[ Next Chapter ] [ Table of Contents ] [ E-mail Author ]
1.4 The Computer/Critical ProcessNone of the foregoing is meant to suggest that the computer should do the actual criticism, only that much of what the computer can do in processing a text is directly compatible with what many formalist critics are already doing. Computer-assisted criticism can best provide a collection of techniques that can profitably be used in conjunction with conventional critical perspectives.
There seem to be at least three distinct modes in which the computer can operate to help the critic.
The first of these is revelatory, and involves bringing to light data that might otherwise have been overlooked. This would primarily occur through the gathering of information that could only be obtained through long and tedious effort if computers were not used. We have already seen that for many of the more complicated formalist analyses computational assistance provides the only real hope for their general application.
The second mode is descriptive, and operates by applying analytical techniques to the text to derive an objective representation of certain features. Some of these techniques have been outlined in the previous section. This mode enables comparison across texts, with the possible benefit of the discovery of useful generalizations about groups of texts or genres.
The third mode is demonstrative. It seeks to statistically prove or disprove hypotheses about a text, or about data derived from the text. This mode is most susceptible to abuse, and therefore must be used wisely and cautiously. It does, however, imply an altered view of what constitutes proof in literary criticism, and how one might demonstrate a hypothesis. Thus "pervasiveness and adequacy" 24 could join authority and citation as modes of demonstration.
In all these modes, the computer only serves to gather information, not to use it. The assimilation and interpretation of that information can only take place in the mind of the critic. It is also the critic who must provide the intellectual context for the study, decide on the techniques used, and formulate the final critical insight.
Bearing this in mind, how might one go about applying computational assistance to the analysis of a given text? There seem to be at least five distinct and necessary phases:
i) One must begin with a question which is worthy of enquiry in its own right and justifiable within conventional critical values. It must not be chosen simply because it is something that the computer can do. The question should be placed in a familiar context, and phrased using conventional terminology.
ii) One must translate the question from substantive terms into operational or functional terms. Great care must be taken to specify any assumptions made during this process, to avoid or minimize the distortion inherent in any translation process, and to make sure that the operative definition is a close fit to the substantive definition, allowing for necessary changes in terminology.
iii) The computer is then used to gather information in accordance with the functional terms, and to explore the implications of the results under the guidance of the critic.
iv) One then assimilates the information gained, places it firmly in context, and synthesizes an accurate and justifiable interpretation with its aid.
v) This interpretation must then be back-translated to express the results in substantive, conventional terms, so that they can be readily understood within the context of the original question.
The end product should be based on solid, demonstrable and repeatable results (one of the most important requirements for a 'science'). The computation should recede into the background, not dominating discussion, but rather acting as a firm and verifiable base for it.
This thesis describes this process as applied to one aspect (characterization) of Virginia Woolf's novel, The Waves.
End of Chapter
Last modified: Monday, 11-Dec-2017 14:42:07 AEDT