My principal contribution comes from the proposal, with my colleague Graham Hitch, that the capacity to think, to learn, and to plan for the future all depend on a temporary memory system that can be divided into a small number of separate components. It was important to us that the resulting theoretical model could then be applied outside the laboratory to help understand a range of issues of practical importance. Our original idea has been more successful than we dared hope, with the term “working memory” having occurred in the title of over 170,000 papers – not all, of course, accepting our own theory, although our initial paper has been quoted more than 10,000 times. The idea of working memory combines two essential features – temporary memory, and its attentional control – both of which are limited in their capacity. Theorists vary in their emphasis on memory or attention, but all accept the need for both.

My own approach began with an emphasis on temporary storage and was strongly influenced by the controversy during the 1960s as to whether human memory involved separate short-term and long-term memory systems, or whether a single system could explain everything.

I began my research career in long-term memory, researching the way in which postal codes could be constructed so that people would find them easy to remember. I was working at the Medical Research Council Applied Psychology Unit in Cambridge, which specialized in linking the development of theory with its practical application. Having worked on postal codes, I was next given the task of attempting to improve methods of measuring the quality of telephone lines. The standard method was simply having people listen to potentially confusable words.

I proposed that having both to discriminate and to remember the words might make the task more sensitive, particularly since my boss, Conrad, had shown that similarity of sound made sequences much harder to recall. For example, a sequence such as b g c t d would be harder to remember in the right order than k w y q x. I decided to use words since this also allowed me to incorporate similarity of meaning, contrasting sequences such as man mat cat cap map or huge big wide long tall. I expected adding noise would make similar items much more difficult to remember in the right order. It did not.