Reference: Insights from Sign Language (ASL and LSF)

Abstract: We argue that sign language data (here: ASL and LSF) can bring crucial insights into mechanisms of reference in language.

1. First, non-indexical pronouns are traditionally taken to come with unpronounced indices that provide them with a referential value. Sign languages have the advantage that these indices are arguably overt (Lillo-Martin and Klima 1990): coreference is typically established by assigning a position (or 'locus') to the antecedent, and by pointing back towards that position to realize the pronoun. Upon closer inspection, however, sign language loci have a dual face: while they display the syntactic and semantic behavior of variables, they can also function as simplified pictures of their denotations (Liddell 2003). We argue that in the end the formal and iconic aspects must be integrated within a 'formal semantics with iconicity', one that is expressively richer than what is standardly posited for spoken language.
2. Second, indexical expressions usually get their denotation from the context of speech. But in spoken language, it has been argued that the context of evaluation can be 'shifted' by some indirect discourse constructions (Schlenker 2003, Anand 2006). In sign language, this operation is arguably overt, and is realized by a shift of the signer's body and/or eye gaze. The typology of context-shifting operations is significantly enriched by sign language data. First, context shift does not just occur in indirect discourse, but also in action reports; and there are fine-grained grammatical differences between the two constructions. Second, both variants turn out to have an iconic component: under context shift, properties of signs must be interpreted 'maximally iconically', in the sense that properties of signs that can be taken to correspond to properties of the denoted situations must be so interpreted. As a result, in indirect discourse, context shift gives rise to near-quotational readings, by way of a mechanism that vindicates the view – expressed for instance in Recanati 2001 – that expressions can be simultaneously used and mentioned.

Note: The first part of the talk (on iconic variables) will cover data discussed in 'Iconic Variables' and 'Iconic Features', available online (see below). The second part of the talk (on context shift) is based on ongoing research.

http://semanticsarchive.net/Archive/zJjZjllO/SchlenkerLamberton_and_Santoro-Iconic_Variables.html

http://ling.auf.net/lingbuzz/001808