5.4 Approaches to document analysis (internal analysis) The preceding criteria (Section 5.3) constitute the cataloguing and credentialising of the document (the external analysis). The following sections explore how a document's content is analysed (the internal analysis).
The various procedures of document analysis involve analysing and interpreting data generated from the examination of documents. What is sought from a document depends on the epistemological stance of the researcher.
A positivistic approach to document analysis is looking for supposed factual evidence or confirmation to support or reject a hypothesis. Does an annotated copy of a ministerial document show that the minister knew about the likely negative outcomes of a policy but chose to disregard them? Does the photograph reveal that the army commander was present at the site of a massacre? Is the detail in a painting evidence that the artwork is a forgery?
The main positivistic approaches include confirmation of fact, content analysis, qualitative document analysis and historism.
A phenomenological approach is concerned with interpreting the meaning of the document: both its surface meaning and any underlying meaning or connotation that the document reveals. This is about going beyond the truth of the document's factual assertions and drawing inferences from the document. Methods of doing this include interpretative analysis, some forms of historical research/ historiography (historicism, historicalism), hermeneutics, narrative analysis and aesthetics.
A critical social research approach locates the document historically, socially and politically and explores how the content reveals the nature of society by digging beneath the surface appearance of the document. Critical social research attempts to understand how the document is related to and informs social processes.
Scott (1990, p. 28), for example, asserted that the ultimate purpose of examining documents is to arrive at an understanding of the meaning and significance of what the document contains. He maintained that the literal meaning of a document gives only its face value meaning. This does not reveal the real significance. This has to be reconstructed through 'an interpretative understanding', which involves relating the literal meaning to the contexts in which the documents were produced in order to assess the meaning of the text as a whole.
For example, information on the growth of both gross domestic product and per capita income is given in factual quantitative terms. On its own, this type of information may lead people to expect amore equitable distribution of income and a higher standard of living. But when this information is read together with information on sectoral and structural changes in the economy, ownership and control of the means of production, employment patterns and income distribution, the picture is more complex. This is because statistics only give face value meaning. Statistics are only 'raw materials', which must be interrogated and their real meaning reconstructed. One can only make sense of this welter of apparently disparate bits and pieces of information, by situating it within a theoretical context. It is the theory that re-orders the data, and inferences come as a matter of interpretation of the raw material informed by theory.
For example, in my research on labour relations in Botswana (Mogalakwe 1994), I realised that although both wages policy and various pieces of labour legislation appeared innocuous, a rigorous reading of these, that is, informed by critical theoretical approach, revealed how the wording of these documents, that is their language, was a form of discourse that subtly ordered people's perceptions of the social structure and could be used to construct specific forms of social relationships and maintain the status quo. Language, whether written or spoken, subtly orders our perceptions of situations, and thus also constructs and creates social interaction. Social texts do not merely reflect or mirror objects, events and categories existing in the social world, but also actively construct a version of those objects, events, and categories (Potter and Wetherell 1987).