How reliable is memory?

Have you ever had a conversation with someone about a shared experience, only to find you remembered it differently? Even more unsettling, have you ever questioned whether something you think you “remember” actually happened at all?

Until the middle of last century, it was assumed that memories were laid down rather like photographs, then retrieved when necessary in their original form. Thanks to the work of Endel Tulving in Toronto, Alan Baddeley in Cambridge and Elizabeth Loftus at the University of Washington, we now know that when we try to remember something, we call up only the essential bits. We then reconstruct — or invent — the details. Furthermore, those details are vulnerable to our circumstances when we’re recounting the memory.

Loftus conducted an experiment to see whether a memory can be altered depending on how the request for recall was worded. She showed participants a short video of a car accident. She then divided them into four groups and asked everyone to recall the crash, using a different verb for each group: “How fast were the cars going when they contacted/ hit/bumped/smashed one another?” The participants who heard the word “smashed” estimated a significantly faster time than the others.

If word choice at recall can alter a memory, Loftus wondered if word choice might even cause us to “remember” events that never took place. She recruited 24 adults and asked their families to describe events from their childhood. She selected three real events for each, but then made up a fourth — a story about getting lost in a shopping mall aged five.

Loftus presented these events to each participant and asked them to describe each one in as much detail as possible. Twenty-five per cent of them were convinced the false memory had really occurred and described it in some detail.

So, next time you wonder whether an event actually occurred, be assured your skepticism is justified. Odds are on your side that it did happen — but probably not quite as you remember it.