Martin (1992: 517):
One way to scale texts along this action/reflection dimension of mode is to take the activity sequences aspect of field as a base line and see to what extent texts are structured with respect to these sequences. Texts can then be divided into those organised primarily with respect to activity sequences (iconic texts) and those organised along different lines (non-iconic texts).
 There is a general blurring here of the distinction between different features of mode (context stratum) and the different registers (text types) that realise situation types. That is, the blurring is of two dimensions simultaneously: stratification (context is confused with language) and instantiation (general potential is confused with registerial variation). This, in fact, summarises the problems with Martin's misconstrual of context as register.
 As previously demonstrated, this notion of 'activity sequences' confuses ideational semantics, meanings of language, with field (the ideational context of language). This misunderstanding arose through confusing the meanings of a conversational text that recounted the procedure of 'parading a dog for judging' (language) with the cultural practice (context: field) of 'parading a dog for judging', in which language plays an ancillary rôle (context: mode).