2.2 Unraveling the writing construct: A review of the literature
A glance at the relevant literature tells us scholars define “writing ability” in various ways. Generally speaking, researchers view the nature of writing mainly from the following three perspectives: writing as a written text, writing as a process of composing a text, and writing as a socialcognitive activity.
2.2.1 Product-oriented view of writing
For many years, the notion of writing was decontextualized, considered independent of particular writers or readers, and viewed as an ideal form capable of being analyzed independently of any real-life uses (Hyland, 2002). The written products were seen as an autonomous object, where various elements were coherently and accurately put together according to a rule-governed system. A writing teacher who subscribed to the product approach was concerned about whether the end product was readable or grammatically correct, or whether it obeyed discourse conventions. The focus in class was exclusively on formal features of texts, and the practice primarily involved copying and imitation, carrying out sentence expansions from cue words, and developing sentences and paragraphs from various models.
The product view regards the act of writing as a process of encoding via employing skills at various levels. After establishing the main idea, a writer encodes in his/her mind. He/She first forms words with letters, and then, according to certain grammatical rules, makes sentences by using these words. Next, the writer composes sentences into a paragraph. Finally, a complete article comes into being via arranging several paragraphs. In essence, this view of writing is structuralist, treating the writing ability as a decomposable, complicated process which contains writing skills at various cognitive levels.
2.2.2 Process-oriented view of writing
In the 1960s, views shifted towards cognitive psychology's conception of learning as a process of discovery and learners as creative hypothesis generators (Grable & Kaplan, 1996). Meanwhile, many researchers dissatisfied the teaching of writing because the common practice of focusing on written products fostered coherence of meaning at the macro-structural level, but failed to address the psychological and cognitive processes intrinsic to writing (Braddock et al., 1963; Rohman, 1965). Therefore, scholars looked beyond written texts to inquire into the nature of composing itself, the process of creating the text. Writing, by nature, is a problemsolving activity that requires high level cognitive processing (Bereiter &Scardamalia, 1987; Flower & Hayes, 1981; Hayes & Flower, 1980). Various cognitive models suggest that the writing process is recursive, iterative, and nonlinear in nature, rather than a set of neat, sequential, and additive steps. In fact, writing-as-a-process has become so popular that it was claimed as “a paradigm shift” (Hairston, 1982) in writing theory.
Later, however, the process theory of writing was attacked fiercely later on by a number of writing researchers (e.g., Horowitz, 1986a) who were opposed to the exclusive focus on the process of writing at the expense of virtually all other essential elements of writing such as how writers responded in writing to a particular communicative setting. After summing up the criticisms leveled against the process approach, Horowitz (1986b: 447) further underscored the importance of the context of writing by asserting that “the form in which a writer expresses meaning owes just as much to the constraints of the writing situation as it does to the writer's mental processes”.
2.2.3 Writing as a social-cognitive activity
Since the mid-1980s, in response to the criticisms of the process theory of writing, a growing number of researchers (e.g., Faigley, 1986; Johns, 1990; Nystrand, 1989; Spack, 1988; Witte, 1992) shifted their focus towards the social context in which the process of composing took place. Writing, in their view, was a socially-embedded act. In contrast to the cognitivist view mentioned in the previous section, this perspective gave greater emphasis to the actual performance of writing in a particular context. Of crucial importance is a notion of context as the “situation of expression” (Nystrand, 1987). The remarkably distinguishing features of context of writing are often regarded as the audience and purpose of writing. Recognizing the importance of context for the act of writing, Odell (1981: 103) proposed to define competence in writing as “the ability to discover what one wishes to say and to convey one's message through language, syntax, and content that are appropriate for one's audience and purpose”. In a similar vein, more recently, Lindemann (2001: 10) perceived writing as a communicative act and explicitly remarked that “writing is a process of communication that uses a conventional graphic system to convey a message to a reader”. In accordance with the view of writing as a socially-embedded act, the linguistic patterns employed in a piece of writing are influenced by context which brings with them a variety of social constraints and choices. The writer's goals, relationship with readers and the content knowledge they want to impart are accomplished by the text forms appropriate to that social context.
It should be noted that emphasis on the social view of writing did not mean abandoning strictly cognitive inquiries (Berkenkotter, 1991;Flower, 1990a). Instead, each of them has a valued place in writing research. Therefore, writing has come to be viewed both as the effort of an individual and as a strongly contextualized phenomenon which should take into consideration the key contextual parameters such as the relationship between the writer and the reader, and the purpose of the writing.
At present, writing is widely regarded as a communicative act, which happens under particular circumstances. In line with current views on the nature of writing, for the test of writing, writing ability cannot be validly abstracted from the context in which writing takes place. The writing construct should, accordingly, be viewed as a social-cognitive act and tested under certain context which simulates factors making up the real world writing. This view is consistent with the socio-cognitive perspective on language testing put forward by Weir (2005: 19), who maintained that“language processing does not take place in vacuum, so testers also need to specify the context in which this processing takes place”. Thus, in writing assessment, test takers should be informed of the setting under which writing is to be undertaken, resembling the distinctive features defining real life writing tasks.