Thursday, 15 December 2016

David Rose Confusing Delicacy With Instantiation And Stratification

Thanks Gordon
It’s nice to see the alternative interpretations set out so clearly. Can I suggest that the source of their incommensurability is the (unresolved) tension between grammar and lexis that runs through all of linguistics? 
In this example, functional interpretations of a ‘ditransitive’ grammatical structure [nom gp + verb gp = nom gp + nom gp] vary with the lexical items instantiating each element
e.g. item instantiating verb gp - she made herself a cup of tea / I gave my love a cherry 
The interpretation of functional difference here assumes some criterial relation between grammar and lexis. So do terms such as ‘light verb’, since ‘verb’ is a grammatical unit and ‘light’ denotes some kind of lexical quality. Are such terms classifying grammar or lexis? 
When an item instantiating a nom gp normally instantiates a verb gp, the tension of grammar/lexis is exposed, and theories struggle to manage it. Halliday’s solution is to propose two strata — grammar/semantics — whose relation may be ‘congruent/incongruent’. (But note the dissonance here with his notion of ‘lexis as delicate grammar’.) Chomsky’s solution is to propose two structures related as 'deep/surface’. Cardiff's solution resembles Chomsky’s, as both require all semantic difference to be handled in the grammar. So a nom gp instantiated by ‘kiss’ is no longer a participant but a ‘main verb extension’. 
Whatever the proposed solution, the problem is the grammar/lexis relation, which is at the heart of linguistics but which no linguistic theory has resolved.

Blogger Comments:

[1] For the "source of their incommensurability", see the previous post.

[2] The relation between grammar and lexis is modelled differently by different linguistic theories. The "tension" that Rose imagines arises from his multiple misunderstandings of the architecture of SFL theory, as will be demonstrated below.

[3] In SFL theory, the relation between grammar and lexis is modelled as the dimension of delicacy.

[4] The four claims that lexis "instantiates" grammar confuse instantiation with delicacy.  It's been 17 years since Halliday & Matthiessen (1999: 15) clarified the distinction:
Note that it is important to keep delicacy and instantiation distinct. In early work on semantic networks, they were sometimes neutralised (cf. Woods', 1975, review). The difference is essentially that between being a type of x (delicacy) and being a token of x (instantiation). Both may be construed by intensive ascription: cf. Chapter 4, Section
To be clear, for the most part, lexical items are said to be specified by combinations of features of the most delicate grammatical systems, in the same way that phonemes are specified by combinations of features of articulatory systems; e.g. /b/ is specified by combinations of the features [voiced, stop, bilabial].  Instantiation, on the other hand, is the perspectival cline between language as potential (system) and language as actual instance (text).

[5] This confuses stratification with delicacy.  Halliday did not stratify the content plane into grammar and semantics in order to solve the "tension" between grammar and lexis.  He modelled grammar and lexis as two ends of the same continuum, lexicogrammar, and stratified the content plane into grammar and semantics in order to account for grammatical metaphor.

[6] As these clarifications demonstrate, the only dissonance here is between SFL theory and Rose's understanding of it.

[7] The closest analogue in SFL theory to Chomsky's deep and surface structure is the cline of instantiation.  That is, Rose claims that Chomsky dealt with the lexis–grammar relation (delicacy) through instantiation (and that Halliday dealt with it through stratification).

To be clear, Rose has confused the systemic scale of delicacy with both the hierarchy of stratification and the cline of instantiation.

See Rose's previous confusion of delicacy with stratification here.