Martin (1992: 102-3):
The realisation of the [presenting/presuming] system and the probability of phoric options being chosen is very much affected by a third system [generic/specific], which cross-classifies [presenting/presuming] and [comparison/–]. Generic reference is selected when the whole of some experiential class of participants is at stake rather than a specific manifestation of that class.
 To be clear, Martin provides no realisation statements in any of networks that feature his [presenting/presuming] system, namely Fig. 3.2 (p102), Fig. 3.3 (p105), Fig. 3.4 (p106), Fig. 3.5 (p110), Fig. 3.6 (p112) or Fig. 3.9 (p120).
 To be clear, the probability of features being selected (instantiated) varies according to register.
 There are two confusions here. On the one hand, the [generic/specific] distinction is a distinction in the referents, not in the means of referring (the system of reference). On the other hand, a cohesive relation between a class and a member of a class is an instance of lexical cohesion, subtype: hyponymy (elaboration + attribution).
 To be clear, 'the whole of some experiential class of participants' confuses the class with the members of the class ('whole', 'participants'), thereby undermining Martin's distinction between generic and specific.