The overall aim of this research study was to establish an objective clustering of Thomas Hardy’s prose fiction texts as a basis for better understanding the associations between the texts, and the development of an objective thematic analysis of Hardy’s corpus that can address the problems of replicability and objectivity in non-computational thematic classification of literary studies. To achieve this, this thesis used vector space clustering based on hierarchical cluster analysis methods. The novels and short stories of Thomas Hardy formed the context of the study. Implications for theory and practice were examined. At this final stage, this chapter is an overall summary of the findings of the thesis. The chapter ends, however, with a note of suggestions and recommendations for improvement and speculating on future directions.
Returning to the question posed at the beginning of this study, it is now possible to state that the novels and short stories can be thematically clustered using in objective and replicable methods. In spite of its limitations (discussed in 9-3 below), this thesis suggests the following significant findings.
• All of Hardy’s prose fiction works exhibit rich thematic concepts. It can be claimed that many thematic discussions of Hardy were unjust in limiting their discussions to the series of novels and short stories he wrote between 1871 and 1895.
• One of the most significant findings to emerge from this study is that cluster analysis methods can be used to empirically derive taxonomies of thematic concepts of the novels and short stories of Thomas Hardy. Equally important, nonetheless, computers and machines cannot be replacements for humans in reading and interpreting literature.
• An implication of this thesis is that computational methods would usefully supplement and extend the qualitative analysis. The computational element in literary criticism provides what Susan Hockey (2000) terms “concrete evidence to support or refute hypotheses or interpretations which have in the past been based on human reading and the somewhat serendipitous noting of interesting features” (2000: 66).
• Exploratory multivariate analysis is proved effective in document classification of literary works.
• The clustering results of this study can serve as a base for future studies and criticisms of the novels and short stories of Thomas Hardy. Besides, it can usefully supplement information retrieval applications. Based on Rijsbergen’s conception “associated documents tend to be relevant to the same request” (Rijsbergen, 1979: 30), our clustering will make the retrieval more effective, because the class once found will contain only the relevant documents.
• The study has gone some way towards issues of authorship attribution. Exploratory multivariate analysis methods proved successful too in identifying authors. In this they are recommended as effective methods for authorship problems.
• The current...