An analysis of spoken grammar:

the case for production

Simon Mumford

Corpus-based grammars, notably Cambridge Grammar of English, give explicit

information on the forms and use of native-speaker grammar, including
spoken grammar. Native-speaker norms as a necessary goal in language
teaching are contested by supporters of English as a Lingua Franca (ELF);
however, this article argues for the inclusion of selected forms for teaching for
production based on an analysis of the usefulness of individual forms. The
forms are analysed in two sections, relating to fluency and appropriacy, since,
while every student can benefit from improved fluency, native-speaker
appropriacy may not be a need for all. The conclusion is that such an analysis
strengthens the arguments for teaching many of these forms for production,
while acknowledging the case for fluency features is stronger than more
context-dependent appropriacy forms. It briefly looks at some possibilities for
teaching the forms.

Introduction Advances in technology have brought a more detailed analysis of native-

speaker language, which has been used as the basis of teaching material,
and we have already seen the publication of course books and dictionaries
informed by corpora. Attention has now been focused on the (native
speaker) Spoken Grammar of English (hereafter SGE). This can be defined
as those aspects of English which are almost always associated with the
spoken language or its written representation, as recorded in new corpus-
based grammars, notably Cambridge Grammar of English (Carter and
McCarthy 2006) (hereafter C GE) and Longman Grammar of Spoken and
Written English (Biber et al. 2000). This article examines selected SGE forms
in C GE and related writings by Carter (1998, 2007) in order to analyse their
role in native-speaker speech, and to attempt to establish what the potential
benefits to learners of using such forms may be.
The analysis and conclusions drawn are set against the significant debate in
English Language Teaching (E LT) over whether to use native-speaker norms
regarding grammar, or to use some other standard. In order to give some
background to the debate, I briefly look at three approaches to native-
speaker spoken forms as a model for learners. The first is that of supporters
of World Englishes/English as a Lingua Franca (EL F), which denies any
need for specifically native-speaker norms. Another position advocates
a passive knowledge of these forms for students. Finally, an argument for

E LT Journal Volume 63/2 April 2009; doi:10.1093/elt/ccn020 137

ª The Author 2008. Published by Oxford University Press; all rights reserved.
Advance Access publication May 22, 2008
 teaching native-speaker forms for productive use based on student need and
interest is set out.
The last position, as expressed by Kuo (2006), will be appropriate, this
article will argue, for at least some forms of S GE, because C G E has now
isolated and described the forms, and enabled us to understand their role in
speech and, therefore, the advantages for learners. We can now pass this
information on to learners and help them use the forms, potentially
benefiting them in two main ways: firstly, in making their language more
fluent and, secondly, more appropriate. This distinction is used as the
criterion both for the analysis and for the basis of selection of forms for
teaching. The rationale for this approach is that while it can be argued that
native-speaker appropriacy is more applicable to students dealing with
native speakers, fluency can be considered a more general need, desirable in
any situation.

Three approaches to The basis of the ELF argument is the fact that many students will never
the native-speaker come into contact with native speakers. Supporters of this view, for example
model Rajagopalan (2004: 114), argue that in an age when native speakers are no
The World Englishes/ longer the largest group of English speakers, using their norms in ELT gives
E L F approach them unwarranted status and precedence, and thus should be avoided.
These writers argue for an English based on ‘intelligibility’, where the
learners’ aim is to speak in a way that can be understood, rather than
precisely follow native-speaker norms, with the result that minor
grammatical inaccuracies and some variation in pronunciation are accepted
as long as mutual intelligibility is maintained.

The passive Other writers have suggested that a native-speaker model would in fact be
knowledge approach useful to learners. One approach that has been suggested includes raising
students’ awareness through listening to recordings of native speakers, and
using scripts and exercises to help them notice features of SGE. Such
a model has been put forward by Timmis (2005: 118), whose research leads
him to conclude that ‘a significant number of learners and teachers’ want to
sound like native speakers as far as possible, including the use of ‘informal
grammar’, and he notes that many others will at least want to understand
native-speaker spoken language conventions. Despite this, however, he
stops short of recommending teaching for production, pointing to lack of
rules for use, stating that ‘it is at least questionable whether we want learners
to produce these forms at any stage’ (ibid.: 120).

The production However, another view proposes that there is a need to go beyond this
approach passive knowledge approach, and teach native-speaker norms for
production. Kuo’s (2006) research reveals that many of her students see
native speakers as a desirable model and that, in an increasingly competitive
world, merely being ‘intelligible’, being able to make themselves
understood is, in itself, insufficient. Where learners are in contact and in
competition with native speakers, lack of practice in native-speaker spoken
norms will lead to a distinct disadvantage, for example, in showing the
appropriate level of formality and politeness. Thus, for many students,
especially those in, or planning to travel to, English-speaking countries or
those who work with or meet native speakers, learning SGE will bring

138 Simon Mumford

 benefits. Furthermore, Kuo (2007) points out that these forms will help
learners communicate in a range of situations where native speakers may or
may not be present.

Analysis of the The potential benefits that the students of Kuo and Timmis and many
features of S GE others feel they would gain from native-speaker grammar seem not to have
been precisely defined or comprehensively analysed so far. The purpose of
this article is to understand the role of SGE and how these forms could help
learners. The following analysis of C G E reveals how British speakers use
certain forms to communicate both fluently and appropriately, and as these
two areas, fluency and appropriacy, seem to be the main purposes of SGE,
they are used as criteria for the analysis, which looks at selected features that
have a clear relation to learner language and implications for teaching/
learning and testing.

Analysis of SGE The forms most likely to be useful to students consist of the following:
fluency features phrasal chains; simple sentence structure; non-canonical use of some
singular and countable/uncountable forms; ellipsis of subjects and
auxiliaries; use of declaratives as questions; flexible word order, including
headers and tails and fronting of objects; use of lexical chunks, fillers, and
placeholders.
n Carter and McCarthy (2006: 168) note that in real-time speech
‘utterances are linked . . . as if in a chain’ rather than built into sentences.
Thus, unless students can learn to speak in phrasal chains, they will be
under a double disadvantage, as they will not only have fewer language
resources than more fluent speakers but will also be setting themselves
the more difficult goal of speaking in sentences. Practice in speaking in
phrases rather than sentences could help students produce a greater
volume of language, and language which sounds more fluent.
n Carter and McCarthy (ibid.: 170) note that native speakers tend to use
coordinating conjunctions (‘and’ and ‘but’) and simple subordinating
conjunctions (‘so’ and ‘because’) in real-time communication, and this
is the kind of language that students should aim at when producing
unplanned speech in situations like oral exams. The combination of
simple conjunctions and phrasal chains in native-speaker speech
suggests that the traditional complex (written) sentence is not natural
in spoken language and therefore not a suitable standard to judge
students’ speech by.
n Native speakers sometimes use ‘ungrammatical’ structures, they are
more likely to say, for example, ‘There’s your pills’ than ‘There are your
pills’ in informal situations when indicating location (ibid.: 95). ‘Less
children’ (instead of ‘fewer’) is another such form produced by native
speakers (ibid.: 103). By eliminating the need to choose between different
singular and plural, and countable and uncountable forms, the cognitive
load is lightened and fluency made easier. If we as teachers insist that
students conform to ‘written grammar’ norms while speaking, we may
again be making their task more difficult.
n Some elliptical forms used by native speakers resemble learners
‘incorrect’ forms. If a learner produces ‘You like ice-cream?’ or ‘Like
ice-cream?’, he or she may be corrected, and prompted to say ‘Would/Do

An analysis of spoken grammar 139

 you like ice-cream?’ However, native speakers frequently omit pronouns
and auxiliaries in informal interrogatives (ibid.: 182). Students may, in
fact, know the full forms, but, like native speakers, under pressure of real-
time conversation, take short cuts, which do not interfere with meaning
since it is clear from the context.
n Declaratives are used for questions by native speakers (ibid.: 533), which,
again, could lead to the phrase ‘You like ice-cream?’ being used as
a question. As mentioned before, learners often use this as a question
form. If teachers fail to consider and allow for this, students will again be
operating under the same double disadvantage of fewer language
resources, yet a harder task in constructing sentences that are
‘grammatically correct’ according to the written language.
n Word order is more flexible in spoken language, since speech is
constructed in real time and follows the order of ideas emerging from
a speaker, which may override grammar rules (ibid.: 172). Learners’
‘incorrect’ word order may, of course, show L1 interference or wrong
learning, but it could equally reflect this S GE feature, and thus, again,
language teachers may demand standards of spoken language that native
speakers frequently fail to achieve. Students may feel less anxious about
word order in speech if they understand the cause of these natural
inconsistencies.
n The use of headers and tails (ibid.: 194), and fronting (ibid.: 779), can also
alter the word order of traditional written grammar, again showing that
spoken language is much more flexible than written forms are. Using
headers may come naturally to some students, because it is often simpler
to mention the topic first and then add a comment, as in the following
example: ‘That old lady, I really like her’. The fronting of objects in
sentences is likely to appeal to students where objects precede other
sentence elements in their native language. Students may be reassured to
know these spoken forms are not necessarily wrong.
n Language is repetitive and fluency is increased by the ability to use chunks
such as ‘I see’, ‘I think’, ‘you know’, ‘kind of ’ (ibid.: 828). Drilling may be
a solution which would help to make these terms familiar and natural
(Carter 2007: 44). Ironically, such mechanical exercises may contribute to
fluency, if, as a result, students can use the chunks effectively and
automatically.
n Pausing and repeating are common, especially at the beginning of
utterances (Carter and McCarthy 2006: 172). Fillers and repetition are
natural features of language and could be encouraged, especially as an
alternative to silence. Students could be shown that the function of a filled
pause (‘err’ and ‘um’) is to draw attention to the thought that the speaker is
giving to choosing the right words (ibid.: 173). This device is extremely
common in the corpus; the fillers ‘um’ and ‘err’ represent the sixteenth
and seventeenth most common word forms, respectively, in native-
speaker speech (ibid.: 12). Like other word forms, these will need
practice in order for them to become automatic.
n Native speakers use words such as ‘thingy’ and ‘thingamajig’ to refer
to items that they cannot think of words for (ibid.: 149). This type of
vagueness has been labelled ‘forced vagueness’, since is it used out of
necessity rather than having a communicative purpose (Trappes-Lomax
2007: 122). Carter and McCarthy (2006: 147–9) note that these

140 Simon Mumford

 placeholders are usually to describe items that are actually present.
However, learners could be trained to make use of this feature to
substitute for unknown words, whether the referents are present or not.
This seems preferable to the alternative, which is to define the unnamed
item, for example: ‘a thing for opening a bottle with’. Such complex
grammatical structures are difficult to produce in real time and are likely
to reduce fluency.

Analysis of SGE Selected appropriacy forms, more likely to be of benefit to students dealing
appropriacy features directly with native speakers, consist of Vague Language, two-step
questions, contracted forms of the verb ‘will’, native-speaker chunks such as
‘you know’, and a specific use of ellipsis.
n One objection to S G E is that it is associated with excessive informality, but
this is not always the case. Vague Language, for example, is the mark of
a skilled user, not an overly relaxed or informal one, according to Carter
and McCarthy (op.cit.: 202), and they point out that this language is not
linked so much to formality as shared knowledge and group
membership. Another function of Vague Language is making the
message less direct and, therefore, its absence may result in language that
sounds more domineering than the speaker intends. (Carter and
McCarthy: op.cit.).
n Two-step questions emphasize indirectness (ibid.: 201), and therefore,
presumably, politeness, and are thus important for learners in or going to
target language countries. This is likely to be especially true of requests,
where the directness can be reduced with a pre-question. As an example,
C G E cites ‘Are you going to the match tonight?’ as a pre-question to
‘Do you mind if I tag along?’
n Carter and McCarthy (ibid.: 632) note that ‘will’ and the contraction ‘‘ll’
may now be recognized as two separate forms, and the implication is that
learners who use the full form when the contraction would normally be
used risk sounding more authoritarian than they may mean to be. The
Cambridge International Corpus shows that the contracted form is much
more common, and can now be regarded as the unmarked spoken form.
These distinct forms may now need to be taught as separate items,
assuming students are in contact with native speakers.
n Carter (2007: 43, 44) points out that certain chunks, for example, ‘sort of’,
‘you know’ mark native speakers, and those who wish for native-like
proficiency should learn these. He notes that those who think they do not
need them or are unable to use them may not be able to represent
themselves in the way they would like when interacting with native
speakers.
n Carter (1998: 49) has previously revealed that, rather than being impolite
or casual, ellipsis is actually more appropriate than full forms in certain
situations, giving the example of service encounters where time is limited
and full forms would cause unnecessary delay, and thus, irritation among
those waiting in queues behind.
The case for teaching Even though there would seem to be clear benefits to students in
SGE for production understanding and applying native-speaker fluency forms, there still
The case for teaching remains the objection that by using native speakers as a model, students are
fluency features being forced to adopt a position that compromises their integrity, since their

An analysis of spoken grammar 141

 learning is based on a model that is not relevant to them, whereas the EL F
model, it is claimed, suits their needs.
Supporters of ELF allow, and even seem to encourage ‘. . . bypassing
redundancy . . . with a view to maintaining intelligibility’ (Alptekin 2007:
267). ‘Bypassing redundancy’ could be said to reflect what native speakers
themselves do all the time while speaking, in regard to omitting auxiliaries,
and using non-standard grammatical forms in noun–verb concord, for
example. Spoken language, whether of the native speaker or language
learner will inevitably involve deviation from ‘written grammar’; however,
the form this deviation takes may be important, since the need to
communicate involves not only intelligibility but also efficiency and
economy, i.e. fluency. Much native-speaker SGE exists for just this reason, to
transmit the message as quickly and efficiently as possible.
When faced with this need to deviate from written grammar, students may
benefit from a model for guidance. Supporters of EL F seem not to have
produced a specific model, and Maley (2006: 5) questions, even if they had,
whether students would accept a ‘reduced version’ of English. Kuo (2006)
notes that native-speaker forms will be the preferred starting point of many
students, since they represent a fuller and more complete view of the
language than any learner language corpus. This is especially true now C GE
has made these forms easily accessible and shows exactly what role they
play.
Learners already benefit from skills training in many areas of English,
training which aims to give them strategies similar to those used by native
speakers, since these are agreed to be the most efficient. There seems to be
no good reason why the attitude to training students for efficiency in
speaking should be any different to other skills, now we know what these
strategies are. One example is ellipsis, which in the case of written language,
as is well known, has definite rules that learners need to know if they are to
write fluently. As can now be seen, ellipsis in speaking also has clearly
defined rules. Students learning these could become better speakers, just as
they become better writers by learning written grammar rules, and thus the
relevancy to all learners can be demonstrated.

The case for teaching Fluency aspects of SGE may prove more generally acceptable to learners
appropriacy features than those of appropriacy since they appeal more to learners’ needs for
flexibility when dealing with the spoken language. Appropriacy features like
Vague Language are likely to be more controversial, since they represent
forms associated with one particular group of native speakers, and thus, it is
claimed, learners who use them will be adopting a false identity.
This argument is again addressed by Kuo (2007: 270), who points out that
for students wishing to take part in real international communication, the
benefits of learning such language include the ‘capacity to adapt to any given
context’, in other words, to be able to use native-speaker forms when and
where appropriate, and she asserts the students’ right to learn the forms and
choose whether to use them or not. In regard to ‘Vague Language’, Cutting
(2007: 240) also proposes raising students’ awareness, and then allowing
them to ‘opt-in’ or ‘opt-out’.

142 Simon Mumford

 Even with appropriacy, there is a case for looking at features on an individual
basis when deciding whether to teach them or not. Evison et al. (2007: 154)
note that some Vague Category Markers relate to the ‘shared knowledge of
all mature, aware human beings’, while others are ‘more locally constrained
and culture-bound’. Thus, the principle of general as opposed to context-
dependent usefulness to learners can be applied within the category of
appropriacy features itself as well as being a major general distinguishing
factor between fluency and appropriacy forms.
Teaching S GE forms for production, needless to say, should be done
sensitively, especially regarding appropriacy features, and should, as far as
possible, take into account learners’ needs and where possible, be negotiated
with the students.

Some teaching It is beyond the scope of this article to put forward a specific programme or
suggestions methodology for teaching S GE. In very general terms, however,
a programme as suggested by Timmis (2005: 199), including noticing tasks
(see also Willis and Willis 2007: 142) could be combined with activities
aimed at helping students produce the language noticed. Current activities,
such as role play, speaking and listening activities, and games can be
adapted to SGE teaching (for some practical ideas, see Mumford 2007).
Such a programme could either be used in a short stand-alone course for
those who were interested in rapidly acquiring the forms, or integrated into
a wider English course. The long-term development of such material will
most likely result from methodologists’ interpretations of researchers’
further findings.

Conclusion Cambridge Grammar of English is a significant contribution to our

knowledge of English because of size of the corpus used and its detailed
analysis. We can now see that S GE consists of a range of different structural
and lexical items with a variety of functions. As it is purely descriptive, it
is not the purpose of C G E to give an indication of whether, which features
of, or how this new information should be taught. However, it gives new
impetus to the productive teaching of SGE by collating and codifying it,
reducing the force of the objection raised by Timmis (op.cit.: 120) that it is
not easy to draw up rules of use for these items.
CGE tells us much about native-speaker spoken grammar, both appropriacy
and fluency. It is the latter which is most likely to benefit students who need
to speak fluently in English. Without knowledge of and ability to use
features such as ellipsis, phrasal chaining, placeholders, headers and tails,
and fronting, native speakers would find communication a great deal
harder. This is the difficulty learners face if denied knowledge of, or
prohibited from, using such forms.
Appropriacy forms are likely to be more context-dependent and therefore
more controversial. However, it may be more profitable to examine all S GE
forms on an individual basis and assess their usefulness to learners, taking
into account their needs, when deciding what to teach for production, rather
than seeing native-speaker norms as a single unified body. When it comes
to classifying SGE forms, the fluency/appropriacy distinction and the

An analysis of spoken grammar 143

 concepts of the general and the culturally specific would seem to be
useful.
Final revised version received January 2008

References Maley, A. 2006. ‘Questions of English’. English

Alptekin, C. 2007. ‘Teaching E L F as a language in its Teaching Professional 46: 4–6.
own right: communication or prescriptivism’. E LT Mumford, S. 2007. ‘The grammar of spoken
Journal 61/3: 267–8. English’. Modern English Teacher 16/4: 27–30.
Biber, D., S. Johansson, G. Leech, S. Conrad, and Rajagopalan, K. 2004. ‘The concept of World English
E. Finegan. 2000. Longman Grammar of Spoken and and its implications for E LT ’. E LT Journal 58/2:
Written English. Harlow, UK: Pearson Education Ltd. 111–7.
Carter, R. 1998. ‘Orders of reality: C AN CODE, Timmis, I. 2005. ‘Towards a framework for teaching
communication, and culture’. E LT Journal 52/1: spoken grammar’. E LT Journal 59/2: 117–25.
43–5. Trappes-Lomax, H. 2007. ‘Vague Language as
Carter, R. 2007. ‘What is advanced level vocabulary? a means of self-protective avoidance: tension
The case of chunks and clusters’ in P. Davidson, management in conference talks’ in J. Cutting (ed.).
C. Coombe, D. Lloyd, and D. Palfreyman (eds.). Vague Language Explored. Basingstoke, UK: Palgrave
Teaching and Learning Vocabulary in Another Macmillan.
Language. Dubai: T ES O L Arabia. Willis, D. and J. Willis. 2007. Doing Task-based
Carter, R. and M. McCarthy. 2006. Cambridge Teaching. Oxford: Oxford University Press.
Grammar of English. Cambridge: Cambridge
University Press. The author
Cutting, J. (ed.). 2007. Vague Language Explored. Simon Mumford has taught for 20 years in Izmir,
Basingstoke, UK: Palgrave Macmillan. Turkey, and in the UK. He currently teaches
Evison, J., M. McCarthy, and A. O’Keeffe. 2007. academic reading and writing, and works in the
‘Looking out for love and all the rest of it: Vague Academic Writing Centre at the Izmir University of
Category Markers as shared social space’ in J. Cutting Economics. His professional interests include
(ed.). Vague Language Explored. Basingstoke, UK: classroom activity design and he edits the Practical
Palgrave Macmillan. Teaching Ideas column in IATEF L Voices. He
Kuo, I. V. 2006. ‘Addressing the issue of teaching organizes a voluntary workshop programme for the
English as a Lingua Franca’. ELT Journal 60/3: British Council in Izmir. He holds the Aston
213–21. University MSc. T ES O L.
Kuo, I. V. 2007. ‘A response to Cem Alptekin’. E LT Email: [email protected]
Journal 61/3: 269–71.

144 Simon Mumford