• Le 15 février 2019
    De 15:00 à 17:00
    Campus Tertre
    Bâtiment Censive, Salle du LLING, C228

  • Vendredi 15 février 2019, 15h
    Responsable : Nino Grillo (University of York) 

    Titre : Kind of perfect: Experience ascription and event kinds across languages

    Résumé : Pseudo-relatives (PRs) are finite embedded constructions that complement perception verbs (brack- eted in (1)). It is commonly thought that the tense of PRs must “match” matrix tense: only imper- fective (=past) under matrix past forms (including perfect as in (1a)); only present under present (1b) (Radford 1977, Guasti 1988, Cinque 1995, a.o.). There are, however, under-reported cases of tense “mismatch”, where present is possible under perfect (1c)

    (1) a. Ho visto [Leo che correva].     b. Vedo [L. che corre]        c. Ho visto [L. che corre].
             I.have seen L. that run.IMPF     I.see L. that run.PRES      I.have seen L. that run.PRES
            ‘I saw L running.’                          ‘I see L. running.’                  ‘I saw L. run.’
    In recent work with Keir Moulton (Simon Fraser University), I argue that Tense-matching PRs (TM-PRs) (1a,b) can deliver an event-token interpretation where as Present-under-Perfect PRs (Tense-mismatching PRs, TMM-PRs) (1c) can only deliver an event-kind interpretation (Portner 1991, Gehrke 2015 a.o.) and force an experiential perfect reading of the matrix predicate. In support for this account, I will discuss the different quantificational properties of the two vari- eties of PRs. I will further discuss how the account derives several asymmetries between TM and TMM-PRs: including the ability to combine with temporal and spatial modifiers, introduce discourse referents, saturate kind-selecting predicates and combine with kind-anaphora. I will de- rive the Kind interpretation of TMM-PRs as a result of two features: (i) the TMM-PR is habitual and so denotes a plurality of events (Ferreira 2005); (ii) whereas TM-PRs are headed by a defi- nite determiner (Moulton and Grillo 2015), TMM-PRs are headed by Chierchia’s kind-forming ∩. When TMM-PRs combine with token-taking verbs (see in (1c)) they combine by Derived Kind Predication (DKP, Chierchia 1998), and as predicted display different quantificational properties compared to TM-PRs. I will argue that cross-linguistic variation in the availability of TMM-PRs is tied to the availability of kind-denoting singular definites in PR-taking environments. Finally, I will discuss novel data from Experiential Constructions and Internally Headed Relatives in Korean and Japanese further supporting the analysis.