conference de michael hegarty le 28 novembre

[Anne Abeillé]

le laboratoire LLF a le plaisir de vous annoncer la conférence de

Michael Hegarty
Louisiana State University

vendredi 28 Novembre

salle 124

   16.00-17.30 hr

30 rue du chateau des rentiers, Paris 13e,  1er etage

  Modalized Dynamic Semantics and Neg-Raising.

Abstract:

Early forms of dynamic semantics, including Discourse Representation  
Theory and Dynamic Predicate Logic, mapped a natural language  
sentence S to a formula in a logical language, and interpreted S by  
simply updating the current information state of the discourse (or  
Discourse Representation Structure) by that formula. As noted by  
Asher and Lascarides (2003) and others, this fails to yield the  
correct interpretation when the intended import of the sentence is  
one which depends on implicit discourse relations and implicatures,  
rather than the literal meaning (or just the literal meaning) of the  
sentence captured by the formula. Building on earlier work by Asher,  
as well as work by Hobbs and Kehler, Asher and Lascarides modified  
DRT to a version of Segmented Discourse Representation Theory (SDRT)  
which corrects these deficiencies of classical forms of dynamic  
semantics. However, this theory still interprets a sentence S as an  
invariable condition on the update of an information state or  
Discourse Representation Structure.

In a further step towards realism, building on SDRT, the talk will  
formulate the dynamic semantic interpretation of S as a modalized  
information state update condition, whereby S is interpreted as  
effecting information state update only in possible worlds which  
accord with the speaker's deontic conditions for information state  
update (presumptively shared with the addressee), and only in a  
subset of those worlds consistent with an implicit evidential modal  
strength of the utterance, reflecting the degree of certitude of the  
speaker. This account of clause interpretation extends naturally to  
clausal complements of propositional attitude verbs, interpreted as  
modalized conditions on update of an information state attributed by  
the speaker and addressee(s) to the referent of the subject of the  
ascription, capturing differences of strength (e.g. between  
'conjecture', 'think', 'be certain') in terms of different modal  
operators, and accounting for attitude ascriptions of emotion (e.g.  
with 'hope', 'wish', 'regret') through bouletic modal conditions on  
information state update (which accompany or replace the deontic  
conditions, depending on the predicate). This yields a highly  
articulated lexical semantic treatment of attitude predicates.

This approach to clause interpretation yields an account of "Neg- 
Raising" in terms of classical scopal interactions between negation  
and modal operators, and successfully accounts for the delimitation  
of predicates with which "Neg-Raising" can be obtained, and provides  
a basis for the observed cross-linguistic and cross-dialectal  
variation in "Neg-Raising".

The talk will then address some facts about the dependence of Neg- 
Raising in Romance languages on the mood and tense of the higher  
clause, as noted by Rivero, Tovena, and others. It will be suggested  
that these can be accounted for in terms of an alternation between  
two types of representations of the interpretation of propositional  
attitude ascriptions and other relevant contexts, one which permits  
the scopal interactions which lie behind the Neg-Raising  
interpretation, and one which does not permit them.



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/parislinguists/attachments/20081027/6f98caaf/attachment.htm>