Martin (1992: 116):
As with superset reference, the nature of the experiential relevance of the presumed participant may or may not be specified. It is not specified when other, same, different, identical and similar function as Post-Deictic or when the same items or such realise the Epithet; it is specified when comparative quantifiers and adjectives are deployed. This opposition is illustrated below.
GENERAL - the other four frogs Deictic Post-Deictic Numerative Thing the four other frogs Deictic Numerative Epithet Thing EXPERIENTIALISED - QUALITY: the four bigger frogs Deictic Numerative Epithet Thing the four equally big frogs Deictic Numerative Epithet (β α) Thing QUANTITY: fewer frogs Numerative Thing equally many frogs Numerative (β α) Thing
 This continues the mistaking of experiential construal for textual reference, and nominal groups for reference items; see previous posts.
 To be clear, Martin's opposition of general vs experientialised is a misunderstanding and rebranding of Halliday's comparative reference opposition of general vs particular on the grammatical stratum as a discourse semantic opposition.