Quick, remember a birthday party.

Got it? Okay, in that memory, what is everyone wearing? You might say "I don't know", or "I can't remember", but they are clothed in the memory. The picture you've recalled contains something to cover each person at the party, and each thing is somehow labeled with a surety, with how much you can trust that part of the image. As memory seems to be largely structural and holistic (more on this later) if you probe any particular detail, any particular part of the scene, you will never find blankness, but rather further elaboration of detail which you, for some reason, don't trust.

Now, its probably best that you have a good idea of what not to trust when you attempt to recall details or events, because most of your actual memory is elaboration as well, and without some limits, you'd have no real basis in reality. For instance, while you might think that you cannot recall what a particular friend was wearing (meaning that you don't trust that part of the memory), you probably believe think you can recall what they looked like. This memory, what your friend looked like, is cross woven with your idea of your friend, and is largely informed by what that friend always looks like (more on that later as well).

This synthesis is the primary function of memory, it provides context specific answers - "Where is the bathroom?", "What is her name?". But how are these memories requested? When we try to remember something, don't we construct a context, and see if the answer which is filled in "trustable"? Aren't we frequently wrong about when and where something happened? (The answer to that one is yes, ask ANY police officer.)

What if we were to deliberately compose contexts which we had never experienced? It is possible to "remember" things which one does not know, provided that one knows a good deal of the pieces; the memory is very quick and assembling them, since those pieces alread share associative relationships. I will call this, seeking answers through false memories, mnemonic reasoning.

Due to the increadibly parrallel nature of recall, there isn't any real delay involved, most recall takes about the same amount of time. The time involved is so short, it is always possible to attempt to "remember" through mnemonic reasoning as the first step in any problem solving process. Unfortunately, this produces many false answers, but there are mittigating factors:

• For many problems, checking an answer is so much faster than creating one, checking a guess as a first step is not expensive.
• For many problems, refinement may be used to improve an initial solution, and so guessing provides quick entry into the space.
• When you are wrong, it means that some set of associations in your memory is incorrect, and it behooves you to track them down. Regular appliaction of this technique will slowly clean out these misconceptions.

So in the end, what are we left with? As many answers as we want, but most of them incorrect? A growing surety that our entire base of knowledge is built upon tissue? A crushing distrust of anything we think we remember?. Some people who walk down this road do just that. They become obsessed with the need to know why they think things (without, for some reason, asking why they are obsessed).

However, I think that there is another way. Language (langue, as Sausure would say) is reasonably stable, while Speech (parole) is ever flexible, and defined only in some definite context. As I do not miss sleep from wondering who "he" or "it" is referring to, so to shall I not fear the many things which I know only within some very specific, and often ephemeral context. Mnemonic reasoning producess a great many things which are locally true, it provides a collection of concepts to guide future search and analysis, and a basis for action when there is insufficient time to perform that analysis.