Martin (1992: 33):
This is an important step as far as interpreting dialogue is concerned since there is no simple relation between general grammatical classes and the rôle they play in structuring a conversation. The most commonly discussed example of this is the realisation of demands for goods and services. Each class considered so far, for example imperative, declarative and interrogative MOOD, can be used to code a Command.
Blogger Comments:
[1] To be clear, this important step is to stratify an already stratified content plane by relabelling Halliday's work as Martin's discourse semantics; see the previous post.
[2] On the one hand, this misrepresents grammatical functions — imperative, declarative and interrogative MOOD — as grammatical classes (of form). On the other hand, and more importantly, it is misleading, because it is not true, since, in the absence of grammatical metaphor, the "simple relation" is:
- the MOOD feature 'imperative' realises the SPEECH FUNCTION feature 'command',
- the MOOD feature 'declarative' realises the SPEECH FUNCTION feature 'statement', and
- the MOOD system feature 'interrogative' realises the SPEECH FUNCTION feature 'question'.
[3] To be clear, the semantic feature 'command' is realised congruently by the grammatical feature 'imperative', but incongruently by the features 'declarative' of 'interrogative', the latter being instances of interpersonal metaphor.
