"The Role Of 'Some Proposals' In Developing A Modern SF Theory Of Syntax"

Fawcett (2010: 166-7):

Let me now summarise this short chapter. "Some proposals" began the work of overhauling the concepts first presented in "Categories" in the light of the requirements of a modern theory of SF syntax — a theory in which the system networks of TRANSITIVITY, MOOD, THEME and the rest are regarded as modelling the level of meaning. As I wrote at the time, "the syntactic categories [...] are those [...] needed to state with the greatest economy the realisation rules that express the options in the semantics" (Fawcett 1974:4-5). The new ideas that it introduced have all proved their value, both in the construction of computer parsing systems (Weerasinghe & Fawcett 1993, Weerasinghe 1994 and Souter 1996) and in the hand analysis of texts by myself, my teaching colleagues, members of a large team of researchers in a study of the language of children aged 6-12, and many generations of students. 

However, later work has shown that in some cases the revisions were not drastic enough (e.g., the retention of the concept of the 'rank scale'). Yet in most cases the concepts established in "Some proposals" have passed the test of twenty-five years of use in various fields of application as a descr[ip]tion of English, as well as in describing various other languages, and they are the central concepts in the theory of syntax for a modern systemic functional grammar that I shall present in Part 2.

Blogger Comments:

[1] To be clear, Fawcett's "Some proposals" addressed Halliday's first theory, Scale and Category Grammar, after Halliday had devised his second theory, Systemic Functional Grammar, in the light of the requirements of a theory of syntax, despite the fact that neither of Halliday's theories is a theory of syntax, as previously explained.

[2] To be clear, in SFL Theory, the system networks of TRANSITIVITY, MOOD, THEME are located at the level of wording (lexicogrammar), not meaning (semantics). As Halliday (1985: xvii) explains:

The relation between meaning and wording is not, however, an arbitrary one; the form of the grammar relates naturally to the meanings that are being encoded. A functional grammar is designed to bring this out; it is a study of wording, but one that interprets the wording by reference to what it means.

[3] To be clear, this is a statement about Fawcett's model only (Figure 4), where realisation rules involving features of meaning are construed as the form that realises systems of meaning.

[4] To be clear, here Fawcett positively assesses his own model.

[5] To be clear, in SFL Theory, the rank scale is the means of modelling what other theories model as syntax.

Fawcett's Two Types Of Sequential Relationship And Recursion

Fawcett (2010: 166):

Two types of sequential relationship between sister elements of structure are recognised in "Some proposals": the segmental relationship of adjacency and the discontinuous relationship that occurs when elements are non-adjacent. These are not explicitly mentioned in "Categories", but every theory of syntax must have adequate ways of handling the various types of discontinuity found in language. Part 2 will introduce these concepts in Section 11.7 of Chapter 11.

Finally, two types of recursion are recognised. The first is embedding, in which a unit is 'rankshifted' (to use Halliday's original term) to fill an element of structure in a unit of the same or a lower rank. The second is co-ordination, in which two or more units fill a single element of structure.

With the exception of the concept of 'rank' — and so the derived concept of 'rank shift' — all of these 'relationships' are retained in the model of syntax to be presented in Part 2, and a small number of additional concepts that are not included in "Some proposals" are added to them. There is no need, therefore, to discuss the concepts of "Some proposals" any further at this point.

Blogger Comments:

[1] To be clear, SFL Theory proposes different structural relations for each metafunction, with 'segmental' describing experiential structures only. Halliday & Matthiessen (2014: 85):

[2] As previously noted, Fawcett's "Some proposals" (1974) was oriented to Halliday's first theory, Scale and Category Grammar, after it had been superseded by Halliday's second theory, Systemic Functional Grammar.

[3] To be clear, in formal theories, unlike SFL Theory, the term 'embedded' covers both rankshift and hypotaxis. Halliday & Matthiessen (2014: 382):

Such items are said to be ‘rankshifted’ – by contrast with ranking ones, which function prototypically as constituents of the higher unit. We may also use the term ‘embedded’, taken from formal grammars; but with the proviso that this term is often used to cover both rankshift (where the item is downgraded as a constituent) and hypotaxis (where the item is dependent on another one but is not a constituent of it. Here we shall use embedded only as an alternative term synonymous with rankshifted.

As will eventually be seen, Fawcett uses 'embedding' to include hypotactic relations between ranking clauses.

[4] To be clear, in SFL Theory, the traditional notion of co-ordination is treated as paratactic extension. Halliday & Matthiessen (2014: 472).

The combination of extension with parataxis yields what is known as co-ordination between clauses. It is typically expressed by and, nor, or, but.

Fawcett (p272), on the other hand, uses co-ordination and embedding to replace parataxis and hypotaxis, depending on the logico-semantic relation involved:

we treat four of Halliday's five types of 'hypotaxis' and two of his five types of 'parataxis' as embedding, and one type of 'hypotaxis' and his three 'expansion' types of 'parataxis' as co-ordination.

The problems with this approach will be identified when Fawcett's model is finally presented.

[5] To be clear, rank is the means by which SFL Theory models the phenomena of language known in other models as 'syntax'.

[6] To be clear, Fawcett's model includes embedding, but not rank-shift.

The Concept Of Rank In Fawcett's Models

Fawcett (2010: 165-6):

The concept of rank from "Categories" is retained in "Some proposals" — at least in relation to the clause and the group, which form the core of the model of syntax in all SFL descriptions (with far less work on the proposed 'morpheme-word' relationship). However, this small syntactic 'rank scale' is interpreted in a very different way from "Categories", because it is seen as the realisation of an equivalent semantic relationship between a 'situation' and the 'things' and 'qualities' that are its 'elements' at that higher level. And, as we shall see in Part 2, later work in this version of the theory was to reduce the role of the concept of a 'rank scale' to the point where it no longer has any status in the theory at all. … In the theory of syntax to be presented in Part 2 the concept of the 'rank scale' is replaced by the concept of probabilities in the relations between elements and units.

Blogger Comments:

[1] As previously noted, Fawcett's "Some proposals" (1974) was oriented to Halliday's first theory, Scale and Category Grammar, after it had been superseded by Halliday's second theory, Systemic Functional Grammar.

[2] This is potentially misleading, because SFL is not a theory of syntax. Halliday (1985: xiv):

[3] To be clear, this aspect of Fawcett's model is included in his Figure 12 (p210):

In SFL Theory, the semantic unit congruently realised by a clause depends on metafunction:

• figure (ideational)
• proposition/proposal (interpersonal)
• message (textual)

and ideationally, the semantic unit congruently realised by a group is an element (Halliday & Matthiessen 1999).

[4] To be clear, SFL Theory it is the rank scale that models the linguistic phenomena deemed 'syntax' in other theories.

[5] To be clear, in SFL Theory, rank is the ordering principle of structure (syntagmatic order), whereas probability is the quantification of system (paradigmatic order). This will be elaborated further in the examination of Fawcett's model in Part 2.

Fawcett's Componence, Filling and Exponence

Fawcett (2010: 165):

8.4 The 'relationships' of "Some proposals" 

The first difference is that "Some proposals" recognises many more 'relationships' than the three 'scales' in "Categories". There are ten of them. 

One of its major innovations is to split Halliday's 'scale' of 'exponence' into three: componence, filling and exponence proper. To cite Butler's excellent summary of these concepts:

Componence is the relation between a unit and the elements of structure of which it is composed. For example, a clause may be composed of the elements S, P, C and A. Each of these elements of structure may be (but need not be) filled by groups. In the specification of a syntactic structure, componence and filling alternate until, at the bottom of the structural tree, the smallest elements of structure are not filled by other units. It is at this point that we need the concept of exponence, as used by Fawcett: the lowest elements of structure are expounded by items', which are [...] more or less equivalent to 'words' and 'morphemes' in Halliday's model. (Butler 1985:95) 

These three concepts, which today still form the basis of the Cardiff Grammar's model of syntax, are clearly exemplified in the top half of Figure 10. These concepts are necessary, in one form or another, in any adequate systemic functional model of syntax, and they will be illustrated, discussed and compared with their antecedents in the relevant sections of Chapter 11.

Blogger Comments:

[1] As previously noted, Fawcett's "Some proposals" (1974) was oriented to Halliday's first theory, Scale and Category Grammar, after it had been superseded by Halliday's second theory, Systemic Functional Grammar.

[2] To be clear, in Halliday's first theory, the term 'exponence' covered what were to become two distinct relations in his second theory: realisation and instantiation. On the one hand, exponence corresponds to the relation between an element of function structure and the class of unit that realises it. Halliday (2002 [1961]: 54, 57):

The exponent of the element S in primary clause structure is the primary class nominal of the unit group. …

The fact that by moving from structure to class, which is (or can be) a move on the exponence scale, one also moves one step down the rank scale, is due to the specific relation between the categories of class and structure, and not to any inherent interdetermination between exponence and rank.

On the other hand, exponence corresponds to the instantiation relation between theory and data. Halliday (2002 [1961]: 57):

Exponence is the scale which relates the categories of the theory, which are categories of the highest degree of abstraction, to the data.

[3] To be clear, 'componence' is the relation of composition, a type of extension. In SFL Theory, composition is modelled as a rank scale of forms, and is distinct from the function structures of the units on the rank scale.

[4] To be clear, one the one hand, Fawcett's 'filling' corresponds to the relation between an element of clause structure and the group that realises it, but treats these two distinct levels of symbolic abstraction, function and form as if both were at the same level. On the other hand, 'filling' suggests an active process, rather than an inert relation, and if this were a model of human language, rather than an algorithm for text generation, the process would be (an aspect of) instantiation.

[5] To be clear, Fawcett's 'exponence' corresponds to the relation between an element of group structure and the word (and morphemes) that realise it. That is, it is a different term for the same relation as 'filling': realisation.

[6] This is misleading, because it is untrue. As can be seen below, while Figure 10 does exemplify Fawcett's notion of componence, a form (clause) composed of functions, it does not exemplify the filling of clause functions by groups or the exponence of groups by words and morphemes. (The 'text' line is the data being analysed, not an analysis of the data at group or word level.)

Systemic Syntax Without Systems

Fawcett (2010: 165):

The last of the three theoretical categories recognised in "Some proposals" (along with 'unit' and 'class of unit') was element of structure. Like 'unit', this term was used in essentially the same sense as in "Categories". However, the concept plays a far stronger part in the theory as a whole than it does in "Categories" because of the mutual dependency, in the present framework, of the definitions of a class of unit and its elements of structure. In "Categories", in contrast, a class of unit is defined by its potential for "operation in the structure of the unit next above" (1961/76:64).

In two of its 'categories', then, "Some proposals" is similar to Halliday (1961/76). While the concept of 'class of unit' is broadly similar, the criteria for recognising the 'class' of a 'unit' are very different, and the concept of 'system' does not appear at this level of description at all. However, there are even greater differences between the 'scales' of "Categories" and the 'relationships' of "Some proposals", as the next section shows.

Blogger Comments:

[1] As previously noted, Fawcett's "Some proposals" (1974) was oriented to Halliday's first theory, Scale and Category Grammar, after it had been superseded by Halliday's second theory, Systemic Functional Grammar.

[2] As previously explained, this downplaying of difference is misleading, because the difference between Fawcett's notion of 'class of unit' and Halliday's notion is indeed significant. Halliday (1961) takes the view 'from above', distinguishing classes of unit according to the functions they realise, whereas Fawcett takes the view 'from below', distinguishing classes of unit according to the structures that realise them. Halliday (2002 [1961]: 50):

What is theoretically determined is the relation between structure and class on the one hand and unit on the other. Class, like structure, is linked to unit: a class is always a class of (members of) a given unit: and the class–structure relation is constant – a class is always defined with reference to the structure of the unit next above, and structure with reference to classes of the unit next below. A class is not a grouping of members of a given unit which are alike in their own structure. In other words, by reference to the rank scale, classes are derived “from above” (or “downwards”) and not “from below” (or “upwards”).

Moreover, as the units 'quantity group', 'quality group' and 'tempering cluster' demonstrate, Fawcett confuses classes of units with their functions.

[3] To be clear, if the concept of system does not "appear" at the level of syntax, then Fawcett's claim that his model is a systemic model of syntax is invalid.

Classes Of 'Cluster' And No Classes Of 'Word'

Fawcett (2010: 164):

Three classes of the new unit of the cluster were tentatively recognised in the 1974-6 edition: the genitive cluster (e.g., her boyfriend's)', the proper name cluster (e.g., Dr Ivy Idle) and the tempering cluster (e.g., much less in much less painful). By 1981 the 'tempering cluster' had been absorbed into the 'quantity-quality group', and later still the data that it covers became part of the evidence for the need to introduce the 'quantity group'. Further classes of 'cluster' are added in the current version of the grammar, as described in Part 2.

There was no discussion in "Some proposals" of the concept of 'class of word', because 'words' are treated there as items rather than as syntactic units (1974-6/81:67). (See Part 2 for a fuller justification for taking the position that the relationship of morphemes to a word is not the same as that between, let us say, groups and clauses.)

Blogger Comments:

[1] To be clear, like the notions of 'quantity group' and quality group, the notion of a 'tempering cluster' mistakes a function of a cluster (tempering) for a class of a cluster.

[2] To be clear, in SFL Theory, 'word' is used for two distinct abstractions: lexical item and grammatical rank scale unit. The word as lexical item is the synthetic realisation of the most delicate lexicogrammatical features — just as the phoneme /b/ is the synthetic realisation of the features [voiced, bilabial, stop]. The word as grammatical unit is (i) a constituent of groups/phrases and realises elements of their function structures, and (ii) composed of morphemes. It is the word as grammatical unit that is differentiated by (grammatical) class.

[3] Fawcett's justification for his position will be critically evaluated in the examination of Part 2.

[4] To be clear, in SFL Theory, the relationship of morphemes to a grammatical word, and of groups to clauses is the same: constituency. It is this commonality that allows Halliday to model syntax and morphology as a rank scale in both theories, Scale and Category Grammar and Systemic Functional Grammar. For example, for Scale and Category Grammar, Halliday (2002 [1961]: 51) explains:

Traditionally these terms have usually referred to “grammar above the word” (syntax) and “grammar below the word” (morphology); but this distinction has no theoretical status. It has a place in the description of certain languages, “inflexional” languages which tend to display one kind of grammatical relation above the word (“free” items predominating) and another below the word (“bound” items predominating). But it seems worthwhile making use of “syntax” and “morphology” in the theory, to refer to direction on the rank scale. “Syntax” is then the downward relation, “morphology” the upward one; and both go all the way. We can then say, simply, classes are syntactical and not morphological.

Later Changes (1981) To Fawcett's "Some Proposals" (1974): Verbal, Quality & Quantity Groups

Fawcett (2010: 164):

However, by the 1981 edition I had made several changes to the model. Firstly, I had strengthened my position on the supposed 'verbal group', writing that even for text analysis the elements of the supposed 'verbal group' "would in my view be best shown as clause elements" (Fawcett 1974-6/81:31) (See Fawcett (2000) and (forthcoming b) for the full set of reasons and Appendix C for a summary of them.) Secondly the 'adjectival' and 'adverbial' groups, which were already shown in the 1974-6 edition as sharing a common structure, were brought together as the 'quantity-quality group' (since they typically express the meaning of a 'quantity' of a 'quality'). And thirdly the 'prepend group' had been re-named the "prepositional group". Later still, as Part 2 shows, I realised that, in order to provide an adequate description of English and other languages, we need to recognise both a quality group and a quantity group. See Tucker (1998) for the fullest explanation of the quality group in English and Fawcett (in press) for an introduction to both classes of group.

 Blogger Comments:

[1] As will be seen in the examination of Appendix C, Fawcett's argument against the verbal group (pp333-6) centres on misunderstandings involving the Finite element and phrasal verbs.

[2] To be clear, the notion of 'quality group' and 'quantity group' confuses class (of group) with function (of group).

[3] As previously noted, Fawcett (in press) is still unpublished, 20 years after the first edition of this work.

The Concepts Of Unit And Class Of Unit In Fawcett's "Some Proposals" (1974)

Fawcett (2010: 163-4):

"Some proposals" recognises a number of units (in precisely the same sense of "unit" as in "Categories"). These units are related to each other on a 'rank scale', and for English they are: clause, group and cluster. The unit 'cluster' is a new unit that is lower on the 'rank scale' than the group; see below for examples. (The concept of 'unit', which is mutually defining with that of the 'rank scale', is not used in the framework to be proposed in Part 2.)

"Some proposals" uses the concept of class of unit in a way that is loosely similar to its use in "Categories". It relates the concept of 'class of unit' to 'unit' essentially as in "Categories", except that the criteria for recognising a given class of unit are the elements of its internal structure, rather than its potential for operation in the unit above, as in "Categories". (For a full discussion of the criteria for recognising classes of unit, see Section 10.2.2 of Chapter 10.) On this criterion, "Some proposals" recognises only classes of groups and clusters (and not clauses or words). Thus it recognises that, at the 'rank' of clause rank in English, there is only one 'class' of clause, so that the concept of 'class of unit' is not applicable to this unit. At group rank "Some proposals" initially recognised four classes. These were the nominal group (e.g., she, Ivy and the man in black), the prepend group (e.g., in black), the adjectival group (e.g., very quick) and the adverbial group (e.g., very quickly). In the 1974-6 edition I further suggested (1974-6/81:31) that "in an introductory analysis it is probably helpful to make use of the traditional S&C concept of the 'verbal group'"(e.g., might have seen).

Blogger Comments:

[1] To be clear, Fawcett's "Some proposals" (1974) was oriented to Halliday's first theory, Scale and Category Grammar, after it had been superseded by Halliday's second theory, Systemic Functional Grammar.

[2] This downplaying of difference is misleading, because the difference between Fawcett's notion of 'class of unit' and Halliday's notion is indeed significant. Halliday (1961) takes the view 'from above', distinguishing classes of unit according to the functions they realise, whereas Fawcett takes the view 'from below', distinguishing classes of unit according to the structures that realise them. Halliday (2002 [1961]: 50):

What is theoretically determined is the relation between structure and class on the one hand and unit on the other. Class, like structure, is linked to unit: a class is always a class of (members of) a given unit: and the class–structure relation is constant – a class is always defined with reference to the structure of the unit next above, and structure with reference to classes of the unit next below. A class is not a grouping of members of a given unit which are alike in their own structure. In other words, by reference to the rank scale, classes are derived “from above” (or “downwards”) and not “from below” (or “upwards”).

[3] To be clear, Fawcett's "Some proposals" (1974) does not recognise classes of words: nouns, adjectives, verbs, adverbs, prepositions, conjunctions.

[4] To be clear, although not acknowledged, this is also the case in Halliday (1961).

The Other Types Of 'Form' In Fawcett's "Some Proposals" (1974)

Fawcett (2010: 163):

"Some proposals" was written as a short description of the syntax of English, and it concludes with a summary of the theory that the description presupposed. The structure of the paper therefore reflects the view that theory should grow out of description — or, more accurately, out of the attempt to apply the concepts of an earlier theory (here "Categories") in a description. In my view, this interleaving of theory and description is one of the more effective ways to improve one's theory — as well as through the use of corpora and, as suggested in Section 1.2 of Chapter 1, large scale computer implementations. (See also the discussion of the relationships between theory, description and application in Halliday & Fawcett 1987b:1f.).

We must begin by relating the model of syntax to the model of language as a whole. Within a model of the sort presented in Figure 4 in Section 3.2 of Chapter 3, "Some proposals" sees syntax as one of three ways in which meanings are realised as forms — the other types of 'form' being items and intonation or punctuation (depending on the channel of discourse).

Blogger Comments:

[1] To be clear, Fawcett's "Some proposals" (1974) presupposed Halliday's first theory, Scale and Category Grammar, after it had been superseded by Halliday's second theory, Systemic Functional Grammar.

[2] To be clear, a linguistic theory is a model of language, and a linguistic description is the application of a theory to a particular language. For example, Halliday's IFG (1985, 1994) presents a theory of language which is applied to a description of English. That is, theory precedes description, and the description of data tests the theory that is designed to account for the data.

[3] To be clear, computer implementations of a theory, large or small, are concerned with adapting a theory of language to the limitations of machines. This is distinct from modelling language as a system that has evolved in the human species.

[4] Reminder:

[5] To be clear, Fawcett's "Some proposals" (1974) locates some aspects of the content plane — syntax and items (words and morphemes) — and some aspects of the expression plane — intonation or punctuation — at the same level of symbolic abstraction: form.

The omissions in these characterisations of phonology (rhythm and articulation) and graphology (orthography) will be examined when Fawcett eventually discloses his current model. For the moment, it can be seen, from Figure 4, where system networks are limited to his level of meaning, that Fawcett does not model phonology and graphology as systems.

"My Theory Of Syntax"

Fawcett (2010: 162-3):

The following words provide the key to understanding my original motivation for developing a framework for representing structure — a framework that was very different from the 'multiple structure' model that Halliday was developing at about the same time:

the syntactic categories [...] are those [...] needed to state with the greatest economy the realisation rules that express the options in the semantics (Fawcett 1974:4-5). 

Over twenty-five years later, my theory of syntax and the consequent description of English syntax have both developed in various ways, but those words still express exactly what I wish to say on this matter. It is because the description of the functional structure is necessarily complemented, in my approach, by a description in terms of a functional semantics that the syntax can — and should — be less "extravagant" (Halliday 1994:xix) than it is in Halliday's IFG. Moreover, I have discovered in the intervening time that this general principle holds just as strongly for the version of the grammar that is used for the computer model of natural language generation as it does for the version used for text analysis. The clear implication of all of this work is that we cannot provide a complete description of a text without providing both an analysis of its functional syntax and an account of the semantic features that have been chosen in generating it. 

Blogger Comments:

[1] To be clear, this is a statement that only applies to Fawcett's own model (Figure 4), in which structure is misunderstood as an "instance" of realisation rules which are, in turn, misunderstood as the form that realises systems. Both of Halliday's theories model syntax as a compositional rank scale on the stratum of lexicogrammar. As will be seen, Fawcett rejects the theoretical value of a rank scale.

[2] To be clear, as previously explained, Fawcett's syntax is only "less extravagant" than Halliday's function structures because Fawcett exports the majority of Halliday's grammatical functions to his level of meaning. When these exported functions are counted in the description, Fawcett's "less extravagant" model of the clause involves at least seven lines of description (Figure 10), compared with Halliday's "more extravagant" three (theme, mood, transitivity).

[3] To be clear, this is a personal recount presented as if reasoned argumentation.

[4] To be clear, this is again a statement of Fawcett's model, not Halliday's Systemic Functional Grammar. In Systemic Functional Grammar, there is no "functional syntax", and what Fawcett terms 'semantic features' are features of lexicogrammatical systems. In the absence of grammatical metaphor, lexicogrammatical features are congruent with (agree with) semantic features. It is grammatical metaphor that motivates the distinction between semantic and lexicogrammatical systems.

A Second Motivation For Fawcett's "Some Proposals" (1974)

Fawcett (2010: 162):

There was a second motivation, however. This was that my attempts to apply Halliday's categories in the analysis of text at the level of form had led me to modify his description in a number of ways — and so in due course to think about the consequences of these descriptive changes for the theory upon which the description rested.

Blogger Comments:

[1] To be clear, neither Halliday's first theory, Scale and Category Grammar, nor his second, Systemic Functional Grammar, postulates a level of form. The fact that Fawcett attempted to apply Halliday's categories at "the level of form" demonstrates that Fawcett did not understand either of Halliday's theories, and that his modifications to Halliday's "descriptions"  — and their consequences — derive from his inability to understand them.

[2] To be clear, as Fawcett acknowledges (p161) his "Some Proposals" (1974) presented his set of revisions to the concepts of "Categories", an outline of Halliday's first theory, Scale and Category Grammar, after Halliday had already devised an early version of his second theory, Systemic Functional Grammar. Scale and Category Grammar is not the theory on which Systemic Functional Grammatical descriptions rested.

The Purpose Of Fawcett's "Some Proposals" (1974)

Fawcett (2010: 161-2):

The purpose of "Some proposals" was to present the set of revisions to the concepts of "Categories" that I considered at the time (1974) to be required, in order to constitute a 'systemic syntax' that would complement Halliday's revolutionary proposal that the system networks of TRANSITIVITY, MOOD and THEME should be regarded as choices between meanings (as described in Section 4.3 of Chapter 4). As Butler wrote in 1985:

Halliday's most recent proposals [...] make it clear that the functional component networks are to be regarded as semantic, but we are given very little idea of what the lexicogrammar [i.e., the level of form] might now consist of [my emphasis] (Butler 1985:94).

"Some proposals" was a first attempt to fill this gap, and it probably reached its widest readership through Butler's generally approving summaries of its contents — first in Butler (1979), and then in Butler (1985:94-102). He makes the key point when he states, "[Fawcett's] approach to the recognition of syntactic categories is dictated by his commitment to the centrality of semantics" (Butler 1985:94). In this approach, then, a large part of the description of a text should be in terms of the semantic features that have been chosen in generating it (e.g., in the way described in Section 7.8 of Chapter 7).

Blogger Comments:

[1] To be clear, the stated purpose of Fawcett's "Some proposals" was to present a set of revisions to Halliday's first theory, Scale and Category Grammar, despite the fact that it had been replaced by a new theory, Systemic Functional Grammar.

[2] As Halliday (1985: xiv) explained, the notion of syntax is inconsistent with the assumptions and mode of inquiry of SFL Theory, for reasons previously cited on this blog.

[3] To be clear, this refers to Halliday's second theory, Systemic Functional Grammar, not his first, Scale and Category Grammar. In the earliest form of Systemic Functional Grammar, the metafunctional systems of the clause were mapped onto the grammatical rank scale of form (see [4] below). This was soon replaced by the relocation of these systems to the stratum of lexicogrammar, largely to provide a systematic account of mismatches between meaning and wording (grammatical metaphor). Because SFG is a functional grammar, the grammar is modelled in terms of its functions, and the function of the grammar is to realise meaning. 

[4] This is misleading, because it is untrue. Even in the earliest version of SFG, where the metafunctional systems were located on the stratum of semantics, Halliday was explicit about what constituted the lexicogrammar. For example, Halliday (1978: 128-9):

Fourth, we shall assume that each component of the semantic system specifies its own structures, as the ‘output, of the options in the network (so each act of choice contributes to the formation of the structure). It is the function of the lexicogrammatical stratum to map the structures one onto another so as to form a single integrated structure that represents all components simultaneously. With negligible exceptions, every operational instance of a lexicogrammatical construct in the adult language — anything that realises text — is structured as the expression of all four components. In other words, any instance of language in use ‘means, in these various ways, and shows that it does so in its grammar.

Fifth, we shall assume that the lexicogrammatical system is organised by rank (as opposed to by immediate constituent structure); each rank is the locus of structural configurations, the place where structures from the different components are mapped on to each other. The ‘rank scale' for the lexicogrammar of English is:

 [5] This misleading, because in all of Halliday's theorising, lexicogrammar is the level of wording, not form; form is a level in Fawcett's model only. For example, in the earliest model of SFG, Halliday (1978: 128) writes:

We shall start with the assumption that the semantic system is one of three strata, chat constitute the linguistic system:

Semantic (the meaning)
Lexicogrammatical (the wording, i.e. syntax, morphology and lexis)
Phonological (the sound)

[6] As demonstrated above, there was no gap to fill. In both of Halliday's theories, grammatical form is modelled as a rank scale. The reason why Fawcett argues against a rank scale, is that it is the most obvious aspect of Halliday's theory that makes his own model redundant.

[7] To be clear, as the critical examinations on this blog demonstrate, contrary to the priorities of Systemic Functional Theory, Fawcett's "commitment to the centrality of semantics" involves prioritising form over function and structure over system.

[8] This is potentially misleading. To be clear, in Systemic Functional Grammar, the metafunctional systems of the clause are located on the lexicogrammatical stratum (rather than semantics), whereas in the Cardiff Grammar they are located at the level of meaning (rather than form). That is, what the Cardiff Grammar describes in terms of semantic features, Systemic Functional Grammar describes in terms of lexicogrammatical features.

Misrepresenting Critical Reactions To Fawcett (1974)

Fawcett (2010: 161):

It is ironic, therefore, that I should have been admonished in the mid-1970s by two senior systemic linguists at the time (though not by Halliday himself) for the changes that I suggested in "Some proposals" . The irony lies in the fact that my proposals were for a syntax whose representations were far more like those of the Scale and Category model than the proposals for representing structure that Halliday himself developed during the very late 1960s and the early 1970s. As we saw in Chapter 7, it was during this period that he shifted his representations of structure away from the model presented in "Categories" (in which the major clause elements were 'Subject', 'Predicator', 'Complement' and 'Adjunct') to the multiple structures used in IFG and widely elsewhere, in which an analysis in terms of such elements plays little or no role. The changes made by Halliday were far greater than the changes that I was suggesting in "Some proposals", and yet the position of my critics at the time seemed to be that I should not be suggesting any changes to "Categories" at all! The fact is that, even though Halliday's changes were already well under way by the mid-1970s, the key role of "Categories" as the founding document of the new theory seems to have placed it beyond criticism for many. On the whole, then, I think that my critics' response shows that the admittedly provocative use of the term "iconoclastic" in the subtitle of "Some proposals" was justified. 

However,"Some proposals" was really not as revolutionary as its critics felt it to be. With the benefit of a further twenty-five years of hindsight, I see now that, far from being too critical of "Categories", "Some proposals" did not go as far as it should in proposing changes, as we shall see in due course.

Blogger Comments:

[1] This is misleading. By the mid-1970s, Halliday had replaced his Scale and Category Grammar with a new theory: Systemic Functional Grammar. In proposing a syntax at that time, Fawcett was inconsistent with both theories. In proposing changes to the first theory, Fawcett was ignoring the fact that that theory had been superseded.

[2] This is very misleading indeed. In the new theory, Systemic Functional Grammar, the elements Subject, Predicator, Complement and Adjunct (± Finite) constitute the interpersonal structure of the clause.

[3] To be clear, the changes made by Halliday were so great that they resulted in a new theory. Fawcett's suggestions for changes to "Categories" were suggestions for changes to the superseded theory.

[4] This is misleading. Scale and Category Grammar was the earlier theory from which Halliday developed his new theory. An old theory (e.g. Newton's Universal Gravitation) is not the founding document of the theory that replaces it (e.g. Einstein's General Relativity).

[5] To be clear, Fawcett's use of 'iconoclastic' couches support for a theory in terms of belief, rather than, say, in terms of explanatory potential, self-consistency, reasoned argumentation and evidence.

[6] This is misleading, because it is the opposite of what is true. Here Fawcett misrepresents his critics as deeming his proposals as revolutionary, despite the fact that, in advocating the old theory, his proposals were reactionary.

Misrepresenting Scale-&-Category Grammar And Systemic Functional Grammar As The Same Theory

Fawcett (2010: 161):

As will by now be clear, my view at the time was that the elevation of the concept of 'system' to the semantics was so revolutionary a change that it should be expected to affect every part of the theory. In a sense, the strongest possible support for this position came when Halliday introduced an entirely new way of representing the structure of clauses. You can see this difference very clearly if you compare Figure 1 in Chapter 2 and Figure 7 in Chapter 7.

Blogger Comments:

This is misleading. To be clear, the changes introduced by Halliday should not "be expected to affect every part of the theory" — Scale and Category Grammar — because these changes were part of Halliday's second theory, Systemic Functional Grammar, which replaced his first theory. This is a motivated misrepresentation, because Fawcett's model has more in common with Halliday's superseded theory than Systemic Functional Grammar, as will become evident when Fawcett eventually begins outlining his own model.

Misrepresenting Scale and Category Grammar As A Model Of Syntax

Fawcett (2010: 160):

At the end of the 1960s and the start of the 1970s there was a spate of text books that functioned as introductions to Scale and Category syntax — rather as there was to be a second spate of introductory grammars in the theory in the 1990s, this time based on Halliday's IFG. Each of those early textbooks had its st[r]ong points, but two stood out because of their clear vision of a model of language in which one component was systemic and presented the system networks that constituted the 'meaning potential' of the language, and another component which provided for the structures — using the familiar Scale and Category concepts (minus 'system'). The first was Muir's two-part text book, with one part on structures and one on systems, and the other was the two-volume work that became the standard introduction to the theory (Berry 1975 and 1977). Berry (1977) is particularly noteworthy for providing, in a book that was essentially designed to enable its readers to analyse texts, a sketch of how a generative version of the model would operate.³ But the key point here is that in both Muir's and Berry's books the picture of syntax that was presented was that of "Categories".

³ This set of books also included Strang (1962/69), Leech (1966), Scott, Bowley et al 1968 (where Bowley was the principal contributor), Turner & Mohan (1970) and Sinclair (1972). One reason why Berry (1975 and 1977) quickly became established as the standard introduction to the theory was that she introduces more of the theory that underpins the description than Muir (1972), including an early picture of how 'realisation' works.

Blogger Comments:

This is misleading. To be clear, here Fawcett again strategically misrepresents Halliday's first theory, Scale and Category Grammar, as a model of syntax, despite the fact that Halliday explicitly regarded 'syntax' as merely one component of grammar, which he modelled as a rank scale. Halliday (2002 [1961]: 51):

The distinction does, however, need a name, and this seems the best use for the terms “syntax” and “morphology”. Traditionally these terms have usually referred to “grammar above the word” (syntax) and “grammar below the word” (morphology); but this distinction has no theoretical status. It has a place in the description of certain languages, “inflexional” languages which tend to display one kind of grammatical relation above the word (“free” items predominating) and another below the word (“bound” items predominating). But it seems worthwhile making use of “syntax” and “morphology” in the theory, to refer to direction on the rank scale. “Syntax” is then the downward relation, “morphology” the upward one; and both go all the way.

Fawcett's "Iconoclastic Approach To Scale and Category Grammar"

Fawcett (2010: 159-60):

I gave "Some proposals" the subtitle of "an iconoclastic approach to Scale and Category Grammar" — a subtitle that was, I have to admit, a little provocative. But I did so for what I thought at the time (and still think) to be a good reason — namely, that too many of those who were working in SFL seemed to look upon "Categories" as a set of concepts whose authority fell not far short of the tablets brought down from Mount Sinai by Moses. And yet there were serious questions, I felt, that needed to be asked.

If, for example, the meaning of the term "system" had changed by becoming a choice between meanings rather than forms, where did that leave the concept of "class"? After all, 'class' was a term which had been interpreted in "Categories", like 'system', as a paradigmatic relationship, such that one might set up 'systems' of 'classes'. So what did it mean, in a model in which the system networks were semantic, to say that a nominal group was a 'class of group'? Since the concept of a 'nominal group' was used to refer to a syntactic unit, it belonged, surely, at the level of form rather than meaning — so at what level of language did the concept of 'class' now belong? Indeed, did any of the original four categories remain as concepts that could be used for the task of modelling what we might term 'pure form'?

In other words, it seemed to me that the elevation of the concept of 'system' to the level of meaning meant that the whole framework for describing language needed to be re-examined. Moreover, if we did retain concepts such as 'class of unit' at the level of form, we needed new terms for the equivalent concept at the level of meaning.² (See Figure 12 in Section 10.2.10 of Chapter 10 for the equivalent terms that I use.)

² The entry conditions for Halliday's system networks for 'meaning potential' have always had labels such as "clause", "nominal group", etc. It is odd to find these terms being used at the level of meaning, because they are the names of syntactic units at the level of form — the outputs from the grammar. This is why, in the Cardiff Grammar, the equivalent features have explicitly semantic labels, such as "situation" and "thing".

Blogger Comments:

[1] To be clear, in this discussion, Fawcett is not concerned with the current model of Systemic Functional Grammar, but with two different superseded theories — Halliday's first theory, Scale and Category Grammar, and the earliest version of Halliday's second theory, Systemic Functional Grammar. Fawcett's assumption throughout is that the second superseded theory must be consistent with the first.

[2] To be clear, in both of the superseded theories, the concept of (grammatical) class is located in the rank scale of the lexicogrammatical stratum. For the first theory, Halliday (2002 [1961]: 49) writes:

The structure is set up to account for likeness between events of the same rank, and it does so by referring them to the rank next below. To one place in structure corresponds one occurrence of the unit next below, and at each element operates one grouping of members of the unit next below. This means that there will be certain groupings of members of each unit identified by restriction on their operation in structure. The fact that it is not true that anything can go anywhere in the structure of the unit above itself is another aspect of linguistic patterning, and the category set up to account for it is the class. The class is that grouping of members of a given unit which is defined by operation in the structure of the unit next above.

For the early version of the second theory, Halliday (1978: 129) writes:

[3] To be clear, the original four categories from Halliday's first theory are unit, structure, class and system (Halliday 2002 [1961]: 41), all of which also figure in the current theory. Of these, unit and class are concerned with form, which SFL Theory, then and now, models as a rank scale.

[4] This non-sequitur (see above) is misleading, because it is untrue. In the superseded version of Halliday's second theory, each formal unit is a structural composite deriving from the (metafunctional) components of the semantics. Halliday (1978: 129)

It follows from the above that each type of unit — clause，verbal group, nominal group etc. - is in itself a structural composite, a combination of structures each of which derives from one or other component of the semantics.

[5] To be clear, because semantics is not a level of form, there is no 'equivalent concept' of class of form at the level of meaning. However, for ideational semantics, Halliday & Matthiessen (1999) propose sequence, figure and element as the semantic counterparts of clause complex, clause and group/phrase, and distinguish sub-types for each of these three 'order of phenomena'.

[6] To be clear, the reason why, in the current theory, units of the grammatical rank scale — clause, group etc. — serve as the entry conditions to "Halliday's system networks" is that such networks are grammatical systems. Here again Fawcett confuses 'meaning potential' (language as system) with 'meaning' as stratum (semantics).

[7] To be clear, Fawcett's Figure 12 (p210) is internally inconsistent and inconsistent with SFL Theory:

It is internally inconsistent because it presents relations between levels — expression, realisation — as levels. It is inconsistent with SFL Theory, because the inclusion of a 'belief system' component takes a transcendent, rather than immanent, perspective on meaning. Halliday & Matthiessen (1999: 415-6):

We can identify two main traditions in Western thinking about meaning (see Halliday, 1977):

(i) one oriented towards logic and philosophy, with language seen as a system of rules; 

(ii) one oriented towards rhetoric and ethnography, with language seen as resource. …

The two orientations towards meaning thus differ externally in what disciplines they recognise as models. These external differences are associated with internal differences as well.

(i) First, the orientations differ with respect to where they locate meaning in relation to the stratal interpretation of language:

(a) intra-stratal: meaning is seen as immanent — something that is constructed in, and so is part of, language itself. The immanent interpretation of meaning is characteristic of the rhetorical-ethnographic orientation, including our own approach. 

(b) extra-stratal: meaning is seen as transcendent — something that lies outside the limits of language. The transcendent interpretation of meaning is characteristic of the logico-philosophical orientation.

Many traditional notions of meaning are of the second kind — meaning as reference, meaning as idea or concept, meaning as image. These notions have in common that they are 'external' conceptions of meaning; instead of accounting for meaning in terms of a stratum within language, they interpret it in terms of some system outside of language, either the 'real world' or another semiotic system such as that of imagery.

Fawcett's 'external' conception of meaning, 'belief system', together with his stratification of language into meaning and form, place his work — unknown to Fawcett himself — in a different intellectual tradition from that of SFL Theory.

Misrepresenting Halliday On 'Syntax'

Fawcett (2010: 159):

As its dates of its first publication indicate, the three-part paper by myself that is to be reviewed here first appeared about fifteen years after "Categories", and it was later republished informally in a lightly revised edition in 1981. By the mid-1970s the new set of concepts associated with Systemic Functional Grammar (described in Chapter 4) were beginning to make themselves felt, but the original concepts presented in "Categories" were still widely accepted among systemic functional linguists (as we were beginning to call ourselves) as fully relevant to the description of the level of syntax or "grammar", as Halliday terms it. I would claim that "Some proposals" reflects the position of syntax in the model much as Halliday presents it in Halliday (1970/76) and (1977/78). 

Blogger Comments:

[1] To be clear, "Categories" (Halliday 1961) outlines Halliday's first theory, Scale and Category Grammar, which differs from Systemic Functional Grammar in significant ways, including the absence of metafunctions, and a different model of stratification, as previously explained.

[2] To be clear, the model of Systemic Functional Grammar in Halliday (1970/76, 1977/78) is an earlier superseded model, as previously explained.

[3] This very misleading, because, even in his first theory, Halliday is explicit that syntax is not the same as grammar, and that it can seen as the complement of morphology, and modelled as a rank scale. Halliday (2002 [1961]: 51):

The distinction does, however, need a name, and this seems the best use for the terms “syntax” and “morphology”. Traditionally these terms have usually referred to “grammar above the word” (syntax) and “grammar below the word” (morphology); but this distinction has no theoretical status. It has a place in the description of certain languages, “inflexional” languages which tend to display one kind of grammatical relation above the word (“free” items predominating) and another below the word (“bound” items predominating). But it seems worthwhile making use of “syntax” and “morphology” in the theory, to refer to direction on the rank scale. “Syntax” is then the downward relation, “morphology” the upward one; and both go all the way.