A General Model That Applies To All Systemic Functional Grammars

Fawcett (2010: 41):

At this point I must make it clear that Halliday sometimes writes in a way that implies a substantial change to the model represented in Figure 4, and the effects of this will be explored in Sections 4.6 and 4.7 of Chapter 4. I shall nonetheless argue there that Figure 4 does indeed represent a general model that applies to all systemic functional grammars.

Blogger Comments:

This continues the discussion of Figure 4:

[1] This is misleading.  The model represented in Figure 4 is entirely inconsistent with Halliday's model, as demonstrated in previous posts.

[2] This cannot be true, since the model represented in Figure 4 is internally inconsistent across both dimensions, stratification and instantiation, as demonstrated in previous posts, such as On 'The Main Components Of A Systemic Functional Grammar'.

The Prototypical Instance At The Level Of Form

Fawcett (2010: 41):

The prototypical instance at the level of form is a 'sentence' — and a sentence frequently consists of a single clause, e.g., I've been discussing that new student with Peter. Since the grammar is part of a fuller model for the generation of texts, we may also refer to the output as a text-sentence, and this has the value of reminding us that sentences do not occur singly, as formal linguists sometimes appear to assume, but within longer texts in which they themselves function as elements. (Note, however, that we can also treat a group of words such as that new student as an instance, exactly as is done in the little grammar in Appendix A.)

Blogger Comments:

Here Fawcett confuses the theoretical categories of 'instance' and 'unit'.  In SFL theory, any unit can be construed in terms of both potential and instance, and in terms of both system (paradigmatic axis) and structure (syntagmatic axis).

The Generation Of Instances Of Form

Fawcett (2010: 41):

In a fuller grammar the first unit to be generated would be a clause, and then one or more realisation rules would specify re-entry to the system network of meaning potential (as shown by the loop-back arrow on the left side of Figure 4), in order to generate one or more nominal groups (or even an embedded clause) to fill the relevant elements of the clause (as described in Fawcett, Tucker & Lin 1993). The 'tree structures' in the bottom right box in Figure 4 are labelled sufficiently richly to express the various functions that each element serves, and they are, it will be clear, the instances at the level of form.

Blogger Comments:

This continues the discussion of Figure 4:

[1] By definition, a realisation rule specifies a realisation (a lower level of symbolic abstraction), and so, not a re-entry to a system network at a higher level of abstraction (meaning) than form.

[2] The claim that 'tree structures are instances at the level of form' is merely a bare assertion, since no supporting argument is provided.  As Figure 4 illustrates, Fawcett incoherently regards tree structures as instances of realisation rules.  As the term 'realisation rule' makes plain, the relation between the rule and what it specifies is realisation, not instantiation.

The Rôle Of Fawcett's Realisation Rules

Fawcett (2010: 41):

In brief, we can say that the role of the realisation rules is to convert the selection expression of semantic features that is generated on a traversal of the network into a layer of the tree diagram representation of the sentence that is being built up. This concept is illustrated in Figure 2 of Appendix A, which has the potential to generate just eighteen different nominal groups. 

Blogger Comments:

[1] This blurs the theoretical distinction between the process of instantiation (the selection of features and the activation of realisation statements) during logogenesis ('the sentence that is being built up') with the realisation relation between the paradigmatic axis ('features') and the syntagmatic axis ('tree diagram').

[2] The illustration of the concept in Figure 2 of Appendix A is as follows:



It can be seen that this largely presents realisation rules as acting on individual features, rather than selection expressions, contrary to Fawcett's claims above, and that individual semantic features include the meanings of individual lexical items.