Memory, we now understand, is not a recording. It is a reconstruction. Each time we remember an event, we do not retrieve a stored file and play it back intact. We reassemble the memory from fragments, and in the process of reassembly, the memory is subtly altered. New information encountered since the original event can be incorporated into the memory, making us genuinely believe we remember things that did not happen as we remember them, or things that did not happen at all. This is not lying. It is the ordinary operation of memory.
The implications for law are profound. Eyewitness testimony has historically been considered among the most compelling forms of evidence in criminal trials. But decades of psychological research have demonstrated that eyewitness accounts are frequently unreliable — not because witnesses are dishonest, but because memory itself is reconstructive. Factors such as stress, lighting, time elapsed, post-event information, and the phrasing of questions during police interviews can all alter what a witness sincerely believes they saw.
The psychologist Elizabeth Loftus pioneered this field with experiments in which subjects were shown footage of a car accident and then asked questions using different language. Subjects asked how fast the cars were going when they 'smashed' gave higher speed estimates than those asked about cars that 'hit' each other. Some subjects later reported seeing broken glass that had not been in the footage at all.
