Goals and Results

  • Roland Hausser
Part of the Symbolic Computation book series (SYMBOLIC)


How do people transmit information with natural-language symbols? The literature in philosophy, linguistics, psychology, and artificial intelligence dealing directly or indirectly with this question is substantial. But it provides us only with partial answers based on a wide range of conflicting assumptions. The linguistic analysis presented in this book will clarify the basic issues, and provide a unified perspective of natural-language communication. Ultimately, it will lead to the design of a natural-language communicating robot, or NLC-robot.1


Natural Language Machine Translation Literal Meaning Linguistic Analysis Finite State Automaton 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 2.
    Provided by the author.Google Scholar
  2. 3.
    The “need to integrate physical actions and linguistic actions into a single planning system” is illustrated in Appelt (1982).Google Scholar
  3. 4.
    Weizenbaum (1966).Google Scholar
  4. 6.
    I.e., from surfaces to literal meanings and from literal meanings to surfaces.Google Scholar
  5. 7.
    This notion is explained in 2.4.1. 8 See examples 2.1.2 and 3.3.4.Google Scholar
  6. 9.
    Violation of the Derivational Order Hypothesis 2.4.2.Google Scholar
  7. 10.
    See Section 3.5.Google Scholar
  8. 11.
    Also, they cannot be falsified, in the sense of Popper (1959).Google Scholar
  9. 12.
    The evolution of the notion “generative” is described by Hymes and Fought (1981), pp. 165 – 174, in the historical context of American Structuralism.Google Scholar
  10. 13.
    Furthermore, the attempt to explain a wide variety of phenomena in terms of the four elements fire, water, earth, and air—as well as the promise of turning base metals into gold—constituted a fascinating intellectual challenge even to Isaac Newton, who devoted the latter part of his life to the study of alchemy.Google Scholar
  11. 14.
    “A11 science may indeed be a mythology, but not all mythology qualifies as a science.” Sowa (1984), p. 355.Google Scholar
  12. 15.
    This question is the premise of Hausser (1984a).Google Scholar
  13. 16.
    Woods (1970).Google Scholar
  14. 17.
    Kuno & Oettinger (1963).Google Scholar
  15. 18.
    Mareus (1978).Google Scholar
  16. 19.
    For example, the Earley algorithm is of complexity n 2 for unambiguous context-free grammars. The set of languages accepted by unambiguous context-free grammars is considerably smaller than that accepted by unambiguous C-LAGs.Google Scholar
  17. 20.
    Markov (1954), Yngve (1955). A brief historical summary of Markov models may be found in Damerau (1971). See also Hutchins (1986).Google Scholar
  18. 21.
    The time-linear derivation order and the computation of possible continuations on the level of “next words” was developed during the attempt to implement the categorial grammar defined in Hausser (1984a) as an efficient parser. The partial similarity with Markov models was realized later.Google Scholar
  19. 22.
    For a summary of the considerable cognitive science literature on this subject see Rumelhart (1977), pp. 219 f. See also Brachman (1979) for a review of data-structures called “semantic nets.” 23 See Salton and McGill (1983)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1989

Authors and Affiliations

  • Roland Hausser
    • 1
  1. 1.Laboratory for Computational LinguisticsCarnegie Mellon UniversityPittsburghUSA

Personalised recommendations