A Second Language Assessment Instrument—Elicited Imitation

The instrument of elicited imitation (EI) has been frequently referred to as a successful instrument for second language (L2) linguistic knowledge in second language acquisition (SLA). However, there is no consensus on its possibility in eliciting L2 learners'implicit knowledge. Its underlying mechanisms are rather complex and have not been systematically studied. It has been supported by many studies as good assessment of implicit knowledge (e.g., R. Ellis, 2005; Erlam, 2006). But still, there are doubts about this method and challenge about the possibility of rote repetition. With the purpose of investigating its capability in testing L2 implicit knowledge as proposed by R. Ellis (2005), the study answered the following questions:1) Does EI require subjects'active utilization of their working memory (WM)?2) If active processing occurs, do the results of EI correlate with that of a measurement of implicit knowledge?3) Do the results of EI correlate with that of a measurement of explicit knowledge?A total of 50 participants completed the battery of test. The experimental test consisted of the following parts: EI task, questionnaire of English learning background, WM task, grammaticality judgment test (GJT) and stimulated recall (SR). Among them, the EI task and WM task were administered through computer, and the other two tasks were both paper-and-pencil tests. The testing materials used in EI and GJT were from the same battery, with two different functors of different difficulty levels embedded. Two weeks after the main experiment, a language proficiency (LP) test was held. Through the data we collected and analyzed, we found that:1) There was a correlation with EI task and subjects'performance on WM test. This result can be regarded as the evidence of the active processing effect played in this instrument.2) Subjects'performance on EI task correlated strongly with the results of the implicit knowledge instrument—a LP test. EI performance correlated higher with listening and comprehension subtests than with tests of writing and vocabulary and structure. This pattern of correlation suggests that EI is capable of eliciting L2 learners'implicit knowledge.3) EI scores and GJT scores differentiated greatly from each other. GJT scores were much higher than that of EI and we did not find significant difference in the results of the two different functors embedded in the stimulus sentences, which are said to be of different difficulty level. But the result from EI showed a quite different pattern. However, a relatively moderate correlation was also found between scores of EI and GJT. The results we found from the stimulated recall add weight to the evidence that the correlation between EI and GJT indicates the involvement of explicit knowledge in EI.From the results we can see that EI, on the whole, is a good assessment instrument for L2 implicit knowledge. However, different from R. Ellis's (2005) claim that it is a task that can be authoritatively regarded as fully efficient in tapping L2 implicit knowledge, EI involves explicit knowledge.