Corpora: Ergo's Patent Publishes

Philip A. Bralich, Ph.D. (annes@htdc.org)
Mon, 15 Mar 1999 17:35:57 -1000

The patent for the tools that create Ergo Linguistic Technologiess’
software has just been published by the U.S. Patent Office. Copies
of it can be obtained from them through the usual channels. Many
have asked me to notify them when the patent is published.

The patent description contains a theory of syntax that is far
simpler and far more general than current theories, and more
importantly, that makes software that individuals working with
other theories can only dream of. For examples of the Ergo
software go to http://www.ergo-ling.com.

Probably the strongest recommendation for reading this
patent and studying this theory is the software that it can
create which can be seen at the Ergo web site. Those of you in
industry may want to try and see if you can create similar tools
and still beat the patent. We believe this is not possible, but we
would encourage all to try in the spirit of good sportsmanship.
In academia the very fact that we have a theory that produces
better NLP tools than any other theory calls into question the status of
all other theories of syntax. This is because every theoretical mechanism
ever proposed for a theory of syntax (ours, Chomsky’s, or anyone else’s)
can, in principle, be implemented in a programming language. Thus, the
clearest judge of the best theory of syntax is the working software
that can be produced from it. I have in the past even argued that until
such time as other theories can do as well or better than we can in this
area, that the Ergo parser should be declared the default standard for
computational linguistics both in academia and in industry. (If anyone
can demonstrate why this should not be the case, I would appreciate seeing
the argument). The Ergo parser provides tools and a parser that can
significantly improve navigation and control devices and question and
answer dialoging software as well as other areas of NLP that require
grammatical analysis. All the demos at the Ergo site are WIN95/8/NT
compatible. Get them and compare them to the software made from other
theories (if they are capable of producing any at all).

Please do not take offense at these rather strongly worded statements.
They are required because the soft sciences do not often have to deal
with clear, incontrovertible evidence of the superiority of other theories
or breakthroughs. In chemistry, for example, if someone creates a better
and cheaper formula for a particular result (say the treatment of a disease),
the new method is adopted and older ones are discarded until such time
as evidence (e.g. the computer program in linguistics) demonstrates
otherwise.

For a discussion and description of standards for the evaluation
of parsers and parsing systems go to
http://www.vrml.org/WorkingGroups/NLP-ANIM. In addition, the
Ergo web site provides examples and a parsing contest for those
who would like to compare different parsing tools.

Here’s an even greater challenge: Get the patent and the Ergo
software and then write some papers that explain why working
software is not a criteria for judging a theory of syntax. Or also
why the theory of syntax cannot produce the software that we
have. It might also be interesting to demonstrate why our theory
of syntax is not to be preferred over others and why this theory of
syntax is flawed (in spite of the unique software development
it offers).

Phil Bralich

Philip A. Bralich, Ph.D.
President and CEO
Ergo Linguistic Technologies
2800 Woodlawn Drive, Suite 175
Honolulu, HI 96822

Tel: (808)539-3920
Fax: (808)539-3924
bralich@hawaii.edu
http://www.ergo-ling.com

Philip A. Bralich, Ph.D.
Ergo Linguistic Technologies
2800 Woodlawn Drive, Suite #175
Honolulu, HI 96822
fax: (808)539-3921
tel: (808)539-3924