The Retrieval Path

Share Cognitive Science Implications for Teaching Mathematics
•
•
•
•
•
•

Stored somewhere in your long-term memory is your social security number.  In order to use that number in any fashion –write it, say it, compare it to another number, or just to think of it – the social security number must be transferred from long-term memory to short-term memory (see caveat below). To accomplish that transfer the information (SS number) must be transmitted via electrochemical processes though millions of neurons, axons, synapses, and dendrites.  The collection of neurons, axons, synapses, and dendrites and the order of involvement is called the RETRIEVAL PATH.

There are probably millions of possible retrieval paths which the brain could use to bring that SS number from long-term memory to short-term memory.  If you practice retrieving your SS number often and at appropriately spaced times, then one of those possible retrieval paths becomes preferred – strengthened.  When a retrieval path is strong, it is easy to recall your SS number quickly.  Note that recalling the last four digits of your SS number is a quite different activity and a different retrieval path is used to recall your “last four”.

Suppose that for a long period of time (a year or two) you are asked only to provide your “last four”.  As a result of lack of use the preferred retrieval path for your SS number becomes weak.  If at the end of two years you are suddenly required to provide your SS number, you will probably falter, be slow to remember the number, or you might even get it wrong. Appropriately spaced retrieval practice – simply recalling or writing your SS number once every two or three months – would have maintained the strength of the originally preferred retrieval path and you would have avoided that embarrassing moment.

Caveat: Perhaps it would be better to claim that the information must be transferred to working-memory rather than short-term memory.  I find the literature to be unclear with respect to the role of working memory.  There does seem to be consensus that working-memory is not completely distinct from short-term memory.  I will use that as my excuse for avoiding the use of working-memory terminology.