Feature

Misremembering Dates

Science says it's perfectly natural to forget the dates of all sorts of important events.

Simon Kemp

If you are asked questions like "when did the Berlin Wall come down?" or "when was Winston Peters sacked from the National Cabinet?" or even "when did you last paint the house?" the chances are that your answers will be inaccurate. This result is interesting because it shows that our minds do not automatically record the time that events took place. Indeed, all the available research shows that we remember very little time information.

Another theoretically important result is that your estimates of when events took place are very likely to be biased. On average, people tend to date more remote events -- especially those that are more than about two years old -- too recently.

This intriguing finding, which has been replicated for a number of different types of events by researchers round the world, has naturally invited attempts to explain it. One explanation is that the finding naturally arises from two rather simple and basic properties of human memory.

The first property, which was known at least as far back as Aristotle, is that memory is often associative. Items of knowledge tend to recall similar or related items. So, events that we remember -- or are asked to date -- tend to recall similar or related events.

The second property is that we progressively forget details of knowledge. The longer it is since you learnt something, the less chance you have of recalling it accurately, regardless of whether it is lists of words, where you last put your keys, or the dates of events. Related to this finding is another one -- the events whose dates we do remember accurately tend to be recent ones.

Now suppose you are asked to date an event. It is possible that you can actually remember the date. If so, you respond correctly, and that is the end of the matter. But if, as is more normal, you have forgotten the date, you must try some other means. The associative theory asserts that you do so by casting about in your memory until you find a similar or related event that does have a date associated with it. You then simply take this date in place of the one you have forgotten.

This theory can be simulated on a computer and produces results which mirror what actually happens in practice. The further back in time, the more inaccurate the dating; and the more remote events tend on average to be dated too recently.

In essence, the theory "works" because the dates that we can actually remember tend to be heavily skewed towards the present. Hence, we anticipate more inaccurate dating of older events because recallable dates are sparse for more remote periods. The bias arises because, on average, the similar or related events for which dates can be recalled tend to be recent.

The theory agrees with people's experiences of consciously trying to date events. For example, you may remember that when you first heard about the Berlin Wall coming down you were working for a particular employer, and you recall that you were working for this firm over Christmas 1989. But the theory does not depend on conscious reconstruction (as the process is sometimes called) -- it does not matter for the theory whether you are aware of the way you search for the date or not.

Theories of this kind, which do not assume that we have a mental timeline but only scraps of reliable information about the time that past events occurred, have implications for the way you can come to intuitive conclusions about causation in the events you heard about or observed. Part of misremembering the dates of events is misremembering their order of occurrence, particularly if the events did not seem to be related. This means in turn that it will be very difficult to look back on the events you witnessed in your life and draw causal conclusions.

There are practical implications here. For example, people who suffer from tinnitus (a continual loud noise in the ears or head) are sometimes advised to keep diaries in which they record daily how bad their tinnitus is, and also what they eat and drink and how much exercise they take. It is only by making this record that enough time information is preserved for people to be able validly to draw conclusions such as "whenever I drink red wine, my tinnitus is worse the next day."

Intuitively, it sometimes feels to us that our minds are stuffed full with information. But much of this information now seems to be reconstructed from stored fragments rather than available to us in its entirety. Studies and theories of how we date events help us to understand how this process of reconstruction takes place and how it may on occasion fail us.

Simon Kemp is Associate Professor of Psychology at the University of Canterbury, Christchurch.