(Soutien à la recherche de l'université de Montpellier:
étude pluridisciplinaire de la quantification :
mathématique, informatique, didactique et linguistique)
Tentative
program
10:00-11:00 Oriol
Valentin
Universitat Polytècnica di Catalunya
Some Investigations on Logics of Scope Taking:
Displacement Calculus versus Hybrid Type-Logical
Grammar
Both hybrid type-logical grammar
(HTLG) and the displacement caIculus (D) can be consider
logics of scope taking. In several papers, Kubota and
Levine (K&L) claim thatD and HTLG are essentially the
same calculi. This is not the case as is proved in this
talk. According to K&L, D has problems for
abstractions of second-order string functionals rendering
D not powerful enough to account such phenomena of scope
as the interaction between determiners and gapping.
Nevertheless, in a 2017 paper Morrill and Valentín prove
that K&L’s account of determiner gapping
overgenerates. Moreover, in this paper a proposal for
extending D, the so-called second-order displacement
calculus, is given, which correctly generates the complex
cases of determiner gapping. This talk elaborates on
the issue of overgeneration of HTLG and the solutions D
can give. Besides this, in this talk a type-theoretic
proposal extending HTLG is shown to give a displacement
calculus point of view of this calculus, while maintaining
the well-behaviour of D and second-order D. In this sense,
we propose an intermediate calculus between D and HTLG.
11:00-12:00 Davide Catta Mehdi
Mirzapour université de Montpellier et LIRMM
Quantifier Scoping and Semantic Preferences
We address the problem of ranking valid
semantic readings of a given multiple-quantifier sentence based
on extending Morrill's account on the syntactic complexity
profiling of the categorial proof-nets. This extension is
inspired by Hilbert's epsilon operator and the syntactic
structure constraints.
12:00-14:00 déjeuner
14:00-15:00 Jakub Szymanik
University of Amsterdam
Quantifier semantics: cognition and
learnability
I will talk about some
of the research we have been developing in the framework of
Cognitive Semantics and Quantities project. The aim of this
project is to advance semantics by focusing on the
underlying cognitive representations. Especially, within the
project we are interested in developing cognitively
plausible semantic for generalized quantifiers. I will
present a number of experiments suggesting that the meaning
can be best captured in terms of collection of computational
procedures constrained by cognition. I will then discuss
what is the range of quantification in natural
language. I will argue that from the given
computational perspective one can view complexity as a type
of semantic universal guiding quantifier variation in
natural languages.
15-00-15:30 pause
15:30- 16:30 William
Babonnaud ENS Cachan
Natural language quantification and Hilbert's epsilon and
tau quantifiers
This talk will present the
subnectors epsilon and tau. They model quantification in a way
that is closer to natural language: in situ
quantification, over binding, underspecification and independence.
soutienRechercheUMretoreQUANTI