James P. Pusack Tasks, Negotiation, and Grammar: The Instructional Roles of Multimedia in the Production of Meaning Abstract. This essay attemps to engage in a model dialogue between established SLA theory and current multimedia practice, using as a point of departure the work of Michael H. LONG. The insights provided by Long's work ! end themselves weil to extrapolation into the field of foreign language courseware design and delivery and clearly demonstrate the fact that a rigorous SLA theory and pedagogy can yield powerful insights for developers and implementers of foreign language instructional technology. In two major, interlocking aspects of acquisition grammar and comprehension - Long's work suggests and justifies significant design features that are relatively new, but not technologically unrealistic. In the realm of course design, however, the concept ofthe task-based syllabus or curriculum pushes the limits of what is currently feasible, especially when it comes to comprehensive intelligent tutoring systems for language learning. Conversely, to the extent that international teleconferencing, Web-based information searches, threaded email dialogues, and complex multimedia reference packages form a component of our target cultures, our assumptions about the available venues for negotiable discourse must be radically broadened. 1. Introduction In justifying the expense and effort needed to develop sophisticated multimedia courseware applications, many who work in this area tend to draw their rationales from quite diverse areas of traditional pedagogy and recent SLA research. Often, we praise the strengths of computer-assisted lessons as antidotes for the weaknesses of in-dass instruction. lt may also happen that computer materials, too often uncontextualized drills, are offered to meet learner needs or learning styles that are neglected by the current classroom approach and syllabus. The advent of digital multimedia environments that offer learners convenient networked access to a full range of tutorials, conversational settings, and authentic documents both text and video should prompt us to seek more satisfying and coherent strategies that clearly help us meet well-articulated instructional goals. This essay attempts to engage in a model dialogue between established SLA theory, on the one hand, and current multimedia practice, on the other. The purpose of such a dialogue should be threefold. First, it should ask and answer questions about the way networked digital multimedia can implement theoretical insights on language acquisition. Second, it should challenge SLA researchers to examine and account for new and emerging forms of instruction that go far beyond traditional classroom interactions. And third, it should suggest new forms of non-instructional discourse that have not yet fallen under the gaze of SLA theory and its derivative pedagogical practice. This model dialog will focus primarily on a single scholar, Michael H. Long, whose work (co-authored at times with Diane Larsen-Freeman or Peter Robinson) represents an extensive, evolving, but coherent body of research-based, theory-oriented scholarship that deals IFJLIIL 28 (1999) Tasks, Negotiation, and Grammar: The lnstructional Roles of Multimedia ... 45 not only with SLA as an object of investigation, but also as a pedagogical arena. The insights provided by Long's work thus ! end themselves extremely weil to extrapolation into the field of foreign language courseware design and delivery. Yet while this fruitful potential is manifest at every turn, neither Long nor his co-authors have ventured into this domain (at least in print) to address the many issues raised by non-traditional instructional modes based on multimedia technologies. Long's work thus sets the stage for a worthwhile dialogue. (Fora more wide-ranging attempt to engage methodologists in this field, see Pusack/ Otto 1991.) For those who are not familiar with the basic outlines of Long's thinking, abrief outline is in order. In a body of theory and research over a period of nearly twenty years, he systematically and often vehemently rejects both the strong naturalistic/ analytic pole of second language learning and the traditional/ synthetic pole with its many eclectic variants. Deploring what they dub "unproductive pendulum swings". between the two poles, Long and Robinson (1998: 21) use a clear line of logical argument based on available research into classroom practices to explain how acquisition occurs in the efforts of learners (and teachers) to accomplish comprehension. Long, who in contrast to many theoreticians does not balk at confronting issues of classroom practice, materials development, and syllabus design, offers his own prescription in the form of the slogan "focus on form", lately captured in the acronym FonF. This theoretical and practical construct explicitly and polemically challenges all methods and syllabi that organize student learning around the systematic treatment of an inventory of grammatical features, that is, a "focus on formS". As a consequence of this emphasis, Long subordinates the teaching of grammatical knowledge to meaningful situations and tasks through which learners must negotiate meaning in order to acquire language. In its most radical formulation, this program dictates abandonment of the textbook-based syllabus and leads to a task-based curriculum that is only now emerging in Long's writings. What can this approach to SLA bring to the development and use of foreign language multimedia? And conversely, what can multimedia bring to this program? These are the two sides of our dialogue. At first glance, the relationship may appear remote, since much of Long's work, like that of other research in this field, is based on studies of conversational interchanges, primarily in classroom situations. The most striking overlap between this aspect of SLA theory and the parallel development of multimedia over the past twenty years lies in the concept of interactivity. Having risen to the status of a buzzword in the early 90s, "interactivity" has perhaps best been exemplified in our field by numerous implementations of interactive listening comprehension tools for working with authentic foreign video. Most recently, "interactivity" seems to have taken a back seat to "connectivity" with the overwhelming predominance ofthe World Wide Web. Long's Interactivity Hypothesis, however, reinforces and theoretically justifies what has become obvious: Without high levels of interactivity, the convenient delivery of media-based documents (text, graphics, or video) resembles a supermarket more than a classroom. Long's approach offers us both standards and concepts for defining multimedia interactivity as it applies to language acquisition, especially to foreign language acquisition where access to native speaker (NS) conversational partners is virtually non-existent. Let us now turn to a more detailed look at his concepts and precepts, as they apply to multimedia in foreign language instruction. IFLllllL 28 (1999) 46 James P. Pusack 2. Just-In-Time Grammar and the Production of Meaning 2.1 Focus on Form Ever since the late sixties and as far back as teletype-based instruction, when rudimentary forms of computer-based language practice became feasible, grammar work has occupied a disproportionate share of the CALL (computer-assisted language learning) stage. This was consistent with the kinds of traditional materials being produced by textbook publishers, while reflecting the technical sophistication (or Jack of it) of the available delivery workstations, which were lineand text-oriented. In that era there were three kinds of CALL: 1) bad, i.e., that which provided no diagnostic help to the learner; 2) mediocre, which used a multiple-choice format with right/ wrong diagnosis; and 3) decent, which provided a helpful level of diagnostic feedback based either on error-anticipation with a corrective verbal message or pattern-marking approach that called attention to student errors without explaining them verbally (see Pusack 1983). Contextualization of such practice (or drill) was virtually non-existent, extending at best to scene-setting or some meaningful progression of sentences within an exercise. In the intervening three decades since those first experiments, technology and computing power have advanced dramatically, yet computer-based grarnmar practice has not fundamentally improved. The graphical user interface and the capability to display images and sound have somewhat improved the potential to provide elaborate contexts for practice, but corresponding advances have not occurred in the kind of artificial intelligence that could simulate a teacher's individualized response to foreign-language production errors. With a few worthy exceptions that did not spark a large-scale revolution (see Underwood 1984), publishers and funding agencies have not invested heavily in this subfield of artificial intelligence. The causes of this neglect are twofold. First, the profession's emphasis on communicative competence has relegated advances in grammar "teaching" to the backwaters of the profession, especially when any significant progress would require major investments. Second, no breakthrough in using computer-based interactions to stimulate acquisition of language structures has been possible without some coherent theoretical framework to situate such efforts. Thus, while teachers and students alike may have yearned for technology-driven fixes to the "grammar problem", a certain stasis has taken hold in this arena. A closer look at Long's assessment of our field's approach to grammar will help forge a way out of this impasse. Long and Robinson's analysis (1998: 16) begins with a rejection of uncontextualized grammar presentations and, by implication, such practice: "Of the scores of detailed studies of naturalistic and classroom language learning reported over the past 30 years, none suggest, for example, that presentation of discrete points of grammar one at a time[ ...] bears any resemblance except an accidental one to either the order or the manner in which naturalistic or classroom acquirers learn those items". Thus, the authors (ibid.) reject any synthetic "focus on formS" offered by lexical or structural syllabi and methods ranging from Grammar Translation to Total Physical Response and their associated classroom practices that "largely ignore language learning processes or tacitly assume a discredited behaviorist model." lFJL\llllL 28 (1999) Tasks, Negotiation, and Grammar: The Instructional Roles of Multimedia ... 47 Likewise, Long and Robinson (1998: 19) reject an analytic "focus on meaning" that exclusively emphasizes "the provision of sufficient quantities of positive evidence about what is possible in the L2", as found in Krashen and Terrell's Natural Approach. The authors see this position as more theoretically coherent, but flawed by virtue of factors such as maturational constraints, nonincorporation of input by advanced learners, unlearnability of some grammatical contrasts from positive evidence alone, and inefficiency (20-21). Long and Robinson (1998: 22) propose to discard the above two poles, presenting in their place a "focus on form" that "attempts to capture the strengths of an analytic approach while dealing with its limitations." This approach harks back to Long's earlier Interaction Hypothesis, which "holds that SLA is a process explicable by neither a purely linguistic · nativist nor a purely environmental theory". According to the Interaction Hypothesis, a crucial site for language development is interaction between learners and other speakers, especially, but not only, between learners and more proficient speakers and between learners and certain types of written texts, including "elaborated" interactions, that is, ones that interactively engage the learner and thus ultimately modify the original text (Long/ Robinson 1998: 22). Particularly important is the negotiation for meaning that can occur more or less predictably in certain interactions, according to the kinds of tasks in which speakers are engaged and the prevailing task condition. These theoretical reflections, based solidly on Long's careful review of the research, rehabilitate certain kinds of formal grammar work in instructed language acquisition, but under the new banner of meaningful context. And, as cited above, Long does not limit bis concept of interaction to conversational situations, but rather extends it explicitly to the domain of interaction with text and, by extrapolation, other media. The one unfortunate aspect of this program is Long's continuing insistence on the awkward distinction between bis own alliterative term "focus on form", captured in the unlovely acronym "FonF", and his term for the traditional "focus on formS". This is just too confusing, or at least misleading, since a teacher's prime directive within Long's "focus on form" would be to focus on meaning. Moreover, for those benighted practitioners who already focus only on formS, the apparent imperative or injunction to focus on form appears either reassuring or redundant. At the risk of upsetting this applecart, I suggest that Long's own formulation be abandoned. A preliminary replacement has been suggested in the title of this section: Just-intime grammar, which is in turn subordinated to the production of meaning. This quasiindustrial allusion has several advantages: a) its non-imperative structure signals that it is not a mandate, but rather a component; b) it takes the bold step of actually bringing back the term "grammar" into the lexicon of lesson planners, although "just-in-time form" would work, too; and c) it implies a larger context of timing and purpose. This proposed formulation tacitly asks "when? " and "why? ". lt does not say that there must be a casual attention to grammar sometimes or that grammar should be ensconced in some reassuring, entertaining context, but rather that there must be an urgent need for it. And it asks how such a need can be met effectively for widely diverse learners. FLI.IL 28 ( 1999) 48 James P. Pusack 2.2 Just-in-time grammar and multimedia In subsequent sections, Long's program for establishing a communicative need for grammar will be explored in detail. In the context of multimedia learning technologies, however, it is important to sketch briefly how current and emerging technologies can embed just-in-time grammar into a range of settings that includes writing tools, chat rooms, listening/ reading libraries, and simulations (see Pusack/ Otto 1997). In any and all of these settings, a language learner at a workstation can be provided with a plethora of highly interactive tutorials and practice formats. Given an identified and meaning-driven need for grarnmar training or review, even some relatively traditional forms of practice may conform to a given learner's cognitive style or personal perceptions of what might be helpful. Animation, sound, photos, and motion video can all be brought to bear on explaining, illustrating, and checking understanding. The limitations here are largely those of imagination and investment of resources to create such materials. This grarnmar utopia rests on several underlying requirements, however, that technology alone may not be able to fulfill. An obvious constraint (that may be overcome in time) is that student production in such a tutorial environment has largely been limited to screen-based or keyboard-based input. While the stimuli and context for learning can simulate reality with sounds and images, effective handling of spoken responses is, frankly, pathetic. Speech recognition software on typical student workstations cannot do much more than guess whether a student has responded appropriately to a prompt via voice input. The longer and more complex the anticipated response, the less satisfactory will be the computer's handling of it. Prospects for improvement here rest on the commercial benefits of voice recognition software. The second major constraint lies in the slow progress of the subfield of artificial intelligence (Al) called language understanding. Be it spoken or typed, student input that is meaningful and contextualized can and will take a limitless number of shapes and forms. Current advances in AI rely on radically lirnited contexts and scripts to provide meaningful "communication" between a human and a machine. Furthermore, use of native language lexical items, hesitations, misspellings and mispronunciations, and last but not least malformed grammar and syntax all promise to tax even the best language understanding systems for many years to come. We therefore cannot expect to employ such research cheaply and easily in instructional settings very soon. As a result of all these limitations, it is clear that any courseware that provides just-intime grammar will not likely accomplish this task in a naturalistic fashion. Learners will not soon find themselves engaging in realistic practice dialogues on a wide range of topics with a HAL-like computer persona who can not only understand their Statements, but also gently correct their grammar. This kind of interaction must still be reserved for student/ student or student/ teacher contact. Even when these practical limitations are accepted, we have not yet defined the precise way that just-in-time grammar delivery can be triggered: "Focus onform refers to how focal attentional resources are allocated. Although there are degrees of attention, and although attention to forms and attention to meaning are not always mutually exclusive, during an otherwise meaning-focused classroom Jesson, focus on form often consists IFLl.lllL 28 ( 1999) Tasks, Negotiation, and Grammar: The Instructional Roles of Multimedia ... of an occasional shift of attention to linguistic code features by the teacher and/ or one or more students triggered by perceived problems with comprehension or production" (Long/ Robinson 1998: 23). 49 Lacking the presence of an instructor, the multimedia environment must confront the issue of just how and when just-in-time grammar instruction is to be invoked. 2.3 Need and learner control No matter whether the setting is synchronous conversation with another learner or native speaker, or interaction with a newspaper article or TV soap opera, the identification of need is a critical component of Long's concept of instruction. The goal of such instruction is not only to provide the kind of knowledge dictated by the individual's need to communicate/ understand, but also to peg such information to the learner's precise developmental readiness (e.g., position in some acquisition sequence). The solution to this dilemma is to implement some combination of two approaches: 1) intelligent tutoring that continuously models each student's interlanguage at some level; and 2) learner-invoked tutoring that is perhaps more globally formulated for the learner's overall proficiency level (as measured by placement devices on-line or off-line). The critical issue here is the way that just-in-time grammar is invoked within a given meaning-driven context. Whether it takes the form of a sudden craving for formal accuracy in the context of on-line chat, or whether it bubbles up out of the frustration of trying to extract information from a foreign Webpage, or whether it emerges during an assignment to take on the role of a tour-guide in a video soap opera the learner's own psycholinguistic situation must be paramount if such instruction is to prove effective: "Focus on form, therefore, is learner-centered in a radical, psycholinguistic sense: it respects the learner' s internal syllabus. lt is under learner control: it occurs just when he or she has a communication problem, and so is likely already at least partially to understand the meaning or function ofthe new form, and when he or she is attending to the input. These are conditions most would consider optimal for learningthe psycholinguistic equivalent of worker control of the means of production" (Long 1998: 41 ). In the subsequent two sections we turn to the ways that communicative potential of networked multimedia can immerse the learner in contexts that make such just-in-time grammar instruction valid. 3. N egotiation for Meaning 3.1 Classroom and written venues for negotiation In his comprehensive report on the "Role of the Linguistic Environment in Second Language Acquisition" Long (1996) pulls together the results of dozens of studies that have looked at the role of conversation, input, and negotiation in the acquisition process. Virtually all of these studies relate to traditional classroom interactions or conversational interactions of various sorts, without considering the new forms of interaction that are afforded by the FlLlJIL 28 ( 1999) 50 James P. Pusack media both in target-culture situations and in learning environments. In naturalistic and classroom settings alike, Long (l 996: 418) posits the central importance of negotiation for meaning, which he defines as: "the process in which, in an effort to communicate, learners and competent speakers provide and interpret signals of their own and their interlocutor's perceived comprehension, thus provoking adjustments to linguistic form, conversational structure, message content, or all three, until an acceptable level of understanding is achieved". For Long (l 996: 451-452), the primary setting in which negotiation for meaning takes place is conversation, be it within or without the classroom: "[N}egotiation for meaning, and especially negotiation work that triggers interactional adjustments by the NS or more competent interlocutor, facilitates acquisition because it connects input, internal learner capacities, particularly selective attention, and output in productive ways". Long ( 1996: 453) further specifies the way various forms of input provide positive evidence, with some possibility of negative evidence, for acquisition of new structures: "Heightened attention makes detection both of new forms and of mismatches between input and output more likely, and such mismatches may also provide at least some of the information a learner needs about what is not permissible in a language. More such incorporations and changes can be predicted, therefore, in learners who receive higher quantities of semantically contingent speech through negotiation for meaning". The opportunities afforded by networked digital multimedia are particularly valuable to instructors in a foreign language setting where access to NS conversation is extremely limited. Moreover, in most instructional settings the small number of scheduled classroom hours available to learners means that technologies which can extend the times and places for conversation could prove critical to learner progress. If acquisition occurs via negotiation and if one can double the amount of negotiation per week, the likelihood of accelerating acquisition is heightened. 3.2 Extrapolation of negotiation to electronic communication Current technologies already allow for the convenient use of chat rooms and structured (or "threaded") conversations. lt is not hard to imagine how emerging digital audio and video technologies can make such conversations among pairs or small groups much more realistic, while overcoming the constraints of communication via typed input. On the surface, this seems especially promising because commercial conferencing software needed by the business world appears readily adaptable to educational settings at little or no additional cost. However, just as most instructors would not simply instruct a group of students to "go to dass and talk", the unstructured environment of a chat (be it typed, spoken, or watched) has many drawbacks. At more advanced instructional levels, these can be solved by the · clever use of collaborative assignments that culminate in a work product such as a travel itinerary. Even here, the temptation for learners to revert to the native language may be irresistible. At elementary levels, such unsupported environments bode ill for the effectiveness of the learning process. This means, then, that specialized software development will FJL\WL 28 ( 1999) Tasks, Negotiation, and Grammar: The Instructional Roles of Multimedia ... 51 probably be required, as discussed in the treatment of Long's concept of task-based instruction below. Long's concept of negotiation for meaning has evolved largely with conversational interactions in mindboth those between instructor and learner, and those between learners and other learners or native speakers. As indicated above, it appears quite appropriate to extrapolate the term "negotiation" to learners' work with texts, including multimedia documents. As we shift to the technological realm, however, some further reflection is called for in the consideration of those forms of technology that simulate oral communication via typed input. Those who have experimented with typed chat-room exchanges among learners have already begun to document its benefits. Kern (1995: 470), for example, studied the quantity and quality of French students' conversations using a networked environment called Deadalus InterChange: "Compared to oral discussions, lnterChange was found to offer more frequent opportunities for student expression and to lead to more language production. Furthermore, students' language output in InterChange was of an overall greater level of sophistication than in oral discussion, in terms of the range of its morphosyntactic features and in terms of the variety of discourse functions expressed. Direct student-to-student interaction stimulated students' interest in one another, contributed to peer lea: rning, and decreased students' reliance on the instructor". Kern also found what he called "unsettling" aspects of the medium, including the observation that "discussions often seem to lack coherence and continuity". lt is clear from his study that a new and interesting kind of discourse emerges under the conditions of networked communication. What is lacking for this type of instructional setting is a sufficiently detailed understanding about the kinds of negotiation that take place under the constraints that keyboard input imposes. So, while many benefits of such communication over even the best computer-based exercise types may be seen in terms of student motivation and the availability of intelligent understanding, several problems remain. One of these is the quality of the typed input, which may parallel spoken input but certainly will not have the syntactic and discourse complexity of authentic NS-composed texts. Although input qualities are already a problem in NS/ non-NS oral conversations, it can be expected that the keyboard input setting only exacerbates the differences between such conversation and NS/ NS dialog. A second concern relates to the ease of negotiation in a setting where non-verbal signals and even many verbal signals such as hesitation, reformulation may be invisible or effaced. Thus, to transfer our expectations for conversational negotiation of meaning to computer chat environments would simply be premature. Fortunately, the ease with which such exchanges can be captured and analyzed guarantees that research can easily discover much more about the embedded levels and types of negotiation found in this new mode of communication. 3.3 Extrapolation of negotiation to multimedia documents When we apply Long's concept of negotiation to archived multimedia documents of various sorts, the metaphor changes (in part) from the learner as diplomat/ negotiator seeking a satisfactory agreement to the learner as chauffeur negotiating a difficult, twisting road, IFLUl! L 28 ( 1999) 52 James P. Pusack possibly up a rather steep slope, possibly in dusk or darkness. Reaching the top may be the ultimate goal, but surviving each turn is vital, as well. Expanding this analogy just a bit more, we can see that while the driver does indeed have a fixed route to follow (perhaps with one or two detours or shortcuts), progress need not always be forward, tempo is variable, some roadsigns may be noticed, various driving aids are available headlights, maps but that unlike the progress of a conversation, the route and destination are fairly fixed. The key concept needed in adapting negotiation for meaning to the multimedia environment is that of "elaboration of meaning": "There appears to be substantial evidence of beneficial effects for various kinds of adjustments on comprehension, with elaborative, or 'interactional structure', modifications being successful, and having the added ad van tage of providing learners with continued access to the very linguistic items they have yet to acquire. Elaborative, or 'interactional structure', adjustments would therefore seem educationally more appropriate than what is commonly offered in current commercially produced materials" (Larson-Freeman/ Long 1991: 139). Examples of such elaborative interaction include repetition of various types, confirmation checks, clarification requests, expansion, lexical changes, and many others (see Larson- Freeman/ Long 1991: 145). Some pedagogical approaches to comprehension ones that simplify the inputare bound tobe counterproductive, because they tend to remove the very features that are needed for acquisition. In relationship to spoken interactions, Long ( 1996: 451) observes that "the comprehensibility brought about by interactional modifications allows the input itself to remain relatively more complex ultimately not modified at all thereby allowing learners access to new target forms, and eventually to the füll target code. Linguistic modifications that simplify the input, conversely, achieve comprehensibility partly by removing unknown forms, thereby improving comprehension at the expense of acquisition potential. [...] The semantic transparency achieved by interactional modifications as speakers negotiate for meaning is important, therefore, not just because it makes input comprehensible, but because it makes complex input comprehensible. Both comprehensibility and complexity are necessary for acquisition". This emphasis on a balance between comprehensibility and complexity clearly opens the door to consideration of elaborate multimedia interactions around authentic text and video documents. Larson-Freeman and Long (1991: 143) further support this conclusion in a review of the linguistic environment for language acquisition: "By suggesting ways that readers can modify the interactional structure of written discourse, we are also acknowledging that learners should not be viewed as passive recipients of input made comprehensible for them by others". Elaborated input can also be understood as the potential to modify "texts" in the widest sense of the word: "Modification of the interactional structure of conversation or of written discourse during reading [...] is a better candidate for a necessary (not sufficient) condition for acquisition. The role it plays in negotiation for meaning helps to make input comprehensible while still containing unknown linguistic elements, and, hence, potential intake for acquisition" (Larsen-Freeman/ Long 1991: 144). IFILd 28 ( 1999) Tasks, Negotiation, and Grammar: The Instructional Roles of Multimedia ... 53 3.4 Negotiation features of multimedia What then, are the features of multimedia that can provide the learner with opportunities to negotiate meaning? Larson-Freeman and Long recognize the fundamental negotiability of written texts and explicitly stress that "an initial choice of material with appropriate content can again be combined with modifications to the interactional structure of, this time, the written discourse, e.g. by varying the pace, and thereby processing time, through controlling reading speed, and by exact repetitions through re-reading phrases and sentences" (Larsen- Freeman/ Long 1991: 142-143). Some of these features of traditional print and tape renditions are carried over into the computer-based environment. These include: • organizational features such as tables of contents and indices; • structural highlighting features such as titles and illustrations; • advanced organizers provided by authors or instructors; and • ability to reread or replay document elements at will. The unique capabilities of interactive multimedia (currently with videodisc or CD-ROM), soon with networked digital video and audio) have been extensively demonstrated: • comprehension aids under learner control, such as glossaries and dynamic highlighting; • alternate media references (hyperlinks) to explain context and deepen understanding; comprehension checking with immediate feedback; • comprehension clues, such as selective replay and text-focusing tools that direct the learner's attention to salient points; • note-taking and writing tools (such as syntax checkers); • reference aids such as dictionaries and encyclopedia; • branching capabilities that move learners to other relevant documents. This incomplete list only hints at the possibilities. If, moreover, the concept of just-in-time grammar is applied, then a whole range of explanatory tools arises for consideration. These would include the ability to investigate or clarify the syntax of any sentence in the target text, to ask for level-appropriate clarification of grammatical features, and even to engage in a quick practice activity whose content would be linked directly to the target text. 3.5 Negotiation and learner control All of these features can most easily be provided under learner control, which means that their use will be subject not only to the learner's real-time comprehension needs, but also to the learner's own perception of his or her own needs. For this reason, it is tempting to imagine computer-adaptive systems that track a learner's acquisitional level and provide either mandatory or optional elaborations of the input. Currently, this kind of functionality represents a daunting task and one that contravenes one's intuitive sense that such courseware should work toward cultivating the learner's own ability to identify such needs and deficits. Thus, although it is widely recognized that learners may be quite inefficient at using grand arsenals of comprehension aids, only a modest level of obtrusive intervention is likely to be effective. Many current multimedia implementations therefore combine exposure to authentic text or video documents with two types of support: 1) control and help features FJLllllL 28 (1999) 54 James P. Pusack under learner control; and 2) selective comprehension-building activities that provide checkpoints along the path to understanding. 4. Task-Based Instruction 4.1 Development of the task-based syllabus The learner's needs, both for just-in-time (i.e., need-justified) grammar instruction and for negotiation opportunities, lie at the heart of Long's principled approach to second-language acquisition. Thus far, in exarnining the multimedia contexts available to meet such needs, we have considered highly interactive environments (chats, multimedia documents); yet these environments might not, in and of themselves, necessarily stimulate such needs within the learner. Long's approach (1996: 448) deals explicitly with this issue by advocating a syllabus design based on tasks, rather than sequenced language features or texts: "Free conversation is notoriously poor as a context for driving IL development for a number of reasons, because the Jack of any fixed topics or outcomes permits rapid, superficial treatment of topics and the dropping of any that cause linguistic trouble. In contrast, tasks that orient participants to shared goals and involve them in some work or activity produce more negotiation work. [...] When working cooperatively on certain kinds of problem-solving tasks (e.g. two-way tasks) that are closed known by participants to have only one or a small number of correct solutions, participants' conversational feet are held to the fire. The nature of the task causes topics and subtopics tobe recycled until solutions are reached, producing more negotiation work [...]". lt goes without saying that few language teachers design lessons or syllabi around significant blocks of unstructured time. (One fashionable exception might be the common expectation that novice learners can browse the World Wide Web effectively and productively in search of relevant information on a broad topic.) Long's proposals and initiatives in this area are radical, not incremental. He makes the task itselfthe fundamental unit of syllabus design: "Syllabus content is a series of pedagogical tasks [...], the justification for which is that the content or tasks are related to the current or future needs of the particular group of learners tobe served" (Lang/ Robinson 1998: 23). lt goes beyond the scope of this article to portray the intricacies of a full-scale task-based syllabus (see Long 1985). 4.2 Necessity of multimedia for the task-based syllabus Of primary interest here are the capabilities of multimedia to deliver the kinds of instruction envisioned by Long (1985: 95), who already recognizes that individual learner needs in a consistently-applied task-based curriculum will tend to explode the boundaries of traditional classrooms: "While the teacher is the basic source of target language data in the Natural Approach, the fact that often quite specialized learner needs are being addressed in task-based language teaching means that alternate sources must be provided, and may even predominate. Thus, utilizing traditional teacher resources, such as taped dialogues, original documents, and simplified versions thereof, models of language use accompanying performance of target tasks will be provided to IFlLrutlL 28 (1999) Tasks, Negotiation, and Grammar: The Instructional Roles of Multimedia ... leamers. There is nothing new in this, of course. Unlike traditional SL teaching, however, the purpose of such models is not to induce accurate replication by the learner. Rather, the models are to serve as a target towards which learners approximate, over time, as the pedagogic tasks the models accompany gradually increase in complexity until they attain füll target difficulty. The focus throughout is resolution of the problems posed by the pedagogic tasks themselves, and learner success on these is judged by task accomplishment, not target-like linguistic production while achieving this". 55 This passage amounts to Long's own incipient multimedia manifesto, for it is clear that the level of individualization demanded by such a syllabus cannot be provided by traditional delivery systems, even when various archived media are brought to bear. Only the inte- · grative potential of digitized media, with the capacity to apply a wide array of instructional resources to constantly changing learner needs within a progression of acquisitionally and pedagogically relevant tasks can deliver what Long wants. 4.3 Features of multimedia task-based syllabus Given an adequately articulated task-based syllabus, the fundamental features of such a delivery environment in the near future would consist of the following: a human instructor to perform needs assessment and face-to-face conversation; learner access to native speakers and non/ native peer learners to engage in collaborative tasks, either in person or via video teleconferencing; a networked digital archive of reading, viewing, and listening materials indexed for their relevance to a task-based syllabus; a database of reference, tutorial, and practice materials (including computer-driven tasks) that could be invoked under learner control andin response to learner need. For many reasons, this is an imperfect, yet realistic system that can be designed and delivered on a small scale, incrementally, with current digital technologies. The path to a more distant and ideal future for such a system requires that many of the above elements that must today, of necessity, remain almost fully under learner control be automated via intelligent tutoring systems that constantly respond to measured or monitored acquisition levels of the learner. As Otto and Pusack ( 1996: 169) have observed, ambitious goals in this area may be impossible to attain. The state of SLA theory and practice and the state of AI do not currently suggest that this is possible on any large scale, although ·the capabilities of the needed media are fully available. 5. Conclusion The three preceding sections have clearly demonstrated the fact that a rigorous SLA theory and pedagogyalbeit one based largely on classroom practice can yield powerful insights for developers and implementers of foreign language instructional technology. In two major, interlocking aspects of acquisition grammar and comprehension - Long's work suggests and justifies significant design features that are relatively new, but not technologically unrealistic. In <l third pedagogical realm, that of course design, the concept of the task-based f]Lll]l, 28 ( 1999) 56 James P. Pusack syllabus or curriculum pushes the limits of what is currently feasible, especially when it comes to comprehensive intelligent tutoring systems for language learning. This is one half of the model dialog proposed above. The other half of the dialog lies in the research agenda that can and must be transmitted back to SLA scholars. At the same time that multimedia implementations can be found for many of Long's tenets, their effectiveness needs tobe studied. Such study should, however, not be limited to judging whether, in general, multimedia represents a cost-effective alternative or enhancement to traditional instruction, for it is clear that in some cases (such as exposure to complex media documents) the technology can provide fullyunique opportunities. More significant will be focused examination of the way just-in-time grammar, negotiation for meaning with learner-controlled facilities, and task-management under control of a tutoring system can selectively augment face-to-face instruction. In particular, we need to discover those kinds of interactions that may be superior to their traditional antecedents. A final stimulus to SLA theory and research comes from the recognition that many of the forms of multimedia interaction envisioned here rely not on sporadic or idiosyncratic instructional projects, but rather on a changing discourse universe. To the extent that international teleconferencing, Web-based information searches, threaded email dialogs, and complex multimedia reference packages form a component of our target cultures, our assumptions about the available venues for negotiable discourse must be radically broadened. lt eventually becomes self-defeating to generate theoretical positions about how learners acquire language based solely upon what people do when they are in a room together or when they hold a book in their hands. Whether or not we can bend multimedia to our instructional purposes, learners will negotiate meaning in multimedia discourse and we must begin to understand its mies. References KERN, Richard G. (1995): "Restructuring Classroom Interaction with Networked Computers: Effects on Quantity and Characteristics of Language Production". In: Modern Language Journal 79, 457-476. LARSEN-FREEMAN, Diane / LONG, Michael H. ( 1991 ): An lntroduction to Second Language Acquisition Research. London and New York: Longman. LONG, Michael H. (1998): "Focus on Form in Task-Based Language Teaching". In: University of Hawai'i Working Papers in ESL 16 (2), 35-49. L0NG, Michael H. ( 1996): "The Role of the Linguistic Environment in Second Language Acquisition". In: RITCHIE, William C. / BHATIA, Tej K. (eds.): Handbook of Second Language Acquisition. San Diego [usw.]: Academic Press, 413-468. L0NG, Michael H. ( 1985): "A role for instruction in second language acquisition: Task-based language teaching". In: HYLTENSTAM, K. / PIENEMANN, Manfred (eds.): Modelling and Assessing Second Language Development. Clevedon: Multilingual Matters, 77-99. LONG, Michael H. / ROBINSON, Peter (1998): "Focus on form: Theory, research, and practice". In: DOUGHTY, Catherine / WILLIAMS, Jessica: Focus on Farm in Second Language Acquisition. Cambridge: Cambridge University Press, 15-41. OTTO, Sue K. / PUSACK, James P. ( 1996): "Technological Choices to Meet the Challenges". In: WING, lFILll.lL 28 (1999) Tasks, Negotiation, and Grammar: The lnstructional Roles of Multimedia ... 57 Barbara H. (ed.): Foreign Language for All: Challenges and Choices. 1996 Northeast Conference Reports. Lincolnwood, Illinois, USA, 141-186. PUSACK, James P. (1983): "Answer-Processing and Error Correction in Foreign Language CAI". In: System 11.1, 53-64. PUSACK, James P. / OTTO, Sue K. (1997): "Taking Control of Multimedia". In: BUSH, Michael D. / TERRY, Robert M. (eds.): Technology-Enhanced Language Learning. ACTFL Foreign Language Education Series. Lincolnwood, Illinois, USA: National Textbook Company, 1-46. PUSACK, James P. / OTTO, Sue K. (1991): "Dear Wilga, Dear Alice, DearTracy, Dear Earl: Four Letters on Methodology and Technology". In: SILBER, Ellen S. (ed.): Critical lssues in Foreign Language Instruction. New York: Garland Press, 80-103. UNDERW00D, John H. Linguistics (1984): Computersand the Language Teacher: A Communicative Approach. Rowley, Massachusetts, USA: Newbury House. lFILIIL 28 ( 1999)