Cue strength in second-language processing: An eye-tracking study

Abstract

This study used eye-tracking and grammaticality judgement measures to examine how second-language (L2) learners process syntactic violations in English. Participants were native Arabic and native Mandarin Chinese speakers studying English as an L2, and monolingual English-speaking controls. The violations involved incorrect word order and differed in two ways predicted to be important by the unified competition model [UCM; MacWhinney, B. (2005). A unified model of language acquisition. In J. F. Kroll & A. M. B. de Groot (Eds.), Handbook of bilingualism: Psycholinguistic approaches (pp. 49–67). Oxford: Oxford University Press.]. First, one violation had more and stronger cues to ungrammaticality than the other. Second, the grammaticality of these word orders varied in Arabic and Mandarin Chinese. Sensitivity to violations was relatively quick overall, across all groups. Sensitivity also was related to the number and strength of cues to ungrammaticality regardless of native language, which is consistent with the general principles of the UCM. However, there was little evidence of cross-language transfer effects in either eye movements or grammaticality judgements.

Keywords:

• Cue strength
• Cross-language transfer
• Eye-tracking
• Unified competition model
• Second-language learning

We thank Erik Reichle for access to the eye-tracking lab and procedure and Katherine Martin for her extensive help with the procedure and data collection. We thank four anonymous reviewers for their helpful comments on our manuscript.

During the writing of this manuscript, Alba Tuninetti was supported by the Behavioral Brain Training Grant, National Institutes of Health (NIH) [grant numbers 2T32GM081760–06 and 5T32GM081760–04], Tessa Warren was supported by NIH [grant number R01 DC011520], and Natasha Tokowicz was supported by NIH [grant number R01 HD075800].

Notes

1 In a few items the adjectives could have been interpreted either as being resultatives or initiating an adjective phrase (e.g., “the shirt striped with blue and white . . . ”). For these few items, a savvy reader might not have detected grammaticality violations until the posttarget region.

2 Ten native English-speaking participants who did not participate in the eye-tracking study rated the naturalness of our items truncated after the first word in the target region—that is, “She pulled skirt” versus “She pulled the”, on a scale from 1 to 7, on which 1 indicated “very natural”, and 7 indicated “very unnatural”. Items were distributed across two counterbalancing lists so no participant saw two versions of the same item. Participants rated sentence fragments ending with nouns as significantly more unnatural than fragments ending with “the” (4.54 vs. 1.58), t(55) = 17.57, p < .01, indicating that the bare noun may have been an additional cue to ungrammaticality.

3 However, it is important to note that Mandarin Chinese precedes its nouns with determiners, which are thought of as functional equivalents of “this” and “that” in English (Robertson, 2000). Definiteness and indefiniteness may be marked by certain prefixes, but this marking is not required (Li & Thompson, 1997). The possibility arises that Mandarin learners of English may build on their knowledge of these determiners to facilitate acquisition of articles in English, and if our learners do this, then the article–noun construction might be better categorized as similar to the L1 rather than unique to the L2. To foreshadow, however, the noun–adjective condition is much more important to our arguments about cross-language transfer than the noun–article condition, so the exact categorization of the noun–article condition ultimately is not critical.

4 Additional analyses including the entire set of participants that was tested showed a similar pattern of findings.