The marked status of ergativity

Abstract

Ergative patterns treat the subject of an intransitive clause (S) similarly to the object of a transitive clause (O), thus singling out the subject of a transitive clause (A). The more familiar accusative pattern treats both S and A alike and differently from O. From an empirical point of view, ergativity is a marked phenomenon. The pattern occurs only in a quarter of the world’s languages, and even those languages displaying it often apply it restrictively. Former analyses have not paid much attention to this fact, as most of them formulate a macro-parameter whose sole function is to distinguish between ergative and non-ergative languages. The present proposal predicts the marked status of ergativity, deriving the pattern from an independently motivated parameter. The main hypothesis is that UG allows one way of base-generating verbal arguments: A is structurally higher than O. In addition, syntactic licensing has universal characteristics. Structural case is a feature on v that can only be checked by an argument in the complement of V. Therefore, it can only be Accusative. Phi-features, on the other hand, only play a role in the IP-projection. Nominative case does not exist. Instead, a caseless argument is licensed by agreement. In languages like English, this causes the familiar accusative pattern. Ergative patterns may occur in languages with so-called pronominal arguments. Every verbal argument in these languages is realized by an obligatory pronoun. According to the present proposal, such a pronominal argument (PA) cliticizes to the predicate. It can be doubled by an independent pronoun or full noun phrase, which adjoins to IP as a lexical argument (LA). The PAs obey the universal rules of syntactic licensing, and the LAs coindexed with them are not licensed separately. However, they may receive additional case marking, which may lead to an accusative, ergative or even tripartite pattern. This is the Second Pattern Hypothesis (SPH). Assuming an analysis of passive constructions which states that only the A-argument is realized as a PA, ergative patterns in Basque, Kurdish, Northwest Caucasian and Mayan can be accounted for. In these languages, S and O pattern together, to the exclusion of A, not only with respect to case marking on the noun, but also with respect to person/number marking on the verb. Assuming that incorporation of the PA does not necessarily trigger agreement, the O-argument of a transitive clause can be licensed by agreement. This hypothesis, called the Ergative as Passive Hypothesis (EPH), is motivated by morphological data from the languages mentioned. Both the SPH and the EPH suggest that an LA c-commands a corresponding PA from an A-bar position. It is well-known that in such a constellation, quantifier-variable readings are impossible. Native speaker-data from various ergative languages suggests that nonreferentially quantified DPs are indeed excluded from the supposed LA-positions. Ergativity, then, is linked to a macro-parameter dividing languages between those that do not allow for PAs and those that do. In languages with PAs, ergative patterns may be further restricted to certain values of functional heads such as I.