Friday, September 30, 2005

Subsets of language and intelligence

Random. unresearched RFC: The play between language and thought is ongoing, but is there a parallel to be drawn in the world of programming? Where's the separation and interedependency between the lexical/syntactic aspect and the semantic aspect of any programming "language"? Are concepts (such as procedures, datatypes, objects, etc) influenced by the keywords as much as the keywords used are influenced by the original concepts set out by the inventors of the language? Is how a language is used (its "deeds", perhaps) partially based on what the language sets out to contain, and partially on how it expresses itself?

By way of explanation, I was wondering about machine intelligence, and intelligent design, for some reason. (AI is, by definition possibly, "intelligent design".) The relationship between concepts and internal representation, and interfacing with non-like organisms such as ourselves, through language, is still in its infancy, I feel. To get workable linguistics, you need a working internal representation - context. But all of the computational contexts are defined in terms of our own language - that's the way computers work currently. Is this, then, a paradox? Can we ever achieve a workable AI capable of linguistics if its "universe" is already defined and restricted via our own linguistics?

No comments: