A witness remembers a meeting but cannot give its date. They recognise the people involved but cannot reproduce their exact words. Years later, their account differs from their first statement. Do these differences mean that the witness is lying?
Assessing testimony requires a distinction between a witness's sincerity and the accuracy of their recollections. A person can honestly describe an event and still be mistaken about its details. Equally, the fallibility of memory does not automatically explain every inconsistency.
The procedural provisions discussed in this article concern Polish civil and criminal proceedings.
What is the Ebbinghaus forgetting curve?
The forgetting curve describes how retention of previously learned material changes over time. It is named after Hermann Ebbinghaus, who began systematic experimental research into memory in the nineteenth century.
He learned sets of syllables intended to have no meaning. This choice was designed to limit the influence of prior knowledge and associations: a familiar sentence may be easier to learn because it makes sense, while an arbitrary sequence of syllables benefits less from that advantage. [6]
After learning a set, Ebbinghaus returned to it after a specified interval and measured the work needed to learn it again. A key measure was “savings” in relearning: the reduction in effort compared with the initial learning.
A simple hypothetical example illustrates the distinction. If initial learning took 20 minutes and relearning took 10, the saving would be 50%. This would not mean that the participant could independently recall exactly half the information. Relearning savings and the number of details freely recalled are different measures.
The findings showed rapid forgetting initially, followed by a slower rate. In a graph, time appears on the horizontal axis and memory performance on the vertical axis. The line first falls steeply and then more gradually. Information is therefore not lost at an equal rate each day. [6]
How did Ebbinghaus conduct his experiments?
Ebbinghaus was both the researcher and the participant. His main experimental series took place in 1879–1880 and 1883–1884; he published the results in 1885. This was not research on a representative group of witnesses or people involved in real events. [6]
- Preparing the material. He constructed sets of syllables generally consisting of a vowel between two consonants. He tried to limit meaning and associations, while acknowledging that differences in ease of learning could not be eliminated completely.
- Learning under controlled conditions. He read and repeated whole sequences aloud at a controlled pace. He sought comparable conditions across trials, including time of day and limited distractions.
- Checking mastery. The criterion was correct reproduction from memory. In the first research period he required two successive error-free reproductions; later he used the first fluent reproduction.
- A delay followed by relearning. After intervals ranging from minutes to many days, he learned the previously mastered sequences again.
- Comparing effort. He compared the work required for initial learning and relearning. The resulting savings indicated the persistence of effects from earlier learning.
The experiment did not count true and false recollections of an event. Its findings describe a particular measure of memory under experimental conditions. That difference matters when discussing courtroom evidence. [6]
Why can the curve not be applied mechanically to witnesses?
Ebbinghaus studied learning under controlled conditions, not the truthfulness of accounts of real-life events. He himself noted the limits on generalising his findings. His research provides no conversion rule for calculating how many details a particular witness should remember after a month or five years. [6]
A court case also requires asking whether a detail was noticed at all. Someone who did not check the time during a meeting may never have known it. That differs from forgetting a time they previously read on a clock.
The forgetting curve helps frame questions about memory; it is not a truthfulness test. It cannot determine whether a particular gap results from forgetting, failure to perceive something, or deliberate concealment.
Sincerity and accurate recollection are different issues
Memory does not function like an unchanging recording. Information can be perceived inaccurately, forgotten, or modified by subsequent experiences. Both the original observation conditions and events between the incident and the interview therefore matter. [2]
Assessing evidence involves two separate questions: is the witness deliberately misleading, and does the account accurately describe reality? Finding a witness sincere does not settle the second question.
Consider a dispute about an oral agreement on remuneration. A witness may remember that payment was discussed but not whether the amount mentioned was a final offer or merely a starting point for negotiations. That distinction may be decisive.
Do inconsistencies disqualify testimony?
Each discrepancy must be assessed against the fact to be proved. A mistake about the weather may be peripheral to a contractual dispute but important when evaluating visibility at an accident scene.
A practical analysis should establish:
- which fact the discrepancy concerns and why it matters;
- when and in what circumstances each account was obtained;
- whether the difference can be explained by the questions asked or the scope of the earlier interview;
- whether other evidence supports the particular version.
A minor discrepancy does not establish a lie, and a general reference to forgetting does not resolve a serious contradiction. The explanation must address the particular account.
A question can affect the answer
In the classic study by Elizabeth Loftus and John Palmer, participants watched films of car accidents. Changing the verb in a question affected their speed estimates. In the second experiment, a particular wording also increased subsequent reports of broken glass that had not appeared in the film. [3]
This illustrates why it is useful to examine both the answer and the question that preceded it.
“What happened after that person entered the room?” leaves room for the witness's own account. “Did he start threatening the defendant when he entered?” introduces a particular version of events.
Article 171 § 1 of the Polish Code of Criminal Procedure provides for a free account first, followed by questions to supplement, clarify or test it. Article 171 § 4 prohibits questions suggesting the content of an answer. [5]
Does the witness remember the event or a later discussion?
Conversations, media reports and materials seen later can influence recollections. A witness may also struggle to distinguish personal observation from information learned afterwards. Subjective confidence alone does not guarantee accuracy. [2]
Useful questions about the source of information include:
- What did the witness personally see or hear?
- What did they learn from others, and when?
- Whom did they discuss the event with before giving evidence?
- Which documents, messages or recordings did they review?
- What is the basis for the date or words they give?
Reviewing a document does not in itself disqualify a witness. Personal recollection must nevertheless be distinguished from information reconstructed from that document.
Practical significance in civil cases
Memory issues arise in oral agreements, financial settlements, and employment, family and inheritance disputes. A general recollection of a situation should be distinguished from memory of the detail on which the outcome depends.
Example: an oral loan. A witness remembers an envelope changing hands but did not see its contents or hear the terms. The account may establish an object being handed over, but does not itself explain the amount, the legal basis for the transfer or the repayment date. The fallibility of memory cannot fill those gaps.
Example: a discussion of remuneration. Years later, a witness gives an exact amount and date. It is useful to ask whether these come from recollection of the conversation, messages reviewed later, or information provided by a party. Each answer points to a different source.
A timeline is helpful: the event, the first recorded account, later statements, and review of documents. It can show when a disputed detail first appeared and how the witness explains its origin.
Article 233 § 1 of the Polish Code of Civil Procedure requires comprehensive consideration of the evidence. Article 271 § 1 addresses the source of the witness's knowledge. The practical implication is that a memory-based challenge should identify a particular fact and its relationship to other evidence. [4]
Saying “the witness could not remember after so many years” does not explain why a particular statement should be rejected. Equally, detail alone is insufficient to establish accuracy.
Practical significance in criminal cases
Memory analysis can be particularly important where a dispute concerns identification, violence, threats, or the sequence of participants' actions.
Example: a brief observation. A witness identifies a perpetrator long afterwards. Relevant questions include the duration and distance of observation, lighting, and whether the person was already known to the witness. Considering the delay before the interview cannot replace examining the original opportunity to perceive the person.
Example: new details in later statements. An initial account describes an argument, but only a later one quotes a specific threat. It is necessary to ask whether the witness was previously asked about those words, whether the earlier account was complete, and what enabled later recall. Omitting a detail and denying that it occurred require separate analysis.
Comparing answers with the questions that prompted them, and establishing whether the witness encountered others' accounts before a later interview, can help verification. These are lines of inquiry, not ready-made grounds for finding testimony false.
Article 7 of the Polish Code of Criminal Procedure requires assessment in accordance with sound reasoning, knowledge and life experience. If doubts that cannot be resolved remain after the evidence has been taken and properly assessed, Article 5 § 2 requires their resolution in the accused's favour. A conflict between accounts does not automatically trigger this rule. [5]
When is expert knowledge needed?
Not every memory gap requires a psychological opinion. In criminal proceedings, Article 192 § 2 provides for questioning with an expert's participation where specified doubts arise, including about perception or reproduction of perceptions. In civil proceedings, Article 272¹ concerns doubts about the ability to perceive or communicate perceptions. [4][5]
An expert supplies specialised knowledge within the scope of the opinion requested. Assessing credibility and establishing facts remain the court's tasks. The Ebbinghaus curve is not a diagnosis of an individual witness's memory abilities.
What should a witness do if they cannot remember?
A witness should clearly distinguish recollection from assumptions and information obtained from others. If they cannot recall an exact date or the words of a conversation, they should say so. Trying to match an answer to the questioner's expectations makes it harder to establish the true extent of their knowledge.
The Ebbinghaus curve helps explain the importance of elapsed time. The value of particular testimony nevertheless depends on analysis of its content, sources and circumstances alongside the other evidence.
A sound assessment examines both what the witness says and how they know it.
Sources
- J.M.J. Murre, J. Dros, „Replication and Analysis of Ebbinghaus’ Forgetting Curve”, PLOS ONE, 2015. Open source.
- National Research Council, „Identifying the Culprit: Assessing Eyewitness Identification”, 2014, chapter 4; the report primarily concerns eyewitness identification. Open source.
- E.F. Loftus, J.C. Palmer, „Reconstruction of Automobile Destruction: An Example of the Interaction Between Language and Memory”, 1974. Open source.
- Polish Code of Civil Procedure, Article 233 § 1, Article 271 § 1 i Article 272¹. Open source.
- Polish Code of Criminal Procedure, Article 5 § 2, Article 7, Article 171 § 1 i 4 oraz Article 192 § 2. Open source.
- H. Ebbinghaus, Memory: A Contribution to Experimental Psychology (1885/1913), chapters III and VII. Experimental method; Results on retention over time.