The LEXICAL System for
Computer-Aided Legal Instruction
In these times, when the roller coaster of artificial intelligence is at one of its zeniths (and before it crashes once again to yet another nadir), computer systems are built in a climate where the concepts of "intelligence," "expert system," and "knowledge" are the currency of academic debate. It becomes difficult to justify a system that attempts to match up to none of these. The process of academic legitimation almost seems to require that our researches are directed toward what is fashionable at this particular point in time. Undoubtedly, artificial intelligence is the current fashion in research in computers and law, and the drive toward achieving the goals of intelligent, decision-making legal systems can be seen wherever research is published in the field.
However, it seems that a plurality of research is a much more valid strategy-- particularly a plurality of research that involves debate between the various factions. Plurality, for one thing, ensures that we are not all jumping onto the same bandwagon in a vain attempt to keep up with fashion (remember Macaulay's pithy comment that fashion is but provincialism transferred from map to calendar). And, for another, plurality means that there is a meeting of ideas out of which we can refine and rethink our own research.
This, then, is the justification for the content of this chapter. On one level it is the description and discussion of the design of a computer-assisted learning (CAL) package. On another level it is a discussion of why, in the design of this package, artificial intelligence (especially expert system) techniques were not used. The goal is to press the point that we should not all follow the path of artificial intelligence in trying to find ways of applying computers to law. It seems important to clarify this, because the field of computer-assisted learning