Human and Machine Understanding of Natural Language Character Strings
March 6-9, 2009 - Arlington, Virginia
Peter G. Tripodes. There is a great deal of variability in the way in which different language users understand a given natural language (NL) character string. This variability probably arises because of some combination of differences in language users’ perceptions of its context-of-use (pragmatics), identity and mode of organization of its meaning bearing parts (syntax), and in the meanings assigned to those parts (semantics). This paper proposes a formalization of the syntax and semantics of NL character strings within a logical framework which is sufficiently flexible to represent the full breadth of possible ways of understanding NL character strings as influenced by different contexts-of use, beyond what can be represented in currently used predicate-logic-based frameworks. While the question of how language users understand NL character strings is ultimately a question in the psychology of language, it appears to us that, for the purposes of AGI, that question needs to be addressed within a logical framework which explicitly identifies the syntactic and semantic components that comprise that understanding, and which account – in formal terms – for differences in that understanding. Such a logical framework would provide a formal basis not only on which to address further psychological issues regarding human language understanding, but also for coherently imparting such understanding to machines.