Search

The intense debate on continuity in child grammars has largely focused on the existence of Functional Categories (FCs) in early grammars despite differences in language use with adults (e.g., omissions, word order errors, lack of case marking). Simplified, the major theoretical explanations for these properties range from Maturational accounts (Radford 1990 and later) claiming an initial absence of FCs, to Weak Continuity views (Clahsen, Eisenbeiss and Penke 1996) claiming one (or more) initially underspecified FCs that are subsequently specified by exposure to input, and further to Strong Continuity views (Poeppel and Wexler 1993) where an adult set of FCs is initially assumed and child language is claimed to be subject solely to performance constraints. A similar debate has raged in SLA (see Herschensohn this volume, for discussion) where an initial absence of FCs has been advocated by many scholars (Vainikka and Young-Scholten 1996, among others). The opposite view, i.e. initial presence of FCs, has been defended by scholars such as Schwartz and Sprouse (1996), who claim that, initially, structural representations of the L2 are based solely on the L1, and by scholars claiming direct UG-access to the FCs (White 1996, Prevost and White 2000b). Most of these studies concern adult SLA. One way to investigate FCs is to study the acquisition of clitics. There is a strong connection between FCs and clitics; clitic pronouns in French have an especially tight relation to the finite verb, which they precede in most cases. Since French is a verb raising language, it follows that the clitic must also move to an FC at spell-out. A common argument, based on this logic, is that a structure of the type je l'entends (I it hear - ‘I hear it’) is diagnostic of the existence of (some) FC in the grammar of a particular learner. These facts have lead researchers investigating FCs in L1 and L2 to analyse the development of clitic pronouns (Hamann et al. 1996 on L1 monolingual French, White 1996 on child L2 French, Herschensohn, this volume, on adult L2 French), and Meisel (1994 on bilingual L1 French) who uses the emergence of subject clitics and finite verbs for determining when AGR is acquired. If scholars agree on the relation “if clitics then FCs“, the inverse relation is much more problematic. Certain data (see below) suggest that, in adult L2 acquisition of French, there may be object pronouns but not clitics. But a lack of clitics does not necessarily imply a lack of FCs. The question mirrors in a certain way the issue of Missing Inflection: if systematic and functional inflection is present, then we can conclude that FCs are accessible, but the lack of inflection does not necessarily imply the absence of FCs (Lardière 1998, Prevost and White 2000b). Now, the syntax of clitics, and more generally cliticisation (pronouns and articles), is in itself a long-standing issue in theoretical linguistics and especially in Romance linguistics. A recently developed theory of pronouns provides new perspectives from which to approach acquisition data. In their detailed analysis, Cardinaletti and Starke (1999) reveal a typology that seems to have been rapidly accepted (see peer comments in van Riemsdijk 1999), where pronouns are classified as either strong, weak or clitics. The distributional and interpretative properties of clitics, weak pronouns and strong pronouns depend on the amount of (functional) internal structure they project. Cliticisation in this view can be seen as a change in structural representation during the derivation, from more to less (i.e., from XP to X0). Since both UG-access and FCs are prerequisites for cliticisation, the study of cliticisation can contribute to a better general understanding of L1 and L2 acquisition. Indeed, the strong consensus on direct access to UG and to early instantiations of FCs in L1 acquisition is not as clear in adult L2 acquisition (see White 2000 for an overview). Furthermore, previous work on clitics in L1 and adult L2 acquisition suggests that there may be differences in the way these are acquired. There is, therefore, a need for further L1 and adult L2 comparative research addressing cliticisation and controlling for general access to FCs. In this paper, we will address the issue of differences between L1 and adult L2 acquisition with respect to cliticisation. Adopting the framework of Cardinaletti and Starke (1999), we will investigate how and to what extent subject and object pronouns and articles become clitics in developing grammars. The data come from bilingual first language acquisition (2L1) and adult second language acquisition (L2). The children are Swedish-French bilinguals and the adults native speakers of Swedish. The fact that Swedish, the “other” language here, is present in both cases allows us to separate transfer from age effects.

This thesis deals with the development of the DP (the Determiner Phrase) in bilingual first (2L1) and in adult second (L2) language acquisition of French. The thesis shows that the children and the adult Swedish learners acquire syntactic categories differently: the children follow a structural economy principle in positing as small categories as possible (Rizzi, 1998). The adult learners aim to establish as few different categories as possible, which leads them to larger syntactic categories. Both principles are claimed to be part of the Universal Grammar (UG). The empirical part contains three interrelated analyses. The first analysis deals with the omissions of determiners and investigates the status of early Functional Categories (FCs). The children and the adults differ with respect to the underlying grammar of DPs at the initial state: only the children go through a stage where determiners are generally absent and only in the children is there a correlation between suppliance of determiners and linguistic development. The results speak against theories in L2 where an initial absence of FCs is postulated. Moreover, we find only a weak correlation between the omission of determiners, Null Subjects and Root infinitives in 2L1. The independent development of DP and CP is compatible with a structural economy approach to omissions but seems to speak against underspecification accounts that predict a parallel development of DP and CP. The second analysis deals with the syntactic status of the definite article, a clitic head (X0). Data from elision and incorporation are used for the analysis. Following our main hypotheses, the syntactic “lightness” of the clitic definite article predicts different acquisition patterns in 2L1 and L2. This is borne out in the data. The children have few problems with elision and incorporation (au/du ventre) whereas the adult learners of the lowest level of development nearly never elide or incorporate with prepositions (but produce *à le/de le bar). The children rapidly perceive the grammaticalised nature of the definite article whereas the adult learners begin with a more “lexical” definite article that is subsequently grammaticalised. We argue that the children soon analyse the definite article as a head under D0 whereas the adult learners first treat the article as an inherently [+definite] XP-element in Spec-DP. The third analysis shows that the features “number” and “gender” are acquired differently: Number is accessible from an early stage whereas gender is absent at the initial state in both groups. We argue that the development of gender in 2L1 is a consequence of the clitic status of the article in combination with morphophonology of the noun. In contrast, the adult learners first use a default gender form of the article. The L2 development is lexical and item based. In conclusion, the main results are accounted for if we assume that 2L1 and adult L2 acquisition are guided by the two different principles of UG mentioned. This difference has consequences for the three domains studied, the structure of DP, the status of definite articles and possibly also for the development of gender.