| Abstract | Models of the writing process are used to design software tools for writers who work with computers. This thesis is concerned with the construction of a model of fiction writing. The first stage in this construction is to review existing models of writing. Models of writing used in software design and writing research include behavioural, cognitive and linguistic varieties. The arguments of this thesis are, firstly, that current models do not provide an
adequate basis for designing software tools for fiction writers.
Secondly, research into writing is often based on questionable assumptions concerning language and linguistics, the interpretation of empirical research, and the development of cognitive models. It is argued that Saussure's linguistics provides an alternative basis for developing a model of fiction writing, and that Barthes' method of textual analysis provides insight into the
ways in which readers and writers create meanings. The result of reviewing current models of writing is a basic model of writing, consisting of a cycle of three activities - thinking, writing, and reading. The next stage is to develop this basic model into a model of fiction writing by using narratology, textual analysis, and
cognitive psychology to identify the kinds of thinking processes that create fictional texts. Remembering and imagining events and scenes are identified as basic processes in fiction writing; in cognitive terms, events are verbal representations, while scenes are visual representations. Syntax is identified as another distinct
object of thought, to which the processes of remembering and
imagining also apply. Genette's notion of focus in his analysis of text types is used to describe the role of characters in the writer's imagination: focusing the imagination is a process in which a writer imagines she is someone else, and it is shown how this process applies to events, scenes, and syntax. It is argued that a writer's story memory, influences his remembering and imagining;
Todorov's work on symbolism is used to argue that interpretation plays the role in fiction writing of binding together these two processes. The role of naming in reading and its relation to problem solving is compared with its role in writing, and names or signifiers are added to the objects of thought in fiction writing. It is argued that problem solving in fiction writing is sometimes concerned with creating problems or mysteries for the reader, and
it is shown how this process applies to events, scenes, signifiers and syntax. All these findings are presented in the form of a cognitive model of fiction writing. The question of testing is discussed, and the use of the model in designing software tools is illustrated by the description of a hypertextual aid for fiction writers. |
|---|