• Le 03 décembre 2018
    De 10:00 à 12:30
    Campus Tertre
    Bâtiment Censive, Salle du LLING, C228

  • Lundi 3 décembre 2018, 10h
    Responsable : Teodora Mihoc (Harvard University) 

    Titre: Modified numerals between polarity and valence

    Résumé: Bare, comparative-modified, and superlative-modified numerals (henceforth BNs, CMNs, and SMNs, respectively) exhibit interesting similarities and differences with respect to scalar implicatures, ignorance, and acceptability of embedding in downward-entailing environments. I present an account that derives all these patterns in an integrated way from exhaustification of grammatically-derived scalar and subdomain alternatives with the silent exhaustivity operator O(nly) (using technical assumptions from Chierchia 2013's system, including the possibility that O may be obligatory for some items, and it may come with a proper strengthening requirement). A crucial advantage of the present account over previous literature is the way it captures the contrast between CMNs and SMNs under negation (I don't have more than 2 / *at least 3 diamonds) and the absence of such a contrast in other downward-entailing environments such as the antecedent of a conditional or the restriction of a universal (If you have more than 3 / at least 3 diamonds, you win). But how clear are these patterns? I present experimental evidence that supports them. The experimental data however also reveal new interesting patterns on which SMNs in the antecedent of a conditional and the restriction of a universal are sensitive not just to polarity but also to what we may call valence -- positivity or negativity in a looser sense (e.g., If you don't have at least 3 diamonds, you lose / *win). In the rest of the talk I investigate a series of patterns of this type and argue that they can be captured from exhaustification of contextually-derived, probability-based alternatives with the silent exhaustivity operator E(ven) (cf. Crnic 2012's solution for similar phenomena in other scalars items). Finally, I discuss how the solution to the polarity facts, which is based on O, might be integrated with the solution to the valence facts, which is based on E.