Iconic Agreement

Two apparently contradictory insights about sign language can be made compatible within a formal semantic analysis: (i) the claim that directional verbs incorporate agreement markers/pronouns (e.g. Lillo-Martin and Meier 2011); (ii) the observation that directional verbs have an iconic component (e.g. Liddell 2003). I sketch formal rules that integrate iconic requirements with standard semantic conditions (be they presuppositional or assertive in nature). In Lillo-Martin and Meier 2011 and in past work, the authors have forcefully and insightfully argued that directional verbs in sign language include agreement markers. At the same time, other researchers (notably Liddell 2003 and Taub 2001) have observed that directional verbs display iconic properties. Lillo-Martin and Meier grant this point, but maintain that it does not contradict their own analysis. However, since their goal is primarily syntactic, they do not lay out a semantic procedure that can make these two insights compatible. We sketch a formal semantics that seeks to do justice to both (we leave for future research a comparison with Zucchi’s recent account of some iconic effects in terms of hidden demonstratives (Zucchi 2011)). 1 The Pronominal Nature of Directional Verbs Lillo-Martin and Meier 2011 provide several arguments for the following theory: (1) a. “An agreeing verb copies the index of its argument, including values for person (first/non-first) and number.” b. “Co-indexing is interpreted as coreference at the meaning level, and is expressed by directing the sign (pronoun or verb) to the same locus.” The agreement theory can be spelled out as follows: (2) Directional verbs have overt or null arguments whose features and referential index they copy. This gives rise to two cases, depending on whether the arguments are overt or null. (i) Overt arguments a-MARY a-INFORM-b b-SUE a-IX PASS TEST (Lillo-Martin and Meier 2011) ‘Mary1 informs Sue2 that she1 passed the test.’