Comparative Semantics for Quantification: Evidence from Verification Studies using "Self-Paced Counting"
Martin Hackl, Pomona College

Formal semantic analyses aim to establish a systematic relation between the truth-conditional (TC) import of an expression and its syntactic/combinatorial properties. How (descriptions of) TCs are used by other systems of the mind - for instance in verification tasks - is typically not seen as something that formal semantics needs to account for or that could help distinguish between competing semantic analyses. An area where this lack of interest yields a particularly wide gap that more complete theories eventually will have to bridge is quantification.

This talk presents a novel experimental technique ("Self-paced Counting")that allows us to gather fine grained timing information about how subjects gather information incrementally in verification tasks that involve counting. It will be shown that this technique can detect different verification profiles for semantically equivalent quantified statements. Evidence of this sort reveals that the semantic primitives of quantification as assumed in Generalized Quantifier Theory are too coarse and supports the idea that measure functions, degree expressions and comparative relations are part of the basic building blocks of quantification in natural language.