Case File 258 · Investigation Psychology

Why Can Two Honest Witnesses Remember the Same Event Differently?

Two honest witnesses can remember the same event differently because memory is not a recording. Each person encodes only a fraction of what happens, filtered by where they were standing, what they wer

By Murder Book Editorial TeamPublished by Gungrounds, the studio making Murder BookPublished 3 October 2026Updated 3 October 2026

Two honest witnesses can remember the same event differently because memory is not a recording. Each person encodes only a fraction of what happens, filtered by where they were standing, what they were paying attention to, what they expected to see and how they felt at the time. After the event, each memory is stored and reconstructed separately, shaped by different conversations, different exposures to media and different intervals before recall. The result is that two truthful, well-meaning people can give accounts that contradict each other on significant details without either of them lying. This is normal and well documented. It does not mean one witness is dishonest; it means memory works differently from how most people assume.

Key takeaways

Different witnesses encode different details depending on their position, attention and emotional state.
Memory is reconstructive: each recall can introduce changes shaped by post-event experiences.
Contradictions between honest accounts are expected, not evidence that someone is lying.
Post-event influences such as conversation, media and questioning can push accounts further apart or pull them together artificially.
Investigators should treat discrepancies as information about memory conditions, not as proof of deception.

Why do witnesses notice different things?

At the moment an event happens, each witness is in a different physical and psychological position. One may be closer, another further away. One may be facing the scene, another glancing sideways. One may be focused on a weapon, missing the face entirely, a phenomenon known as weapon focus. These differences in viewpoint and attention mean that each person's brain encodes a different subset of the available information.

Beyond physical position, prior experience shapes what gets noticed. A person with mechanical knowledge might register the make of a car. A parent might focus on a child in the scene. A person who has experienced violence might attend to threat cues and miss peripheral details. None of this is deliberate; the brain selects what to encode based on relevance, expectation and arousal, and those filters are different for every individual.

How do memories change after the event?

Even if two witnesses encoded similar information at the time, their memories can diverge afterwards through several well-documented processes:

Forgetting. Memory fades over time, and it does so unevenly. Some details drop out faster than others, as covered in how does time affect eyewitness memory?. Two witnesses interviewed at different intervals may give differently complete accounts simply because one has forgotten more.
Post-event information. News coverage, social media, conversations with friends and questions from investigators all introduce new details that can be absorbed into a witness's memory, as described in can new information change an old memory?. Two witnesses exposed to different post-event information may remember different versions of the same moment.
Memory conformity. If witnesses discuss the event with each other, their accounts can converge on a shared version that may not match what either of them originally saw. Conversely, if they do not discuss it, their accounts may remain different simply because each reflects a genuinely different experience.
Reconstruction. Each time a memory is recalled, it is rebuilt from fragments and filled in with plausible details. The filling-in process draws on general knowledge and expectations, which differ between people. Two witnesses recalling the same scene may reconstruct different versions without realising they have added anything.

Does stress affect this?

Yes. High stress at the time of an event narrows attention, a response that can sharpen memory for central details while weakening it for peripheral ones. But what counts as "central" depends on the individual. One witness under stress may focus on the attacker's face; another may fixate on the weapon or on an escape route. The result is that stress does not uniformly improve or impair memory. It reshapes it, and it reshapes it differently for different people, as explored in does trauma make memory more or less accurate?.

This means that two equally stressed witnesses to the same event can have strikingly different memories, both shaped by genuine stress responses but focused on different aspects of the scene.

Are contradictions a sign of lying?

Not on their own, and treating them that way is a common mistake. Research on deception detection shows that inconsistency between accounts is a weak cue to dishonesty. Honest witnesses contradict each other regularly because their memories are partial and reconstructed. In fact, perfectly matching accounts can be more suspicious than divergent ones, because they may suggest rehearsal or co-witness contamination rather than independent recollection.

The more useful approach is to ask why the accounts differ. Is the discrepancy in a peripheral detail that one witness might not have encoded? Does it concern something that happened quickly, in poor conditions, or under stress? Has one witness been exposed to information the other has not? These questions can explain the difference without assuming bad faith, a principle covered in why can truthful accounts be inconsistent?.

What should investigators do with conflicting accounts?

Rather than choosing one account and discarding the other, best practice treats each account as partial evidence:

Record each account independently, before witnesses have a chance to compare notes.
Note the conditions under which each witness observed the event: distance, lighting, duration, stress, obstructions.
Ask about post-event exposure: what coverage, conversations or other information the witness encountered before giving their statement.
Look for convergence on well-supported details while treating unsupported details with proportionate caution.
Seek corroboration from non-testimonial evidence: physical evidence, recordings, documents.

Discrepancies are not problems to be eliminated. They are data about what each witness experienced and how their memory has been shaped since, which is more useful than a forced consensus. An investigation that pressures witnesses into agreeing may produce a tidier narrative, but it sacrifices the diagnostic value of the original differences. Those differences can reveal which details were strongly encoded, which were marginal, and where outside influence may have played a role.

How does Murder Book use conflicting witness accounts?

Murder Book: The Art of Interrogation is built around the idea that people tell different versions of the same events, and the player's job is to work out what the differences mean. The game, developed by Gungrounds, an independent studio in Zagreb, Croatia, draws on real police footage and psychological interviews. It does not hand players a reliable narrator. Instead, it presents unreliable stories, narrative gaps and contradictions, and asks players to weigh them against the evidence.

That reflects the research: two conflicting accounts do not automatically mean someone is lying. They might mean witnesses saw different things, remembered at different times, or were influenced by different sources. Cases are designed for focused 5–10-minute sessions. Murder Book is in development for PC (Steam) and mobile, with release currently planned for the end of Q2 2027.

Sources