Tense, agreement and defaults in child Catalan: An Optimality Theoretic analysis

At the earliest stage of acquisition, children learning languages such as German, English or Swedish often have an “optional infinitive” stage. During this stage, child productions include both verbs used correctly as well as infinitival forms used as matrix verbs (e.g., Weverink 1989, Platzack 1990, Wexler 1994, Schutze 1997). It has been claimed that children learning null subject or richly inflected languages like Spanish, Catalan, or Italian, however, rarely use non-finite root forms (NRFs), or use them only at early stages and for a very short period of time (Guasti 1994, Torrens 1995, Bel i Gaya 1998). Recent work on the acquisition of null subject/richly inflected languages suggests that their developmental pattern is more complex than previously thought. One line of research indicates that “default”-like forms that do not exhibit tense or agreement may not be limited to the NRFs found in the speech of children acquiring relatively poorly inflected languages. As originally observed by Ferdinand (1996), in addition to using root infinitives and bare participles, French speaking children also overgeneralize the third person singular present indicative (3S-PI) form. Based on the large proportion of 3S-PI forms found in child speech, Grinstead (1998) suggests that Catalan and Spanish-speaking children probably also use the 3S-PI form as a default. Varlokosta et al. (1998) noted that since Modern Greek does not have an infinitive form, NRFs could not be observed in child speech, but they reported that a form ambiguous between the third person singular and the active participle was instead used as a default by Greek children. A second line of research has shown that during the acquisition of functional projections, children’s use of NRFs decreases gradually, not discretely, as they approach an adult-like state (Phillips 1995). Despite this observation, previous accounts of the optional infinitive stage do not contain a mechanism to account for the actual proportions of finite and default forms observed in the child’s speech. In this paper, data from three Catalan-speaking children is used to extend this research. We argue that there are default usages of the 3S-PI forms that can be separated from either correct or ambiguous productions and quantified in order to determine what proportion of children’s utterances are either NRFs or defaults. In order to understand the role of defaults in the bigger picture of the acquisition of functional projections, the development of tense and agreement is