviated JOL, estimates the likelihood that information studied now will be recalled or recognized later. Researchers may ask participants to study paired words, give a predicted probability of remembering each target, and later take a memory test. The difference between those predictions and observed performance reveals how informative the judgments were.

A critical issue is when the prediction is requested. Nelson and Dunlosky's classic delayed-JOL experiments found that judgments made after a short delay could predict subsequent recall more accurately than immediate judgments under their experimental conditions. Delaying the judgment may encourage an attempt to retrieve from the cue rather than relying on the recently visible answer. The mechanism remains important: Spellman and Bjork highlighted that delayed judgments can themselves involve retrieval practice, which can change what is being measured.

Immediate versus delayed judgments

Immediately after rereading a paragraph, comprehension and familiarity may feel unusually strong because the text is still present and fluent. A later question without the text available often tests a different ability. A delayed, cue-only attempt can offer a more diagnostic indication of whether the material is retrievable. The research supports using retrieval-based checks when future retrieval is the goal, but it does not establish that one fixed delay or one judgment protocol is optimal for every topic, age, or testing format.

Judgments of learning are a distinct research topic within metamemory. Their methods, immediate-versus-delayed comparisons, and calibration procedures warrant fuller treatment under the future EN-29 article on judgments of learning; the present page explains their role in memory self-knowledge.

Feeling of Knowing: When an Answer Seems Just Out of Reach

A feeling-of-knowing judgment occurs when someone cannot currently recall an answer but estimates whether it could be recognized or recovered with additional cues. For example, a reader may fail to name a particular scientist while believing they would identify the name from a list. In experiments, such predictions can be assessed against subsequent recognition outcomes.

Koriat's accessibility model proposes that the feeling often arises from accessible fragments, related information, or the ease of partial retrieval. The brain does not need a special, infallible view of a hidden memory trace to produce the experience. Available cues may genuinely predict recognition, but partial familiarity can also come from incorrect associations.

The feeling of knowing overlaps with tip-of-the-tongue experiences, which are often accompanied by a strong sense that a missing word or name is imminent. However, the two terms describe different measurements and experiences. A feeling-of-knowing judgment is a prediction commonly studied by comparing a failed recall attempt with later recognition; a tip-of-the-tongue state includes a particular subjective feeling during retrieval failure. Neither establishes that the desired answer must exist intact in memory.

Retrospective Confidence: How Sure Are You About a Memory?

After answering a question or recalling an event, people can rate their confidence that the response is correct. This retrospective judgment is another form of metamemory monitoring. The subjective certainty can guide whether a person reports an answer, checks it, or withholds it. The scientific question is whether confidence reliably distinguishes correct memories from mistakes.

Confidence can correlate with memory accuracy under some conditions, yet the relationship varies with tasks, evidence, retrieval cues, and the effects of misinformation or repeated exposure. Research on retrieval fluency demonstrates that a readily accessible response can encourage confidence even when the same accessibility is an unreliable predictor of later recall. Therefore, a vivid or easy recollection is not independently sufficient to verify a factual detail.

This issue is especially important where a precise quotation, numerical claim, event chronology, or another consequential fact matters. A stronger practice is to compare the recollection with independent documentation rather than using the strength of the feeling as proof. That recommendation concerns verification of information; it is not a claim that people should distrust every ordinary memory.

Why Metamemory Sometimes Works—and Sometimes Misleads

The cue-utilization account

One influential explanation is that people make memory judgments by interpreting cues available during study or retrieval. In Koriat's cue-utilization experiments, influences on learning judgments included intrinsic properties of the material, aspects of the study conditions, and experiential cues such as perceived processing fluency. Some cues are more diagnostic of eventual performance than others. As a result, confidence can shift even when actual retention changes much less, or retention can improve without a commensurate increase in predicted performance.

Intrinsic cues include the apparent difficulty of the item itself. Extrinsic cues include study time, repetition, or presentation conditions. Mnemonic cues include the subjective experience of ease, familiarity, or retrieval success. These are categories used in a particular research framework, not three independent storage systems in the brain.

Familiarity, fluency, and the illusion of knowing

Repeated exposure can make material feel familiar and easy to process. That subjective ease is real, but it may be an inadequate predictor of whether the person can reconstruct the information unaided days later. Likewise, information retrieved quickly today can feel as though it will remain readily accessible. Benjamin, Bjork, and Schwartz's experimental work shows conditions in which retrieval fluency systematically misled predictions about future remembering.

Fluency is useful when it tracks the capacity the later test requires. Problems arise when it reflects other factors, such as a currently visible answer or a very recent encounter. A learner may therefore become more confident through exposure while making limited progress toward durable independent retrieval.

Why accurate self-evaluation varies across tasks

Memory predictions are shaped by what question is asked. Recognizing a multiple-choice answer, recalling a name without cues, reproducing a definition, identifying a source, and reconstructing the timing of an event impose different demands. Confidence about one task need not predict performance in another. Dunlosky and Thiede emphasize that the usefulness of a cue depends on whether it predicts the specified criterion task.

That principle explains why the familiar instruction to “check whether you know it” is incomplete. A meaningful self-check requires specifying what remembering should look like and under what conditions it will be tested.

How Scientists Measure Metamemory

Predictions and performance tests

A typical study pairs subjective predictions with an observable outcome. Participants learn words, concepts, pictures, or stories; they predict future recall or recognition; and they complete an assessment at a defined interval. Another design elicits confidence after each answer. A feeling-of-knowing study asks for predictions after failed recall attempts and then measures recognition. These paradigms produce evidence about specific kinds of monitoring, not a direct measurement of all aspects of personal memory knowledge.

Calibration, discrimination, and bias

Calibration describes how closely numerical confidence estimates correspond to observed success rates. If many items assigned 80% confidence are answered correctly about 80% of the time, those ratings are well calibrated for that set and task. A person can still be poorly able to distinguish which particular items will be correct.

Discrimination, sometimes termed relative accuracy or resolution in metamemory research, concerns how effectively higher versus lower judgments separate more and less successful items. It can be examined using within-person associations, ranking methods, or analyses adapted to the task. Global overconfidence and poor item-level discrimination are different possible patterns. A mean confidence score alone cannot settle either question.

Measurement requires care because the memory task has its own difficulty, number of items, response bias, timing, and scoring rules. Measures such as correlations, calibration error, or signal-detection-based metacognitive indices answer related but nonidentical questions. Their interpretation changes with the data-generating conditions. The broader construct of metacognitive sensitivity and the technical use of meta-d-prime are treated more fully within EN-29's planned measurement node rather than being assigned the same definition as metamemory.

Examples of different outcomes

Suppose two people each answer 7 out of 10 questions correctly. One predicts 70% success across the test and another predicts 95%. Their overall accuracy is identical, but the second person shows stronger global overconfidence. If the first gives nearly equal confidence to all ten answers, while the second assigns greater confidence to the seven correct answers, the second may nonetheless demonstrate better item-level discrimination. Accurate memory, calibrated predictions, and discrimination between correct and incorrect responses must be evaluated separately.

This simple example illustrates how a seemingly intuitive term such as “good memory awareness” contains multiple experimental criteria. Interpretation should always identify which criterion an actual study measured.

From Monitoring to Control: How Judgments Change Study and Retrieval

People use their own judgments to determine what deserves effort. A learner who identifies uncertain material may schedule another practice session, while someone who feels a topic is mastered may move on. These decisions are adaptive when the judgment is sufficiently informative and the resulting strategy actually improves retention. They can be costly when fluency-based overconfidence leads to premature stopping.

Dunlosky and Rawson's experiments provide evidence that inaccurate monitoring can impair learning by influencing study decisions. However, a perfect confidence estimate is not the sole condition for successful learning: available time, the effectiveness of available strategies, motivation, task importance, and the form of the final test also matter.

Retrieval practice as information and intervention

Trying to answer a question before checking the answer supplies evidence about what is retrievable. In many learning settings, successful retrieval practice also strengthens later retention. These functions are analytically distinct. The testing attempt provides diagnostic information for monitoring, while the retrieval process itself can contribute to learning. Karpicke, Butler, and Roediger investigated students' self-directed use of retrieval practice and discussed gaps between laboratory evidence and common study habits.

The broad evidence review by Dunlosky and colleagues rated practice testing and distributed practice as particularly useful across many studied conditions. The evidence supports these learning methods in appropriate contexts; it does not support a claim that any single self-assessment exercise will permanently improve every person's metamemory capacity.

Metamemory, Remembering, and Related Concepts

Metamemory versus memory ability

A person can have strong recall but poor insight into which facts are at risk of being forgotten. Another may have relatively weak recall but predict that weakness accurately and compensate through external aids. These patterns distinguish first-order memory performance from second-order monitoring and memory-related control. Metamemory research studies the relationship between them rather than assuming they are interchangeable.

Metamemory versus metacognition

Metacognition covers judgments and regulation across cognition. Metamemory is its memory-specific domain. Monitoring comprehension, detecting an error in mathematical reasoning, and evaluating a perceptual decision may all be metacognitive, but only judgments about learning, remembering, retrieval, or forgetting belong centrally to metamemory. A separate English Hub article on metacognition in the age of AI discusses an AI-mediated application of cognitive monitoring; its search intent differs from the human memory-monitoring construct explained here.

Metamemory versus source monitoring

Source monitoring concerns determining where information came from: an eyewitness event, a conversation, a book, a dream, or imagination, for example. It contributes to memory evaluation and may interact with confidence judgments. Yet an error in identifying a memory's origin is a property of retrieval and attribution, while a metamemory judgment estimates one's knowledge or confidence about that origin. Research designs must specify whether they measure source-memory correctness, confidence in it, or both.

Metamemory versus self-awareness

Self-awareness encompasses perceptions and representations of the self, one's feelings, body, behavior, and attributes. Metamemory is a narrower inquiry into beliefs, judgments, and choices about memory. Introspection into a sense of certainty can contribute to metamemory, but being reflective about one's personality or identity is not equivalent to accurately predicting a memory test.

Metamemory versus prospective memory

Prospective memory means remembering to carry out an intention later, such as taking a document to work or calling someone after lunch. Metamemory concerns the belief that one might forget such an intention and the strategy chosen to manage the risk. Scheduling a reminder may be an act of memory control, whereas successfully remembering to act at the correct time is prospective-memory performance. The two processes interact but answer different scientific questions.

Metamemory Across Development and Aging

Metamemory develops with experience and changes across life, with different trajectories for general beliefs, item-by-item monitoring, strategy knowledge, and objective memory performance. Children can learn that certain tasks are harder to remember and gradually refine self-assessment and strategy selection, but developmental findings depend on the age group, memory task, and specific construct measured. Adults likewise differ in whether they rely on broad self-beliefs or current evidence from actual retrieval attempts.

Older adults may report concerns about memory while performing competently on some tests, or may exhibit changes in specific judgments alongside changes in memory. Hertzog and Curley's research synthesis emphasizes the multidimensional nature of metamemory in cognitive aging and the need to separate beliefs about personal memory from on-task monitoring. Age group membership by itself is therefore insufficient to infer the accuracy of a particular memory judgment.

Worries about memory can also be shaped by stress, mood, sleep, health, and everyday demands. Neither a single subjective complaint nor a single laboratory confidence score establishes the presence of a neurocognitive disorder. Persistent, worsening, or functionally significant memory changes warrant a properly contextualized clinical assessment rather than interpretation through a self-test alone.

What Neuroscience Tells Us About Metamemory

Metamemory draws on brain systems involved in memory retrieval, evaluation, and cognitive control. The neural evidence comes from tasks in which individuals both perform an objective memory operation and make judgments about that performance. Brain activity associated with confidence is therefore interpreted relative to the experimental task and its controls.

A coordinate-based meta-analysis of 47 neuroimaging studies found partly shared neural associations across metacognitive judgment domains, including medial and lateral prefrontal areas, precuneus, and insula, together with some domain-dependent differences. A separate meta-analysis of retrospective memory confidence synthesized 19 eligible functional MRI studies and identified convergent associations involving parahippocampal and frontal areas, among others.

These findings support a distributed account in which metamemory depends on interacting systems rather than a single “memory awareness center.” Neuroimaging correlations do not establish that any one region independently produces subjective certainty or that modifying it would reliably improve memory in daily life. Small study samples, differences in tasks and contrasts, reverse inference, and the distinction between confidence magnitude and monitoring accuracy limit strong localization claims. The overall evidence is meaningful but does not yield a simple neural diagnostic marker of metamemory.

Clinical Relevance and the Limits of Interpretation

Metamemory is relevant to clinical research because awareness of one's memory capabilities can influence reporting, rehabilitation planning, and decisions to use compensatory strategies. A review of metamemory experiments in neurological populations documents diverse task designs and outcomes. A systematic review of metamemory monitoring in Alzheimer's disease found variation by judgment type: participants sometimes assessed recent performance more accurately than they predicted future performance. These patterns underscore why clinical interpretations require task-specific measures and contextual information.

It would be incorrect to treat everyday overconfidence, tip-of-the-tongue experiences, or forgetting as clinical signs in isolation. Neurological disorders, psychiatric symptoms, and medication or sleep effects can influence memory and monitoring in different ways, and a diagnosis requires an appropriate history and professional examination. For readers facing new, persistent, worsening, or safety-relevant memory difficulties, professional evaluation is more informative than an online estimate of metamemory skill.

Can You Improve Metamemory?

The most defensible practical goal is to make specific judgments better informed and to use them for useful actions. A person studying for a test can periodically close the book and explain a concept without cues, compare that performance with prior confidence, and review the items that were unexpectedly difficult. That method creates an explicit reference point for updating predictions. It also incorporates retrieval practice, whose effects on retention have been evaluated in controlled research Dunlosky et al..

A simple evidence-aligned routine starts by defining the desired memory outcome. If the goal is to reproduce a definition unaided after one week, confidence should be judged against that task rather than against the ease of rereading it immediately. After an interval, attempt retrieval without consulting the answer, record uncertainty, check the response, and plan further practice based on the discrepancy. When accuracy matters beyond learning, verify important facts against reliable external records.

Avoid assuming that deliberate reflection alone creates accurate metamemory. It may help someone articulate an impression without making the impression better calibrated. Equally, no general evidence establishes that repeated use of a generic “brain training” app transfers into universally reliable memory self-awareness. Specific monitoring prompts and test formats are best evaluated for the population and goal where they will be used.

Practical examples

For studying, replace “I recognize this chapter” with “Can I answer an appropriately difficult question about it tomorrow without the chapter?” For an important meeting, replace “I remember the exact number” with “I can give the approximate range, and I will check the minutes for the precise figure.” For a recurring intention, replace “I will certainly remember” with a timed reminder if past experience shows that the task is easily missed. These choices connect memory judgments with observable outcomes and practical control.

Research Strengths, Limitations, and Open Questions

The central phenomena of memory predictions, feelings of knowing, retrospective confidence, and study regulation are supported by extensive experimental research. The monitoring–control distinction is a useful organizing model, and cue-based theories provide tested explanations for both insight and error. Evidence for retrieval practice and distributed learning is relatively strong in many educational contexts. More precise questions remain unsettled, including how individual differences in monitoring generalize across task types, how introspective reports relate to neural processes, and which training effects transfer durably to naturalistic memory decisions.

Many laboratory experiments use word pairs, short lists, and tightly controlled test intervals. These designs isolate mechanisms, yet complex everyday memories include social context, emotional significance, delayed reconstruction, source ambiguity, and competing goals. Results from student samples or brief laboratory tasks cannot automatically establish what happens in clinical settings or years-long autobiographical memory. Researchers also need to distinguish a change in stated confidence from a change in memory accuracy and from a change in subsequent behavior.

Measurement may itself influence what is being studied. Asking people to predict later recall can prompt retrieval attempts or strategy adjustments, making metamemory judgment both an observation and a potential intervention. The delayed-JOL debate illustrates this interpretive challenge. Strong future work requires preregistered comparisons where suitable, transparent measurement of confidence and task difficulty, sufficiently powered samples, and more representative outcomes across time.

Frequently Asked Questions

What is metamemory in simple terms?

Metamemory is your understanding and assessment of your own remembering. It includes knowing that you may forget something, predicting whether you can recall information later, estimating confidence in a recollection, and choosing strategies based on those judgments.

Is metamemory the same as memory?

No. Memory refers to encoding, storage-related processes, and retrieval. Metamemory refers to beliefs, predictions, confidence, and regulation concerning those processes. Someone can remember well without accurately estimating future recall, or remember less well while predicting their performance accurately.

What are examples of metamemory?

Examples include deciding to review a definition because you cannot recall it without notes, believing you will recognize a forgotten name, checking the source of a vivid recollection, and setting a reminder because you anticipate forgetting an appointment.

What is the difference between a judgment of learning and a feeling of knowing?

A judgment of learning usually predicts future remembering of material that has been studied. A feeling of knowing often estimates whether currently unrecalled information could be recognized or recovered with a cue. They arise at different points in learning and retrieval and can use different diagnostic cues.

Is confidence a reliable guide to memory accuracy?

Confidence can carry useful information, especially when it is linked to cues that genuinely predict performance. It can also be distorted by familiarity, fluency, and task conditions. The appropriate test is whether confidence and correctness correspond under the specific circumstances, rather than whether confidence feels strong.

Is the tip-of-the-tongue feeling proof that I know the answer?

No. A tip-of-the-tongue experience can accompany partial retrieval and may indicate some accessible information, but it does not guarantee that the correct target can be recalled or recognized. Its mechanisms and reliability depend on the task and the available cues.

How can I test my metamemory?

For a particular learning goal, predict whether you will remember each item after a defined delay, complete an appropriate recall or recognition test, and compare your predictions with actual outcomes. Several items and repeated occasions provide a more informative pattern than one question. This is an informal self-check, not a standardized clinical assessment.

Can metamemory be trained?

People can practice making predictions based on retrieval tests and use feedback to adjust studying. The effectiveness of any training depends on what is practiced and what outcome is tested. Evidence for improved study behavior should not be generalized to a permanent improvement in every dimension of metamemory.

Does poor metamemory mean a neurological disorder?

No. Everyone can misjudge memory under some circumstances. Clinical interpretation requires information about objective performance, daily functioning, course of symptoms, medical history, and appropriate assessment. Ordinary inaccurate predictions are not a diagnosis.

How does metamemory relate to AI tools and external memory aids?

People decide whether to rely on notes, calendars, search results, or AI-generated summaries partly through beliefs about what they know and what they might forget. Verifying an important claim against an original source remains necessary even when the generated explanation feels familiar or plausible. These are human monitoring and control decisions; they do not establish that an AI system has human-like metamemory or subjective experience.

Related Articles

Memory Retrieval: Recall, Recognition, and Cues — the first-order memory processes and tests that metamemory judgments attempt to evaluate.

Why Do We Forget? Causes and Theories of Forgetting — a broader account of retrieval failure, interference, encoding, and changes in accessible memory.

How to Improve Memory: Evidence-Based Strategies That Help — an EN-24 guide to retrieval practice, spaced review, and other memory strategies, with their evidence and limits.

Metacognition in the Age of AI: How to Monitor Your Thinking When AI Helps You Think — an existing English article on monitoring one's thinking when AI systems assist cognition. The EN-29 parent article on metacognition and the dedicated article on judgments of learning remain planned editorial connections and will be linked when their English pages are published.

References

Benjamin, A. S., Bjork, R. A., & Schwartz, B. L. (1998). The mismeasure of memory: When retrieval fluency is misleading as a metamnemonic index. Journal of Experimental Psychology: General, 127(1), 55–68. https://doi.org/10.1037/0096-3445.127.1.55

Brandt, M., Carvalho, R. L. S. de, Belfort, T., & Dourado, M. C. N. (2018). Metamemory monitoring in Alzheimer’s disease: A systematic review. Dementia & Neuropsychologia, 12(4), 337–352. https://doi.org/10.1590/1980-57642018dn12-040002

Dunlosky, J., & Rawson, K. A. (2012). Overconfidence produces underachievement: Inaccurate self-evaluations undermine students' learning and retention. Learning and Instruction, 22(4), 271–280. https://doi.org/10.1016/j.learninstruc.2011.08.003

Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students' learning with effective learning techniques: Promising directions from cognitive and educational psychology. Psychological Science in the Public Interest, 14(1), 4–58. https://doi.org/10.1177/1529100612453266

Dunlosky, J., & Thiede, K. W. (2013). Metamemory. In The Oxford Handbook of Cognitive Psychology (pp. 283–298). Oxford University Press. https://doi.org/10.1093/oxfordhb/9780195376746.013.0019

Hertzog, C., & Curley, T. (2018). Metamemory and cognitive aging. Oxford Research Encyclopedia of Psychology. https://doi.org/10.1093/acrefore/9780190236557.013.377

Karpicke, J. D., Butler, A. C., & Roediger, H. L., III. (2009). Metacognitive strategies in student learning: Do students practise retrieval when they study on their own? Memory, 17(4), 471–479. https://doi.org/10.1080/09658210802647009

Koriat, A. (1993). How do we know that we know? The accessibility model of the feeling of knowing. Psychological Review, 100(4), 609–639. https://doi.org/10.1037/0033-295X.100.4.609

Koriat, A. (1997). Monitoring one's own knowledge during study: A cue-utilization approach to judgments of learning. Journal of Experimental Psychology: General, 126(4), 349–370. https://doi.org/10.1037/0096-3445.126.4.349

Martín-Luengo, B., Zinchenko, O., Dolgoarshinnaia, A., & Leminen, A. (2021). Retrospective confidence judgments: Meta-analysis of functional magnetic resonance imaging studies. Human Brain Mapping, 42(10), 3005–3022. https://doi.org/10.1002/hbm.25397

Nelson, T. O., & Dunlosky, J. (1991). When people's judgments of learning (JOLs) are extremely accurate at predicting subsequent recall: The delayed-JOL effect. Psychological Science, 2(4), 267–271. https://doi.org/10.1111/j.1467-9280.1991.tb00147.x

Nelson, T. O., & Narens, L. (1990). Metamemory: A theoretical framework and new findings. Psychology of Learning and Motivation, 26, 125–173. https://doi.org/10.1016/S0079-7421(08)60053-5

Pannu, J. K., & Kaszniak, A. W. (2005). Metamemory experiments in neurological populations: A review. Neuropsychology Review, 15(3), 105–130. https://doi.org/10.1007/s11065-005-7091-6

Schwartz, B. L. (1994). Sources of information in metamemory: Judgments of learning and feelings of knowing. Psychonomic Bulletin & Review, 1(3), 357–375. https://doi.org/10.3758/BF03213977

Spellman, B. A., & Bjork, R. A. (1992). When predictions create reality: Judgments of learning may alter what they are intended to assess. Psychological Science, 3(5). https://doi.org/10.1111/j.1467-9280.1992.tb00680.x

Vaccaro, A. G., & Fleming, S. M. (2018). Thinking about thinking: A coordinate-based meta-analysis of neuroimaging studies of metacognitive judgements. Brain and Neuroscience Advances, 2, 2398212818810591. https://doi.org/10.1177/2398212818810591