Doran (2021):
Blogger Comments:
To be clear, Doran's claim that the "strongly bound types", multivariate vs univariate structure, "are often found wanting in more complex or less prototypical constructions" is a bare assertion, unsupported by evidence.
Although Doran provides no evidence that different types of structure are described differently by same authors in different situations, the most relevant example of this is Halliday himself. At the time of his now superseded theory, Scale & Category Grammar, Halliday (1965) classified a Head°Modifier structure as multivariate, whereas by the time of Systemic Functional Grammar, Halliday (1985) classified a Head°Modifier structure as univariate.
Similarly, Halliday (1965) — Scale & Category Grammar — defined a multivariate structure as comprising 'a specific set of variables each occurring once only', whereas Halliday (1985) — Systemic Functional Grammar — defined a multivariate structure as 'a constellation of elements each having a distinct function with respect to the whole', with no specification of the number of occurrences.
It will be seen, as this blog unfolds, that Doran adopts the superseded definition of multivariate structure, and makes it pivotal to his argument.