Dear all,
Two remarks concerning the relationship between RT and other theories:
(1)
It would be interesting to discuss how relevance theory interfaces with
psychological and psycholinguistic theories language and discourse
production, that is to say theories that try to account what happens between
intention and articulation (to paraphrase the title of a book by Levelt). To
my knowledge RT has been developed mainly as a theory of comprehension (an
account of the way hearers (1) enrich/loosen logical forms and (2) derive
implicatures), but I guess relevance should play a role also in the
production process. I wonder if there is any relevance theoretic work on
production.
More generally, as far as I know, in linguistics the few who have paid much
attention to the production process tend to subscribe to a 'holistic',
non-modular view of language (perhaps I should say they're "functionalist",
but I think that this term tells more about the sociology of the field than
about the content of the theories): I think in particular of the Hallidayian
school, Mel'chuk's Meaning-Text model, or, more recently, of the
Functional-Pragmatic Grammar approach proposed by Nuyts (which is an
offspring of Dik's Functional Grammar).
Moreover, some of these theories tend to be accounts of an 'encoding'
process: how meaning/intentions come to be expressed by linguistic
structures (this is certainly true of Mel'chuk's model, and at least in part
of Halliday's).
So, I wonder what it will look like an account of production where speaker's
intentions aren't translated/encoded in linguistic strings but give rise
instead to a complex strategy which exploits the resources of the
linguistic module "within" acts of ostensive-inferential communication that
rely on the search for relevance of the hearer.
(2)
Concerning theories which put "pragmatic functions in the grammar" vs RT.
RT is usually seen as a theory which extends the domain of pragmatics at the
expense of linguistically encoded semantics: the classic example is
'enrichment' where pragmatics intrudes in the truth-conditions. This can
lead to caricaturing RT as a theory where everything is pragmatic and the
code does next to nothing: I've seen an article - by Michael Stubbs - which
opposed RT to approaches like "Construction Grammar" where the focus is on
the way in which specific linguistic constructions impose fine grained
constraints on context (!!!). This is ironic, because the idea of "semantic
constraints" on the construction of context is precisely what is behind the
RT notion of "procedural encoding". If you look at the original Construction
Grammar articles by Kay & Fillmore you find that Sperber, Wilson and
Blakemore are credited for the idea, toghether with Oswald Ducrot -whom I
think influenced early RT more than it is usually thought.
In fact, if we emphasize the role of procedural meaning, RT can be seen also
the other way round as a theory where (a certain type of) linguistic units
intrude pragmatics by constraining the search for relevance.
My question is: how well does the idea of procedural encoding fit in the
standard Chomskyan picture of language as a computational system which
developed independently of communication and whose core isn't influenced in
any relevant sense by communication? Obviously procedural units make sense
only in a system that is used as a complement of ostensive-inferential
communication - the idea that a purely internal computational/representation
system should contain such units strikes me as utterly weird.
It may well be that as long as procedural units are limited to a handful of
lexical items such as discourse particles and pragmatic connectives, they
can be relegated to the periphery of the linguistic system, as something
whose functioning doesn't pertain to the core of linguistic competence -
just like our knowledge of social connotations of words doesn't belong to
it. But what happens if we assume, as it is the case in much current RT
research, that a great number of morphemes such as tense or mood has a
procedural meaning? And if we assume that particular focussing/ topicalizing
constructions encode procedural meanings? (See what Robyn Carston, 2000 says
about Prince's work on these topics).
It may well be that RT ends up being more "functionalist" than what is
generally presumed.
Regards
Andrea
Reference
----------------
R. Carston (2000), The relationship between generative grammar and
(relevance-theoretic) pragmatics, "Language & Communication", 20: 87-103
This archive was generated by hypermail 2b29 : Wed May 29 2002 - 10:10:39 GMT