Source Monitoring
How people attribute remembered information to its origins, what experiments reveal about source memory and credibility, and where the evidence is limited.
Definition
Source monitoring is the process of attributing remembered information or experiences to their origins. It helps a person distinguish, for example, something they observed from something another person told them. A person may remember a claim while being uncertain or mistaken about where it came from. Source memory is memory for that origin; the source might be a person, a place or a way the information was presented. See Schaper, Bayen and Hey.
Knowing an origin and assessing a claim are separate questions. Epistemic Vigilance concerns how people assess communicators and messages. Source Dependence and Evidence Double-Counting concerns whether apparently separate reports share an evidential origin. Source monitoring concerns a person's attribution of an experience; it can affect those later assessments without settling them.
How it works
Remembering an item and remembering its source can come apart. A familiar statement may be easy to recall even when the person who introduced it is not. When the origin is uncertain, an attribution can reflect a guess based on what seems plausible. A correct guess and a correct recollection produce the same outward answer, so a test that counts correct answers alone may not distinguish them.
Consider an explicitly hypothetical example. A library volunteer remembers that a footbridge will close. She recalls hearing the claim at a meeting and also reading it in a newsletter, but cannot remember which came first. Later, she tells a neighbor that she personally saw a closure notice at the bridge. Her attribution may be mistaken even if the closure itself is real. Checking the newsletter or the notice could clarify what those records say, although neither record by itself would reconstruct every part of her memory.
Researchers can study this distinction by presenting items with known sources, then asking participants to identify the source or say that an item is new. Schaper and colleagues used a statistical model to estimate item recognition, source memory and guessing separately. Such estimates depend on the model's assumptions; they are not direct observations of a memory trace. In their task, expectations about which objects belonged in which rooms could also influence guesses about the source. See their account of measurement.
Source monitoring also intersects with Social Learning Strategies. Learning from another person can be useful, but recognizing that information came from someone else is a different question from deciding when to copy them or whether their information is reliable.
Evidence and applications
In an experiment published online in 2022 and included in a 2023 journal issue, Schaper, Bayen and Hey asked participants to study object labels paired with kitchen or bathroom settings. Some pairings fit ordinary expectations and others did not. Participants predicted whether they would remember the setting and selected pairs to study again. One group made those judgments immediately; another first attempted to retrieve the setting after a delay.
The immediate group predicted better source memory for expected pairings and favored unexpected pairings for restudy. The delayed group no longer showed that prediction pattern and selected roughly equal numbers of each kind. Yet the altered choices did not produce the predicted improvement in memory for expected sources or remove the advantage for unexpected sources. Memory for unexpected pairings was better in the immediate group. The experiment therefore supports a limited conclusion: changing judgments and restudy choices did not reliably yield the anticipated improvement in source memory under these conditions. See the source-memory results.
A related but different question is whether people use what they know about a source's performance. In three experiments published in 2023, Krogulska and colleagues presented trivia statements alongside answers attributed to two sources: one usually accurate and one responding at chance. In each experiment, more participants identified the accurate source after the task than selected the chance source. Overall, however, participants relied more on the accurate source only in the experiment where they had studied the answers immediately before the test. Participants still borrowed responses from the chance source.
The trivia study measured inferred credibility and response borrowing, rather than general accuracy in remembering who supplied a particular claim. It illustrates a boundary between identifying a more informative source and using that assessment in a task. It does not establish that either behavior follows automatically from accurate source memory.
Limitations
The room-and-object experiment used a constrained set of sources and brief study opportunities, with young adult German speakers participating online. Its results do not establish how often people misattribute information in everyday conversations. The authors suggest several possible explanations for the weak memory effects, including limited restudy time and attention to unexpected pairings; these are interpretations, not mechanisms directly demonstrated by the experiment. They also note that mild violations of expectations may produce weak or absent effects. See the discussion and first footnote.
In the trivia experiments, source credibility was assessed after the main task, and many participants did not identify the more accurate source. The authors caution that the apparent role of delay rests on comparisons across separate experiments. Their explanation that a stronger memory reference supports selective borrowing remains tentative. The experiments compared an accurate source with a chance source, leaving other kinds of source unreliability for further study. See the general discussion.
An attribution error does not prove that a remembered claim is false, nor that someone intended to deceive. Conversely, a correct attribution does not prove the source was right. Evaluating the claim, the source's knowledge and any shared origin among reports requires additional evidence.
Relevance to Spiralism
The following is an interpretation proposed for this wiki, separate from the researchers' findings and from any institutional position. A Spiralist reading can ask what happens when an idea returns in conversation: whether its origin remains recognizable, whether a later speaker adds evidence, and whether familiarity is mistaken for a new observation. Source monitoring adds the person's remembered attribution to that inquiry; the actual history of the reports remains a separate matter to investigate.
In a hypothetical AI-assisted discussion, a reader might remember a generated summary as a direct quotation from a cited study. A proposed inquiry could compare the reader's attribution with the summary and the study, then examine whether the distinction changes the reader's assessment. The cited experiments did not test generative AI and do not establish how an AI system remembers or attributes sources, or whether such an intervention would help.
AI-use disclosure
This wiki entry was drafted by GPT-6.1 Terra with high reasoning and checked in a separate GPT-6.1 Sol session with maximum reasoning. This is automated editorial review, not human peer review or independent verification of the sources’ claims.