ple tense (past, present, future), aspect (indefinite, complete, continuing, etc.), logical quantifiers (every, one, none), non-logical quantifiers (many, most, some, few, etc.), multi-sentence discourse, modalities, beliefs, generics, anaphora, speech acts, meronymy, synonymy, hyponymy, polysemy, and others.

Secondly, the notion of user-friendliness must be addressed more precisely. While no evaluation methodology has yet been determined, several options are being considered. One option is users evaluation, where a selection of computer scientists would become users, and would be asked to formulate a number of meanings. Time taken to formulate a response, its accuracy with respect to the intended meaning, and general user observations would all become evaluation criteria. The task could also proceed in reverse: users would be given a semantic expression and asked to summarize its meaning using natural language. Again, time, accuracy, and user observations would become evaluation criteria. Another option is a study involving expression of complex narratives, where the number of semantic primitives introduced is used to measure complexity. Too many primitives make a cumbersome formalism; the extreme case being a primitive for each completion. Too few primitives would also exert a negative effect by adding a layer of notational complexity for complex phenomena and possibly impacting expressiveness. Here, the ratio of completions to primitives would provide insight into user-friendliness. Regardless of the evaluation methodology, a clear and precise metric is needed.

Finally, since our goal is to map this formalism onto a full generation environment, the actual process of transforming semantic expressions into natural language utterances must also be explored.

Conclusions

No current system is truly user-friendly. While first order logic and Montague’s intensional logic are being used to express a number of linguistic phenomena, they do so at the expense of creating complexity for the user. Unfortunately, this complexity continues to grow as more complex linguistic phenomena are explored.

This paper explored the form of a new system of semantic representation which has considerable expressive power, while maintaining a more intuitive form to its users. While many feel a formalism’s expressive power is more important than user friendliness, developing an intuitive formalism could lead to wide-spread use in natural language generation tasks and possibly even to the larger natural language processing community.

Indeed, to verify any automated task, a human being must be present to review and interpret any semantic intermediate form created during the debugging process. Simply put, the simpler the semantic form becomes, the easier the verification process will be. In terms of natural language generation, this formalism coupled with a natural language system such as VINCI [ 13] could develop into a new and simple way of providing a wide range of users the ability to generate multi-language utterances with a minimal amount of effort.

References

1. Bird, R. 1988. Introduction to Functional Programming Using Haskell 2nd Ed. Prentice Hall Series in Computer Science. Prentice Hall.

2. Blackburn, P. and Bos, J. 2003. Computational semantics. Theoria 18, 46. 27-45.

3. Dowty, D. R. 1981. Introduction to Montague Semantics. Synthese Language Library 11. D. Reidel Publishing Co., Dordrecht, Holland.

4. Donald, M. C. 2005. A metalinguistic framework for specifying generative semantics. Master’s thesis, Queen’s University.

5. van Eijck, J. 2006. Discourse representation theory. In Encyclopedia of Language & Linguistics, K. Brown, Ed. Elsevier. 660-668.

6. Fillmore, C. J. 1968. The case for case. In Universals in Linguistic Theory, E. Bach & R. T. Harms, Eds. Holt, Reinhart and Winston, New York, NY. 1-90.

7. Jackendoff, R. 1983. Semantics and Cognition. Current Studies in Linguistics Series. The MIT Press, Cambridge, MA.

8. Jackendoff, R. 1990. Semantic Structures. Current Studies in Linguistics Series. The MIT Press, Cambridge, MA.

9. Jacobs, J. 1987. The story of the three little pigs. In Classics of Children’s Literature 2nd Ed., J. W. Griffith and C. H. Frey, Eds. Macmillan Publishing Co. 815-817.

10. Kasper, R. T. and Rounds, W. C. 1986. A logical semantics for feature structures. In Proceedings of the 24th Annual Meeting on Association for Computational Linguistics. Association for Computational Linguistics, Morristown, NJ. 257-266.

11. Kay, M. 1984. Functional unification grammar: A formalism for machine translation. In Proceedings of the 10th International Conference on Computational Linguistics and 22nd Annual Meeting on Association for Computational Linguistics (ACL- 22). Association for Computational Linguistics, Morristown, NJ. 75-78.

12. Lehmann, F. 1992. Semantic networks. Comput. Math. Appl. 23, 2-5. 1-50.

13. Levison, M. and Lessard, G. 2004. Generated narratives for computer-aided language teaching. In eLearning for Computational Linguistics and Computational Linguistics for eLearning (COLING’04), Geneva, Switzerland. L. Lemnitzer, D. Meurers, and E. Hinrichs, Eds. 26-31.

14. Lønning, J. T., and Oepen, S. 2006. Re-usable tools for precision machine translation. In Proceedings of the COLING/ACL Conference on Interactive Presentation Sessions. Association for Computational Linguistics, Morristown, NJ. 53-56.

15. Montague, R. 1974. The proper treatment of quantification in ordinary English. In Formal Philosophy, R. Thomason, Ed. Yale University Press, New Haven, CT.

16. Moss, L. S., and Tiede, H.-J. 2006. Applications of modal logic in linguistics. In Handbook of Modal Logic, P. Blackburn, J. F. van Ben-them, and F. Wolter, Eds. Elsevier Science. 1031-1076.

17. Reichenbach, H. 2004. The tenses of verbs. In Semantics: A Reader, S. Davis and B. S. Gillon, Eds. Oxford University Press, Oxford, UK. 526-534.

18. Schank, R. C. 1972. Conceptual dependency: A theory of natural language understanding. Cognitive Psych. 3, 4. 552-631.

19. Tennent, R. 1991. Semantics of Programming Languages. Prentice Hall International Series in Computer Science. Prentice Hall International, New York, NY.

20. Winograd, T. 1982. Language as a Cognitive Process: Syntax. Addison-Wesley Longman Publishing Co., Inc., Boston, MA.

 

Biography Craig Thomas ( [e-mail]) is a PhD candidate in the Computational Linguistics Laboratory in the School of Computing at Queen’s University in Kingston, Ontario, Canada. This research was conducted with funding assistance from the Natural Sciences and Engineering Research Council of Canada.