Text Structure as a Window on the Cognition of Writing
How Text Analysis Provides Insights in Writing Products and Writing Processes
Ted J. M. Sanders and Joost Schilperoord
Many teachers believe that the best and the worst essays written in class differ in organization. The best one is clearly structured, whereas the worst one is hard to follow. What can writing research tell us about this observation? Not much, until recently, because text structure was hardly studied seriously in the context of writing research. Traditionally, writing researchers have paid relatively little attention to the role of text analysis. In Flower and Hayes's (1981) cognitive approach, for instance, the characteristics of the text (e.g., style or structure) are almost completely neglected. However, Bereiter and Scardamalia (1983, 1987) have shown the way to an interesting interaction between psychological models and text linguistic research, which, in our view, is of crucial importance to writing research. They argued that "research is needed to discover what rules less skilled writers actually use and how these rules differ from those of experts." They even pointed to the remedy for this deficiency in studies of writing: text analysis (Bereiter & Scardamalia, 1983, pp. 11, 23). By analyzing the texts writers have produced, the rules they use in composing can be discovered.
In this chapter we show how text analysis can be used to identify such composition rules. We argue that an adequate analysis of the product of writing—the text—reveals significant traces of the writer's cognitive representation, even of online text planning. In order to gain insight into the writer's representations, we analyze the structure of texts and try to give a cognitive interpretation of this analysis. Furthermore, we show how data of online writing processes corroborate this cognitive interpretation of text structure.
In his overview of the psycholinguistic processes involved in language production, Levelt (1989) distinguishes between conceptual processes (e.g., macroplanning) and linguistic processes (e.g., syntactic formulation). The few linguistically oriented writing research studies that were carried out, often focused on "lower" levels of text, such as syntax (Hunt, 1970) or the presence of connectors (Lintermann-Rygh, 1985; see Van Wijk, 1992, for further discussion). However, if we want to gain insights into the cognition of writing, that is, both in the processes involved and in the writer's representation, we need to focus on conceptual processes. More specifically, if we use text analysis as a research method, it is imperative that it deal with the essential characteristics of a text and its un-