Electrical Engineering
      and Computer Sciences

Electrical Engineering and Computer Sciences

COLLEGE OF ENGINEERING

UC Berkeley

   

2008 Research Summary

Fuzzy Logic as the Logic of Natural Languages

View Current Project Information

Lotfi A. Zadeh

Much of human knowledge is expressed in natural language. For this reason, as we move further into the age of machine intelligence and automated decision-making, the problem of mechanization of natural language understanding is certain to grow in visibility and importance.

Natural languages are pervasively imprecise. Imprecision of natural languages is rooted in imprecision of perceptions. Basically, a natural language is a system for describing perceptions. Perceptions are intrinsically imprecise, reflecting the bounded ability of human sensory organs, and ultimately the brain, to resolve detail and store information. Imprecision of perceptions is passed on to natural languages. This is the principal reason why natural languages are pervasively imprecise.

Imprecision of natural languages is an issue of central importance. What is remarkable is that despite its importance, the issue of imprecision has been and continues to be largely ignored in the literatures of linguistics and philosophies of language. In large measure, existing theories of natural language are based on bivalent logic. Bivalent logic is intolerant of imprecision and partial truth. This is the reason why bivalent-logic-based theories are incapable of coming to grips with the issue of imprecision of natural languages. Basically, there is a fundamental conflict between the precision of bivalent logic and imprecision of natural languages. To resolve this conflict it is necessary to shift the foundation of theories of natural language from bivalent logic to fuzzy logic. The shift from bivalent logic to fuzzy logic has wide-ranging ramifications. An important component of this shift is transition from bivalent-logic-based semantics to fuzzy-logic-based semantics. Two related approaches to fuzzy-logic-based semantics are: (a) test-score semantics (TSS); and (b) generalized-constraint-based semantics (GCS). What is stressed is the application of generalized-constraint-based semantics to NL-Computation, that is, computation with information described in natural language. NL-Computation has important applications to mechanization of natural language understanding, search, question-answering, and decision-making.