David Rose Confusing Grammar With Lexis And Context With Register

Shooshi Dreyfus wrote to sys-func and sysfling on 10 February 2016 at 7:48:

Yeah the ones with grammatical metaphor like this are proving to be the most tricky!

and David Rose replied on 10 February 2016 at 11:58:

One reason they [clauses with grammatical metaphor] are tricky is that what’s being classified are lexical items. We are still unable to say what a lexical item is. We call these items ‘verbs’, but a verb is a structural unit in grammar. A glance at a dictionary tells us that most lexical items can be realised as various word classes. Grammatical metaphor increases this potential. 

Relations between lexical items and grammatical units like ‘verb’ are really probabilistic, depending on register variables. So interpreting the ‘meaning’ or ‘sense’ of items really has to start with register. Hence the dictionary definitions in Eugeniu Costetchi’s word lists, or David Banks’ ‘conceptual’ analyses. This fact is problematic for models that would like to subsume lexical items within grammatical categories like process types, as though they were register neutral. … 

PS By register variables I mean field/tenor/mode

Blogger Comments:

[1] One thing that makes transitivity analysis, in general, difficult is not taking a trinocular perspective — the perspective on which SFG was theorised.  One thing that makes transitivity analysis difficult, in the specific case of grammatical metaphor, is not unpacking the metaphor, at least enough to see the junction of meanings being construed by the metaphor.  See, for example, Metaphor As Junctional Construct and Grammatical Metaphor: A Junction Of Perspectives.

[2] Despite being unable to say what a lexical item is, Rose is able to say that lists of verbs that can serve as specific process types are classifications of lexical items.

[3] As Halliday & Matthiessen (1999: 568) point out:

The folk notion of the “word” is really a conflation of two different abstractions, one lexical [lexical item] and one grammatical [word rank].

When we are talking about words that can serve as a process type, we are talking about the grammatical abstraction, not the lexical.  Rose follows Martin (1992) in confusing these two abstractions, as demonstrated in the critiques Misconstruing The Difference Between Lexical Item And Grammatical Word and Confusing Lexis With Grammar. 

[4] A verb is not 'a structural unit in the grammar'.  A verb is a classification of form at the rank of word, just as a verbal group is a classification of form at the rank of group/phrase.

[5] The relations between verbs and lexical items are mapped out by the architecture of the theory. For example, verbs are related to processes by a chain of realisation along the rank scale, and processes are related to lexical items in terms of delicacy.  Halliday & Matthiessen (1999: 198-9):

… we can differentiate both processes and participants into finer and finer subcategories, until we reach a degree of differentiation that is associated with the choice of words (lexical items). Note that it is not (usually) the lexical items themselves that figure as terms of the systems in the network. Rather, the systems are systems of features, and the lexical items come in as the synthetic realisation of particular feature combinations.

[6] Here Rose follows Martin (1992) in confusing context with register; see some of the critiques here. In SFL, context refers to the culture as a semiotic system whose expression plane is language.  Context is more abstract than language.  Language realises context.  Register, on the other hand, is a functional variety of language.  It is not more abstract than language, it is language.  Whereas context is a level in the stratification hierarchy, register is a point (of variation) on the cline of instantiation between language as potential (system) and language as instance (text).

[7] The so-called 'conceptual' analysis is the view of transitivity 'from above' — that is, viewing the wording in terms of the meaning it realises.

[8] The model that 'subsumes lexical items within grammatical categories like process types' is called Systemic Functional Linguistic Theory.  This subsumption is the systemic dimension termed 'delicacy'; see [5] above.  Registerial variation according to context is modelled in SFL in terms instantiation and stratification: in terms of different situation types being realised by different registers.

[9] The notion of 'register neutral' comes from Martin (1992: 289) and is based on misunderstandings of Systemic Functional Linguistic Theory, as demonstrated, for example, in the critique Relocating Lexis Outside Language.