When languages don’t: null arguments and coherence relations

Abstract

Studies on null arguments have long debated if some of these constructions are elliptic rather than pronominal. As such, most recent claims have dichotomized null-argument languages into two types; those with pro-forms, and those with argument ellipsis (AE), with the idea being that pro-forms do not lead to sloppy identities, whereas AEs do (Roberts, 2022). However, this idea of (un)availability of sloppy identities in seeming pro-form situations does not appear to hold true across all cases, with context being a strong indicator of this type of resolution.

This paper argues that studies on null argument resolution should look into the innerworkings of discourse and information structure in addition to syntax. This claim is supported with data from Turkish, a language that has been previously claimed to display both pro-forms and AE (Şener & Takahashi, 2010). Using Kehler’s (2002) Coherence Relations, it is shown that the findings from syntactic literature are a result of the differences in information-structural elements of discourse and the coherence relations between the source and the target, indicating that the availability of sloppy identities is not a direct indicator of the presence of pro-forms or AE.