Misconstruing Instantial Probabilities As Structural Relations

Martin (1992: 372-3):

Like reference chains and conjunctive structures, lexical strings were treated as covariate structures; message parts are typically both depending and depended on.  However, it was pointed out that the lexicalised message parts are not phoric and that the retrospective dependency that characterised phoric reference items and conjunctions was not relevant.  Rather message parts enter into relationships of mutual expectancy in text, fulfilling predictions deriving from preceding text at the same time as they create expectancies about what will follow.  In this respect lexical relations are more field-like than particulate in nature, and it was suggested in 5.4.1.6 above that they might be better modelled as a kind of lattice rather than as a string…

Blogger Comments:

[1] As previously argued, covariate structures are not structures. They are related units (X related to Y) not structured units (X structured as A^B).

[2] In SFL theory, interdependency is a logical system.  In discourse semantics, a message part is an experiential unit.

[3] 'Lexicalised message parts' are the lexical realisations of discourse semantic message parts.  Here they are construed as if they are both at the same level of symbolic abstraction (stratum).

[4] In SFL theory, the different kinds of pointing, phora, are features of cohesive reference, a non-structural resource of the textual metafunction.  Conjunction is construed here as a structural resource of the logical metafunction.

[5] In SFL theory, this type of 'prediction' during logogenesis is modelled as instantiation probabilities of register.

[6] The argument here is:

• message parts are related by mutual expectancy,
• therefore the relations are more like fields than particles,
• therefore, the relations would be better modelled as a lattice than as a string.