Collaborative Writing in L2 Classrooms

1   Introduction

The Aims of the Book

Writing is generally perceived as a solitary, individual activity. Writing in pairs or small groups is a novel activity and there are reported observations of teachers’ reluctance to implement such activities (e.g. McDonough, 2004). Some of this reluctance may stem from the perception of writing as an individual act as well as from assessment practices that tend to measure individual achievement. It may also stem from a lack of awareness of the potential benefits of collaborative writing for language learning or a lack of knowledge of how best to implement such writing activities. However, collaborative writing is likely to increase given developments in Web 2.0 technology, and particularly the use of wikis and Google Docs – new collaborative writing platforms. Ortega (2009a) argues that in our technologically driven world, the inclusion of computer mediated activities in language classes is no longer a choice but an imperative. Research on the use of wikis in second language classes suggests that, as in the case of face-to-face collaborative writing, online collaborative writing activities need to be carefully designed.

Thus this book has two overarching goals. The first goal is to encourage language teachers to consider implementing collaborative writing activities in their classes. The book attempts to provide a theoretical, pedagogical and empirical rationale for the use of collaborative writing activities in second language (L2) classes as well as some guidelines about how to best implement such activities in both face-to-face and online modes. The second goal is to encourage researchers to continue investigating collaborative writing activities. The book critically reviews the available body of research on collaborative writing and identifies future research directions. It should be noted at the outset that throughout the book the term second language (L2) is used as an umbrella term to refer to both second and foreign languages, although I acknowledge that there are important differences between second and foreign language contexts in terms of exposure to the target language and learners’ need and motivation to write in the target language (see Manchón, 2011a).

What Does Collaborative Writing Mean?

Let me begin by defining collaboration, the central term in this book. Collaboration means the sharing of labour (co-labour) and thus collaborative writing, in its broadest sense, means the co-authoring of a text by two or more writers. Some writing scholars (e.g. Bruffee, 1984; Harris, 1994) assert that all writing is collaborative to some extent. Individual writers composing with a certain reader in mind or seeking assistance from others at some stage of their writing can be said to engage in collaborative writing. Under such a broad definition, peer editing or peer planning would also qualify as collaborative writing.

An alternative view of collaborative writing is offered by Ede and Lunsford (1990). The authors identify three distinguishing features of collaborative writing: (1) substantive interaction in all stages of the writing process; (2) shared decision-making power over and responsibility for the text produced; and (3) the production of a single written document. From this perspective, collaborative writing is a distinct process and product. The process is one where participants work together and interact throughout the writing process, contributing to the planning, generation of ideas, deliberations about the text structure, editing and revision. This process is not merely an exchange of ideas but negotiations which often arise as a result of a struggle to create a shared understanding and shared expressions (Schrage, 1994). The product of the collaborative writing process is the jointly produced and shared text, a text that cannot easily be reduced to the separate input of individuals (Stahl, 2006). As such the text produced is also jointly owned, with all writers sharing in the ownership of the text produced.

On the basis of this definition, peer planning or peer editing (often referred to as peer response) do not qualify as collaborative writing because the interaction occurs only at one stage of the writing process (planning or editing) and the process of writing remains a private act. More importantly, ownership of the text produced rests with the individual writer rather than being jointly owned. Hirvela (2007) uses the term ‘collaborative approaches to writing’ to describe peer planning or peer editing, rather than ‘collaborative writing’. Collaborative writing also excludes editing tasks where the learners are asked to amend a text that they did not compose, or a text-reconstruction task where learners have to reconstruct a text based on given content words (see Storch, 1998a, 2001a).

Although I have previously referred to such tasks as collaborative writing tasks, on reflection I think that these kinds of grammar-focused tasks, where learners are not involved in constructing a text, should be labelled collaborative editing or reconstruction tasks rather than collaborative writing tasks.

Similarly, group projects, a frequent form of assessment at universities (Leki, 2001; Strauss & U, 2007) which are said to emulate the kind of writing prevalent in the workplace (Ede & Lunsford, 1990; Lay & Karis, 1991; Mirel & Spilka, 2002), do not necessarily qualify as collaborative writing activities. Here Dillenbourg et al.’s (1996) distinction between cooperation and collaboration is useful. Whereas cooperation involves the division of labour between individuals in order to complete a task, collaboration involves individuals in a coordinated effort to complete a task together. Research (e.g. Ede & Lunsford, 1990; Lay & Karis, 1991; Leki, 2001) has shown that in group projects, responsibilities are often divided, either by negotiation or by an assigned group leader, with each member of the group having a defined role. These roles may include the drafting of one discrete section or the editing of the entire document once it has been completed. Thus, what such an activity describes is cooperative writing (Dillenbourg et al., 1996), a form of coauthoring which involves the production of ‘a singular text by multiple authors’ (Ede & Lunsford, 1990). In collaborative writing, roles and contributions to text creation are not split up. Instead, there is mutual engagement and a coordinated effort by all members of the group or pair throughout the composing process.

Thus, in this monograph, collaborative writing describes an activity where there is a shared and negotiated decision making process and a shared responsibility for the production of a single text. In the L2 class, the text produced may be a composition or a report, but can also include more language-focused tasks such as a dictogloss, where students work in small groups or pairs to reconstruct a text based on notes taken from a dictated text (Wajnryb, 1990). However, it excludes grammar exercises such as joint editing, cloze or text reconstruction. In such tasks, students do not compose a text, rather they ‘reprocess language previously produced by others’ (Manchón, 2011b: 76). Nevertheless, I will refer to studies reporting on learners completing such grammar tasks in pairs as their findings are of relevance to a discussion of collaborative writing as a site for language learning.

It should be noted that the outcome of a collaborative writing activity is not just the jointly produced text. It is also collective cognition, emerging when two or more people reach insights that neither could have reached alone, and that cannot be traced back to one individual’s contribution (Stahl, 2006). In the context of L2 learning, it is cognition related to language learning, including, for example, learning new vocabulary, improved ways of expressing ideas, gaining a greater understanding of certain grammatical conventions or greater control over the use of a particular grammatical structure.

Outline of the Book

Chapter 2 presents the theoretical and pedagogical rationale for the use of collaborative writing tasks in L2 classes. The chapter also includes a brief review of collaborative writing in first language (L1) composition literature, where collaborative writing is relatively well established. However, as noted above, L1 scholars promote collaborative writing as a vehicle for developing good writing skills. In L2 contexts, the rationale for collaborative writing is generally to develop language skills. Manchón (2011b) distinguishes between using writing activities as the means to develop writing skills; that is, learning to write (LW) and activities which use writing to learn language (WLL). Using this distinction, the rationale for collaborative writing in L1 is predominantly couched in terms of learning to write (LW); in L2 it is writing to learn language (WLL).

Chapter 3 reviews empirical research on collaborative L2 writing showing that such tasks provide learners with language learning and language practice opportunities. It presents excerpts from a range of studies on collaborative writing showing what learners focus on when they deliberate about language, and how they use language in their deliberations. The chapter also discusses extensively the unit of analysis used in this research, the language-related episode (LRE).

Chapter 4 discusses the factors that may impact on the number and quality of the LREs found in the talk of learners when they write together. Here I include reference to studies where the learners completed grammar-focused tasks in pairs. The factors discussed include the type of task, the learners’ L2 proficiency and the relationships they form. Chapter 5 reviews the relatively small body of research investigating the outcomes of collaborative writing activities. The outcomes considered are in terms of the nature of the coauthored text and evidence of longer term language learning.

As mentioned previously, language teachers may hold some reservations about using collaborative writing tasks for language practice or assessment purposes. Learners have also been observed to be reluctant to participate in collaborative writing activities. Chapter 6 discusses the language learning beliefs and concerns that underpin teachers’ and students’ attitudes towards group and pair work in general, and by implication to collaborative writing tasks. The chapter then presents the findings of a relatively small number of studies which have elicited learners’ evaluations of collaborative writing once they had experienced such activities.

Whereas the previous chapters focused mainly on face-to-face collaborative writing activities, Chapter 7 focuses on collaborative writing that is computer mediated. It discusses briefly research on collaborative writing using text-based online communication, but the main focus of this chapter is on wikis, the new collaborative writing platforms. Wikis have a number of features which facilitate the creation of collaborative texts by multiple authors. The chapter describes these features and then reviews the main strands of research on wiki collaborative writing in both L1 and L2 contexts.

Chapter 8 concludes with a summary of the main themes covered in the book, reiterating the main reasons for implementing collaborative writing activities, both in face-to-face and online modes. It then identifies the decisions that L2 instructors need to make before implementing collaborative writing activities in their classes and the challenges they may face. Some suggestions are put forward for how such writing activities could be implemented in order to maximise the language learning opportunities they offer. Throughout the book, I note the dearth of research on a number of aspects related to collaborative writing. The final section of the chapter thus identifies future research directions.

2   Theoretical and Pedagogical Rationale for Collaborative L2 Writing

Introduction

Collaborative writing involves learners interacting in pairs or small groups on a writing task. Thus the two key components in collaborative writing are verbal interaction and writing. Verbal interaction has been identified as fundamental in both cognitive and sociocognitive theories of second language (L2) learning. The act of writing also has language learning potentials. The cognitive processes that occur in the production of oral language also occur in the production of written language and in fact some research suggests that writing may be superior to speaking as a site for L2 learning.

The first section of this chapter discusses the importance of interaction from both theoretical perspectives. It describes the evolution of Long’s (1983, 1985, 1996) interaction hypothesis and Swain’s (1985, 1993, 1995) output hypothesis, leading cognitive theories of second language acquisition (SLA). It then presents arguments and some research evidence which suggest that tasks which combine speaking and writing may be better than speaking only tasks in promoting interaction with a focus on language. A discussion of sociocognitive perspectives, and in particular Vygotsky’s (1978, 1981) sociocultural theory of mind, follows. Vygotsky’s sociocultural theory is, strictly speaking, not a theory of second language learning, but rather a psychological theory that explains the development of complex human cognitive abilities. The ability to acquire a second language is one such cognitive ability. It is relatively recently that sociocultural theory has gained recognition within Applied Linguistics research and has been used to explain second language learning processes. The work of scholars such as Lantolf (2000), Lantolf and Thorne (2006), Ohta (1995, 2001), and particularly Swain (2000, 2006, 2010), provides a radically different conceptualisation and analysis of interaction. Swain’s notions of collaborative dialogue and languaging are pertinent here; they provide a cogent argument in support of tasks that combine speaking and writing.

The second section presents the pedagogical rationale for the use of collaborative writing tasks. The discussion here first focuses on communicative and task-based approaches, considered to be best practice approaches to L2 instruction. These approaches promote interaction on tasks that require learners to reflect on language form while still being oriented to meaning. Collaborative writing tasks combine a focus on language form and meaning. The discussion then turns to the pedagogical rationale for collaborative writing as the means to advance not only language but also writing abilities. Here I draw mainly on the work of composition scholars (e.g. Bruffee, 1984; Dale, 1994a, 1994b) who have long championed the benefits of collaborative writing in general education and first language (L1) writing classes. Their arguments are equally applicable to L2 writing contexts.

Theories of Language Acquisition and Learning

Although a number of theories have attempted to explain how humans acquire or learn a second language, the two perspectives discussed here are cognitive and sociocognitive. Long’s (1983, 1996) interaction hypothesis and Swain’s (1985, 1993) pushed output are considered cognitive theories of SLA. Such theories view language acquisition as primarily a cognitive process and thus focus on what triggers learner internal cognitive processes. These processes include noticing, hypotheses testing and how the mind stores and retrieves information. In these theories, the learner’s existing mental capacity is the source of their own learning.

Vygotsky’s (1978, 1981) sociocultural theory, on the other hand, represents a sociocognitive view of learning. From this perspective, language learning has both social and cognitive dimensions. Sociocultural theory views the learner as a social being, and all cognitive development (including language learning) as essentially embedded in social interaction. Swain’s notions of collaborative dialogue and languaging (2000, 2006, 2010) are informed by sociocultural theory. They explicate how the use of language in social interaction mediates language learning.

Cognitive theories: Long’s interaction hypothesis and Swain’s output hypothesis

One of the most influential theories of SLA is Long’s (1983, 1985) interaction hypothesis. The genesis of this hypothesis is Krashen’s (1981, 1982, 1985) comprehensible input hypothesis, but Long’s hypothesis diverged from Krashen’s in terms of what makes the input comprehensible. Krashen claimed that the necessary and sufficient condition for SLA is for learners to be exposed to language that they can understand (comprehensible input) and which contains linguistic structures which are just beyond their interlanguage, their current L2 knowledge (i + 1). Krashen (1982) explained that what enables a learner to comprehend input beyond their interlanguage is the learner’s existing L2 knowledge and extralinguisitic knowledge. Long (1983, 1985) accepted that comprehensible input is key to L2 learning but claimed that a more consistently used and prevalent way of making input comprehensible is via interactional modifications during conversations. It is this claim that first highlighted the importance of verbal interaction for language learning.

Long’s (1983) claims were based on his own study which compared the conversations of native-speaker (NS) and non-native speaker (NNS) dyads with those of dyads composed of two native speakers. Long found that what distinguished the conversations between these two types of dyads were certain conversational moves that were used by the interlocutors to keep the conversation going. In NS–NNS dyads, the conversational moves included clarification requests, confirmation and comprehension checks. These moves were grouped under the umbrella term negotiation for meaning. Long found that when linguistic input was not comprehensible, and there was the potential for a communication breakdown, speakers engaged in negotiations to avoid or repair such breakdowns. Such negotiations led to modifications of the input (e.g. rephrasing) making it more comprehensible. Thus Long argued that verbal interactions in the form of negotiations for meaning promote acquisition because the interactions provide learners with the comprehensible input they need for acquisition to take place.

However, as a number of case studies of unsuccessful language learners (e.g. Schmidt, 1983) showed, exposure to comprehensible input is insufficient for successful L2 learning. More convincing evidence for the insufficiency of comprehensible input for L2 learning came from reports on the Canadian immersion programmes (Harley & Swain, 1984; Swain, 1985). Learners in these programmes were found to be able to use the L2 fluently but not necessarily with native-like accuracy, despite many years of exposure to presumably comprehensible second language input in the classroom. These findings led Swain (1985) to propose a comprehensible output hypothesis. Without denying the essential role of input in SLA, Swain argued that producing language (output) also plays an important role in SLA. Swain pointed out that comprehension and production involve different processes. Comprehension can be achieved by processing language input for gist by relying on understanding key content words or on general knowledge. Production requires learners to process language syntactically. Thus for successful L2 learning, learners need to be not only exposed to comprehensible input but also to produce spoken or written language.

In 1993 Swain revised her claim for comprehensible output calling for pushed output. Swain argued that just speaking and writing may provide learners with practice opportunities and hence develop their fluency but it may not develop their accuracy. For accuracy to develop, learners need to be pushed to produce language that is not only understood, but that is also grammatically accurate and which stretches their linguistic resources. Swain’s (1993, 1995) arguments in support of pushed output were based on important cognitive processes activated during language production. While attempting to produce in the L2, learners may notice gaps in their knowledge, gaps which arise when they encounter difficulties in expressing their intended meaning. This noticing may prompt learners to look for ways to address their gaps, including exploring their own internal linguistic resources and testing various hypotheses, processing language input more deeply (syntactically rather than semantically), and reflecting on their language use.

The study by Swain and Lapkin (1995), using think-aloud protocols of learners writing in their second language, showed evidence of learners engaging in the kind of cognitive processes identified by Swain’s pushed output hypothesis. Studies by Izumi (2002) and his colleagues (Izumi & Bigelow, 2000) further substantiated Swain’s claims. The studies, which compared learning gains arising from input and output conditions, showed greater learning gains from the output condition. Learners who were exposed to input immediately after they were required to produce output significantly outperformed the learners who were only required to process the input for comprehension. Izumi (2002) proposed that the need to produce output pushes learners to process language more deeply and thus establish a more durable memory trace. It should be noted that in Izumi’s studies, learners were required to produce written output.

Long’s (1996) revised interaction hypothesis took account of Swain’s arguments for the importance of pushed output for language acquisition, but also drew on the work of Schmidt (1990, 1994) on the importance of noticing (selective attention) for language learning, and on the work of Pienemann (1989) on processing capacity. As Long writes (1996: 451–452):

I would like to suggest that negotiation 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. (original emphasis)

Long conceded that comprehension alone was insufficient for SLA, particularly for the development of native-like L2 proficiency. Thus in subsequent writing, the importance of negative feedback for SLA was highlighted (e.g. Long, 2007; Long & Robinson, 1989). It is negative feedback which is said to draw learners’ attention to gaps in their L2 knowledge, either in their comprehension or their production. Such negative feedback can be provided during negotiations. For example, a clarification request is an implicit form of negative feedback. Another form of negative feedback is a recast which involves the correct reformulation of an incorrect utterance. The recast generates a negotiation of form. In sum, the argument put forward was that negotiations and negative feedback provided during such negotiations will lead to acquisition but only if the learner pays attention to the feedback (Schmidt, 1990, 1994), the learner is developmentally ready to attend to such feedback (Pienemann, 1989), and the learner attempts to incorporate the feedback in their modified or pushed output (Swain, 1985, 1993).

Although the interaction hypothesis focused on the language learning opportunities created in negotiations between a learner and a native-speaker or with a more competent peer, others (e.g. Gass, 1997; Pica, 1994) suggested that the benefits could also accrue to pairs of students of similar L2 proficiency. During such pair interaction, if the lack of comprehension stems from an ill-formed message, the request from the listener acts as negative feedback and encourages the speaker to pay attention to the grammatical accuracy of their utterance, making connections between grammatical form and meaning, and pushing the speaker to modify their output to make it more target-like. These negotiations allow for language hypotheses to be tested and refined in response to feedback. If the negotiation is triggered by a perceived or imminent misunderstanding, comprehension and confirmation checks may require the speaker to repeat, elaborate or rephrase the original message, providing the speaker with additional speaking practice. The listener also gains because presumably the input they are exposed to is now more comprehensible.

Researchers informed by Long’s original hypothesis set out to identify how to best pair students and what tasks to use in order to maximise negotiations. The assumption underlying this research was that the more negotiations learners engage in, the better. Thus it was found that if learners share the same L1 or are at a similar level of L2 proficiency, they are likely to engage in fewer negotiations than if they come from a different L1 backgrounds or