Misunderstanding Halliday On Linguistics As Ideologically Committed Social Action [New]

Martin (1992: 2):

There are any number of uses to which the analyses presented here can be put. In Australia they have evolved in two main contexts: (i) as a means of exploring the relation between text and context — between text and register, genre and ideology (see Chapter 7 below); and (ii) as one foundation for the development of an educational linguistics, which has been used in particular to focus on literacy development (e.g. Painter and Martin 1986, Hasan and Martin 1989). From the start, this work has been pursued within the framework of critical linguistics (inspired by Fowler et al. 1979, Kress and Hodge 1979, Chilton 1985, Kress 1985/1989, Fowler 1987, Fairclough 1989) — a linguistics which "deconstructs" texts in such a way as to draw attention to the semiotic systems they instantiate, with a view to critically evaluating the ideologies they construe. Observational, descriptive and explanatory adequacy aside, this means that English Text has been written as a contribution to the linguistics envisioned by Halliday (1985e:5) as "an ideologically committed form of social action." For this reason it needs in part to be read in the context of projects oriented to de-naturalising hegemonic discourses and, concomitantly, facilitating intervention in political processes (e.g. Poynton 1985/1989, Martin 1985b/1989, Threadgold et al. 1986, Christie 1991, Giblett and O'Carrol 1990, and the new journal Social Semiotics).

Blogger Comments:

[1] To be clear, in SFL Theory, the relation between text and context is the realisation relation between an instance of language (text) and an instance of context (situation).  Martin, however, misunderstands varieties of language, register and genre, as not language, but the context of language. This is analogous to claiming that varieties of dog, such as Rottweiler and Retriever, are not dogs, but the context of dogs.

In SFL Theory, register is the midway point on the cline of instantiation — for the content plane of language — viewed from the system pole, and genre (in sense of text type) is the same point viewed from the instance pole. That is, register and text type are complementary perspectives on the same phenomenon: functional varieties of language.

Ideology, on the other hand, is a different matter. Ideology is realised in language, and as such, is located at the level of context, where 'context' means the culture as semiotic system. Ideationally, then, ideology is modelled in terms of field, with different ideologies characterised by different combinations of FIELD system features. And interpersonally, the fact that ideologies tend to be realised more by proposals than propositions suggests that modelling them also involves TENOR system features.

Martin, however, here models ideology as a stratum above genre and register, though he has since withdrawn ideology from this hierarchy.

[2] From the perspective of SFL Theory, this confuses both levels of symbolic abstraction and poles of the cline of instantiation. To be clear, the semiotic systems that texts instantiate are those of language, whereas the ideologies that texts construe, are characterised by the selection of contextual features that specify the instance of context: the situation.

[3] This is a very widespread misunderstanding of Halliday (1985e: 5). What Halliday actually wrote was:

… linguistics cannot be other than an ideologically committed form of social action.

That is to say, any linguistics is 'an ideologically committed form of social action', whether linguists realise it or not. The question then is 'Which ideology?'. For example, SFL Theory foregrounds the notion of choice, whereas Formal linguistic theories foreground rules, government and binding.

[4] To be clear, it will be seen that Martin (1992) provides nothing in this regard that SFL Theory, properly understood, does not already provide.

Misunderstanding The Natural Relation Of Grammar And Semantics [New]

Martin (1992: 1-2):

Halliday's Introduction to Functional Grammar (1985a) outlines the grammar of English which realises the discourse semantics developed in English Text (for a very useful elaboration of this grammar, see Matthiessen 1992). As Matthiessen and Halliday (in press) point out, in functional linguistics semantics is naturally (not arbitrarily) related to grammar. It follows from this point that functional approaches to discourse systems and structures will be enriched to the extent that they are able to draw on comprehensive, semantically oriented grammars which interface in a responsible way with textual considerations. English Text evolved in tandem with the richest extant English grammar of this kind, as developed by Halliday, and later Matthiessen, in the Department of Linguistics at the University of Sydney. 

For many linguists, the richness of Halliday and Matthiessen's grammar has appeared excessive (see for example Huddleston's (1988) dismissal of Halliday (1985a)). It is important to stress here however that what might appear extravagant from the point of view of "syntax" is more often than not fundamental to the point of parsimony from the perspective of discourse analysis (see Matthiessen (1989) for an alternative perspective on Halliday (1985a) and Matthiessen and Martin (1991) for a reply to Huddleston (1988)). English Text has been designed to complement in a solidary way a non-parsimonious grammar of this kind. It has been designed in other words to be grammatically responsible, interfacing with a grammar that is equally responsible to textual considerations.

Blogger Comments:

[1] This is misleading. Martin's discourse semantics is largely a rebranding of textual lexicogrammar (Halliday & Hasan's systems of COHESION), together with a rebranding of interpersonal semantics (Halliday's system of SPEECH FUNCTION). Moreover, even ignoring this, Martin provides no realisation statements specifying how features in his systems are realised in lexicogrammatical systems.

[2] To be clear, this point had already been made in Halliday (1985: xvii):

[3] To be clear, what Halliday means by this is that the relation between experiential meanings and grammatical forms is not arbitrary: process and verbal group, participant and nominal group, circumstance and adverbial group/prepositional phrase.

Martin, however, mistakes this to mean the relation between his discourse semantics and lexicogrammar interpreted 'by reference to what it means' (rather than grammatical forms).

Confusing Strata [Revised]

Martin (1992: 1):

Like Cohesion In English [Halliday & Hasan], English Text uses systemic functional grammar to ask questions about text structure, and complements the grammar by developing additional analyses which focus on text rather than clause.  Cohesion In English organises this division of labour as the opposition between grammar and cohesion (between structural and non-structural resources for meaning).  English Text organises this division of labour in a different way — stratally, as an opposition between grammar and semantics (between clause oriented and text oriented resources for meaning).

Blogger Comments:

[1] This misunderstands the most basic premise of Cohesion In English. Halliday & Hasan (1975: 6,7):

[2] In SFL, cohesion is not in opposition to grammar — it is within the grammar, and within the textual metafunction, as its non-structural component (Halliday & Matthiessen 2004: 524-85).  As such, cohesion is a lexicogrammatical means of realising textual meaning.  That is, cohesion is construed as a lower level of symbolic abstraction than semantics.

Symbolic abstraction is the principle on which linguistic strata are organised.  The relation between strata is one of realisation: lower stratum realises higher stratum.  A transitivity analysis of this shows that the relation between strata is identifying and intensive (elaborating):

lexicogrammar (wording)	realises	semantics (meaning)
Token	Process: identifying: intensive	Value

[3] Clearly, construing the opposition between the non-structural textual grammar and the remainder of the grammar — textual, interpersonal, experiential, and logical — as a stratal opposition is inconsistent with the meaning of strata: as different levels of symbolic abstraction.

Since Martin distinguishes the content plane strata as an opposition "between clause oriented and text oriented resources for meaning", it is reasonable to conclude that he formulated his semantics on the basis of the size of units, rather than on the basis of symbolic abstraction.  This is a serious inconsistency with the original model.


Conclusion: The founding rationâle for establishing a stratum of 'discourse semantics' is inconsistent with the theory in which it is intended to be situated.