You may care to look at:
Sager, J.C. & L'Homme, M-C. (1994) A model for the definition of concepts.
Terminology V1 N2, John Benjamins, pp351-373.
While this takes the point of view of how good analytical definitions
could be constructed (also to aid retrieval), it is moreover a good
source of analytical defining patterns.
A few example applications:
An earlier version of the rules proposed here was used by UMIST
to define some 75,000 non-specialised concepts for the Japan EDR project
(interestingly, in the absence of associated headwords, thus the
definitions had to be clear, concise and able to allow sufficient
differentiation among concepts).
UBielefeld adopted the rules of this article to define the terminology
for the supplement to the EAGLES Handbook of Standards and Resources for
Spoken Language Systems (EAGLES=Expert Advisory Group on Language Engineering
Standards, cf.: http://www.ilc.pi.cnr.it/EAGLES/home.html)
The ESPRIT CONCERTO project (P29159) on conceptual indexing, querying and
retrieval of digital documents, which involves among other things building
domain ontologies, is currently investigating the use of these rules to
define the concepts of the ontologies.
Thus these rules can be relevant for both general and specialised
analytic definition and have been tested out, in some cases on a rather
large scale.
Anyone considering embarking on the construction of analytic definitions
for term banks, general dictionaries, ontologies, etc., would benefit
from reading this article.
A wider discussion of the role of definition in terminology and
lexicography, including numerous examples of different types of definition
and different patterns, can be found in Chapter 2 of:
Sager, J.C. (1990) A Practical Course in Terminology.
Amsterdam/Philadelphia: John Benjamins.
Jock
--John McNaught jock@ccl.umist.ac.uk Department of Language Engineering UMIST PO Box 88 Sackville Street Manchester, UK tel: +44.161.200.3098 (direct) M60 1QD fax: +44.161.200.3099