A GRAMMAR FOR THE
NATIONAL ENGLISH CURRICULUM
TRADITIONAL, FUNCTIONAL, STRUCTURAL, OR WHAT?
This paper explores some possible approaches to grammar teaching and its role in the study
of language as prescribed by the framework of the National English Curriculum (NEC). This
framework provides a common base from which state education systems and schools can
develop a curriculum along national guidelines, for years 1 to 12. At the time of writing, thre
frameworks for each area of learning, including English, were still under consideration for
incorporation into state curriculums. The paper divides belt into six sections. Section 1
outlines the topic and scope of the paper; section 2 sets the background and provides some
general criteria for evaluating the grammars under consideration; sections 3, 4, 5 respectively,
examine three approaches to grammar, traditional, functional and structural; section 6
discusses the findings and suggests directions which may best serve students’ interests.1
The content document for the NEC, A statement on English for Australian Schools
(Statement), states that “students should know about the structure and features of written
spoken and visual language and the ways in which the use of English varies according to
situation and social or cultural context” (Australian Council [AEC] 1994:10) Through the
texts2 studied and composed, knowledge about language structure and its use, is to be
formally taught at the point of need. The grammar advanced should have the capacity to
support every relevant facet of the curriculum.
it is a state responsibility to formulate guidelines, from which individual schools can make
so-called choices. However, implementing a new curriculum with considerable focus on
knowledge about language’, requires of English teachers a considerable linguistic knowledge
many of them do not possess (Hannan 1992; Dowse 1988; Piper 1983). Investigating the
background to English language teaching practice in Australia, Piper (1983:9) observes “the
general lack of training in any form of systematic language study of most English teachers”.
Personal communication suggests the current situation is little changed. Until intensive in-
service teacher education becomes a widespread reality, this factor must have an influence on
decisions concerning the implementation of the NEC, both at the school and state level.
a model of language study based primarily on systemic functional grammar, can meet the
demands of the linguistic component 01 the NEC. This paper aims to critically evaluate some
options, which might challenge that notion. Section 2.1 sets out some general criteria for
assessing the grammars under review.
2.1 Criteria determinations for assessing the grammars
The Statement outlines general linguistic requirements of the curriculum. I have identified
those requirements principally associated with grammar, discourse and language variation,
and translated these into criteria by which the grammars under discussion can be assessed for
1 This paper is an expanded and revised version of a paper presented at the 1994 Melbourne University
Postgraduate Linguistics Conference. Special thanks are due to Kate Burridge for her good-humoured
advice and painstaking comments for revision.
2 Text is defined broadly as “any communication, written, spoken or visual, involving language” (AEC 1994:6).
potential use in the school curriculum. I or example, the term ‘grammar’ taken from the NEC,
translates into this paper as the criterion ‘category’ and the criterion ‘function’. The criterion
‘category’ is used for all grammatical classes and other features such as tense and agreement,
and the criterion ‘function’ represents grammatical function; they are listed respectively as
criteria (a) and (b) below, The Statement makes it clear that learning associated with
grammatical knowledge must be linked to language use, including strategies for making texts
cohere and recognition of ways in which language is manipulated. These entail such skills as
reorganisation of the syntactic structure of sentences to produce different thematic meaning
as part of the creation of texture. It follows then, that any framework posited for school
grammar must have the capacity to support features of discourse analysis as listed in criterion
(c) below. Deliberate placement of thematic reordering3 within this criterion extends the
boundaries of sentence grammar into discourse or text analysis. The NEC also requires of
students the ability to “discuss and analyse the linguistic structures and features of texts in
relation to their use” (AEC 1994:12) The grammar must therefore accommodate language
variation. In particular students must learn, “the linguistic structures and features of written
standard English” (AEC 1994:12) and “know about the structure and features of written [and]
spoken ... language and the ways in which the use of English varies according to situation and
social or cultural context” (AEC 1994:10) These requirements are encompassed by criterion
(d) below. The focus will be on elements that might present problems in the school context.
Categories - word classes (parts of speech), phrases, clauses, tense, agreement
Functions - subject, object, complement, determiner, modifier, auxiliary
Discourse analysis - textual cohesion: reference, conjunction, ellipsis
- thematic variation and reordering: voice: active and passive
- extraposition, fronting, cleft construction, nominalisation
Language variation – style - formal, informal, mode spoken, written
- standard Australian English.
Traditional grammar takes the word as the bask unit, and for that reason is considered a word
based grammar. The principle of a word based grammar, according to Blake (1988:8), is that
the important constituents of a language are the words, which join together to form larger
units, such as phrases, clauses and sentences. Depending on their function in higher
groupings, words can be divided into classes, the set of terms known in traditional grammar
as ‘the parts of speech’. They are the elements common to all three grammars under review.
Garner (1989:218) refers to them as ‘a sort of lingua franca used by linguists subscribing to
different theories, when discussing language. Huddleston employs many of these terms in his
account of modern structural grammar, stating “that from a pedagogical point of view the
traditional scheme arguably provides the best starting point for discussion” (1984:98).
Halliday (1985a:28) also sees merit in many of these terms “that have become familiar in
With the move away from traditional concepts of language education in the late 1960s and
early 1970s, traditional grammar was no longer a compulsory unit of the English syllabus of
Australian secondary schools (Piper 1988:7). However, Piper (p.86) finds that although most
3 This term is used by Huddleston (1984); examples appear in 2.1.1 (c)
teachers do not accord grammar the central place it had thirty years ago, there is still reliance
on the same or similar textbooks. This claim is supported by Huddleston’s (1989) review of
English grammar texts currently published and presumably used in Australian schools. This
being so, and taking into account teachers general lack of linguistic training as established
above in section 2, the question is, whether these longstanding experiences with traditional
grammar can be exploited in the implementation of the NEC.
Traditional grammar typically has eight parts of speech, as set out in Figure 1:
FIGURE 1: Traditonal parts of speech
boy, woman, cat, apple, truth
I, he, everyone, nothing, who
he, become, come, die, believe, jump, take
big, happy, old, wooden, some, few
quickly, very, here, afterwards, nevertheless
at, in, on, by, for
and, but, because, although, while
ouch, oh, alas, grr, psst
(based on Huddleston 1984:90)
Traditional grammar: criteria (a) and (b): criticism
A number of criticisms of traditional grammar pertain to the parts of speech and other
categories and functions. Many apply directly to the definitions.
3.1.1 Notional definitions
The traditional approach to defining categories is a notional one. Definitions are typically
based on meaning, that is, semantic properties. It is important to distinguish between these
notional definitions found in most pedagogical ‘school’ grammars, and those definitions
found in more scholarly works of Traditionalists (Long 1961, LaPalombara 1976, Blake
1988, and Garner 1989), that include grammatical properties in their Criteria.
In his survey of school grammar texts, Huddleston (1989), found that notional definitions
predominated for the traditional parts of speech representing word classes, a well as other
grammatical categories, such as tense, phrase and syntactic function (p.4). Cited here are
some typical examples from one of the texts reviewed, Living English (1985). A noun is the
name of a person, animal, place or thing; A verb expresses an action or state”; “The part of a
sentence which tells us who or what does the action is called the subject”; “The complement
... completes the sense of the sentence”; The possessive case shows ownership; Tense means
Criticism of the semantic definitions of traditional grammar is not new. Quirk, Greenbaun,
Leech & Svartvik (1985:74), Garner 1989:24), LaPalombara (1976:54), and Huddleston
(1984, 1989) all consider these notional definitions inadequate. How are they inadequate?
This question will be taken up in 3.1.2, points (i) to (v) below As to why they are inadequate,
the answer is well documented in Huddleston (1994, 1989). Notional definitions are useful
for identifying the ventral members of a class for cross-linguistic purposes, after
classification for each language has been assigned through grammatical properties. School
grammar is not concerned with comparing classes across languages. Its focus is surely on the
language-particular level, for learning about the English language or any other language. And
at the language-particular level notional definitions are not well suited to identifying classes
from particular languages. For example, the notional definition of noun does not include the
many nouns that fall outside the semantic criteria (person, animal, place, thing). nouns such
as acceptance, knowledge, jump as in That jump was a winner.
The inadequacy of the notional definitions is manifested in at least five ways. To illustrate, a
small but significant sample adapted from Huddleston’s (1984: 1989) accounts is set out
3.1.2 Inadequacy of notional definitions
Notional definitions are inadequate for recognizing categories because they fail
to provide clear criteria for assigning words to the class designated for traditional
A common definition offered for a verb is that ‘a verb denotes action or being, or ‘action or a
state of being. Consider the first of these two definitions, ‘action’, in (1). A valid
interpretation must include not only the verb destroyed, but also the noun destruction. Bold
type is used throughout this paper for elements under scrutiny.
they destroyed the city unnecessarily
destruction of the city was unnecessary
The second definition, ‘state of being’ includes the verbs know and like but does not exclude
the adjectives knowledgeable and likeable.
Again, if adjectives are said to describe nouns, what criteria decide that in Joyriders are
idiots the word idiots, which can be said to describe the noun joyriders, is a noun and not an
adjective like idiotic, as in Joyriders are idiotic?
(ii) The definitions can be misleading because the relation between grammatical form
and meaning is rarely one-to-one.
Traditionally, the possessive construction with its marker, the apostrophe, indicates
possession or ownership; yet, by including examples such as a month’s training, or the firm’s
failure which clearly don’t express ownership, the traditional definitions discourage any
investigation of the range of interpretations that can be associated with the possessive
Other examples include: “Tense refers to time” (Winch & Blaxell 1994:27). This, and similar
definitions, categorically correlate the category present tense with present time, which is not
always the case. Although primary past and present tense refer to past and present time
respectively, as an example of secondary use, present tense commonly denotes future time I
go tomorrow (cf. 4.1 iii ) (Huddleston 1989:11).
(iii) The definitions for each category are not all of the same kind.
Some categories, such as noun, are defined independently by their inherent semantic
properties as a class, as in: ‘a noun is the name of a person, place or thing’ (Winch & Blaxell
1994:7), while others, such as adjective, are defined in terms of their semantic relation: ‘an
adjective describes a noun’ or, Bernard (1975/1986:42),4 ‘an adjective modifies or adds to the
4 In this paper the reprint publication date for school texts is significant, because it signifies more current
usage than the earlier edition date. Accordingly, the edition date and the reprint date will be linked by /, for
meaning of nouns’. This type of definition, dependent on a knowledge of other categories, is
also applied to pronouns, adverbs and prepositions, and unless introduced in logical
progression, can lead to circularity that the independent definitions avoid. Bernard
(1975/1986:11) also draws attention to this same lack of consistency in the system of
classifying in traditional grammar.
(iv) The definitions are not all mutually exclusive.
Consider the noun phrase: the woman senator. For this and similar examples Huddleston
(1984:9 3) argues: woman as the name of a person, satisfies both the definition of a noun by
denoting a person, and also the definition of an adjective by modifying senator. With further
analysis, such a applying adverbs of degree, we can say the rather worthy senator but not
*the rather woman senator (rather modifies adjectives but not nouns: and a sequence of two
adjective modifiers allows the worthy new senator but not *the woman new senator. For our
purposes it is sufficient to say that here, woman differs in class from the adjective worthy in
the worthy senator, and both should be assigned to different classes, while still performing
the same function, that of modifier to a noun.
Notional definitions for grammatical functions are generally underdeveloped..
Definitions for subject, object, determiner or auxiliary are either indeterminate, such as for
subject, or nonexistent, as for determiner.
Consider the grammatical function subject. Winch & Blaxell’s definition (1994:138): “The
subject is the subject of the verb” is very indeterminate. Clutterbuck (1989/1990:103) defines
it thus: The subject “tells us who or what performed the action of the verb”, which only
provides part of the semantic function, and is therefore misleading Bernard (1975/1986:59)
has the subject as: “the words which tell us the thing, the topic being discussed.” Huddleston
(1984:58) considers that the standard notional definition “equates it with the ‘topic’ of the
sentence, that is, what the sentence is about.” It does not acknowledge that identification of
the topic is often dependent upon a wider context than the clause or even the sentence,
whereas the subject is not. Consider (2) below, where she stands for Liz and him stands for
(2) She shot him
The topic may vary according to different views of the event. Responding to the question
What did Liz do?, the topic would be she (Liz), which is the subject, but in answer to What
happened to the intruder?, the topic may become the intruder which is the object him. Again,
it is possible to find many sentences whose subject is unlikely to be recognized as topic, or
what the sentence is about’, exemplified in the following:
(3) Nobody likes John
(4) It turned out that she was related to him
(5) There should be at least one student on the committee
(6) The remaining issues they have left till the next meeting
The bolded subjects of the examples above do not convey information on what the sentence is
about. For instance, in (3), it is unlikely that the sentence is about nobody, similarly for the
‘dummy’ subject there in (5); and in (6), rather than they, the topic is surely the fronted element
the remaining issues. In 5.1 .1 it will be shown that independent criteria are more reliable for
identifying all subjects, including ones that are not topic-like, in examples (3) to (6).
3.1.3 Membership of categories
Further criticism applies to membership of the categories. The membership of the parts of
speech in traditional grammar is in many cases too heterogeneous. This is true of the (lass ot
adverbs, prepositions, adjectives and others. Consider the class of adjectives. Some grammars
include in this class, what are known in traditional grammar as articles, a(an) and the. Blake
(1988:14) notes that articles were not considered at all in some earlier grammars because
there was no class of articles in Latin. However, even in some current school grammars, such
as Winch & Blaxell (1994:40), articles appear as “adjectives of a special kind”, which, cannot
be said to describe a noun. Other words such as this, each, some, few, which also appear in
the adjective class along with prototypical ones like big, grumpy, real, are more like the
articles, and as Huddleston (1984:97) suggests, could be assigned to a class of their own.
Concept of phrase
A further important area for criticism concerns the concept of some classes, such as phrase.
The concept of phrase represents a complex area in ‘:o,ial grammar and deserves particular
study because of the difficulties it presents for the school curriculum. For this paper a brief
look at the class phrase will illustrate some of the inexactness in this area, which is continued
on through school texts.
Firstly the definitions for the class phrase are found to be inconsistent: Waldhorn & Zeiger
(1967/1986:46): “A phrase is a group of words which acts as a single part of speech”;
Clutterbuck (1989/90:81): “A phrase is group of words without a verb”; Winch & Blaxell
(1994): “A phrase is a group of words that has no finite verb”. An examination of the validity
of these definitions is the scope of this paper.
Secondly, the classificatory system hound in current school grammar texts is confusing.
Phrases can be classified by two different criteria. The first is by function, that is, phrases
classified according to what is believed to be their functionaal equivalent in the word class, as
exemplified in (7-9).
with large sunglasses visited our school
We often eat our lunch on the school bus
I love eating baked beans
Winch & Blaxell 1 994:49)
The prepositional phrase a with large sunglasses in (7) is traditionally classified as an
adjectival phrase because it functions as modifier to a noun (filmstar), a function
characteristically performed by an adjective. In (8) the prepositional phrase on the school bus
is classified as an adverbial phrase because it is said to function as modi8er to the verb (eat),
a function characteristically performed by an adverb. In (9) the gerund (or more properly the
structure containing it) eating baked beans is classified as a noun phrase because it is said to
function as a noun.
The second criterion for classification is according to ‘form’, a term which follows Waldhorn
& Zeiger (1967/1986:49), who claim that phrases can be classified according to their
“introductory or pivotal word as prepositional, participial, infinitive or gerund”, which “does
not contradict the classification according to use”. By this criterion the prepositional phrase in
(10) is so classified because it begins with a preposition; in (11), the reduced or non-finite
clause of standard modern classification is classified as a participial phrase because it begins
with a participle, and so on with (12) and (13):
(10) Let us eat in silence
(Waldhorn & Zeiger 1967/1986:49)
(11) The boy kicking the football is my nephew
(12) To do is to learn
(13) Reading books enlarges ones horizons
(Waldhorn & Zeiger 1967/1986:49)
Some texts, such as Winch & Blaxell (1994:48), recognise phrases by both form and
function, with function usually taking precedence over form. Others, such as Clutterbuck
(1989/1990:81) use a mixture of the two, marking the so-called verbal phrases of (11-13) by
their form, and the prepositional phrases of (7-10) by their function, It should be noted that
classifying by form bears no relation to the head-based system of analysing phrases, in which
a noun acts as head constituent in a noun phrase, and an adjective heads an adjective phrase.
To further complicate the classificatory system, many of the more scholarly works on
traditional grammar, such as Blake (1988) and LaPalombara (1976), choose to classify a
phrase by its head constituent, thus applying the standard modern approach.
Given that this is a reasonably consistent method, with the exception of prepositional phrases,
this could be a partial solution to the problem. However, there still remains the problem of the
verbal phrases discussed above. Blake (1988:111 f.) himself suggests a solution. Many of
these constructions appear to occupy the same role as that filled by clauses; they can take
objects (reading books) and even subjects (anyone writing graffiti). Therefore Blake
advocates reclassifying them as non-finite clauses even within the framework of traditional
Further problems exist within the class of clause, but the only revision recommended at this
stage is the transfer of verbal phrases to the clause Class as non-finite clauses.
Criteria (a) and (b): assessment
The main recommendation arising from the discussion in 3.1 to meet criteria (a) and (b) is a
revision of the definitions from four aspects. Firstly, the definitions should follow a
consistent set of criteria. Secondly, the criteria should ensure consistent distinction between
syntactic class and syntactic function. Thirdly, a system that largely defines categories
independently of others could be explored. Fourthly, the system must provide definitions that
are clear, have mutual exclusivity, and must be appropriate to the understanding of school
students. Further recommendations involve revision of the classes, such as the class of
adjectives, adverbs and phrases, to establish more homogeneity. This could also entail re-
assignment of some class members to other classes, such as pronouns to noun class and
‘verbal’ phrases to clause class; others, such as articles, would be assigned to new classes, in
this case, the more modern class of determinative. Blake (1988) shows that modifications
founded on traditional features, but amplified by modern terms, are possible. This is apparent,
as we have noted above in 3.1.4, in his more modern structural approach to the class of
phrase. The head-based structure he advocates, relates phrase class to word class, thus
extending the traditional word-based grammar to now include such phrases as the N(oun)
P(hrase) with the functions determiner and modifier, and the V(erb) P(hrase) with modal
auxiliaries, functioning as a modern predicator.
Further criticism arises with discussion of each of the remaining criteria.
Criterion (c): assessment
Incorporating the components of the discourse analysis criterion (c) into the curriculum,
through the medium of traditional grammar, requires an innovative approach, since there is
little historical evidence of any such endeavour. It is generally agreed that traditional grammar
does not reach beyond the sentence (Blake 1988; LaPalombara 1976). However would like to
suggest a basis from which an analysis may evolve. The analysis of texts for school students
may require the introduction of some additional discourse terms such as topic, referent, text,
context, connector. To minimise the size of the metalanguage I envisage a discussion of texts
using a style that is not truly formalised. For example, the various cohesive devices such as
conjunction and reference, could be served by a common term such as connector or link, and by
class names such as pronoun, nouns (related or generic), conjunctions, adverbs and so on, for
individual examples. For younger students, attention could be drawn to cohesion in texts
through the lexical items themselves. Similarly, for more senior students, a simple
metalanguage of discourse can be progressively built for discussing topics such as thematic
reordering, for example ‘fronting’, ‘clefts’ and ‘highlighting’ (cf.5.3.1).
Garner (1989) has speculated on the notion of traditional grammar being applied to discourse
analysis. He suggests a number of potential areas of study: sentence combining through
clause embedding and nominalisation (p.1 94), sentence linking through grammatical devices
such as adverbs however (p.) 96), phrases on the other hand, connectors but, and, adjectives
as connectors the following previous ..., former ..., later ... (p.1 97), reference pronouns -
anaphoric and cataphoric, and ellipsis indicated by pro—adverbs here and ‘pro-verbs’ do
(p.200). Although Garner (1989:200) believes traditional grammars application above the
sentence to be limited, from the instantiation above, there seems to be some potential for
further development towards meeting criterion (c).
3.4 Criterion (d): assessment
The NEC requires students to gain a knowledge of language variation according to user,
particularly written standard Australian English. The standard dialect is described simplistically
as “the language of formal spoken communication, the education system and professional life”
(AEC 1994:lOf.). We are also told it is distinguishable by its grammar. Thus we can identify it
through grammar analysis, and style and mode variation discussed below.
The style of language varies according to use. The speaker/addressee or writer/reader
relationship may produce formal or informal style; it may be in spoken, written or even visual
mode; and the function of the text or (what it is about), will influence the style. For example,
a chemistry report will have a different form from a narrative, with different structural
consequences. One of the main criticisms levelled at traditional grammar has been its
restriction to formal style with its inherent prescriptiveness. Christie et al. (1991:106)
describe this restrictive account as “rules about what not to do”. Gleason (1965), Blake
(1988) and Huddleston (1989) suggest that traditional grammar need not be prescriptive, that
it can be as prescriptive or descriptive as style warrants. A formal written style may reflect
the most prescriptive features of At the other end of the continuum, by not applying certain
rules, the grammar could recognise informal spoken language, traditional grammar at its least
prescriptive. A further criticism raised by Christie (1994:18) traditional grammar’s
preoccupation with the written language mode. However, there is no linguistic impediment to
its use in the analysis of the features of spoken language The phonological features of spoken
language, such as intonation, are not directly within the sphere of traditional grammar, but
other characteristics can be accessed through the grammar, lexis and discourse. These may
encompass such features as clause complexity, noun and verb phrase structure, verb
incidence and ellipsis.
It has been shown that by employing an extended metalanguage, which might include the
terms style, mode, formal, and informal, words understood outside of linguistics, traditional
grammar could, for the most part, fulfil the requirements for the teaching of language
variation. In this way, the study of language variation according to use will cover language
variation according to user, including standard Australian English, thus meeting criterion (d).
The functional grammar which will engage the attention of this paper is the one most cited in
education documents and journals, that is, Halliday’s account (1985a), which he sums up as.
the interpretation of the grammatical patterns in terms of configurations of functions’ (p.x).
Hudson (1986:795) views it as a serious attempt at a fully integrated analysis. The theory
behind it is systemic, that is, a theory of meaning as choice. It moves from general features
though interlocking options: “either his or that, or the other”, or “either more like the one or
more like the other” through to the ever more specific (Halliday 1985a:xiv). In its purpose of
analysing text, the analysis can proceed as far as is needed or as capabilities allow Halliday’s
system, which has attracted educationalists, particularly in Australia and and in Britain,
manifests a number of anomalies, many of which have been recounted, notably by Hudson
(1986) and Huddleston (1988). Discussed here, are some that relate to the NEC, and which
are likely to cause confusion in the teaching and learning of functional grammar in schools.
4.1 Functional grammar: criteria (a) and (b): criticism
Halliday’s account suffers from ill-defined categories.
Hudson (1 986:799) offers the following key categories as a sample:
‘subject’: the element that is held responsible, in which is vested the success of the
clause in whatever is its particular speech function (p.36); something by
reference to which the proposition can be confirmed or denied ip.76).
‘given’: information that is presented by the speaker as recoverable (p.277); what you
the listener, already know about or have accessible to you (p.278).
‘reference’: something which can be taken as a reference point for something which
Throughout the analysis similar instances abound: “a Complement5 is an element within the
Residue that has the potential of being Subject but is not” (p 79). This interpretation will be
discussed below in (iv-b).
Such vague notional semantic descriptions for functional and relational terms would be
difficult to consistently apply to text analysis, from a pedagogical approach.
(ii) Halliday (1985a) provides debatable analyses with little or no justification.
One example is the category noun or nominal which contains the following word classes -
common noun, adjective, numeral, and determiner. Participles, such as the present participle
in a losing battle and the past participle in a lost cause are also included in this category
(p.164). Thus, as Huddleston (1988:144) points out, in this analysis the predicative
complements brilliant and a genius of the sentences she is very brilliant and she is genius are
both assigned to the nominal group class rather than to the traditional classes, adjectival and
nominal, respectively. The constitution of the nominal group and the lack of an adjective
group altogether, are difficult to explain or understand, particularly at school level.
5 Following Halliday, an initial capital will be used to indicate functions in section 4.
However, Halliday’s (1985a:188f.) concept of word class groups can be justified, to some
extent. The main word groups in functional grammar, are nominal, verbal, adverbial, and
prepositional. He has replaced the head-based phrases of formal grammars with these word
groups, with the addition of the prepositional group to account for complex prepositions. In the
prepositional group (p.188f.) the complex prepositions such as instead of, far as, roughly
correspond to the word extensions of the other groups, such as the verb group will have arrived,
or the noun group those very last trains, which are extensions of their respective heads arrived,
and trains. This allows the creation of a quite separate unit, the P(repositional) P(hrase), which
is non-headed,6 as Quirk, et al. (1985:60f.) and Newby (1987:49) attest. However even
Halliday (1985a:189) admits that the line between prepositional group and PP becomes fuzzy
when considering those complex prepositions, such as in front of, which he claims, have
evolved from the PP. Halliday’s justification for positing a PP rests on his treatment of the
preposition as a minor verb, thus creating a link between PP’s of the type regarding the
proposal, concerning the matter, and the non-finite clause, which he also claims, overlap in
classification. Nevertheless, whatever the analysis, it serves to distinguish the PP from the
headed phrases of modern standard analysis. This area of analysis might prove productive for
the logical treatment of the phrase category of English grammar for the school curriculum
Further criticism concerns speech roles. Halliday (1985a:69) recognizes two fundamental types,
described by Hudson (1986:797) as related on two dimensions: (i) giving versus demanding,
and (ii) goods-and-services versus information. Together these define four primary speech
functions OFFER, COMMAND, STATEMENT, and QUESTION. Halliday classifies OFFER and
QUESTION as maximally different, that is, they differ on both dimensions, and yet in contrasting
them with other types, they are both realized, as interrogative clauses, shown in (14-15):
(14) Would you like this teapot? OFFER Giving
(15) What is he giving her?
The matter is further complicated: the semantic function of a clause in the exchange of
information is a PROPOSITION which refers to a statement or question; the semantic function
of a clause in the exchange of goods and services is a PROPOSAL” (Halliday 1985a:71). Thus,
by both arguments, an OFFER, which cannot also be a QUESTION, and is not a PROPOSITION,
would appear to contravene the analysis if it is realized as an interrogative. Although
Halliday believes that propositions have a clearly defined grammar, while proposals do not,
the possible implications are not always obvious, as in the following case cited. Christie,
Martin & Rothery (1991:73), in their proposed unit on teaching functional grammar for pre-
service preparation of teachers, claim that the grammar of English “codes these different roles
as a set of choices with specific structural consequences. They give as examples, imperative
for structures that demand goods-and-services, declarative for structures that give
information, and interrogative for structures that ask for information. Although Christie et al.
concede that grammatical metaphor7 can extend the choice of structure available for these
speech roles, nevertheless, an offer giving goods-and-services, which can be realized by all
three structures mentioned, as in (16), would seem to challenge the notion of ‘specific
6 The prepositional phrase, which contains two obligatory elements (Preposition + NP), is called nonheaded, for
example, in the race, *the race, *in, whereas headed phrases have only one obligatory element, the head, such
as the noun race in the NP the fast race/fast race/fast (Quirk, Greenbaum, Leech $ Svartvik 1985:60).
7 See 4.3 (A)1 for an explanation of grammatical metaphor.