inter-stratal tension

So paraphrasing field as social activity does not recognise the verbal action aspect of social activity. Verbal actions such as those of explaining, defining, narrating, reporting, chronicling, lecturing and a myriad of others that I would describe as verbal actions are treated as a matter of mode in the current SFL models of context. From this perspective, the distinction between field and mode is suspiciously reminiscent of the distinction between the what and the how, the content and the style, which has always been popular in literary criticism.(270)

…the notion of social activity must be reconceptualised to cover both action and locution, both material and verbal action, and my recent exploration of the relations of context and text appears to support this position. (271)

Table 10 suggests that linguistic models of context which treat material action as a matter of field and verbal action as a matter of [ancillary] v. [constitutive] mode might have greater difficulty in identifying those environments where textual integration or co-location might be at risk (280-281)

3.4 Ancillary and constitutive verbal action: field or mode ?
… In agreement with other systemicists’ views, I too have typically treated the distinction between constitutive and ancillary as a matter of mode, even despite occasional misgivings (see Section 2.3 and footnote 58): I suggested, in fact, that the two terms refer to the two endpoints of a cline, viz., the role of language (see Hasan 1980, 1985b etc; for further discussion and development, see Cloran 1994). In SFL, persuasion, explanation, definition, etc, are described as categories of rhetorical mode.
…the role language is playing or what it is doing in the social situation (cf Halliday’s remarks in Ch1) are not aspects of mode, nor is rhetorical mode really a phenomenon that belongs in mode: rather, these various cases of speaking, viz., persuading, explaining, joking, narrating are cases of verbal action. There seems no reason for suggesting that instead of verbal action, they are just a modality or mode for bringing that action about, especially since the actions in question are un-doable any other way except verbally…

as the description of the field of discourse progresses in delicacy, it would become possible to identify the specific lexicogrammatical and semantic domains at risk in the realisation of specific choices from the systems of field: such work is in fact well under weigh (see Halliday & Martin 1993; also Matthiessen, in press). This claim takes us to a higher level concept, to what I have called genre specific semantic potential elsewhere (Hasan 1985b:98ff.) which consists of the meanings and wordings that are crucial to the identity of a register, which is naturally related to the notion of domain of signification. (287)

From the point of view of instantiation, situation is to culture as text is to language; the first term of each proportion instantiates the second:

situation: culture :: text: language

From the point of view of realisation, language is to situation; the first term of each proportion realises the second:

language: culture :: text: situation

Current SFL models possess no satisfactory means of contextual and/ or registral changes which do not disturb the unity of the text, nor can they specify where i.e., in what kind(s) of social situation, such changes are most at risk.