Who has not got into a silly dispute with a loved one or friend about a minor point of memory - who said or did what, and exactly when? Generally, people are so attached to their version of events - and are so certain of their verity - that they dispense disproportionate amounts of energy on the smallest point, such as whether something happened on a Tuesday or Wednesday. We defend our memories with such energy because they represent important elements in our sense of reality. If we are wrong, it means our grasp of life is not as firm as we would like to think. Psychologists who study memory, on the other hand, learn to have a very modest view of their own memory abilities because they witness daily how profoundly flawed is the process of remembering. In fact, the common view - consulting a sort of reference library or massive hard disk - has long been discredited in academic circles. What people actually do when they remember, so current thinking goes, is closer to creation than retrieval. Memory is one of the most studied areas in psychology because it is one of the most crucial aspects of the mind. Theorists and researchers grapple with the vast and complex subject in a variety of ways. One approach is to describe memory in terms of how it appears to be organised. A second is to look at the many sorts of processes it seems to encompass. A third is to study its capabilities - the various things memory apparently can (and cannot) do. These three characterisations inform three popular models of memory, which underpin virtually all the research done. These models, or ways to think about memory, are interesting because they reveal the progression of expert thinking on the subject. But together, they also offer an explicit example of what for most people remains a fairly fuzzy notion: how technology deeply influences the whole tone of an era's thinking, even with respect to the most intimate areas of our mind. The first model of memory, and probably the most influential to date, is the serviceable box-and-arrow image which psychologists call the modal model. It still looms large in lay conceptions of memory and this is how memory is typically portrayed in the media. Memory, according to this model, is like a giant mechanical conveyor belt. Information is processed and moved from one location to another, say, from short-term memory to long-term storage. But as technologies advanced, the mechanical and the linear gave way to the electronic and the global. Hence, the more recent neural net model which stresses all-over connectedness and explains memory in terms of how connections emerge and change. A third paradigm is the functional approach, which simply focuses on what memory can do without recourse to a grand theory. This mirrors, perhaps, a greater acceptance of fragmented and parallel modes of thinking and more tolerance for gaps in knowledge. The metaphors that correspond to these three models of memory show how technology shaped the thinking of their respective times: the first is mechanical, the second computer-like and the third avoids metaphor by operating on a phenomenon-by-phenomenon basis. There is comfort in the mechanical; it represents certainty and knowability. This contrasts with the relative impenetrability of concepts like neural networks. I believe this is one of the reasons why the box-and-arrows model endures in the popular mind. Humans will never tire of being able to take apart a machine and discover how it works - without a degree in a relevant specialism. This does not mean that memory works in a mechanical way. It almost certainly does not. To get back to the argument about Tuesday or Wednesday, what actually happens is that a complex matrix of influences combines to influence a recreation of the past. An event is not lodged in some more or less accessible part of the brain. The ingredients of the memory process are not as neutral as that suggests. They include confounding factors like an individual's personal agenda (to score points in an argument or unconsciously portray themselves in a flattering light) along with other very weighty factors, such as past patterns of events, rather than the individual event in question. The popular appeal of an academic model beats intellectual coherency when it comes to real life, of course. But it may make sense the next time you get into a memory argument to recall just how imprecise and biased your memory can be. Jean Nicol is a psychologist specialising in issues of cultural identity and change in an era of globalisation everydaypsychologist@yahoo.com