Scale structure and genericity: Understanding generics using stochastic comparison
Ashwini Deo, Mokshay Madiman
January 2015

This paper presents an analysis of generic sentences that takes into consideration the scalar properties of the predicates that they contain. It is shown that the variability observed in the real-world distributions that support or fail to support the truth of generic sentences is not arbitrary when we examine the structural and linguistic properties of the scales introduced by their predicates. The analysis is closely guided by work in degree semantics — particularly the notions of scales, comparison classes, and standards of comparison — and analyzes generic sentences as predicating a (non-)gradable property directly of a kind. This contrasts with the dominant approach in the literature that takes them to be tripartite quantificational structures that express relations between sets. The tools that allow the measurement of the degree to which a kind expresses a (non)-gradable property and its relation to a standard of comparison are drawn from probability theory — cumulative distribution functions and stochastic comparison. On the proposed analysis, any generic sentence As are B may be judged true iff two conditions are satisfied: (a) the cumulative distribution function (CDF) of the random variable that gives the value of a random element of A along the scale introduced by B satisfies salience; and (b) the stochastic process associated with the sentence satisfies stationarity.
Format: [ pdf ]
Reference: lingbuzz/002364
(please use that when you cite this article)
Published in: manuscript; comments welcome
keywords: generics, gradability, scale structure, comparison class, kind predication, probability, stochastic comparison, semantics
previous versions: v4 [January 2015]
v3 [January 2015]
v2 [January 2015]
v1 [January 2015]
Downloaded:2147 times


[ edit this article | back to article list ]