Property-type Objects and Modal Embedding

Verbs in West Greenlandic require a special piece of morphology, the antipassive, in order to take narrow-scope indefinite objects. Opaque objects of intensional verbs require the same treatment. This paper develops a semantics for the antipassive morpheme in West Greenlandic that shifts the verb’s object position to a propertytype, providing for the object’s narrow scope, while introducing modal embedding. The modal embedding provides for the interpretation of opaque objects of intensional verbs, in a way syntactically constructing the intensional construction. The modal embedding of property-type object constructions is visible not just in West Greenlandic antipassives but also in Hindi-Urdu and even English, suggesting a generalized modalization in the combination of verb with property-type object. 1 Narrow scope indefinite objects Many languages mark a distinction between indefinite objects which must take narrow scope with respect to operators such as negation, modals and verbs with intensional object positions, and those which may take wide scope with respect to such operators. This is the pattern described by Bittner (1987) for the West Greenlandic antipassive construction, where we see an antipassive suffix on the verb and an oblique case marker on the object ((b) examples below). (Canonical objects are provided for contrast in the (a) examples). (1) Modal operator(Bittner, 1987, ex 29) a. atuartut of.students ilaat one.of.them.abs ikiur-tariaqar-pa-ra help-must-tr.indic-1sgE/3sgA I must help one of the students ∃x[x is one of the students & it is necessary that (I help x) ] b. atuartut of.students ilaa-nnik one.of.them-obl ikiu-i-sariaqar-pu-nga help-antip-must-intr.indic-1sgA ∗Thanks to Angelika Kratzer, Rajesh Bhatt, Chris Potts and the SuB 12 audience members. Research presented here was supported in part by an NSF graduate research fellowship. Grønn, Atle (ed.): Proceedings of SuB12 , Oslo: ILOS 2008 (ISBN 978-82-92800-00-3), 92–106. Amy Rose Deal Property-type Objects and Modal Embedding I must help one of the students It is necessary that ( ∃x[x is one of the students & I help x]) (2) Negation (Bittner, 1987, ex 38) a. suli yet uqaasia his.utterance.abs puiu-nngi-la-a forget-neg-indic-3sgERG/3sgABS Hei had not yet forgotten hisj utterance ∃x[x is an utterance of hisj & not yet ( hei has forgotten x) ] b. suli yet uqaasia-nik his.utterance-obl puiu-0-nngi-la-q forget-antip-neg-indic-3sgABS Hei had not yet forgotten hisj utterance not yet ( ∃x[x is an utterance of hisj & hei has forgotten x]) (3) Verb with intensional object position (van Geenhoven and McNally, 2005, 892) a. Juuna-p J.-erg atuagaq book.abs ujar-p-a-a. look.for-indic-trans-3sg.3sg Juuna is looking for the book/a specific book. b. Juuna J. atuakka-mik book-obl ujar-lir-p-u-q. look.for-antip-indic-intrans-3sg Juuna is looking for any book The same type of morphological contrast also holds even when no scope bearing operator is present in the clause, however, both in West Greenlandic and in other languages marking similar scopal contrasts in object position (e.g. Hindi-Urdu, Mohanan (1995), Dayal (2003)).1 (4) Antipassive alternation without higher scopal operator (Bittner, 1987) a. Jakku-p J.-erg ujarak stone.abs tigu-a-a take-trans.indic-3sgE/3sgA Jacob took a/the stone b. Jakku J ujaqqa-mik stone-obl tigu-si-vu-q take-antip-intrans.indic-3sgABS Jacob took a stone The glossing here reflects the view that antipassive objects are restricted to indefinite interpretations. Bittner (1987) argues against this generalization based on cases where names, pronouns and demonstrative phrases appear as antipassive objects, as e.g. in (i). i. Jesusi-mik Jesus-obl taku-si-vu-q see-antip-intrans.indic-3sgA He saw Jesus (Bittner, 1987, 196) Cases with names as narrow-scope-only objects can also be constructed in Nez Perce, a language that like West Greenlandic lacks articles; these names are crucially not interpreted referentially, receiving instead a ‘somebody called X’ reading. It may be that the West Greenlandic examples that pose a problem for a straightforward account of antipassive objects as indefinite merit a similar analysis, as also suggested by Manning (1996, 94) (though the extension to pronouns and demonstrative descriptions will require further investigation). Michael Fortescue (p.c. to Manning) notes that the object of (i) is interpreted by Greenlanders referring not to Jesus but to the “concept of Jesus”, suggesting some fundamental problem with definite/referential interpretations of antipassive objects.