Givenness and Locality

Abstract

The ProblemConstituents that encode information that was made salient by prior context often remain unaccented.This paper presents evidence that in order for a constituent to be marked as given by deaccenting, it is not sufficient that it is given, it must be given relative to something else.In particular, I will argue that constituents have to be given relative to their sister.This first section outlines the problem that earlier approaches to givenness face: sometimes, but not always, it appears that the sister of a constituent is relevant in deciding whether it can be marked as given. Marking GivennessConsider a context in which the direct object is deaccented:(1) Mary sat at her desk.John walked in.What happened next?a. # She kissed JOHN.b.She KISSED John.Context induces a deviation from the otherwise expected stress pattern.The last accent in (1b) falls on the verb, not on the direct object, as would be expected had there been no mention of John in the context.This paper focuses on cases of deaccentuation that change the 'nuclear stress', i.e. the location of the last pitch accent, and consequently I will only mark the last accent of a sentence by capitalization. 1 Recently this phenomenon has often been framed in terms 'givenness' or 'anaphoric destressing ' (cf.Williams 1997, Schwarzschild 1999, Sauerland 2004, Reinhart 2006), based on notions of 'givenness' and 'recoverability' discussed a.o. in Halliday (1967), Chafe (1976), Prince (1981).In contrast to focus-based approaches (e.g.Chomsky 1971, Jackendoff 1972, Rooth 1992b)), it is not the focused but the given constituent that imposes conditions on the context.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 137 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.