[parislinguists] Conf. Geurts : **ERREUR DE DATE** (lire *JEUDI* 16/12 et non vendredi)

[Benjamin Spector]

ERREUR DANS MON PRECEDENT DE MESSAGE :

la conférence de Bart Geurts aura lieu le JEUDI 16 DECEMBRE, 
et non le vendredi 16, qui n'existe pas.

Abstract ci-dessous.



BART GEURTS, professeur à l'université de Nimègue,
donnera une conférence

le **JEUDI** 16 décembre de 10h30 à 12h30

à l'ENS, en salle W (45 rue d'Ulm, 4ème étage, aile Rateau),

**Numerical quantifiers in semantics and psychology**.

ABSTRACT

In the first part of my talk I present a simple semantic measure of the
complexity of quantifiers and quantified expressions, which is based on
monotonicity properties. The account predicts, among other things, that
a sentence whose quantifiers are all upward entailing is easier to
process than one whose monotonicity properties are mixed. Furthermore,
it is predicted that downward entailing quantifiers built from numerals
(like 'at most five sheep') are harder to process than other
quantifying expressions. I present experimental data that confirm these
predictions.

In the second part of the talk, I will have a closer look at numerical
quantifiers. It turns out that there are surprising differences between
quantifiers of the 'at least/most' type and quantifiers of the
'less/more than' type, and I propose to account for these by assuming
that quantifiers of the first type are in fact modal expressions. This
proposal yields a number of predictions concerning acquisition and
reasoning, which still are to be tested experimentally.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/parislinguists/attachments/20041212/c473f3ea/attachment.htm>