Thus, a number of theoretical propositions have been introduced in [Chapter 5] and these will be developed and augmented, to a certain extent, in the present one; to what extent is the textual analysis which is to follow a test of these propositions? I can address this question, initially, by referring again to the process by which the language of description was generated. Essentially, this involved, on the one hand, an engagement with literature relating to social theory and to textual analysis. This literature constitutes the theoretical antecedents which are represented by the work discussed in the [Chapter 5]. On the other hand, there was an engagement with an empirical text. The language of description emerged in a complex synthesis of deductive theoretical construction and inductive empirical reading. The theory and the description which it produces are therefore produced together, so that their separation is itself an analytic exercise. At this stage, it is possible, as I have suggested above, to forge a distinction between the propositions and the specific details of the language of description as, respectively, internal and external languages. The former would be more closely, but not exclusively, associated with the theoretical antecedents, the latter would be more closely associated with the empirical text. This would produce the [constructive description machine].

It will be apparent that ‘the empirical world’ is introduced at both the uppermost and lowermost points of the [constructive description machine] and that each level is articulated to subjacent and superjacent levels by downward deduction and upward induction. It will also be noticed that a distinction is being made between ‘theoretical antecedents’ and ‘empirical text’ each as ‘work’ or as ‘text’. This latter distinction corresponds to the distinction made by Barthes (1981a) whereby the text is to be understood as that which is produced by the reading. The work is taken up and produced as a text in its reading. To this extent, reading is always writing. The present work must establish its position within a field of academic production. This field is constituted by this positioning as more general than the work itself, so that my particular product is to be derived from that field. However, the field itself is also to be inaugurated by the work. In other words, I must act selectively upon a notional set of available discursive resources, just as the reader must impose selection principles of some kind in taking up a book. There is, therefore, an inductive relationship between my emerging text and the field which I am constituting and which is constituting my text. Thus, in producing my theoretical antecedents as text, I am also constituting an exteriority, a work of which my product is a reading.

The lowest articulation of the [constructive description machine] constitutes an equivalent dialectic. The language of description must be capable of describing more than a single text, so that it is at a higher level of generality. In the inauguration of my description of the empirical text I must impose more or less rudimentary selection principles of recognition of what is to count as an object and of realisation of how that object is to be realised as data. There is, in other words, a deductive relationship between the description of the empirical text-as-text and the empirical text-as-work. However, to constitute this relationship as purely deductive would be to construct the present work as pure theory which would dissolve into solipsistic musings if, in addition, the deductive link with theoretical antecedents as work were to be denied.

The specificity of the empirical text and, indeed, of the academic field, as work must be constituted by the approach being developed here. However, this is not to succumb to naive realism. To the extent that an epistemological positioning may be instructive, it is more appropriate to recontextualise the neo-Kantian philosophy of Jean Piaget.

(Dowling, 1998; pp. 125-7)