John F. Sowa
Natural languages are the ultimate knowledge representation languages. Everything that can be expressed in any artificial language or notation can be expressed with equal precision in a natural language. For many applications, however, artificial languages can be more concise or better tailored to the requirements. A key to understanding the relationships between natural languages and artificial languages is Wittgenstein's theory of language games: the semantics of a natural language consists of the totality of all possible language games that can be played with a given syntax and vocabulary, but an artificial language is designed for a single language game. The lexical ambiguities of natural languages result from the option of using and reusing the same words in multiple ways in multiple games. Those ambiguities can be eliminated by restricting the discourse to a single language game or a limited set of closely related games called a sublanguage. This talk will present examples that illustrate these points, discuss their implications, and suggest an approach that can relate Wittgenstein's insights to current research in linguistics and knowledge representation.
Subjects: 13. Natural Language Processing; 9. Foundational Issues
Submitted: Feb 2, 2006