Unequal election of morphemes in adult Second Language Acquisition

Morpheme accuracy and acquisition order is a frequent subject in the Second Language Acquisition research literature (Bailey et al. 1974; Krashen 1981; VanPatten 1984, 1995). However, the treatment is largely descriptive. That is, researchers offer few explanations for the data discussed. The most frequently invoked explanation puts the burden on first language acquisition theory because the observed orders are similar to those observed in first language acquisition (Krashen 1981; Flynn 1984, 1987; White 1985, 1989). In contrast, this paper explains observed accuracy orders on the basis of a model of morpheme classification originally proposed to account for other bilingual phenomena, the 4-M Model (Myers-Scotton and Jake 1999a). The paper bases its analysis on a cross-sectional analysis of acquisition of English by Japanese and Chinese first language speakers. The results indicate that not all functional elements have the same accuracy order. The paper argues that this order reflects the fact that functional elements differ in how they are activated. The model of morpheme classification assumes that there are three types of system morphemes (functional elements) as well as content morphemes. One type of system morpheme is indirectly elected at the same time that content morphemes are directly elected by the speaker's intentions. The other two types of system morphemes are activated later in the production process. Their abstract lexical-conceptual structure is not relevant to conveying the speaker's intentions, but rather is required by the grammatical frame of the target language. For this reason, the 4-M Model predicts that they are harder to acquire and the data reported in this paper support that hypothesis. That is, the data indicate an implicational hierarchy of morpheme acquisition: content morphemes are acquired before any system morphemes and early (indirectly elected) system morphemes are acquired before later system morphemes.