CASE STUDY Nielsen's account and review of the steps of Bhatia's (1993) analytic method:
Step 1 "Placing the given genre-text in a situational context" ...[includes] drawing on internal clues in the text, previous experience with similar texts and with the professional community, and encyclopaedic knowledge.
Step 2 "Surveying existing literature" is...important for those "who do not belong to the relevant speech community".... The literature that should be surveyed comprises among other things linguistic analyses of the genre in question, tools/methods/theories of genre analysis, practitioner advice, guide books, manuals etc. relevant to the speech community, and literature on the social structure, interactions, history, beliefs, goals etc. of the professional community.
Step 3 "Refining the situational/contextual analysis" requires [inter alia] identification of the network of surrounding texts and linguistic traditions that form the background to this particular genre-text and the extra-textual reality which the text is trying to represent, change or use. Thus, defining, describing and confirming or disproving the first intuitive findings from step 1 is the aim of step 3. ...I would nevertheless like a more strict definition of and distinction between discourse, register, genre, sub-genre and variety....
In step 4 "Selecting corpus" one should define the genre in question well enough to distinguish it from closely related genres. Here Bhatia "admits" that a definition of genre cannot necessarily always exclusively be based on the communicative purpose alone (Ibid.), but also can be based on the situational context(s), some distinctive textual characteristics or some combination of these.... Furthermore, the criteria for defining the genre should be clearly stated, and criteria must be decided on for selection of examples: a long single typical text, a few randomly chosen or a large statistical sample. It is gratifying to see that, contrary to an unfortunately still common belief and widespread misunderstanding, corpus' may consist of one, a few or a large number of text. ...One must only bear in mind that the size of one's sample has to correspond to the method one is applying: one text for detailed analysis, a few texts for exploratory investigation, and a large statistical sample for an investigation of a few features.
Step 5 "Studying the institutional context" recommends that the analyst studies the institutional context, including the system, in which the genre is used and the linguisitic, social, cultural, academic, professional rules and conventions. This step becomes important especially if the data are collected from a particular organization, which often has its own organizational constraints for genre construction.
Step 6 "Levels of linguistic analysis" is divided into three sub-levels, the explanatory force of which increase with each sub-level. Level 1 "Analysis of lexico-grammatical features" is basically a statistical analysis of frequency of syntactic properties that provides empirical evidence to confirm or disprove intuitive statements that one makes. However, it only describes how genres are written, it does not explain why they are written the way they are. Level 2 "Analysis of text-patterning or textualization" deals with the tactical aspect of genre by focusing on how certain linguistic features are used in different genres in order to achieve different things. Level 3 "Structural interpretation of the text-genre" deals with the cognitive aspect and aims at the identification of the genre-text's cognitive move-structure (...which is the linear and hierarchical structure of a genre-text)...: each move serves a typical communicative intention which is always subservient to the overall communicative purpose of the genre.
Step 7 "Specialist information in genre analysis" suggests that the analyst consults a specialist informant, typically a practising member of the discourse community, who confirms, rejects, validates or corrects the analyst's findings and provides him [sic] with supplementary information. Here it might be added that a specialist informant is even more likely to put constraints on a genre (individual constraints) than an organization that submits a genre to organizational constraints. It would therefore possibly be advisable to consult more than one specialist informant.