USAGES OF NATURAL LANGUAGE. PART I. BASIC PRINCIPLES. PART II. DISCUSSION.

Abstract

Research is summarized that is directed toward the goal of developing a precise model of the notation 'able to speak a language.' A language is defined as that relationship between the inputs to and the outputs from a machine that determines the outputs that would be considered suitable by a native speaker. Requirements are discussed for a linguistic theory that designates a class of suitability relationships, a learnability theory that designates a training program and a learning algorithm will subsequently be formulated to permit the learner to determine which of the languages allowed by the linguistic theory is being presented. Several linguistic theories are developed based on three principles: (a) that information about utterances and context be in the form of a large number of 'units of information' or descriptors; (b) axioms and rules of inference provide the rules of grammar; and (c) a method of discretization performed by a finite set of predicates provide for continuous-valued functions of time, time also being continuous valued. A body of axioms designated sensory axioms provide initial sensory information and are added to the rules of grammar to permit the derivation of pattern recognition, syntactic, and semantic information. (Author)

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1966
Accession Number
AD0644521

Entities

People

  • E. Mark Gold

Organizations

  • Stanford University

Tags

DTIC Thesaurus Topics

  • Algorithms
  • Grammars
  • Language
  • Learning
  • Linguistics
  • Natural Languages
  • Notation
  • Pattern Recognition
  • Recognition
  • Social Sciences
  • Training

Readers

  • Mathematical Modeling and Probability Theory.
  • Neural Network Machine Learning.
  • Speech Processing/Speech Recognition.

Technology Areas

  • AI & ML
  • AI & ML - Machine Translation