This thesis investigates whether there are individual constancies in writing style, that is, whether there are any features in an individual's writing which remain the same over a large number of samples of their texts. This question is examined first by an analysis of the Cumulative Sum technique; a method which has recently been developed for purposes of authorship identification, and claims to be able to distinguish with 100% accuracy, whether a piece of text has been written by one, or more than one individual. This claim is explored thoroughly, as well as other related claims which concern the techniques' application to different genres. The analysis of this specific method, is followed by a systematic analysis of some surface variables in the text, those measured by a stylechecker. The relationships existing among these stylecheck variables are examined. In addition, the relationship between the stylecheck variables and personality is examined using Cattell's 16 P.F. Finally, the relationships between some more subjective or idiosyncratic features of the text are investigated for any individual differences or constancies. The results showed that there was no scientific basis to the claims behind the cusum technique. The fundamental assumption of the technique was proved to be wrong. No evidence for the unique separation of individuals using either the stylecheck variables or the subjective variables was found. However, some interesting correlations were detected between the stylecheck variables and the personality factors. These results implicate certain constancies within individuals but these constancies are not enough to distinguish individuals uniquely.
【 预 览 】
附件列表
Files
Size
Format
View
Individual Constancies in Written Language Expression