How important do you think that it is to recognize when you are about to make an error? If you rate it as pretty important to you, then you will agree that that sort of recognition would be something very meaningful. The very act of distinguishing the error is of perceptual significance and personally meaningful.
Wouldn’t it be nice if a semantic search algorithm can distinguish a bad or false hit (an error) from a good or positive one just as we can?
Recent research (you can read about over at Science Daily: Why We Learn From Our Mistakes) shows that our brains are built for recognizing errors.
Science Daily — Psychologists from the University of Exeter have identified an ‘early warning signal’ in the brain that helps us avoid repeating previous mistakes. Published in the Journal of Cognitive Neuroscience, their research identifies, for the first time, a mechanism in the brain that reacts in just 0.1 seconds to things that have resulted in us making errors in the past.
This is also so universally applicable to human nature that human language has a built in semantic domain to identify, distinguish, communicate and control sense datum such as an error or a deviation.
Some people may be surprised to learn that there are also inheritance rules that require consistency of the semiotic signals as they propagate the properties, characteristics and all variant interpretations of the sense data through all possible symbolic compositions and permutations. The symbols “e r r o r” spell out the property or characteristic to convey, linking the sense datum to the lexical element of the English language in this case. This may seem preposterous without knowing the theoretical basis for that link, but that does not make it so.
Think about it this way: Failures, errors, faults, deviations, in body, mind, health, work, opinion, belief, character: Does the interpretation of these sorts of things occupy any part of your thoughts? Do you talk about them, communicate with others: your spouse, your preacher, your friend, your dog? Do we debate them socially, nationally, culturally? Do they go away if you do nothing or do they stay with you, even grow and propagate? If you don’t get that, consider this: It is clearly an innate function of interpersonal communication, among all species, to observe and alert others of error, deviation, or danger.
It is what we talk about, if you take some time to consider it. Language rests on the ability to interpret such sense data and we have highly refined symbols, semiotics and communications systems to propagate such signals.
The syntactic mechanics and semantics for distinguishing and interpreting sense data makes a fine basis for a parsimonious and scalable computational model of human language. Perhaps it also plays a significant role in the psychological and social basis of meaning and interpersonal communications. Because this particular semantic mechanism has been developed into a functional computational model it presents the possibility of a new direction for research on functional models of language and cognition.
Unfortunately, the research scene is tenured and business and government research and development is incestuously imitative. It is almost unheard of to break from the recent past, in computation, in linguistics, and in philosophy and logic. I started blogging about semantic search mainly because I have direct experience developing and fielding semantic search applications since before the Internet was much more than a figment of people’s imaginations. That is a long time– not as long as John Sowa perhaps, but a very long time nonetheless.
For as long as I have been in the business, the semantics of personal meaning and perception have always been a kind of nebulous subject in the computing sciences where they invent and use formal programming languages with formal logic and accompanying semantics.
Nevertheless, without the slightest consideration of interpersonal meaning and human perception, it seems every young computer science graduate thinks they have the stuff to process natural language statements, messages and texts, simply by building a few arrays and parsers and doing some table look up. Those that try fail.
They fail for many varied reasons, but mostly they fail for lack of a well-grounded and universal semantic theory of natural language and human cognition. It has been that way since before computer science began being offered as a course at major universities. It is that way today.
I guess there is hope and we can have some faith that people are built with the faculties to eventually realize they have made a mistake and orient in a new direction. What do you think?
