The semantic and aspectual properties of child L 2 root infinitives

One of the objectives of investigating child second language acquisition (SLA) is to find out to what extent it is similar to first language (L1) acquisition. Assuming that L1 acquisition is guided by principles of Universal Grammar (UG), it is often thought that second language (L2) acquisition by children is of the same nature, as it seems so easy for young learners to acquire a foreign language (Johnson and Newport, 1989; Bley-Vroman, 1990). However, the question of whether UG is accessible to L2 learners is debated mostly in the case of adults, who seem to struggle in the learning process. There exists little research on the exact nature of interlanguage grammars developed by children (see Hilles, 1991; Lakshmanan, 1991; Schwartz, 1992). It is this gap in the literature that the present paper wishes to address. The focus is on the nature of root infinitives (RIs) produced by children learning an L2, namely root declarative clauses whose main verb is either a past participle or an infinitive, whereas a finite form is required in the target language (such as papa partir 'daddy leave-INF' instead of papa part 'daddy is leaving'). Previous research suggests that there exists a period, starting in the early phases of acquisition, during which RIs are produced by child learners (Prévost, 2003; Prévost, 1997; Prévost & White, 1999). Evidence comes mainly from early L2 French and L2 German data. Research has crystallised over the issues of finiteness and the underlying structure of RIs. Are such clauses finite or nonfinite? Do they involve the projection of functional categories, and if so, what are the properties of these categories? Two main hypotheses have been put forward to account for the RI phenomenon in child L2 acquisition: the Missing Surface Inflection Hypothesis (MSIH) and the Truncation Hypothesis (TH). According to the former, RIs, while superficially nonfinite, are in fact finite (Haznedar and