Review of Jin (2009) Third Language Acquisition

A critical review of Jin (2009) Third Language Acquisition of Norwegian Objects: Interlanguage transfer or L1 influence?

Hypothesis and research questions

This study investigates the acquisition of objects in L3 Norwegian by learners whose L1 is Chinese and who are advanced learners of L2 English. The main research question of this study asks whether the early Norwegian interlanguage grammars of these learners derive, through CLI, from L1 (Chinese) or from the ‘typologically’ proximate L2 (English).  In so doing, she aims to test Leung’s (1998) Interlanguage Hypothesis, which states that at the outset of L3A, grammatical CLI will come from the language already known that is typologically the closest to the L3.

Theoretical assumptions:

Chinese has SVO word order and allows null objects in both matrix and embedded clauses. English has also SVO word order but differs from Chinese in that null objects are not allowed in the standard variety. Norwegian has both SVO and V2 word order, and generarlly does not allow null objects in its standard variety, yet there are three main exceptions in this language. First, in the case of topic-linked null objects, topics can be dropped only in the sentence initial position in a V2 structure. The second exception is where null objects have arbitrary reference as for the pronoun one in English. Finally, an object can drop if it is in a co-ordinate sentence.

Jin agrees with Huang (1984, 1989, 1991) and Zushi’s (2003) claims that the mechanism responsible for the distribution of null/overt objects is the [±zero] setting of a Topic parameter: whereas, Chinese is a discourse-oriented language which allows the occurrence of null objects licensed by a zero topic (the [+zero] setting of the Topic parameter), English and Norwegian; in contrast, are sentence-oriented languages, and therefore do not allow null objects or zero-topic settings.

Given these assumptions, Jin aimed to test two hypotheses. The first is that group results would converge between the learner’s performance in English and Norwegian with regard to the resetting of the zero-topic parameter. Second, at the individual level, each learner should reject the ungrammatical sentences with null objects in almost equal percentage in both English and Norwegian versions of the task (Na Ranong 2009).

• Participants and methodology:

A total of 54 participants took part in this study. They were grouped as follows: 14 native Norwegian speakers served as a control group and 40 L1 Chinese-L2 English-L3 Norwegian learners. The latter were of three different proficiency levels in their L3 Norwegian: 14 beginners, 12 low intermediate and 14 upper-intermediate. Based on their score in the Test of English as a Foreign Language (TOEFL) and a questionnaire regarding their language learning experience, Jin concluded that all L3 participants were highly advanced L2 English speakers.

Participants were asked to do a grammaticality judgement task and a sentence correction task in both versions: Norwegian and English. Sentences in the Norwegian test had the same structure as the ones in the English test but they were not direct translations of the former.  Sentences in each task involved two types of sentence construction (null objects in root clauses and embedded clauses) and in each condition animate and inanimate distinctions were used, below are some examples cited from Jin (2009) which best illustrate the properties tested:

i)                    Null objects in root clauses

When you finish using the computer, please let use for a while.

Mary’s bike had gone wrong. I am going to repair for her tomorrow.

ii)                   Null objects in embedded clauses

       Mary lost her bike last week, but John said the police had found for her.

       John said those students were in the library, but I told him I didn’t find there.

• Results

Results show that Chinese L3 Norwegian learners rejected the sentences with null objects in English more accurately (at a rate of 72%) than in the Norwegian task. Jin considers the fact that the L3 learners failed to reset the Topic parameter in the L3 to the [-zero] setting, whereas in the L2 they succeeded in doing so as a good indication that such learners transferred the [+zero] setting from their L1 to their L3 rather than from the steady state L2 grammar. Hence, Jin claims that her work corroborates the “L1 transfer” hypothesis. However, she maintains that her study does not support Leung’s Interlanguage transfer hypothesis, since English and Norwegian are typologically closer than Chinese and Norwegian, and yet it is Chinese that appears to give rise to CLI. 

• Limitations of the study 

Although Jin’s study deals with an interesting area of syntactic knowledge, it does have some limitations.  The first lies in the experimental design of the work. The study included only 5 experimental test items, with only one to two tokens per condition (embedded/matrix clauses and animate/inanimate objects). This number is too small to draw generalisations. Besides, Jin’s study involves only one L3 group and one Norwegian native control group. Ideally, a group of L1 Chinese-L2 Norwegian participants without English background would have been important to check that test items in the L2 are not problematic. However, we all know that such groups are practically impossible to find. It is highly unlikely that a Chinese speaker fetching up in Norway would not have learned English first. An alternative would be a second group with L1 English-L2 Chinese-L3 Norwegian to investigate the source of transfer. Finally, in this study, participants were not independently tested for their L3 level of proficiency; rather Jin considered length of residence and length of exposure to the language to be sufficient indicators of proficiency level.