R. Morales, H. Pain, S. Bull & J. Kay (eds), Proceedings of Workshop on Open, Interactive and Other Overt Approaches to Learner Modelling, International Conference on Artificial Intelligence in Education 1999, 1-8. Cooperative Writer Modelling: Facilitating Reader-Based Writing with Scrawl Susan Bull & Simon Shurville School of Languages, University of Brighton, Falmer, East Sussex, BN1 9PH, UK. email: s.bull@brighton.ac.uk, sjs16@itri.bton.ac.uk Abstract: This paper introduces SCRAWL, a system which uses a cooperatively constructed Writer Model to represent a student's writing strategies. SCRAWL encourages audience awareness in writers, and offers advice on creating a reader-based text, to fit with a student's own writing strategies. An initial evaluation is presented to demonstrate the feasibility of this approach. 1 Introduction One of the many things an author must consider when writing is the requirements of the intended readers. Despite its importance, this is often overlooked by student writers (Kroll, 1993). Flower (1993) urges learners that: The first step in designing your writing to be read is to understand the needs, attitudes, and knowledge of your particular reader, and to help that reader turn your written message into the meaning you intended. Flower (1993:193) However, advice on creating a reader-based text must be given at the right moment, and this will differ for different authors, since writers approach writing in a variety of ways. For example, some plan extensively initially (either mentally or externally), while others start writing directly. The latter writer may be increasing their understanding through writing, while initial planners may be high self-monitors whose ideas do not get further developed during the writing phase (Galbraith, 1996). Torrance et al (1994) state that 'planning is neither a necessary nor a sufficient condition for writing success', and Piolat (1999) suggests that sentence generation coupled with revising, if well-balanced, may be as effective as planning. It is possible that some previous findings suggesting the importance of planning for effective writing may be related more to the effect of increased time on task, brought about by planning (Sharples, 1999). Many writers do not find it easy to alter the strategies they are accustomed to using (Wyllie, 1993), often preferring one set of approaches over others (Snyder, 1993). Indeed, even if strategy change were achieved, this does not imply that new strategies adopted by an individual will suit them. If instructed to follow particular methods, students could be forced into using less productive strategies (Cumming, 1995). In addition, authors may be writing in an already familiar genre, using well-developed strategies which work effectively for them (Torrance, 1996). Since experienced writers also use a variety of approaches (Chandler, 1993; Snyder, 1993), there appears little justification in expecting learners to adapt the way