John, (01)
Response below. (02)
Leo (03)
-----Original Message-----
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of John F. Sowa
Sent: Saturday, March 19, 2011 8:40 PM
To: ontolog-forum@xxxxxxxxxxxxxxxx
Subject: Re: [ontolog-forum] Fwd: Re: Using controlled natural languages for
ontology (04)
... (05)
> JS:
> As late as 1970, Montague claimed that there is no essential
> difference between natural languages and formal languages.
> His version of language had a perfectly Fregean foundation,
> but it was useless for practical NLP. (06)
Actually Montague semantics enabled a real resurgence in NL semantics that
continues today. Prior to Montague, NL semantics was mostly based on semantic
"features" and "markers", and the use of semantic "projection rules" and was
relatively ad hoc (witness the late-60s "war" between "generative semantics"
and "interpretive semantics", notions of "Markerese" and "the language of
thought"). And Montague's work, based on a truth-conditional semantics of NL
(i.e., model-theoretic semantics) spawned solid extensions that included
Discourse Representation Theory, File-change semantics, Dynamic semantics, and
computational semantics. So the effect was to begin to create a science of
natural language semantics, as opposed to theory and engineering hacks. Of
course, the latter continued and continues. For good overviews, I would suggest
[1], and more extensively [2]. And there is the very readable citation by the
Wikipedia article on "Formal Semantics": [3]. I don't advise [4] unless you
have already had a healthy dose of linguistics and logic. Other references: [5,
6, 7]. [8] provides an introduction to computational semantics. (07)
The power of formalization is that it does make things more precise, which
lends itself to explicitly developing hypotheses/theories for refutation,
convergence on disputed terminology, and amenability to computational
implementation. (08)
Also, it probably depends on what you mean by "practical NLP". I can hack
anything up, and that will be practical for my current project, but there is
limited sustaining value. Having a theory or at least a model for what you
intend to do is necessary, if you want to avoid endless hacking. Good
engineering needs good science. In computational linguistics over the past 20
or so years, as an example, witness the loss of knowledge in the use of
"stemming" as opposed to "morphologizing". The former mostly tries to cut words
left and right (or internally) either haphazardly or with the programmer's
limited understanding of morphology for the particular language. The latter
tries to apply morphological rules for the language, typically based on some
theory or set of theories, in a systematic fashion, as in finite state
morphology [9], initially developed as two-level morphology by Kimmo
Koskenniemi many years ago. (09)
John, rather than under-emphasizing the semantics of natural language, and
thence being "useless for practical NLP", I think that the model-theoretic
semantics of NL largely ignited by Montague's work has greatly advanced the
science of the semantics of NL, and thereby made semantics eminently
implementable for practical NLP. (010)
Science, as we all know, is distinct from engineering, and its advancement
enables engineering. Do I think that the model-theoretic semantics of NL as
currently expressed represents all of the semantics-pragmatics of NL? No,
obviously not. But (similar to the case in pre-logical AI) formal semantics
(like formal AI) enables you to advance the science, and not just compare the
often ad hoc, theory-free engineering approaches. And, by the way, most good
engineering aspires to good science, if that science is still incomplete. In
that sense, a sound engineering approach can be considered an emerging theory,
and over time may become one. (011)
Thanks,
Leo (012)
[1] Partee, Barbara. 1996. The Development of Formal Semantics in Linguistic
Theory. In Lappin, ed., The Handbook of Contemporary Semantic Theory,
Blackwell, Oxford, UK, and Cambridge, MA, 1996, pp. 11-38.
[2] Partee, B.H., with H.L.W Hendriks, eds. (commentator: T. Janssen). 1997.
Montague Grammar. In Van Benthem, Johan, and Alice ter Meulen. 1997. Handbook
of Logic and Language. Cambridge, MA: The MIT Press, Part 1, Frameworks, pp.
7-91.
[3] Abbott, Barbara. 1999. The formal approach to meaning: Formal semantics and
its recent developments. In: Journal of Foreign Languages (Shanghai), 119:1
(January 1999), 2-20. https://www.msu.edu/~abbottb/formal.htm.
[4] Dowty, David R., Robert E. Wall & Stanley Peters, eds. 1981. Introduction
to Montague Semantics. Dordrecht: D. Reidel.
[5] Kamp, Hans; Reyle, Uwe. 1993. From Discourse to Logic: Introduction to
Model-theoretic Semantics of Natural Language, Formal Logic and Representation
Theory, Part 1 and 2, Kluwer Academic Publishers, The Netherlands.
[6] Gamut, L.T.F. 1991. Logic, Language, and Meaning, Volume 2: Intensional
Logic and Logical Grammar. The University of Chicago Press, Chicago and London.
[7] Portner, Paul H. 2005. What is Meaning: Fundamentals of Formal Semantics.
Wiley-Blackwell.
[8] Blackburn, Patrick; Johan Bos. 2005. Representation and Inference for
Natural Language: A First Course in Computational Semantics. Stanford, CA: CSLI
Publications.
[9] Karttunen, L. ; Beesley, K. R. 2005. Twenty-five years of finite-state
morphology. In Inquiries Into Words, a Festschrift for Kimmo Koskenniemi on his
60th Birthday, CSLI Studies in Computational Linguistics. Stanford CA: CSLI;
2005; 71-83. (013)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (014)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (015)
|