Ergativity

Abstract

Ergativity refers to a system of marking grammatical relations in which intransitive subjects pattern together with transitive objects (“absolutive”), and differently from transitive subjects (“ergatives”). This ergative alignment pattern may be manifest, for example, in terms of morphological case marking on nominals, or patterns of agreement on the predicate. This contrasts with the more commonly discussed nominative-accusative–type alignment, in which both transitive and intransitive subjects pattern alike (“nominative”), and differently from transitive objects (“accusative”). This article uses the common abbreviations “A” for transitive subject, or most agent-like nominal; “P” for transitive object, or most patient-like nominal; and “S” for the single argument of an intransitive. While counts vary, some surveys estimate that ergativity occurs as a major alignment type in approximately one-quarter of the world’s languages (see General Overviews). However, as many authors have noted, it often does not make sense to speak in terms of ergative languages but instead in terms of ergative patterns or constructions. This is because ergative languages are frequently “split”: they show ergative in some portion of the grammar but nominative-accusative patterning in another (see Split Ergativity). Even canonically nominative-accusative languages may show ergativity in some constructions, such as nominalizations. Work on ergativity has increased steadily over the years as more research has been conducted on ergative languages, many of which are underdocumented. Important questions arise as to the notion of “subject” in an ergative system, since transitive subjects are treated differently from intransitive subjects, at least at a morphological level. Some morphologically ergative languages also display patterns of Syntactic Ergativity, in which some syntactic operations, for example A-bar extraction, is sensitive to the distinction between ergative and absolutive arguments. Other morphologically ergative languages appear to make no syntactic division between A versus S/P arguments, raising further questions about degrees of ergativity. Other research focuses on the existence of a single “ergativity parameter”; as more languages are investigated, many researchers have converged on the idea that “ergativity” is not a single unitary phenomenon with one underlying source but may instead be better characterized as a pattern arising by various mechanisms.