Martin (1992: 562-3):
Martin 1985/1989, working along lines similar to Longacre's, developed a preliminary classification of "factual" genres drawing on field and mode. The basic field opposition was between texts which were focussing on activity sequences (e.g. narratives, recipes, manuals) and texts which were not (e.g. descriptions, expositions); the basic mode opposition was between texts which generalised across experience and those which referred to a specific manifestation of a culture. Generalised texts were further divided into those which function to document information and those which explain. … In later work Martin and Rothery divided explaining texts into those which considered more than one point of view, Discussions and Explorations, and those which presented only one position, Exposition and Explanation. This genre matrix is outlined in Table 7.19.
Table 7.19 Cross-classification of factual genres
– generalised generalised:document explain:resolve debate – activity structured description report exposition discussion + activity structured recount procedure explanation exploration
 This is inconsistent with Martin's (p564) opposition of factual and narrative genres.
 This confuses ideational semantics (activity sequences) with ideational context (field). The confusion is thus along the dimension of stratification.
 This confuses ideational semantics (construing experience) with textual context (mode). The confusion is thus simultaneously metafunctional and stratificational.
 This confuses interpersonal semantics (heteroglossia vs monoglossia) with textual context (mode). The confusion is thus simultaneously metafunctional and stratificational.