ON 'SUBJECTIVE' AND 'OBJECTIVE' AGREEMENT IN HUNGARIAN Huba Bartos Hungarian Academy of Sciences In this talk I investigate the distribution of the so-called 'subjective' and 'objective' conjugations in Hungarian, with the aim of determining the factor(s) governing the choice between the two inflectional paradigms. The paradigms involve different person/number agreement suffixes on V, depending on some sort of agreement vs. non-agreement with the object. The suffixes encode person/number information on the subject, plus some property of the object. The matter of investigation is what this property of the object really is. Traditionally, definiteness of the object has been noted as the decisive factor: definite object NPs trigger objective agreement, while indefinite objects stand with the subjective agreement suffixes. Intransitive Vs invariably occur with the subjective paradigm. See examples in (1). (1a) J\'anos l\'att-a a fi\'ut. John saw-3sg:obj the boy-acc `John saw the boy.' (1b) J\'anos l\'atott egy fi\'ut. John saw-3sg:subj a boy-acc `John saw a boy.' (1c) J\'anos l\'atott. John saw-3sg:subj `John could see.' Definiteness of the object, however, is not the proper trigger. If the object NP is a possessive construction, V will bear objective conjugation, even when the object is indefinite (2); in other cases, both subjective and objective conjugation can be used with the same (indefinite) object, and the choice will result in specific vs. non-specific interpretation (3a, b); 1st and 2nd person object pronouns (as opposed to 3rd person ones) trigger subjective conjugation (4); eventually, if the issue boils down to object definiteness, why should intransitives pattern with the indefinite-object case? (2) J\'anos l\'atta n\'eh\'any emberedet. John saw-3sg:obj some man-2sgPOSS-acc `John saw some of your men / some men of yours.' (3a) J\'anosnak olvast-uk n\'eh\'any vers\'et. John-dat read-1pl:obj some poem-3sgPOSS-acc `We have read some of John's poems.' (specific obj.) (3b) J\'anosnak olvast-unk n\'eh\'any vers\'et. John-dat read-1pl:subj some poem-3sgPOSS-acc `We have read poems by John.' (non-specific obj.) (4a) J\'anos l\'atott engem / t\'eged / minket / titeket. John saw-3sg:subj me / you:sg-acc / us / you:pl-acc `John saw me / you / us.' (4b) J\'anos l\'atta P\'etert /\"ot / \"oket. John saw-3sg:obj Peter-acc / him / them `John saw Peter / him / them.' Based on examples like (3), it has been proposed that object definiteness should be replaced by object specificity as the trigger, but this view still leaves (2) and (4) without a convincing account, and fails to address the case of intransitives. Moreover, objects with specific quantifiers (Enc 1991, E. Kiss 1993) usually stand with subjective conjugation. In the light of the above facts, I propose an entirely new approach to the problem, the essence of which is that subjective conjugation occurs when the verb is intransitive in some sense, while objective conjugation marks actual transitivity. Thematically intransitive verbs are now a clear case. Verbs standing with indefinite objects count as intransitive in the following sense: indef. object NPs do not enter into a Case-relationship with V. More specifically, they do not access the spec of an assumed AgrO projection-the locus of both object agreement (V+AgrO) and accusative Case assignment/checking (V+AgrO to XP in spec,AgrO). For technicalities, I assume Chomsky's minimalist framework (1995). Thus indefinite objects will be Case-invisible for V, and no spec-head agreement obtains in AgrO. Up to this point, the proposal seems to simply rephrase the definiteness account. However, I argue that `definiteness agreement' is epiphenomenal: I propose a categorial distinction among object nominals. Let us assume the projectional economy of Grimshaw (1991), and claim that a DP layer is projected on nominals only if some lexical material (e.g. an overt D0, or something in spec,DP) forces it. Case being a feature of D, nominals not projecting a DP-layer (i.e. NPs, NumPs, etc.) do not engage in Case-relations. On the other hand, definiteness is a function of D0s, so less-than-DPs cannot be definite, and that is why definiteness appears to `trigger' objective conjugation. The so- called indefinite determiners constitute a separate category, distinct from D0 (Szabolcsi 1994), thus they will not give rise to DP. Specificity is also partly dependent on D0, and on the DP- domain, hence the `specificity' phenomena. Interpretational (i.e. not lexically or syntactically determined) specificity is correctly predicted to play no role here. Finally, the only problem cited above that remains to be explained is the behavior of personal pronouns-they can either be treated along the lines of Farkas (1990), or handled by positing an interesting case of `ergative split'. References: Chomsky, N. 1995. The Minimalist Program. The MIT Press, Cambridge, MA. Enc, M. 1991. `The Semantics of Specificity' Linguistic Inquiry 22 (1-25) E.Kiss, K. 1993. `Wh-movement and Specificity' NLLT 11 (85-120) Farkas, D.F. 1990. `Two Cases of Underspecification in Morphology', Linguistic Inquiry 21, (539-550) Grimshaw, J. 1991. `Extended projection' ms. Szabolcsi, A. 1994. `The Noun Phrase' In: Kiefer, F. & K. E.Kiss (eds.), Syntax and Semantics (vol. 27): The Syntactic Structure of Hungarian. Academic Press, N.Y.