Learning Japanese as a Second Language: A Processability Perspect ...

stressed that his ‘research on the teachability of second language is not related to any new or existing teaching method’ (p. 76), it has provided indispensable information for improving the effectiveness of second language teaching.

Despite the model’s contributions to SLA, it has received some criticism. First, there is not much information about how learners actually overcome constraints, that is, learners’ processing strategies (Larsen-Freeman & Long, 1991). Second, the status of the variational feature is unclear, which weakens the ability to falsify the model, if needed. The model has also been criticised for using speech processing strategies (i.e., COS, IFS, SCS) as an explanatory principle because this approach is based on transformational assumptions that have been proven to be psychologically implausible in recent grammatical theory (Pienemann, 1998b).

Predictive Framework¹

Pienemann and Johnston (1985, 1987a, 1987b, and later works) were the first to apply the ZISA research group’s multidimensional model to other languages, particularly to the acquisition of ESL. The predictive framework proposed an explanation for the implicational order of language acquisition in terms of a set of ‘universal speech processing constraints’, rejecting the transformation-based ‘speech processing strategies’ that the ZISA Group employed. The framework also aimed to predict the acquisitional stages in any language. This theory is situated at the transition point between the multidimensional model and the more elaborate PT (Pienemann, 1998b).

The empirical base for the application of the multidimensional model to the acquisition of English was provided by the extensive research of Johnston’s (1985) Syntactic and Morphological Processing in Learner’s English (SAMPLE) Project. In this project, Johnston looked at the development of L2 English cross-sectionally as well as partly longitudinally, with 12 Polish and 12 Vietnamese immigrants after their arrival in Australia. He found there was a fixed order of acquisition of English, regardless of the learners’ L1 background. This fixed order was also observed for modal verbs. In addition, Pienemann and Johnston’s (1985)