Ming-Wei Ernest Lee
University of Cambridge
The implementation of subject-verb number agreement in English is characterised by a striking singular/plural asymmetry: Under laboratory conditions, the probability of English speakers producing a number agreement error on the verb immediately following a complex subject NP is influenced by the mismatch in number between the head and 'local' nouns in the subject NP when the mismatching local noun is plural ((b), compared to (a)), but not when the mismatching local noun is singular ((d), compared to (c)) (Bock & Cutting, 1992; Bock & Miller, 1991; see Eberhard, 1997 for further theoretical study).
a. The proposal by the expert (Singular Head, Head/Local Match)
b. The proposal by the experts (Singular Head, Head/Local Mismatch)
c. The proposals by the experts (Plural Head, Head/Local Match)
d. The proposals by the expert (Plural Head, Head/Local Mismatch)
This asymmetry has also been found in native speakers of other Indo-European languages including Spanish (Vigliocco et al, 1996).
The present study examines whether this processing characteristic can be found in agreement implementation by competent users of English as a second language. It investigates, in particular, the 'learnability' of native-like agreement implementation by learners whose first language is or is not similar to English in terms of subject-verb agreement and morphological marking of plurality on nouns.
Advanced Chinese and Spanish learners of English were tested on subject NPs like (a) to (d) in a spoken sentence completion task conducted in English. Participants were screened according to their performance in the Head/Local Match conditions (i.e., (a) and (c)) in order to ascertain that they had acquired, and were using fairly reliably, the basic subject-verb number agreement rule in English. The singular/plural asymmetry was found in the Spanish participants, who produced more agreement errors in (b) than in (a) but were no more likely to produce errors in (d) than in (c). In contrast, there was little sign of the asymmetry in the Chinese participants' performance in that the number mismatch between the head and local nouns affected both the singular- and plural-head conditions: (b) and (d) elicited significantly more errors than (a) and (c) respectively.
It is impossible to conclude from the present data whether the Spanish participants were using their L1 processing system to deal with their L2 or they were using a functionally and/or neuroanatomically independent L2 processing system (cf. Kim et al, 1997), to which the processing characteristics of the L1 system had been transferred. What is clear is that proficient Chinese learners of English (whose L1 has no subject-verb agreement and no morphological marking of plurality on nouns) implemented subject-verb number agreement in a qualitatively different manner from native speakers of English. This result will be discussed with reference to recent proposals about the acquisition of 'functional' categories in the literature on the generative approach to second language acquisition.
References
Bock & Cutting (1992). JML, 31, 99-127
Bock & Miller (1991). CogPsy, 23, 45-93
Eberhard (1997). JML, 36, 147-164
Kim et al. (1997). Nature, 388, 171-174
Vigliocco et al. (1996). Cognition, 61, 261-298.