Published online by Cambridge University Press: 27 June 2016
Current linguistic theory offers a highly detailed account of what linguistic competence consists of, as well as an indication of how that competence is acquired by L1 learners, via an innate Universal Grammar (UG). In second language (L2) acquisition, a major issue is the nature of the L2 learner’s competence and the degree to which it is similar to or different from the competence attained by native speakers. A theory of linguistic competence is essential to our understanding of what L2 competence might consist of, and should inform L2 acquisition research (Gregg 1989). In this paper, I will adopt certain recent proposals concerning verb movement and the functional categories Agreement and Tense (Chomsky 1989; Pollock 1989), and investigate whether these proposals can provide a suitable explanation of the linguistic behaviour of French learners of English.
1 I am grateful to Kevin Gregg and Roger Hawkins for comments. This work was carried out with the support of the following research grants: Government of Quebec FCAR research grant 88 EQ 3630 (to Lisa Travis and Lydia White), SSHRCC Canada Research Fellowship 455-87-0201 (to Lydia White), SSHRCC research grant 410-87-1071 (to Lisa Travis and Lydia White), SSHRCC research grant 410-90-0523 (to Lisa Travis and Lydia White), SSHRCC research grant 410-90-0097 (to Nina Spada, Patsy Lightbown and Lydia White).
2 In (2a), the verb has raised to COMP. French has other ways of forming questions as well, which do not involve subject-verb inversion. In such cases, the verb moves to T.
3 In Pollock’s structure, pas is the specifier of NegP. The head of NegP in French is ne, which raises to T, where it appears as a clitic before the verb. In English, not is the head of NegP. I omit these details here, as they are not crucial for my analysis.
4 In other analyses, the relative positions of T and AGR are reversed, i.e., T occurs lower in the tree than AGR (Belletti 1991; Chomsky 1989). For the purposes of the issues to be discussed here, it is important that there be a landing site in the lower position; whether it is AGR or T is not crucial.
5 For further details and other aspects of these studies, see White (1991a, 1991b) and White et al (1991).
6 However, correct usage of do often took the form of expressions like Do you have ... , How do you say . . . , How many . . . do you have?, so it might be argued that these are routines and that subjects do not yet have full command of do-support.
7 Since most of the cases of main verbs with 3rd person singular subjects occurred in questions, an alternative analysis of the lack of agreement on main verbs is that these questions contain a deleted auxiliary which carries the agreement.
8 In these examples, the sentence with verb raising is shown in (a) and the sentence without raising in (b). In the tests administered to the subjects, the orders were randomized. Furthermore, the test also included items where both sentences of a pair were incorrect, as well as cases where both were correct.
9 Statistics were calculated on the basis of the error scores. The difference between the mean scores on the three sentence types is significant for the Q group (F (2, 42) = 148.36, p = 0.0001) and the A group (F (2, 15) = 421.8, p = 0.0001). Scheffée tests show that mean scores on questions and negatives are not significantly different, whereas the scores on questions and negatives differ significantly from adverbs (p < 0.05).
10 The difference between the mean scores on the three sentence types is again significant for the Q group (F (2, 55) = 148.76, p = 0.0001) and for the A group (F(2, 13) = 3.93, p = 0.0323). Scheffé tests show that Q group’s mean scores on questions and negatives are not significantly different, whereas their scores on questions differ significantly from adverbs, and negatives also differ significantly írom adverbs. For the A group, on the other hand, mean scores on questions versus negatives and adverbs versus negatives are not significantly different. The difference in mean scores on questions versus adverbs is significant (p < 0.05).