LINGUIST List 3.782

Sat 17 Oct 1992

Qs: Subjects; Ontology; Change

Editor for this issue: <>

Directory

1. , Request for info: English subject inversion
2. anoop sarkar, unification-based grammars and ontology theory
3. , Language changing before our eyes

I would like references to recent (since early 1980's) work on Subject
Inversion in English. I am especially interested in work done in any
type of functionalist or cognitive framework. Please send replies
directly to me.

I have one query concerning "unification- based" or a "constraint- logic
based" grammar: instead of using grammatical constraints, such as tense,
agreement, etc., has anyone worked on the use of ontological constraints
to check for the well formed structure of NL sentences.

Leaving aside implementational approaches which use class hierarchies in
an unification- based system, I'm more interested in the formal work
concerning the compositionality problems of interfacing a conceptual
network to a theory of grammar for the recognition or generation of NL.
Any such theory will also have to deal the checking of constraints
relating to space & time in NL ( space & time themselves are tough
ontological problems).

Apart from "unification- based" theories, has any work been done on
linking "principle- based" theories such as GB theory to ontological
theories, especially in the discharging of thematic roles in the theta
criterion.

I know of only two treatments of this problem:
Inferential Semantics or Rhema Graphs by A.F. Parker Rhodes which
combines mathematical lattice theory with semantic, syntactic and
even tonal analysis of sentences;
And the work on Knowledge- based NL parsing by Elaine Rich, et al.
which is a largely implementational look at the problem under the Cyc
ontology project at MCC.

If any references are available that deal with the above problems please
send their details to me at my email addr. i will summarize for the list.

In the hope of sparking off yet another debate I wish to post the
following rhetoric as well:
Won't it be better to abandon grammatical constraints altogether for
an approach that is embedded in a knowledge- based or ontological
theory as this will provide a more "intelligent" analysis of NL rather
than the flat enforcements of equality constraints that embody an
unification- based approach? Is this an approach that everybody would
readily adopt but for the lack of formalization of intensions in
knowledge- based theory?

Anoop Sarkar
Centre for Development of Advanced Computing
KBCS Group, Pune University Campus,
Ganesh Khind, Pune 411 007, India
email: anoopparcom.ernet.in

I have been noticing for some time now that no-one seems to say
or write "homogeneous" any more - the word has become "homogenous"
in both pronunciation and spelling. I presume that people got used
to "homogenised" milk and reduced "homogeneous" to four syllables by
a sort of analogy. The practice has now permeated the Linguist list
(but is that really how Postal spelt it? - I haven't had the chance
to check it out).

Comments and intuitions please, fellow linguists. Is the change already
absorbed into American English and spreading across the Atlantic? Have
any dictionaries noticed it yet? Should we tell them?

Sue Blackwell
University of Birmingham, U.K.