The output hypothesis and cognitive processes: An examination via acquisition of Japanese temporal subordinate conjunctions

Seiko Horibe, Purdue University

Abstract

This study investigates the validity of the output hypothesis put forth by Swain (1985) and cognitive processes that output triggers through the acquisition of target structures by adult learners of Japanese as a foreign language. As an experimental study, it compares two types of instructional treatments, input only and input + output conditions, in five classes of 50 minutes in order to examine the effects of opportunities for output on the acquisition of the target forms. The learners' thought processes during spoken output were elicited through retrospective interviews. The participants were 31 university students in a second-year Japanese course. Syntactic structures were selected as target forms in order to address the noticing and syntactic processing functions of output in language learning. The three intact classes were assigned to three conditions: input only (Input Group), input and output (Output Group), and no instruction on the target structures (Control Group). After a brief grammar explanation, the learners in Input Group engaged in comprehension of reading and listening materials containing the target structures. The learners in Output Group were given the same grammar explanation and the same reading and listening materials. However, they received opportunities to produce the target structures in addition to the input. Control Group subjects worked on materials without the target forms. The participants took a pretest and two posttests. Each test consisted of four components in order to examine the learning of the forms from various aspects. The results of the tests indicate that there is no statistically significant difference between Input Group and Output Group in terms of the acquisition of the target forms, while both groups maintained the learning effects until the delayed posttest administered nine weeks later. Both experimental groups outperformed the Control Group. The qualitative data, though limited, provide an insight into the learners' thought processes during output. A possible reason for the non-significant difference between the experimental groups may be due to a cognitive overload for the learners in Output Group who had to produce the target structures in addition to processing the input.

Degree

Ph.D.

Advisors

Berns, Purdue University.

Subject Area

Linguistics|Language arts

COinS