Throughout the Above Discussion of Enhanced Input in CALL, It Was Impossible to Concentrate

Chapelle, C.A. (forthcoming). Interactionist SLA theory in CALL Research. In Petrie, G. (Ed.) Research perspectives on CALL. Mahwah, NJ: Laurence Erlbaum Associates. [DRAFT Feb, 2004]

Interactionist SLA theory in CALL Research

Carol A. Chapelle

Iowa State University

I’m not really sure what interactivity is myself. (Rose, 2000, p. xii)

This statement was reportedly uttered by a salesperson who had been demonstrating interactive educational software to a client. The story goes that several employees of an educational software company were pushing their product using the typical language of educational computing. When one of them admitted not knowing what the language was referring to, the others followed, saying that they did not know either what “interactivity” really meant. The boss assured them that it didn’t really matter because “interactivity” means lots of different things to different people. In other words, the quality of the software that they were highlighting and praising did not have any meaning in particular. The story sparked my interest because of the significance of interaction in second language learning and the way in which interaction and interactivity are terms used to express the positive qualities of CALL. But like the salespeople in the story, would CALL researchers have to admit that they really don’t know what interaction means? When pressed, would they have to say that interaction really could mean just about anything, and that they have no idea why interaction is supposed to promote second language acquisition (SLA)?

Anyone who has worked with CALL has considered the meaning and value of interactivity and interaction, and many of the researchers who have been concerned with developing a better understanding of the potential benefits of CALL have looked to theory in SLA that explicitly hypothesizes the value of interaction for SLA. This paper briefly summarizes the basic tenets of a broad theory of interactionist SLA, illustrates how this theory has been incorporated into discussion of CALL pedagogy and research, and outlines some challenges associated with the application of interactionist SLA theory to CALL.

Overview of Interactionist SLA

The way that I have used interactivity and interaction above suggests that these might be distinct but related terms. I will return to this issue in the final section of the paper, but at this point I will use the term interaction as the superordinate concept that includes any type of two-way exchanges. Such exchanges might be enacted through the use of linguistic or non-linguistic means and would include events such as the sales transactions learners engage in at a fast food restaurant, over the telephone, or over the Internet. In all of these cases, the learner has a goal of making a purchase, and uses the linguistic and non-linguistic means necessary to complete the process by requesting and responding appropriately. Insofar as language is involved in constructing the meaning in such exchanges, interactionist theory would predict that they have greater potential for language development than activities in which interaction does not take place and it hypotheses why, how and when acquisition is expected to take place during interaction. The specific psycholinguistic processes involved in acquisition through interaction are explained by Gass (1997) and many other books and papers have described other perspectives on interaction in SLA.

Ellis’ (1999) broadly-conceived perspective of interactionist SLA is a good starting point for applying the concepts of interactionist theory to CALL. He notes that interaction is generally “used to refer to the interpersonal activity that arises during face-to-face communication. However, it can also refer to the intrapersonal activity involved in mental processing.” (p. 3). These categories are useful for CALL as well but in CALL interpersonal interaction takes place not only in face-to-face conversation but also electronically over a computer network. The interaction in CALL also needs to include interaction between the learner and the computer, probably what the salespeople at the beginning of the paper were calling “interactivity.” Ellis describes not only the benefits of interaction through the interactionist perspective but also from other perspectives as well, but here I have focused on the benefits as outlined in interactionist theory, as summarized in Table 1. Whereas Ellis included only “interpersonal,” meaning “between people,” I have added “between person and computer,” and added the logical predictions.

Table 1. Benefits hypothesized by the interaction hypothesis of three types of interaction

(adapted from Chapelle, 2003)

Based on hypotheses about how linguistic input can be transformed into the learners’ interlanguage knowledge (Gass, 1997), interactionist theory suggests that the process of interaction provides good impetus for acquisition. Although the benefits of the various types of interaction would not be expected to be mutually exclusive, the three types of benefits might be characterized as opportunities for negotiation of meaning, obtaining enhanced input, and directing attention to linguistic form.

The linguistic benefit to be obtained through interaction among learners is hypothesized to come about through learners’ negotiation of meaning. The idea is that negotiations that prompt conversational adjustments which help learners to make connections between form and meaning should be beneficial (Long, 1996; Pica, 1994). The research on negotiation of meaning, until recently was conducted by investigating the language of face-to-face conversation in the classroom. The idea of an adjustment that allows learners to make essential form meaning connections, can be extended to learner-computer interaction as well. During reading or listening, if the learner stops the input to request help (e.g., in the form of vocabulary help, rephrasing, or text transcripts) similar benefits might be expected (Chapelle, 2003) through these types of interactions. In both of these cases, the learners’ language or behavior reflects “language related episodes” (Swain, 1998) that may provide the opportunity to strengthen the aspects of the learners’ linguistic system that are fragile. Research on CALL has found evidence supporting the hypothesis that such learner-computer interactions are beneficial to acquisition of vocabulary (Plass, Chun, Mayer & Leutner, 1998, p.30). Intrapersonal interaction that takes place in the learner’s mind is expected to engage the type of deep mental processing that may promote acquisition. This is the type of cognitive activity that might be engaged by observable negotiation or requests for modification, but may consist of unobserved processes as well.

One might sum up the benefits proposed by interactionist theory as means of prompting learners to direct their attention in useful ways to linguistic input. In this sense, the interaction hypothesis is related to the “noticing hypothesis” (Schmidt, 1992), which hypothesizes the value of attention directed toward the key linguistic features during L2 tasks. All three types of interaction can occur during CALL tasks, but it is important to keep in mind that these are the current hypotheses about the benefits to be attained through interaction that stand as applied linguists seek evidence for the extent to which they are justified. The interaction hypotheses provide plenty of suggestions for CALL pedagogy, some of which have been the object of investigation in research.

Interactionist theory and CALL

Whereas ten years ago some imagination was required to find links between research on CALL and theoretical or empirical approaches to second language acquisition, CALL researchers today more frequently frame discussion of both pedagogy and research from the perspective of SLA. A variety of SLA perspectives are evident in CALL research today, but probably the most predominant comes from an interactionist perspective to SLA.

Interactionist perspectives in pedagogy

Discussion of CALL pedagogy frequently refers to the principles and examples that assume an interactionist perspective. For example, in many reviews of CALL software, the quality of the materials is discussed in part with reference to principles of interactionist SLA. For example, Byrnes’ (2000) review of Reading German criticizes the software because “it does not take into account the sophisticated pedagogical interventions that are now being advocated on the basis of cognitive-interactionist SLA research into the role of attention, awareness, and noticing of formal elements within an otherwise meaning-focused engagement with language, such as in reading” (p. 25). Similarly, Tatsumi’s (2001) review of Real English sites the lack of opportunities for learners to engage in meaningful target language use, noting that this limitation is at odds with interactionist SLA theory.

More generally, links between interactionist perspectives are made by Chapelle (1998), who identified some of the concrete aspects of instructional design that would ideally be built into software design to support the types of interactions that interactionist theory would predict to be beneficial to language learners. The following seven features of instructional design illustrate how the theoretically-motivated principles of beneficial interaction can be translated into guidelines for instructional design:

1. Make key linguistic characteristics salient.

2. Offer modifications of linguistic input.

3. Provide opportunities for "comprehensible output."

4. Provide opportunities for learners to notice their errors.

5. Provide opportunities for learners to correct their linguistic output.

6. Support modified interaction between the learner and the computer.

7. Provide opportunities for the learner to act as a participant in L2 tasks.

Expanding on these specifics of instructional design, Doughty and Long (2003) illustrate how principles of task-based language learning, which are based on cognitive and interactionist SLA theory, can be used to guide decision-making for task development in CALL, particularly in distance learning. The principles they describe are the following:

1. Use tasks, not texts, as the unit of analysis.

2. Promote learning by doing.

3. Elaborate input (do not simplify; do not rely solely on "authentic" texts).

4. Provide rich (not impoverished) input.

5. Encourage inductive ("chunk") learning.

6. Focus on form.

7. Provide negative feedback.

8. Respect "learner syllabuses"/develop-mental processes.

9. Promote cooperative/ collaborative learning.

10. Individualize instruction (according to communicative needs, and psycholinguistically).

The guidelines for instructional materials rely on a theoretical understanding of how language is acquired through interaction. Because they are based on theory rather than proven fact, and because they extrapolate from the theoretical articulation to a statement about a defensible course of action for materials development, these types of guidelines can also serve as a basis for research hypotheses about learning.

Interactionist perspectives in research

Pedagogical suggestions from interactionist theory such as those outlined above have served as research hypotheses for investigation of CALL. Interactionist-based CALL research can draw upon these to test out suggestions such as whether or not it is useful to provide negative feedback in a particular activity for certain students. In addition, CALL researchers can draw upon the more basic hypotheses about the benefits of interaction outlined in Table 1, operationalizing these in a variety of ways. Table 2 summarizes examples of CALL research that have drawn upon interactionist SLA to design learning tasks and analyze learners’ performance.

Table 2. The role of interactionist SLA theory in CALL research

Hsu (1994) drew on interactionist SLA theory to hypothesize the benefits of learners’ requests for help in an ESL listening comprehension task. From an interactionist perspective, she interpreted learners’ requests for help in comprehension as a form of request for modified input, which when received helped the learners to comprehend something that had initially not been understood. At the same time, these requests signaled segments where the learners had difficulties. She assessed outcomes through pre- and post-tests which had been constructed specifically for the research to include the lexical phrases in the input, and found a relationship between requests for modifications and comprehension of those forms that had not previously been comprehended. In a second ESL listening task, Kon (2002) attempted to discover whether incidental vocabulary acquisition would take place through a Web-based listening activity which required learners to listen to an academic lecture. In addition to hearing the lecture, the learners saw a video consisting of the lecturer, overhead transparency slides, and picture slides, which were followed by written multiple choice questions. In addition to a variety of descriptive and evaluative data, the researcher attempted to identify vocabulary that had been acquired during the listening activity, and to see to what extent their acquisition could be accounted for by aspects of the input and interaction. In addition to the overall positive finding of incidental vocabulary acquisition (as measured by improvement in listening comprehension for the words), the modes of presentation of the input also seemed to make a difference—the more modes of presentation the better.

The study by Pelliterri (2000) investigated Spanish learners’ negotiation of form and meaning during completion of text chat tasks while they were working on a variety of tasks intended to produce such negotiations. In a classroom setting, she developed tasks that research on classroom SLA had found to be successful in promoting negotiation of meaning. With the intention of identifying any instances of language related episodes during task-based negotiation, she found many instances of both negotiation of meaning and form. The learners actually corrected themselves and each other in during task performance that was primarily centered around the meaning of the tasks. Unlike examples of chat conversation outside the language classroom, the text chat seems to promote learners’ attention to their language in ways that the interactionist theory would suggest is positive for SLA.