Feedback for language transfer errors the garden path technique

In this study we compared two methods for correcting language transfer errors in the foreign language classroom. Thirty-two English-speaking college students enrolled in two sections of an introductory French course served as subjects. Eight commonly encountered English-to-French transfer errors were identified and randomly assigned to one of two teaching conditions for one class section; each error was assigned to the opposite condition for the other section. In both teaching conditions students began by translating English sentences into French. The sentences were such that an L1 (first language) transfer strategy produced correct translations (e.g., using savoir for some uses of “to know”). A sentence for which the transfer would not produce an adequate translation (e.g., a sentence requiring connaître) was then introduced in one of two ways. In one condition—what we have termed the Garden Path condition—students were given the new sentence and asked to translate as before. Their inevitable transfer error was then immediately corrected by the teacher. In the control condition students were simply given the correct French form and told that it differed from the English pattern (they were not given the opportunity to commit a transfer error). Student learning of the non-transferable form was assessed three times throughout the course of the semester, and at all time points performance was better in the Garden Path condition. We interpreted this finding as support for a cognitive comparison model of second language acquisition. © 1989, Cambridge University Press. All rights reserved.

Duke Scholars

Author Michael Tomasello Psychology & Neuroscience