As far as I understand, there are at least two possible sources of
incompatibility, which are, in my opinion, quite independent:
- One has to do with the fact that RT subscribes to an autonomous syntax,
while CG seems quite extreme in advocating an opposite view. I do not think,
however, that this is particularly central for RT (or at least for what 'I'
see as the central claims of RT!). In my opinion, one could come up with a
RT without autonomous syntax, and that would be still a version of RT in
many relevant respects.
- Another, more serious source of incompatibility, concerns the role of
logic, truth and reference. If I have not severely misunderstood CG, the
type of semantics advocated by CG does not seem to have a place for these
notions. The RT approach to pragmatic inference relies crucially on
(substantially classical) logic. In order to (non-demonstratively) deduce
implicatures you need propositions or better mental representations that are
propositional forms, and in order to have these propositions you have to
pragmatically enrich some sort of logical forms. So, in order to have RT
interface with CG you have to suppose that all those pictorial
representations, spatial & force dynamic schemata of CG can be assigned
truth-conditions, or can be enriched to some sort of propositional,
truth-conditional, representations.
This is certainly not what the representations used in CG were originally
designed for, since the issue of truth conditions is not addressed in CG.
And if you leave aside the idea that pragmatic processes are logical,
deductive, inference processes performed on propositional representations I
don't think you can speak of RT anymore.
In principle, it should be possible to develop a truth-conditional semantics
for representations similar to those used in CG (truth-conditional semantics
for diagramming languages exist, and I remember one article in Linguistics &
Philosophy which developed a truth-conditional semantics for Jackendoff's
Conceptual Semantic representations, which are not that far from Langacker).
But I don't think this would be of any help in capturing the best insights
of CG, which concern, in my opinion, the iconicity of grammatical
constructions.
Maybe you can treat the iconically manifested meanings of grammatical
constructions as a more or less conventionalised version of the implicatures
connected with style and poetic effects, so that a certain construction not
only contributes to logical form but also carries a large number of weakly
communicated assumptions. But this isn't CG anymore.
In sum, I think that because of the issue of truth-conditions and logical
inference you cannot interface RT and CG without profoundly altering either
RT or CG, or both.
Best regards,
Andrea
> -----Original Message-----
> From: "Laine, Päivö" [mailto:Paivo.Laine@seamk.fi]
> Sent: giovedì, 4. luglio 2002 11:18
> To: relevance
> Subject: RT and cognitive grammar
>
>
> I have followed with interest the conversation about the
> relationships between the Relevance Theory and syntactic
> theories. In my own work my intention is to apply cognitive
> grammar to the syntactic-semantic analysis of the coded
> communication and use the relevance theoretic model to
> explain user interpretation in the context of my research
> topic. I wonder if there incompatibility other than the claim
> of cognitive grammar to be able to account for pragmatic
> considerations too. I feel that the principle of relevance
> has added value compared to the purely cognitive grammar
> model in the area of pragmatics, but does it conflict with a
> grmmatical analysis based on cognitive grammar?
>
> Pdivv Laine
> University of Vaasa, Finland
> paivo.laine@seamk.fi
>
This archive was generated by hypermail 2b29 : Thu Jul 04 2002 - 13:40:21 GMT