LINGUIST List 7.1115

Tue Aug 6 1996

Qs: Minimalist program, Relevance theory, Software package

Editor for this issue: Ann Dizdar <dizdartam2000.tamu.edu>

Directory

1. J Kalita, The minimalist program and SOV languages
2. Murray Garde, Ethnography and Relevance Theory
3. Kirk Sullivan, software package for linguistic analysis

I was wondering if someone could point me to any work
on explaining the word order in various constructs
in SOV languages in the context of the minimalist program.
Has there been any work on any of the languages
of India?

Thanks!

Jugal Kalita
kalitapikespeak.uccs.edu

To Linguist List recipients,
Does anyone know of applications of Sperber and Wilsons' Relevance
Theory to ethnographic/linguistic anthropological studies. I am
particularly interested in the language of 'institutionalised' or
customary joking relationships.

Murray Garde
Anthropology Dept., Northern Territory University
Australia

I am putting this on the list for a friend. Please
e-mail him directly at Jurgen.Borstlercs.umu.se


We are looking for a software package, which linguistically analyses
sentences. We know that there exist hundreds or more such
packages. The problem is that all these packages systems are
stand-alone systems implemented in Prolog or Lisp. We want something
we can use together with our system, i.e. something with an interface
in an imperative language (Pascal, C/C++, ...) or something that
produces textfiles, which we can then read and interpret ourselves.
We describe our system and needs in more detail below.

Requirements for software systems are usually described by more or
less formal notations (graphical and/or textual). Since these
notations often require considerable knowledge in information/computer
science, it is almost impossible for the end-users themselves to
describe their requirements in these notations. This often leads to
misunderstandings and results in systems which do not fulfil the
original (i.e. the end-users') requirements.

 We are well able to "prove" that the system we built fulfils some
 (formally described) requirements, but we cannot prove that these
 formal requirements are identical to the end-users' requirements.

For obvious reasons the end-users should therefore be enabled to more
directly participate in the requirements acquisition and definition
process. We therefore need easy to use and easy to learn tools.

A useful notation should be informal and must not restrict the users
in the order they do things. But the notation must still be formal
enough to facilitate machine processing of the requirements.

The format we use are a variant of so-called use cases. A use case
describes a situation how a system can be used (a kind of scenario
description). The main part of a use case description consists of a
sequence of actions, which are simple sentences (see the following
example). The words in curly brackets are defined in a domain specific
dictionary or may be detailed in further (sub-) use cases (like
actions 1. and 5.).

Use case: Register new card

Actions:
 1. {Enter_supervisor_mode}
 2. The {system administrator} {enters} a new {card}
 into the {cardreader}.
 3. The {system} {validates} the {card} with the
 {cardreader}.
 4. The {system} {registers} the {card} in the
 {database}.
 5. {Enter_new_code}

We want to translate these sentences in an object-oriented model. This
is a model consisting of objects (persons, things, etc., not to be
confused with linguistic objects) with certain properties (e.g. that a
card is associated with a code; this can only be seen if we go into
the details of action 5.) and behavior (e.g. that
the system can validate and register cards). There are also certain
(often domain-specific) relationships between objects, e.g. the system
must have some "knowledge" of the existence of a database (otherwise
cards could not be registered).

As stated above what we are looking for is a software package,
which can do the required linguistic
analysis of the sentences (or pointers to relevant research).

Any idea?

Many thanks in advance.

Jubo