Abstract

Null objects have been attested in early language production cross-linguistically, with differences in the frequency but also in the length of the period of target-deviant object drop (Costa and Lobo 2006, 2010; Perez-Leroux, Pirvulescu and Roberge 2011; Wexler, Gavarro and Torrens 2003). One of the questions arising from this phenomenon is how to distinguish between null objects and clitic drop: are there different constraints regulating each phenomenon (discourse-related factors, input properties, markedness issues; Schaeffer 1997; Perez-Leroux et al 2008) or is clitic drop a subset case of null objects in developing grammars? A related issue concerns the verbs which appear to allow for null objects in child and adult grammars. I will argue that children make a distinction between verbs in the type of complements they produce (bare nouns, clitics, DPs, null) and to some extent in the frequency of null objects they allow them to show. The relation between null objects and clitic drop will be questioned by reference to a comparison between Greek typical L1 acquisition (spontaneous vs. elicited production) data and a set of data from Dutch L1 acquisition (spontaneous production only; Thrift 2003). The aim of the comparison is to discuss the availability of null objects cross-linguistically and the question of whether they constitute a linguistic and/or a pragmatic default.