SUO: Re: Time, context, and relations
Jack and Pat,
Sanskrit is one of the oldest preserved versions of the
Indo-European branch of languages, and its system of case
endings is better represented than most of the other I-E
languages. Following is from the Encyclopedia Britannica:
Sanskrit grammar is similar to that of other older
Indo-European languages, such as Latin and Greek; it is
highly inflected and complex. Sanskrit has three genders
(masculine, feminine, and neuter), three numbers (singular,
dual, and plural), and eight cases (nominative, accusative,
instrumental, dative, ablative, genitive, locative, and
vocative), although only in the singular of the most common
declension does a noun show different forms for each case.
Adjectives are inflected to agree with nouns. Verbs are
inflected for tense, mode, voice, number, and person.
For the reference, see
http://www.britannica.com/bcom/eb/article/2/0,5716,67242+1+65552,00.html
The modern Slavic languages have almost as complete a case
system. Russian, for example, has seven cases, with the
functions of the ablative case merged with the instrumental
case. It also has three genders and remnants of the dual.
Latin has six cases, with the locative and instrumental folded
into the ablative case. Classical Greek and modern German fold
the locative, instrumental, and ablative into the dative case.
Other modern languages, such as Finnish and Georgian, have an
even richer set of case distinctions. But the same kinds of
distinctions can be found in the systems of prepositions in
languages that do not have case endings, such as English and
Chinese.
Aristotle's four aitia (translated into Latin as "causae" and
misleadingly translated as "causes" in modern English) can be
viewed as the old Stagirite's interpretation of the Greek case
system: the efficient aition is usually expressed in the
nominative case, the material aition is in the genitive case,
the final aition is in the dative case, and the formal aition
is in the accusative case.
James Pustejovsky has written a book, whose major theme is
the interpretation of Aristotle's four aitia as "qualia" of
nouns, and I have combined Aristotle's interpretation with
some of the more modern linguistic work (mostly by Roman
Jakobson, Chuck Fillmore, M. A. K. Halliday, and others) to
form the system of case relations or thematic roles that I
use in my KR book:
http://www.bestweb.net/~sowa/ontology/thematic.htm
Bottom line: I believe that there is a great deal of important
semantic information buried in the linguistic structures that
have evolved over the centuries, and we should look at that
material as a source for the SUO relations.
Sidelight: When people say that natural language X is the ideal
language for doing knowledge reprsentation, they often make some
interesting points, but you have to take their recommendation
with a large grain of salt. A few years ago, a Japanese
linguist gave a lecture at IBM Yorktown to propose the ideal
language for NLP, which just happened to be Japanese. He made
some interesting comments about Japanese linguistics, but the
audience (including me) was not sufficiently convinced to
adopt the proposal.
John Sowa