Agreement is a type of relationship between two linguistic elements, often characterized as an asymmetric relationship where one element, the controller, determines the agreement features of another, the target, within a particular syntactic domain. Although according to prescriptive grammars, agreement relationships are stable and deterministic, usage-based data reveal considerable variation. Building on data retrieved from heTenTen 2014, a billion-token web-crawled Hebrew corpus, we present and discuss two types of agreement alternations: (1) agreement targets which alternate between exhibiting feminine vs. masculine gender, full vs. default agreement, and formal vs. semantic agreement, and (2) controller competition, where an agreement target is controlled by one of two possible controllers. Naturally, this perspective on agreement highlights the exceptions and overlooks the regularities, yet we argue that an examination of such alternations provides clues as to the true nature of the agreement relation.
|Title of host publication||Usage-Based Studies in Modern Hebrew Background, Morpho-lexicon, and Syntax|
|Number of pages||44|
|State||Published - 2020|