Misconstruing 'Multivariate' & Metafunctional Inconsistencies

Martin (1992: 156):

NEGOTIATION was approached from both a multivariate and covariate perspective in Chapter 2.  The notion of constituency between exchange and move was developed to account synoptically for sequences of up to five moves.  These moves are mutually predicting as set out in their constituency structures.  In [3:100] the initiation expects a response and the response presumes its initiation:

[3:100]

K2 Who won?
K1 — Ben did.

Blogger Comment:

[1] This confuses multivariate structures (of a rank) with constituency (rank hierarchy).  The constituency relation between exchange and move is the relation between levels in a rankscale.  A multivariate structure, on the other hand, is 'a configuration of elements each having a distinct function with respect to the whole' (Halliday & Matthiessen 2004: 331).

Further, as Halliday & Matthiessen (2004: 384) point out:

… the relationships among the elements in a multivariate structure can be characterised as segmental from an experiential point of view but as prosodic from an interpersonal point of view and as cumulative from a textual one.

That is, on the SFL model, the relation between elements in the multivariate structure realising interpersonal systems, like NEGOTIATION, are prosodic.  Since this is a proposition Martin elsewhere (1992: 549) supports, Martin is here inconsistent with his own model.

[2] To be clear, the notion of 'covariate structure' derives from Lemke (1985).  In Lemke (1988: 159) he acknowledged that 'covariate' is not a type of structure:

My own 'covariate structure' (Lemke 1985), which includes Halliday's univariate type, is for the case of homogeneous relations of co-classed units, and should perhaps be called a 'structuring principle' rather than a kind of structure.

[3] To be clear, these are merely statements about two alternative features in a closed system.