The processing of subject-verb number agreement in L2 English by native Turkish speakers

Abstract

The present study aims to investigate how end-state Turkish learners process subjectverb agreement in L2 English as a second language (L2). The constructions under investigation are embedded noun phrase (NP) constructions made up of a head NP and daughter NP. The numbers of the two nouns are manipulated (i.e., the toy(s) for the kid(s)). The processing routines of the end-state L2 English speakers are investigated in comparison to those of native English speakers. Studies on first language (L1) sentence processing have revealed that adult native speakers of English tend to experience a processing difficulty in production and comprehension when the subject NP contains a singular head noun and a plural local noun (i.e., the (S)ingular-(P)lural condition) (e.g. the toy for the kids); but not when the head noun is plural and the local noun is singular (i.e., the PS condition) (e.g., the toys for the kid). This difference between the two conditions is called the mismatch asymmetry (Pearlmutter, Garnsey, & Bock, 1999, Vigliocco & Nicol, 1998). Two sentence preamble completion tasks were administered. The first one required declarative sentence production and the second one required question formation. The aim was to test whether subject-verb agreement computation is based on syntactic or linear distance. In addition, a word-by-word moving window technique was used as a self-paced reading task to compare and contrast the processing behaviors of the L2 speakers and the native controls in sentence comprehension. No difference was found between L2 speakers and the native controls either in the production tasks or in the comprehension task. The number of agreement errors and distributional trend across the experimental conditions in the declarative sentence production task was identical in both groups. The same mismatch effect was observed both in L2 speakers’ and the native controls’ agreement errors. In the question formation task, no mismatch effect was observed for either group and no difference was found between the L2 group and the native controls. The results of the comprehension task were parallel to those in the declarative sentence production task. That is, there was a mismatch effect in both groups. Yet, in sentence comprehension, the L2 group was slower than the native speakers with respect to their reaction times. These findings are in line with previous research not only in production but also in comprehension (Bock & Cutting, 1992; Bock & Eberhard, 1993; Pearlmutter et al. 1999).