The witness who was sure

Few things in a courtroom are as persuasive as a witness who points at the defendant and declares, without hesitation, "That is the man." Jurors tend to take such certainty at face value, and it is easy to see why. We like to think of memory as a kind of video recording: imperfect, perhaps, and occasionally blurred, but fundamentally a faithful copy of what happened. Decades of psychological research suggest that this picture is not merely incomplete but, in important respects, mistaken.

Memory is better understood as a reconstruction than as a recording. Each time we recall an event, we rebuild it from fragments, filling the gaps with assumptions, expectations and details absorbed after the fact. In a classic study from the 1970s, the psychologist Elizabeth Loftus showed participants a film of a car accident and then asked how fast the cars had been going when they "hit" or "smashed into" each other. Those who heard the word "smashed" gave higher estimates of speed and, a week later, were more likely to report having seen broken glass, although there had been none in the film.

Such findings would be of purely academic interest were it not for their consequences. In the United States, analyses of convictions later overturned on the basis of DNA evidence have found that mistaken eyewitness identification played a part in roughly seven out of ten cases. In many of them, the witness was not lying; they were sincere, confident and, as it later turned out, entirely wrong. It follows that sincerity alone is no guarantee of accuracy.

Confidence, in particular, is a treacherous guide. A witness who is hesitant at first may grow steadily more certain in the months before a trial, especially if a police officer offers a word of encouragement after an identification, or if the witness repeatedly sees the suspect's face in the media. By the time they reach the witness box, their certainty may owe more to repetition than to the original event. Paradoxically, then, the very process of preparing a case can make a witness more confident and less accurate at the same time.

None of this means that eyewitnesses are worthless, nor that memory is hopelessly unreliable. Most of what we remember about our own lives is broadly accurate, at least in outline, and witnesses often supply crucial information that no camera recorded. The devil, as so often, is in the detail: a face glimpsed for a few seconds, the exact order of events, words spoken under stress. Unfortunately, these are precisely the details that human perception handles worst, and yet they are often the ones on which verdicts turn.

On the strength of such research, some legal systems have begun to change their procedures. Line-ups are increasingly conducted by officers who do not know which person is the suspect, so that they cannot unconsciously steer the witness. Witnesses are asked to state how confident they are immediately, before that confidence can be inflated by feedback. In some jurisdictions, judges now instruct jurors on the limitations of memory, although the evidence that such warnings work is mixed: a brief caution from the bench may struggle to compete with the emotional force of a confident witness.

More broadly, the implications reach far beyond the courtroom. If memory can be reshaped by a single word or a reassuring nod, then all of us are, to some degree, unreliable narrators of our own lives. That need not be cause for despair. It is, however, a good reason to hold even our most cherished recollections a little more lightly, and to give others the same benefit of the doubt that we would hope to receive ourselves.