Transfer of verb placement in L2 English and L3 French Evidence from on-line and off-line measures

Abstract

The present thesis deals with transfer in L2 English and L3 French by speakers of L1 Norwegian, targeting finite verb placement in certain sentence types where speakers’ previously acquired language(s) display(s) (non-)correspondence with the target language regarding word order. The thesis focuses on the question of whether task types that vary regarding the degree to which speakers may access their metalinguistic knowledge (i.e., on-line vs. off-line measures) affect the direction of the results in experiments. The thesis consists of three research papers, each focusing on different aspects relating to transfer of verb placement in Ln and/or the impact of different task types in Ln research.

The first paper investigates whether L1 Norwegian intermediate/advanced speakers of L2 English exhibit residual optionality with respect to verb placement in main and relative clauses with a sentence-medial adverb, and whether speakers’ accuracy and response bias in acceptability judgments are affected by task type (timed vs. untimed judgments). The study found no evidence of transfer of Norwegian verb placement in this population irrespective of task type. However, the population tested was found to have a salient no-bias, which may have masked the potential presence of transfer. These results suggest that speakers in this population are highly proficient in their L2, but perhaps also over-confident, as they tend to reject grammatical structures. The second paper investigates the acquisition of finite verb placement in L3 French by L1 Norwegian L2 English speakers at university level, focusing on speakers’ performance on three different sentence types in an untimed judgment task. Speakers were found to perform better with non-subject-initial main clauses, where English word order corresponds to French, than they did with subject-initial main clauses with a sentence-medial adverb, where Norwegian word order corresponds to French. Furthermore, speakers’ performance was no better with the subject-initial main clauses than with relative clauses with a sentence-medial adverb, where neither of their previously acquired languages corresponds to L3 surface structure. These findings suggest that in L3 French learners at this proficiency level, influence primarily comes from L2 English. The third paper investigated the same speakers’ transfer patterns in L3 French in self-paced reading, an on-line measure which disallows metalinguistic access. The study views the results in light of previous findings from off-line measures, focusing on the untimed judgment task reported in the second paper. The study found that speakers exhibited the same transfer patterns (namely that influence primarily comes from L2 English) as what has been found previously in off-line measures. The overall implications of the findings from the three papers are discussed in the cover article, focusing on three main points: Firstly, it is argued that the high English proficiency in this population, along with their salient no-bias, is indicative of high automation of, and confidence in, the language, suggesting that their L2 is to a large degree procedurally based (Bardel & Falk, 2012, p. 70). Secondly, I propose that the apparent influence from speakers’ L2 English in the acquisition of L3 French is most likely explained by the linguistic similarity between the two languages, rather than the L2 status of English. I also suggest that there could be an interaction with cognitive economy in terms of verb movement, along the lines of Busterud et al. (2023). Finally, the fact that different task types did not reveal any differences in speakers’ performance suggests that the different experimental methods used tap into the same type of knowledge, i.e., implicit knowledge.