No theory exists in a vacuum; there will always be other theories
dealing with similar or complementary areas of investigation. RT is no
exception. Theories deling with (a subset of) the same areas of
investigation as RT will either complement or contradict RT to varying
degrees. Theories dealing with topics beyond the scope of RT will be
either compatible or incompatible with RT, to different degrees. Some
theories partially overlap in scope with RT; Fodor's Mental Models
Theory is one such in my opinion. The question I have for the list
concerns the relation between RT and theories covering (approximately)
the same ground (pragmatic theories), and theories covering different
topics (specifically syntax). Syntax first:
In the past, the general assumption seems to have been that RT can
quite easily 'plug in' to the most current model in the Chomskian
program. This assumption was based on the following factors: 1) The
Chomskian approach sets strict limits on the scope of syntactic
processes, thereby fitting in neatly with the Fodorian modular
approach which RT also adopts; 2) This program has been dominant in
syntax for a long time thereby helping to situate RT in the linguistic
mainstream; 3) Deirdre Wilson studied under Chomsky and as a result
she and her students have been influenced by his approach.
These are all perfectly valid reasons for selecting the Chomskian
program as the syntactic foil to RT, but the situation in syntax at
the moment seems to be a lot more fluid than it was in the mid-80s
when this arrangement began. For a start, the Chomskian program is no
longer the default syntactic approach in many institutions. The
minimalist program (MP) was developed in response to perceived
problems with GB and P&P, but this has recently come under attack. The
following is an extract from Robert D. Levine's review of Johnson &
Lappin (1999) Local Constraints vs. Economy, in J. Linguistics 38
(2002) 177-179:
"The MP emerges from Johnson & Lappin's tightly reasoned, pitilessly
clinical examination as computationally expensive to the point of
intractability, empirically deficient, stipulative at best and
incoherent or inconsistent at worst in its internal architecture, with
metaphysical and teleological foundations that run counter to the
methodological premises undergirding the success of the hard scinces.
[...] until and unless there is a coherent and substantive response to
Johnson & Lappin's critique, the MP will have marginal credibility
outside the increasingly narrow domain in which it is simply the
unquestioned 'house framework'." (178)
It would be a shame if RT became part of this "increasingly narrow
domain", especially if the claim that the MP is "computationally
expensive to the point of intractability" proves to be well-founded.
We should also take into account the critique of GB and P&P accounts
of the various binding constraints in Huang's "Anaphora" (but more of
that later).
I am not a syntactician, so I would value advice on where I should be
looking for a plausible syntactic theory that is consistent with RT.
Cognitive Grammar is obviously at odds with RT over the scope of
grammar, as are various other functional approaches. LFG at one time
looked promising, at least with respect to Bantu languages, but I have
seen a couple of convincing accounts recently of topics in Bantu
morpho-syntax where GB is clearly superior to LFG. OT seems to be at
an early stage and I am unsure whether it even constitutes a coherent
program yet. I'd be interested to know what readers of the list think
of some other less well-known approaches such as Robert Van Valin's
Role and reference Grammar (he has worked with Randy LaPolla, who I
think was once a member of the RT list).
Turning to pragmatics, the field is a lot smaller, with the main
players being RT and the neo-Gricean approach. Although I disagree
with much of the philosophical underpinning of the neo-Gricean
approach and believe that it does not pay adequate attention to issues
of cognition in language use, I have seen a few neat analyses where
the interplay of the Q-, I-, and M-principles seems to make specific
predictions which a RT analysis would not be able to make (as far as I
can see). This is the kind of level of falsifiabilty that critics of
RT have been looking for. (Briefly the Q-principle states 'Say as much
as you know', the I-principle states 'Say as little as necessary', and
the M-principle states, 'Whatever you say, say it clearly'.
Implicatures generated on the basis of these principles are ranked in
the order Q>M>I; that is a Q implicature 'defeats' an M implicature
and so on.)
The combined strengths and weaknesses of the neo-Gricean approach
vis-a-vis RT are apparent in the following ecxtract from a review of
Yan Huang's "Anaphora" which I wrote for Notes on Translation:
"A more fundamental problem exists, I believe, with the neo-Gricean
approach to implicature. In this framework, meaning can be broadly
divided into `what is said' and what is implicated. Even though
reference assignment (which is what occurs when an anaphor is linked
to its antecedent) is clearly not a (totally) semantic and syntactic
process, it seems counter-intuitive to say that reference assignment
is the result of identifying an implicature, particularly when in the
Gricean tradition, `what is said' is defined as the semantically
determined aspects of meaning plus disambiguation and reference
assignment. Alternative pragmatic approaches which allow pragmatic
aspects of what is said (that is, inferentially determined meaning in
addition to reference assignment and disambiguation which contributes
to truth-conditions) should, in theory, be able to provide a more
coherent account. One such theory, which Huang discusses, is relevance
theory (Sperber and Wilson 1995). Unfortunately, I doubt whether
relevance theory as it currently stands is capable of providing an
empirically falsifiable account of anaphora resolution, (although
Matsui 2000 provides a start in her account of bridging reference).
This is a challenge that needs to be met."
What I am wondering is whether RT researchers have avoided adopting
and adapting relevant insights from our biggest rival simply because
we want to be seen to be different. There is no doubt in my mind that
RT is on more secure theoretical ground than neo-Gricean pragmatics,
but it does lack specificity of the Q-, I-, and M-principles. These
principles and the ranking of the implicatures they generate, as
presented by Levinson and Huang, are simply stipulations at the
moment. However, if it could be shown that they follow from the more
general cognitive principles that underpin RT then there could be no
question of RT stealing the neo-Griceans' clothes, rather the
neo-Gricean approach could be largely subsumed within RT.
As many of you know, I enjoy playing devil's advocate, so please treat
this in the spirit of open enquiry in which it is intended.
*******************************************
Steve Nicolle,
P.O. Box 12,
Kwale (via Mombasa),
KENYA
This archive was generated by hypermail 2b29 : Mon May 27 2002 - 11:39:02 GMT