Note that generalisation increases down strata (not up), Martin & Matthiessen 1991...The most powerful generalisation is, arguably, that there is a tendency towards generalisation as we move down either the stratal organisation or the rank scale. That is, to put it crudely, a given number of semantic systems is realised by a smaller number of corresponding lexicogrammatical ones; similarly, a given number of lexicogrammatical systems is realised by a smaller number of phonological ones. And the same holds for the rank scale; for example, a given number of clausal systems is realised by a smaller number of verbal group systems. The reason for this state of affairs is easy to see. A stratal descent or a rank descent always entails a generalisation across contexts and this generalisation is reflected in the relatively smaller number of realising systems.
Blogger Comments:
To be clear, here Rose unwittingly provides further evidence against his own previous claim (here) that
[Halliday's] grammatico-semantic model assumes a bijective or one-to-one relation between LG features and semantic features…
