Instruction and Developing Second Language Pragmatic Competence

The primary purpose of the present study was to investigate the efficacy of output in the context of instruction on developing second language (L2) pragmatic competence. The study focused on the ways in which opportunities for output raise adult ESL learners' awareness of pragmalinguistic forms, thus facilitating the acquisition of those forms. The setting of the study was an intensive English program (IEP) in the United States, and the research was conducted within the broad framework of Swain's (1985, 1995, 2005) Output Hypothesis. In the area of developing L2 pragmatics research, efforts to understand a possible role for Output have been extremely limited. In fact, developing L2 pragmatics research that focuses on acquisitional processes is a young subdiscipline, having emerged on the scene of pragmatics research only within the last two decades (Bardovi-Harlig, 1999; Kasper & Roever, 2005). Also, the focus of much developing L2 pragmatics research has been the effects of instruction on learners' developing L2 pragmatics (e.g., Ohta, 2001; Rose & Ng, 2001; Takahashi, 2005b). Although such efforts have implications for the present study, they have done little to make clear the possible contributions of Output to learners' developing L2 pragmatics. Studies with greater implications for the exploration of Output (e.g., Bardovi-Harlig & Salsbury, 2004; Liddicoat & Crozet, 2001; Martìnez-Flor & Fukuya, 2005) are explored in the review of literature, but are shown to have many limitations as well. For the most part, the construct of Output as a causative variable in developing L2 pragmatic competence has gone unexplored in any meaningful way. The research questions addressed in the present study were: 1. Is there an effect for Output in the context of video-based pragmatic instruction (including output-focused tasks) on the developing L2 pragmatic competence of adult ESL learners, as evidenced by the acquisition of pragmalinguistic forms (pragmatic speech acts) in appropriate contexts? 1a. Is there an effect for Output on pragmatic perception on a pragmatic acceptability judgment task (PAJT; Garcia, 2004) from pretest to posttest? 1b. Is there an effect for Output on written pragmalinguistic production on a written discourse completion task (WDCT) from pretest to posttest? 1c. Is there an effect for Output on oral pragmalinguistic production on an oral discourse completion task (ODCT) from pretest to posttest? 2. Is there a differential effect of Output on the two different types of pragmatic speech acts of directives and expressives? 2a. Is there a differential effect for Output on the perception of directives and expressives on a pragmatic acceptability judgment task (PAJT) from pretest to posttest? 2b. Is there a differential effect for Output on the production of directives and expressives on a written discourse completion task (WDCT) from pretest to posttest? 2c. Is there a differential effect for Output on the production of directives and expressives on an oral discourse completion task (ODCT) from pretest to posttest? 3. Is the proposed rubric a reliable instrument for the assessment of the pragmatic acceptability of the responses of adult ESL learners on discourse completion tasks (DCTs)? The variables tested in the present study included Output (two groups: + Output and – Output); time (pretest to posttest); learners' perceptions of pragmalinguistic forms (pretest-posttest); and the production of pragmalinguistic forms in oral and written modes (pretest-posttest). In addition, following Yamashita (1996), the rubric proposed in the present study for rating the participants' responses in terms of pragmatic acceptability was tested for reliability. This examination of the rubric itself allowed the first two research questions to be addressed with more confidence with respect to the validity and reliability of the measures. The findings were that, with respect to Research Question 1, the + Output instructional treatment had a significant effect on the + Output group participants' pretest-posttest performance on the PAJT. Also, a large effect size was calculated for the + Output treatment group. No significant effect was observed for the – Output group, and the effect size calculated was below medium in size. No significant effects were identified for either the + Output or – Output group on the WDCT task, although a relatively large effect size was calculated for the + Output group. Finally, on the ODCT task, there was a significant effect observed for the – Output treatment group and not for the + Output group, although both groups had relatively large effect sizes associated with their ODCT pretest-posttest performances. With respect to Research Question 2, the findings of the study revealed that there were some significant differences in the way the instructional treatment impacted the learners' perception and production of the two different types of speech act tested (i.e., directives and expressives). For the PAJT task, the + Output treatment had a significant effect on directives, but not on expressives. On the WDCT task, the learners in the + Output group demonstrated significant change from pretest to posttest for the expressive speech act items. Large effect sizes were also calculated for both instructional treatments on the expressives. There were no significant effects identified on the directive items. Finally, on the ODCT task, there were significant pretest-posttest effects with large effect sizes calculated for both the + Output and – Output instructional treatments with respect to the expressive items. Research Question 3 addressed the reliability and internal consistency of the 4-level rating rubric used to assess the acceptability of the participants' responses on the DCT instruments. The findings indicated acceptably high inter-rater reliability levels for three of the four instruments tested (the WDCT pretest and posttest and the ODCT pretest), whereas the consistency of the rating scores on the ODCT posttest were just below what is normally desired in social science research. The limitations facing the study are acknowledged. First, the ability to judge pragmatic responses is a subjective, human endeavor, and so is subject to error. In order to control for this potential source of error, multiple raters were employed for the scoring of responses, and consistency and reliability promoted. Also, the ability to isolate Output as a causative variable is no easy task. The study followed previous research (e.g., Beebe, Takahashi, & Uliss-Weltz, 1990; Izumi, 2002) in research and task design as far as possible in order to promote the validity and reliability of all aspects of the study. Another possible contribution of the study to the field of developing L2 pragmatics research, suggested in the second research question, is the addition of support to the growing numbers of studies that have proposed developmental stages for L2 pragmatics (e.g., Bardovi-Harlig & Salsbury, 2004; Ohta, 2001). The targeting of two categories of speech act in the present study represented an effort to understand whether different pragmalinguistic forms respond differently to the output-focused video pragmatic instruction featured in the treatment. If a developmental stage exists for pragmalinguistic forms, it is reasonable to expect that some forms might be acquired earlier than others.