Memory Is Not a Recording: What Investigators Need to Know

One mistake can distort an entire interview: treating memory like a recording.

If memory worked that way, the rules would be simple. If the account is consistent, the witness must be honest. If the account changes, the witness must be hiding something. If the account is detailed, it must be accurate. If the witness is confident, the account must be reliable.

But memory does not work that way.

Memory is reconstructed each time a person recalls an event. Later conversations, documents, assumptions, questions, and the passage of time can all affect what the witness remembers and how the witness describes it. The same witness may give different accounts of the same event in two conversations. Those differences matter and should not be ignored, but they are not, by themselves, proof that the witness is lying. Often, they are simply evidence of how memory works.

This article is about ordinary memory, not traumatic memory. Trauma raises different issues and deserves separate treatment.

Memory Is Rebuilt, Not Replayed

Many inexperienced investigators think of memory as something stored and retrieved. The witness saw the event, the memory was saved, and the interview simply pulls it back up.

That picture is too neat. Recall is constructive. The witness rebuilds the event from pieces: what she saw, what she heard, what she later learned, what she assumed, and what the question itself suggests. ^[1]

A witness who gives the same account in October and February is not automatically more reliable. She may have repeated the story enough times that it now feels settled.

A witness whose account changes is not automatically dishonest. She may have remembered something new, lost a detail, or absorbed later information without realizing it.

Detail and confidence do not solve the problem. Both can increase as a story is told and retold. A detailed account is not always accurate, and a confident witness is not always right.

Why Witness Accounts Change

When a witness’s account changes, the investigator should not jump straight to credibility. Changes in witness accounts generally fall into three practical categories.^[2]

1. Recovery. A witness remembers a detail that was not retrieved in the first telling but emerges later. This may happen because another question, document, or conversation triggered the connection. Recovery is normal. It is not, standing alone, suspicious.
2. Degradation. A witness loses detail over time. That is also normal, especially when the witness has talked about the event repeatedly and the original memory becomes harder to separate from later retellings.
3. Substitution. The witness replaces one detail with another. This deserves closer attention. It may be ordinary memory reconstruction, but it may also mean the account has been shaped by something outside the original event: another witness, a document, a lawyer’s framing, or an inference the witness now treats as memory.

Recovery and degradation are an unavoidable aspect of interviewing human beings. Substitution is different. It does not prove dishonesty, but it does require follow-up. What changed? When did it change? What happened between the first account and the later one?

Questions Can Contaminate Memory

Investigators do not just retrieve memory. Their questions can shape it.^[3]

A leading question can introduce a detail the witness did not independently recall. A repeated question can turn uncertainty into apparent confidence. A loaded question can change how the witness describes the event.

Ask a witness three times whether a manager was angry, and by the third answer the manager may sound angrier than he did at first. The witness may not be trying to please the investigator. She may believe she is remembering more clearly. But the investigator may have helped create the account now being recorded.

The better approach is to begin with open-ended recall. Ask what happened. Ask what the witness saw, heard, said, did, and understood at the time. Then narrow the questions without supplying the answer. “What did the manager say next?” is usually better than “Was the manager angry when he said that?”

And respect “I don’t remember.” That answer is not a failed interview. Sometimes it is the most accurate answer the witness can give. Pressing past it often produces a guess. Once the guess is spoken, it can harden into something that feels like memory.

The Investigator’s Discipline

The best account is not always the longest or most confident one. It is the account obtained through careful questions before the investigator’s words, assumptions, or pressure have reshaped it.

The point is not to distrust memory. The point is to handle it carefully.

A good investigator asks open questions, listens closely, documents accurately, and avoids forcing certainty where none exists. Memory is evidence. Treat it that way.

---