Syntactic variation and spoken language
Jenny Cheshire
Queen Mary, University of London
Introduction
Variationists and generativists share the goal of understanding the structure of
linguistic systems, so it should be possible, as Chomsky (1999:34) has recently
argued, for 'internal' and 'external' investigations to be mutually supportive rather than
conflicting. For this to happen, however, researchers in each of these fields need to
agree on the nature of the phenomena they are attempting to explain. In the case of
syntactic variation this is not as simple a matter as it may at first appear, because there
is a fundamental problem in identifying the syntactic forms that in variation with each
other. The problem stems from the nature of spoken language, which so far neither
field of research has taken seriously into account.
Generative theory, of course, has never needed to consider the structure of
spoken language. The theory aims to account for our innate knowledge of language
structure and the appropriate data for the development of the theory can derive from
intuitions. Researchers working in the variationist tradition, on the other hand, use the
data of ‘externalised’ language – usually working with audio-recordings of
spontaneous spoken language – but the methodology requires them to focus on one
linguistic variable at a time, with tokens extracted from the conversational contexts in
which they occurred. Here too, then, the structural organization of spoken language
has been ignored. Neither group of researchers has questioned the analytical
frameworks within which syntactic structure is conventionally analysed. If we are to
combine the insights of the two approaches, however, we need to consider the extent
to which these frameworks can apply to spoken, externalised language. The analysis
of spoken language is a relatively recent phenomenon, made possible by the
availability of electronic corpora of transcribed speech. Some linguists assume there is
a shared common core between spoken, written and – presumably – introspective
language (see, for example, Leech 2000), but others have shown how the syntax of
spoken language differs from our intuitive understanding (see, for English, Biber et al
1994; for English and Russian, Miller and Weinert 1998; and for French, BlancheBenveniste 1997). It has only recently been recognised that the structure of spoken
language may reflect interactional factors such as conversational management (see,
for example, the research reported in Ochs et al 1996, and Couper-Kuhlen and Selting
2001), as well as general social principles for co-operative interaction such as those
shown in politeness strategies. Although these social principles lie outside the
organization of language they can have a pervasive effect on its structure (Levinson
1983: 97).
In this paper I will use three features of present-day spoken English to
illustrate some issues that need to be addressed when analysing syntactic variation in
spoken language. The issues come to the fore when trying to use the (socio)linguistic
variable to analyse syntactic variation. I have previously argued (Cheshire 1987) that
it is inappropriate to analyse syntactic variation in this way, and the problems I
describe here confirm my view. In this case, however, the issues derive specifically
from the nature of spoken language, and it is that I will focus on here. I hope that the
discussion will have some relevance for the topic of this volume, as we consider what
2
new insights might be achieved by integrating social and biological approaches to the
analysis of syntactic variation.
Pronoun tags
Pronoun tags occur in several northern varieties of British English. Examples are
indicated by the arrows in (1) and (2) below. The extracts come from an analysis of
variation in the speech of 14-15 year old working-class and middle-class adolescents
in Hull, England (Cheshire, Kerswill and Williams 1999). In our data these pronoun
tags are used only by the working class adolescents. There is social variation here,
then, and the tags seem to be candidates for a variationist analysis.
(1) a. Charlie: the only time I drink is like at parties or =
b. Matt:
= yeah.. not one of the things
you do every day really is it..daft
c. Charlie:
don’t like smoking or anything like that ..no that’s disgusting
→d. Matt:
I used to me..well I tried it
→e. Charlie: I haven’t even tried it me
f. Matt:
my mam wouldn’t say nowt
g. AW:
do your parents smoke?
h. Charlie:
my mam does
i. Matt:
all of them do..got my real dad my step dad and my mam
→j. Charlie:
I don’t like it me
(2) a. AW:
right what about a favourite singer then?
3
→b. Kay:
Peter André me
c. Ruth:
Peter André’s allright but
→d. Kay:
he’s got a real nice chest him
e. AW:
has he? Is it hairy?
f. Kay:
no it’s real brown and greasy
g. Ruth:
cos he has baby oil smothered on him
The well-known definition of a linguistic variable sees it as a structural unit
with two or more variants involved in co-variation with social variables. Crucially, the
variants are considered to be semantically equivalent: in other words, they are
alternative ways of 'saying the same thing' (Chambers and Trudgill 1998: 50).
Semantic equivalence can be established easily for phonological variables, where the
form-meaning relationship is at its most arbitrary, but there has been much
controversy about whether it can also be established for syntactic variation. The issues
were much discussed during the 1970s and 1980s (see, for example, Lavandera 1978,
Cheshire 1987, Levinson 1988, Romaine 1980, Weiner and Labov 1983), and debate
has continued since then (see, with reference to French, Blanche-Benveniste 1997,
Coveney 1997, Gadet 1997; and for general discussions Cheshire, Kerswill and
Williams in press, Cornips and Corrigan in press, Coveney 2002, Milroy and Gordon
2003). A tacit consensus seems to be that the condition of strict semantic equivalence
can be relaxed for syntactic variables, so that a variable can be set up on the basis of
an equivalence in discourse function (Dines 1980, Coupland 1983). The first stage in
a variationist analysis of pronoun tags, then, involves determining their discourse
function and deciding whether forms such as I haven't even tried it me (in 1e) ‘mean
the same' as the corresponding form without the tag (here I haven't even tried it).
4
In 1d, 1e, 1j and 2d the tags are co-referential with the subject pronoun
in the preceding clause. They are unnecessary, in the sense that the clauses are
perfectly grammatical without the tag, so it seems reasonable to say that their function
is simply to foreground the subject. This fits with previous research describing these
tags as emphatic pronouns (Macaulay 1989) that highlight or express intensity
(Macaulay 1991). It also fits with Biber et al’s description of the function of a similar
form, demonstrative pronoun tags (e.g. it was a good book this), which are also
considered to add emphasis, though this time to the entire proposition (Biber et al
1999: 958). Tags are typical of spoken language, so the fact that they have an
emphasising function is unsurprising, given the prevalence of emotive expressions in
conversation, as Biber et al point out.
Emphasis is often appealed to as the discourse function of a syntactic variant.
Multiple negation in English, to give just one example, is sometimes seen as emphatic
(so that I don't want nothing is a more forceful form of negation than the
corresponding I don't want anything). Emphasis does not affect truth conditions, so
there is no problem in considering the clauses in which pronoun tags occur as
semantically equivalent to the corresponding clauses without tags. If emphasis is the
function of pronoun tags we do not need to relax the conditions that apply when using
the variable to analyse phonological variation.
Unfortunately, however, this does not lead to a satisfactory analysis. Although
emphasis is frequently appealed to when analysing syntactic variation, it is an illdefined concept, lacking theoretical rigour. It does not provide a basis for predicting
which variant will be preferred on any one occasion, nor even where it is possible for
variation to occur (Schwenter in press); and to say that emphasis highlights a
particular entity in the discourse – in this case, the co-referential subject pronoun – is
5
too subjective to be useful if we are looking for language universals (Myhill 1992:3).
Sells et al (1996:174) point out that unless the status of emphasis can be clearly
specified in the grammar, along with the extent to which it can affect the form and
function of different linguistic phenomena, the very ubiquity of appeals to this type of
affective meaning may reduce its analytic value. The concepts of ‘intensity’ and
‘emotive expression’ are equally vague and subjective.
Furthermore, the forms of spoken language typically accomplish a range of
functions simultaneously, so it is important to consider the potential influence of
interactional and other factors on the use of the tags. In the Hull data the tags
sometimes occur with a form referring to an entity that becomes a conversational
topic: in 2b, for example, Kay’s Peter André is picked up by Ruth in the next turn.
Similarly, in 2e Ann Williams picks up the topic of Peter André's nice chest, proposed
by Kay in the preceding turn. In both these examples, then, the tag seems to have a
function in conversational management. Note that in 2e the fact that the tag is coreferential with the subject pronoun is irrelevant in terms of its interactional function:
it is the entire proposition (Peter André and his chest) that is proposed as a topic.
Elsewhere in the data the tags sometimes mark an explicit contrast between the
content of the utterance in which they occur and the content of the previous turn. This
is the case in 1d. Matt and Charlie are discussing smoking with the interviewer, Ann
Williams. Charlie is a keen anti-smoker, and his first utterance about smoking, in 1c,
makes it clear that he does not like it (don’t like smoking or anything like that…no
that’s disgusting…). Matt, with his I used to me in 1d appears to infer,
understandably, that Charlie does not smoke – or perhaps he already knows this to be
true. He claims, in contrast, to have smoked himself. Charlie’s I haven’t even tried it
me, in 1e, then emphatically contrasts his own lack of experience with Matt’s, and in
6
1j he repeats the assertion he made in 1c, this time contrasting his own dislike of
smoking with the behaviour of Matt’s family, which has been described in the
preceding two turns. In Charlie’s turns he also expresses his stance towards the
content of his clause; the tags may draw attention to this.
The emphatic and interactional functions are presumably inter-related. Perhaps
speakers use emphasis to highlight certain interactional functions and in that way to
facilitate conversational management. In any event, to see the function of the tags as
providing emphasis not only lacks theoretical rigour but also precludes analysing in
sufficient depth their discourse functions. It is not clear, in other words, what we
would learn from a variationist analysis: we would certainly not advance our
understanding of how speakers use the tags in their spoken discourse.
So far I have been assuming that pronoun tags can illustrate some relevant
issues involved in the analysis of syntactic variation, but in fact these forms may be
more appropriately analysed as an example of pragmatic variation. As mentioned
above, in the Hull data the pronoun tags are used exclusively by the working-class
adolescents. For a sociolinguist this raises the question of how speakers who use the
tags rarely or never perform the discourse functions served by the tags. Macaulay’s
research in Ayrshire (1989, 1991) is of interest here. Macaulay found that the lower
class speakers in his Ayrshire study used constructions that brought personal
pronouns into prominence in a way not found in the middle class data (1991:81).
These constructions included not only pronoun tags but also left dislocation, NPfronting, and it –clefts, all of which were used more frequently by the lower class
speakers to express intensity. The middle class speakers, on the other hand, conveyed
intensity through the use of adverbials. This previously unsuspected finding has
interesting sociolinguistic implications, but it underlines the difficulty of using the
7
linguistic variable for this type of feature (and Macaulay did not attempt to do so).
The variable may well be a heuristic construct that does not necessarily map directly
onto the units of linguistic structure (Wolfram 1993) but to include left dislocation,
say, and it-clefts in the same analytic unit as adverbials is surely stretching the
concept of the variable beyond all credibility. Analysing pronoun tags as a variable,
then, would exclude adverbials and other unrelated forms in the analysis, and e
patterns of sociolinguistic variation within the community would be ignored. This
may not be of any consequence for the generativist tradition, but for sociolinguists it
is unfortunate, to say the least, given that the variable was initially conceived
precisely in order to analyse language in its social context.
In summary, there seem to be three possibilities for analysing emphatic
pronoun tags. First, they could be analysed purely in terms of their form, as a
construction used by some speakers but not others, with no assumption that they are
involved in syntactic variation and with no reference to their interactional functions.
This, I assume, is the approach typical of the generative tradition: the tags could be
seen as an instance, perhaps, of right dislocation. The disadvantage for sociolinguists,
of course, is that the interactional functions of the forms will be neglected: the
analysis will explain only how the tags can be generated by the grammar. A second
possibility is to analyse the tags as a linguistic variable, as variants of the canonical
clause structure. This involves accepting that the discourse function of the tags is to
express emphasis (and that this, in turn, can serve a range of interactional purposes
such as nominating a potential topic). As we have seen, this has the disadvantage for
generativists and variationists alike of lacking analytical rigour. Third, the tags could
be analysed from an explicitly sociolinguistic perspective, to establish their social
distribution within the speech community and, if used by different age groups, their
8
potential involvement in language change. In our Hull data it is only working class
speakers who use the tags, so the analysis would then focus on the linguistic forms
used by other social groups to accomplish the same interactional function. The
analysis could be quantitative, but the concept of the variable would not be invoked. If
middle class speakers in Hull use the same forms as the middle class speakers in
Macaulay's Ayrshire study, this broad brush approach would find emphatic pronoun
tags alternating with adverbials and, perhaps, with other elements that are unrelated
syntactically.
Each approach is likely to reveal different aspects of the phenomenon of
pronoun tags. The tags illustrate neatly the problem of combining internal and
external investigations, for the question is not, of course, which of these approaches is
the right one, but how we can integrate them to achieve a fuller understanding of both
the cognitive and the interactional aspects of their use.
Pivotal lone when phrases
Like some subordinate adverbial clauses, lone when phrases are introduced by when;
unlike subordinate clauses, however, there is no accompanying main clause. As
examples, consider the forms indicated by the arrows in (3) and (4). They are taken
from conversations between 12–16-year-old working-class adolescents in Reading,
Berkshire, recorded in adventure playgrounds during a nine-month period of
participant observation (see Cheshire 1982).
(3) (the boys are talking about one of their teachers, who was married to a
someone I knew)
9
Nobby: yeah Miss Threadgold she ain’t bad
Rob:
yeah she . she went camping with us
Jenny: yes he told me she’d been camping
→ Nobby: when we went camping
Rob:
she’s a good laugh
Jenny: is she?
Nobby: yeah
(4) Jenny:
you have to do horrible jobs if you’re a nurse .. all the bed pans
All:
<LAUGHTER>
Jenny:
have you ever been in hospital?
Valerie: [I have
Christine: [oh yeah I have
Valerie: I got run over by a car
Christine: I fell off a gate backwards <LAUGHS> and I was unconscious
→ Tommy: oi when I .. when I went in hospital just for a little while …
Valerie:
sshh
Tommy: cos my sister and my cousin they bent my arm .. they twisted it right
round
It has often been pointed out that languages with a strong written tradition,
such as English, have a well-developed descriptive and analytical framework derived
from intuitions that are likely to be heavily influenced by the written language (see,
10
for example, Brown and Yule 1983). This may mislead researchers into analysing
spoken forms as realisations of a conventional syntactic category when it is
inappropriate to do so. This is the case for the lone when phrases. They have the form
of an adverbial clause, albeit with no surface main clause, and they share some of the
functions of adverbial clauses, as we will see, but their main interactional function is
different.
Unattached subordinate adverbial clauses do occur in spoken English, but they
are relatively infrequent. For example, Mondorf (2000) found only 6 per cent of
adverbial clauses with no main clause in the London-Lund corpus (259, out of 4462
clauses). In Ford's (1993) corpus of American English only 3 per cent of the temporal
adverbial clauses had no main clause (2 out of 63 temporal clauses). In the Reading
data, however, unaccompanied when clauses are more frequent: here 25 per cent (28)
of the 105 when clauses were unattached. The unattached clauses were uttered with
level tones on every syllable except the last: this has a falling tone and is slightly
drawled. Interestingly, they were used only by the male adolescents.
In Mondorf's and Ford's studies unattached temporal clauses were introduced by a
range of time adverbs of which when was just one. Both authors consider the
unattached clauses as adverbial clauses, despite the fact that adverbial clauses are,
according to conventional frameworks, part of a main clause (Biber et al 1999: 194,
Quirk et al 1985: 1047). In Ford's and Mondorf's studies it was, apparently, possible
to infer a main clause from the context in which the unattached adverbial clauses
occurred. It is difficult, however, to unambiguously infer a main clause for the lone
when clauses in extracts 3 and 4, and perhaps this should alert us to the possibility that
this is not the best way to analyse them. However, their discourse functions
correspond in some respects to those of conventional adverbial clauses. For example,
11
Ford (1993: 29, 32) finds that conventional initial when clauses can be explicatory
when they follow a semantically broad term such as thing or then. In her data the
explication occurs within an extended speaker turn; she argues, in fact, that the use of
the semantically broad term contributes to the projection of an extended turn. In the
Reading data, similarly, four of the lone when forms have an explicatory function.
This can be seen in (5), where Rob explains, in answer to a question, how Britt (one
of the playground leaders) tries to control her mind. The lone when clause provides a
time frame for a specific situation given as an illustration of Britt’s behaviour.
(5) Rob:
and Britt she’s queer =
Nobby:
Rob:
= she’s trying to learn to control her mind
= yeah =
whatever that means
Jenny: is she?
Rob:
[yeah
Nobby: [yeah
Jenny: oh how is she going to what is she doing to con
Nobby:
→ Rob:
I don’t know
when you look at smoke and that you know fire =
Jenny:
= yeah
Nobby: she looks at a flame she’s . you can look at . she’s trying to look at a
flame until it burns right out
Jenny: and then w . how does that control your mind?
Rob:
I don’t know
12
The explicatory lone when clauses in the Reading data, however, do not elaborate a
semantically broad term, nor do they project an extended turn. They do, on the other
hand, clarify a semantically problematic concept that the emerging discourse shows to
be ambiguous or too vague for present purposes: in (5) this is the idea of controlling
your mind, which is initially unclear to all the participants, as indicated by Rob’s
whatever that means and Nobby’s response (I don’t know) to my question about how
this can be done.
Ford also found that conventional adverbial clauses in initial position could form
pivotal points in the development of talk, projecting an extended turn and presenting
explicit background for material that follows (1993: 62). This is the case for the 21
remaining lone when forms in the Reading data. For example, in (4) the other
speakers interpret Tommy's lone when phrase, prefaced by his attention-getting oi, as
an indication that he intends to take a projected turn; this is shown by Valerie
compliantly telling her younger sister to be quiet. Usually the extended turns are
narratives of personal experience; thus in (4) Tommy went on to tell the story of his
stay in hospital. Both explicatory and pivotal lone when phrases, then, share some
aspects of the interactional function of conventional adverbial when clauses. As
mentioned above, there is a social variation in that the forms without an
accompanying main clause are used only by the male adolescents. They would seem
to be candidates for a variationist analysis of a when clause variable with two variants,
one with and one without a main clause.
However, this approach would miss an important discourse function of the lone
when phrases. In every case, the narrative that follows the phrase is about events that
are familiar to the other speakers, either because they have heard the story before, or
because they participated themselves in the events that are recounted. The narrative is
13
a form of joint reminiscing – a discourse event with an important role in reinforcing
group membership (Edwards and Middleton 1986). In the playground conversations
these narratives were especially significant in the construction and reinforcement of
group friendship patterns amongst the male adolescents. The main function of the
lone when phrases, in other words, is as a story opener, marking the upcoming story
as a shared reminiscence Female adolescents constructed friendships on a more
individual basis, telling stories mainly as monologues. Their different narrative style
was reflected their preferred story opener, which as Table 1 shows was a temporal
subordinate clause, clearly situating the story in the past (for further details see
Cheshire 2000).
INSERT TABLE 1 HERE
When the lone when phrases are considered in their full interactional context, then,
it becomes clear that they cannot be analysed as variants of conventional initial when
clauses, since they are not, in fact, functionally equivalent. They have a specific
discourse function as a story opener. A variationist analysis could, perhaps be
performed on the range of story openers that are used to introduce narratives or,
specifically, sequences of joint reminiscing (those shown as the first group in Table
1), but this would be of no interest for the study of syntactic variation.
The lone when phrases, together with the other story openers that mark an
upcoming shared reminiscence, may in fact be better seen as a conversational routine:
a sequence of words that appear to have syntactic structure but that are produced and
processed as a more or less prefabricated phrase (Aijmer 1996). The when of the lone
when phrases may be a reduced form of the other when phrases in this group of story
14
openers (what about when, you know when or remember when). The routines are not,
of course, completely fixed in their form and they differ, therefore, from prefabricated
phrases such as how do you do. Yet they have more in common with a prefabricated
lexicalised form such as this than with completely new clauses that have been
generated by the grammar. They seem to consist of a frame (when + NP + VP, with
the verb in the past tense) within a fixed intonation contour. The words that constitute
the NP and the VP are repeated from the preceding discourse, and this, of course,
facilitates their function as a way of taking the floor: thus in (3) Nobby’s went
camping echoes the words of the preceding three turns, and in (4) Tommy’s in
hospital echoes the question have you have ever been in hospital? What initially
appeared to be an instance of syntactic variation, then, is more appropriately seen as a
conversational routine with an interactional function in turn-takingi and a social
function in indexing group solidarity (as we have seen, it is used only by the boys,
along with other story openers that mark an upcoming shared reminiscence).
It is relevant to note that other forms used as story openers are also better analysed
as prefabricated phrases than as forms generated by the grammar. For example,
Reading English, like other south-western varieties of English, allows a variable -s
suffix with non-third singular present tense verbs. The playground conversations
contain many tokens of the non-standard verb forms, as in (6) – (8) below:
(6) I loves Elvis...he’s great
(7) you knows him don't you Nod?
(8) he says to me "look here and I see if I knows you"
15
In (6) and (8) know occurs with the non-standard verbal suffix. When you know is
used as a story opener, however, as in (9) and (10), it never occurs with the –s suffix:
(9) you know that hill down there? I rode down that with no hands on the handlebars
(10) you know your mum….you know that bike she had
There was X as a story opener is also invariant: was always occurs,
irrespective of whether the subject is singular or plural (see Cheshire 1999 for
discussion).
There is an important implication here for the integration of social and generative
approaches to analysing syntactic variation. As many researchers working on spoken
language have stated, linguists have tended to over-emphasise the creative aspect of
language. There is no doubt that we can produce and understand an infinite number of
sentences that we have never heard before but, as Bolinger 1975: 297) pointed out, the
fact that we can do this does not mean that we do. It would be counter-productive in
spontaneous face-to-face communication to constantly produce brand new sentences,
and speakers use many prefabricated expressions to help them cope with the demands
of fast speech production. Estimates of the proportion of ready-made chunks of
unanalysed language in large-scale corpora of spoken language range from 30 per
cent (Biber et al 1999) to 70-90 per cent (see Aijmer 1996:31). The difference reflects
the ways in which the chunks are defined; but even 30 per cent is a sufficiently high
proportion for their existence to be taken seriously. It raises the question of whether
spoken language is better conceptualised as linear and sequential in structure rather
than as hierarchical. Sinclair (1991: 68) pushes this idea to its limits, predicting that
16
the traditional domain of syntax will be invaded by lexical hordes, leading to the
eventual demise of syntax. Skehan (1998:37) puts forward a more moderate view,
suggesting that producing speech involves improvising on a clause by clause basis,
using lexical phrases and lexical sentence stems wherever possible in order to
minimize processing demands and only as a last resort generating language that is not
part of our memorised lexicon. Whichever view we adopt, it is important to remember
when analysing spoken language that what may appear to be an instance of syntactic
variation may instead be a chunk of ready-made memorised language. Conventional
frameworks of analysis, in other words, may predispose us to classify a form as
syntactic when in fact it is not.
Get-passives in English
The get-passive in English is illustrated in (11):
(11) Josephine got run over by a bus
It can be seen as an alternant to a conventional passive, as in (12):
(12) Josephine was run over by a bus.
Both (11) and (12) can be seen as alternants to a corresponding active clause, as in
(13a) and (13b):
(13a) A bus ran over Josephine
17
(13b) A bus ran Josephine over.
Weiner and Labov (1983:43) claimed that a shift to the get-passive is one of
the most active grammatical changes taking place in present-day English.
Variationists, therefore, would like to know how the form is distributed throughout
the speech community, in order to chart the diffusion of the change and even, perhaps,
to locate the source. As with pronoun tags and lone when phrases, there is no initial
problem in using the variable to analyse this alternation: the alternants in (11) – (13)
have the same verb and the same entities as arguments, so we can assume that they
are semantically equivalent.
Passive constructions are relatively infrequent in speech. It is not surprising,
therefore, that there have been few quantitative studies of the English passive.
Macaulay (1991), however, analysed passive constructions in middle class and
working class speech in his Ayrshire study. Although there were no significant social
class differences in the overall use of passive clauses, get-passives were used more
frequently by working class speakers. Macaulay further reports that the get-passive
occurred almost exclusively with animate subjects and that these, in turn, were also
more frequent in the working class interviews. Get-passives are eventive aspectually,
and this presumably contributes to the animacy effect; events are usually controlled
by an actor, and animates are more likely to be able to control eventsii. Animacy has
an effect on a wide range of structures in different languages, though as far as I know
its overall impact on the spread of language change has not been assessed. I do not
know what light, if any, the generative model of syntax could shed on this
conditioning factor, nor whether there could be any implications for syntactic theory.
18
Most work on the effect of animacy on language structure has been done by
typologists, whose perspective and methodology is different from that of both
generativists (Croft 1994: 153) and variationists.
Carter and McCarthy (1999) add a further dimension to our understanding of
get-passives: their corpus-based analysis reveals that the get-passive highlights the
stance of speakers towards the grammatical subject and the event encoded in the verb
phrase – a stance that usually indicates their judgement that the circumstances are
adverse, problematic or otherwise noteworthy. We cannot necessarily assume that the
Ayrshire speakers and the speakers in Carter and McCarthy’s study are affected
equally by these factors, but the two studies at least suggest the possibility that the
shift to the get-passive in present-day English is led by a group-specific discourse
preference for encoding syntactically the speaker's stance towards actors and the event
referred to by the verb. Unfortunately, however, we now run into the same problems
as with appeals to emphasis: although a wide range of syntactic features have been
said to express the speaker’s stance towards the content of their utterances (see, for
example, those listed in Ochs and Schieffelin1989) it is difficult to give a rigorous
definition of the concept and it has no status within a generative theory of language.
These two studies of the get-passive raise the further question of whether the
social distribution found by Macaulay reflects a distinctive habitual pattern of
interaction for the working class group of speakers that is not shared by the middle
class group. Syntactic variation and syntactic change seem here to be intimately and
inextricably part of the social construction of discourse. Carter and McCarthy’s
findings point to a similar conclusion: as these authors say (op.cit: 55), judgements
about adversity, noteworthiness and the like are socio-culturally founded and
emergent in the interaction rather than inherent in the semantics of verb choice or the
19
selection of voice or aspect. In order to address these issues, then, and to fully analyse
patterns of variation and change in the use of English passive constructions, it is
essential to add a qualitative, interpretive dimension to the analysis.
Even what some have called low-level variables (see, for example, Cornips
and Corrigan in press) call for this kind of interpretive discourse analysis. Levinson
(1988:166) made this point in relation to English ain't, asking whether working class
speakers who use ain't frequently do so because for them ain't is a marker of group
identity, or because it is a more emphatic form of negation than isn't, aren't, hasn’t
and haven’t. If the latter, does this reflect the habitual patterns of social interaction of
the social group to which the heavy ain't users belong? They might, for example, utter
emphatic denials more frequently than other social groups in the community because
they more often receive accusations. Thus in order to understand how and why
speakers use variation, and the effect that their usage has on language change, we
cannot simply analyse the simple alternation of forms: we must also perform
qualitative analyses to see how these forms are used in social interaction and find a
way of incorporating this dimension into an explanatory theory of language structure
and language change.
Conclusion
As I mentioned at the beginning of this paper, the main data for variationists has
always been spoken language (partly, of course, because the main preoccupation of
variationist research has been phonological variation). The work that has been done
on syntactic variation within this tradition has tended to be conservative and
unadventurous, with researchers analysing the same linguistic variables over and over
20
again. The reason is partly because the favourite variables tend to occur with the high
frequencies that are necessary for quantitative analyses (consider English past and
present tense verb forms, for example, or multiple negation; but it is also because their
relation to the ideology of the standard makes them more noticeable to prescriptivists,
laypersons and professional linguists alike. Variationists have worked almost
exclusively on languages that have been heavily standardised, so the potential
influence of the standard ideology has been high (see Cheshire and Stein 1997 and
Milroy 1999 for further discussion). Despite the emphasis of modern linguistics on
the primacy of speech, the field of language study that has perhaps worked longer
than any other on the empirical analysis of spontaneous spoken language has,
paradoxically, discovered rather little about the syntactic features that characterise it.
A hope, from my sociolinguist corner, is that a new alliance between
variationists and generativists might lead to generative linguists becoming more
interested in externalised, performed language – not simply in order to find new data
against which to test the theory but also to apply the rigour of the generative approach
to discovering the structure of spoken language. There are many ways in which our
understanding of spoken language might benefit. First, it may be possible to bring
some theoretical rigour to the concepts of emphasis and stance, which seem so
important to speakers and which are so often appealed to in descriptions of language
use. Second, as pointed out by Henry (2002: 277) generativists might help
variationists determine on a more systematic basis those syntactic structures that
should be considered variants of a single form: this would help assess the universal
dimension of any internal constraints on variation (Cornips and Corrigan in press) and
may also prevent us from being unduly influenced by the standard ideology. This
might even help resolve the issue of whether the variable should be used to analyse
21
variation in syntax. Third, it may be possible to identify some language universals
governing the form of spoken language – though for this to be possible we would
need to broaden our perspective and look for social or interactional principles in
addition to the principles governing innate structures. Perhaps we will need to work
with researchers from other fields of linguistics too. For example, I mentioned
language typology in the previous section; this perspective has already been very
successfully applieded to the findings of social dialectology (see, for example,
Anderwald 19 , Kortman 2002).
Generativists and variationists might also work together to develop a
methodology that could determine when the phrases of spoken language have been
generated by the grammar and when they are prefabricated sequences produced from
memory. As we have seen, when subject-verb agreement is involved this can be
straightforward to determine, but often there are no overt clues. Furthermore, the
distinction is unlikely to be clear-cut: work on large-scale corpora has shown that
conversational routines can be arranged along a continuum from completely fixed
forms through semi-fixed forms (e.g. I’m so/really/very sorry) and frame and slot
forms (e.g. could I have X) to ‘mini-grammars’ (see Aijmer 1996).
Finally, it might be possible to decide on a principled basis when it is more
profitable to think in terms of pragmatic variation rather than syntactic variation.
Some researchers who have worked a great deal on syntactic variation have decided
that pragmatic factors have priority over internal factors: Winford (1996: 188), for
example, suggests that perhaps most of so-called syntactic variation is motivated by
pragmatic factors alone. Perhaps the most pressing question, if generativists and
variationists are to benefit from each other's insights, is to establish, for any given
syntactic feature that seems to be variable, how much of the variation is most usefully
22
considered as syntactic (in other words, internally motivated by the grammar) and
how much is better seen as pragmatically determined.
Many of these issues will require both a qualitative and a quantitative
dimension to the analysis, with a combination of methodologies including corpus
analysis and the elicitation of intuitions. If we are to gain insights from such different
research traditions we need to be aware that the forms of spoken language may result
as much from interactional and social factors as from biological factors. In this way
we may at last succeed in understanding how the cognitive and the social aspects of
language are integrated as part of the human experience.
References
Aijmer, K. (1996). Conversational Routines in English: Convention and Creativity.
Harlow: Longman.
Biber, D. Conrad, S. and Reppen, R. (1994). Corpus-based approaches to issues in
applied linguistics. Applied Linguistics 15,169-187.
Biber, D., Johansson, S., Leech, G., Conrad, S. and Finegan, E. (1999). Longman
Grammar of Spoken and Written English. Harlow: Pearson Education Ltd.
Blanche-Benveniste, C. (1997). La notion de variation syntactique dans la langue
parlée. Langue Française 115, 19-29.
Bolinger, D. (1975). Meaning and memory. Forum Linguisticum 1, 2-14.
Brown, G. and Yule, G. (1983). Discourse Analysis. Cambridge: Cambridge
University Press.
Carter, R. and McCarthy, M. (1999). The English get-passive in spoken discourse:
description and implications for an interpersonal grammar. English Language
and Linguistics 3, 41-58.
Chambers, J.K. and Trudgill, P. (1998). Dialectology. Cambridge: Cambridge
University Press. 2nd edition.
Cheshire, J. (1982). Variation in an English Dialect: a Sociolinguistic Study.
Cambridge: Cambridge University Press.
Cheshire, J. (1987). Syntactic variation, the linguistic variable and sociolinguistic
theory. Linguistics 25, 257-282.
Cheshire, J. (1999). Spoken standard English. In T. Bex and R. J. Watts (Eds.),
Standard English: The Widening Debate (pp. 129-148). London: Routledge,
Cheshire, J. (2000). The telling or the tale? Narratives and gender in adolescent
friendship networks. Journal of Sociolinguistics 4, 234–262.
Cheshire, J., Kerswill, P., and Williams. A. (1999). The Role of Adolescents in Dialect
Levelling. Final Report submitted to the Economic and Social Research
Council, June 1999. Ref. R000236180.
23
Cheshire, J., Kerswill, P. and Williams, A. (in press). On the non-convergence of
syntactic and phonological variation. In P. Auer, F. Hinskens and P. Kerswill
(Eds.) Dialect Convergence and Divergence in a Changing Europe.
Cambridge: Cambridge University Press.
Chomsky, N. (1999). Derivations by Phase. Unpublished MSS. Boston, MIT.
Cornips, L. and Corrigan, K. (in press). Convergence and divergence in grammar. In
P. Auer, F. Hinskens and P. Kerswill (Eds.) Dialect Convergence and
Divergence in a Changing Europe. Cambridge: Cambridge University Press.
Couper-Kuhlen, E. and Selting, M. (Eds.) (2001). Studies in Interactional Linguistics.
Amsterdam: Benjamins.
Coupland, N. (1983). Patterns of encounter management: Further arguments for
discourse variables. Language in Society 12, 459-476.
Coveney, A. (1997). L'approche variationniste et la description de la grammaire du
français. Langue Française 115, 88-100.
Coveney, A. (2002). Supplement: A critical survey of recent research on grammatical
variability, negation and interrogation in French. Variability in Spoken French.
Bristol: Intellect.
Dines, E. R. (1980). Variation in discourse – "and stuff like that". Language in
Society 9,13-33.
Edwards, J. and Middleton, D. (1986). Joint remembering: constructing an account of
shared experience through conversational discourse. Discourse Processes 9,
423-459.
Ford, C. E. (1993). Grammar in Interaction: Adverbial Clauses in American English
Conversations. Cambridge: Cambridge University Press.
Gadet, F. (1997). La variation, plus qu'une écume. Langue Française 115, 5-18.
Henry, A. (2002). Variation and syntactic theory. In J. K. Chambers, P. Trudgill and
N. Schilling-Estes (eds.) Handbook of Language Variation and Change (pp.
267-282). Oxford: Blackwell.
Kortmann, B. (2002). New prospects for the study of dialect syntax: Impetus from
syntactic theory and language typology. In S. Barbaiers, L. Cornips and S. van
der Kleij (Eds.) Syntactic Microvariation. Amsterdam.
Lavandera, B. (1978). Where does the sociolinguistic variable stop? Language in
Society 7, 171-182.
Leech, G. (2000). Grammars of spoken English: The new implications of corpusoriented research. Language Learning 50, 675-724.
Levinson, S. (1988). Conceptual problems in the study of regional and cultural style.
In N. Ditttmar and P. Schlobinski (Eds.) The Sociolinguistics of Urban
Vernaculars (pp.161-190). Berlin: Walter de Gruyter.
Levinson, S. (1983). Pragmatics. Cambridge: Cambridge University Press.
Macaulay, R. K. S. (1989). 'He was some man him': Emphatic pronouns in Scottish
English. In T.J. Walsh (Ed.) Synchronic and Diachronic Approaches to
Linguistic Variation and Change (pp. 179-187). Washington DC.:
Georgetown University Press.
Macaulay, R. K. S. (1991). Locating Dialect in Discourse: The Language of Honest
Men and Bonnie Lassies in Ayr. New York: Oxford University Press.
McCarthy, M. (1991). Discourse Analysis for Language Teachers. Cambridge:
Cambridge University Press.
Miller. J. and Weinert, R. (1998). Spontaneous Spoken Language: Syntax and
Discourse. Oxford: Clarendon Press.
24
Milroy, L. (1999). Standard English and language ideology in Britain and the United
States. In T. Bex and R. J. Watts (Eds.) Standard English: The Widening
Debate (pp. 173-206), London: Routledge.
Milroy, L. and Gordon, M. (2003). Sociolinguistics: Method and Interpretation.
Oxford: Blackwell.
Mondorf, B. (2000). Sex differences in English syntax. Unpublished doctoral
dissertation, Düsseldorf University.
Myhill, J. (1992). Typological Discourse Analysis. Oxford: Blackwell.
Ochs, E., Schegloff, E.A. and Thompson, S.A. (Eds.) (1996). Interaction and
Grammar. Cambridge: Cambridge University Press.
Ochs, E.and Schieffelin, B. 1989. Language has a heart. Text 9, 7-25.
Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. (1985). A Grammar of
Contemporary English. Harlow: Longman.
Romaine, S. (1980). The relative clause marker in Scots English: Diffusion,
complexity and style as dimensions of syntactic change. Language in Society
9, 221-247.
Schwenter, S. (in press). No and Tampoco: A pragmatic distinction in Spanish
negation. Journal of Pragmatics.
Sells, P., Rickford, J., and Wasow, T. (1996). Variation in negative inversion in
AAVE: an optimality theoretic approach. In Arnold, J, et al (Eds.)
Sociolinguistic Variation: Data, Theory and Analysis. Selected Papers from
NWAVE 23 (pp. 171-1760. Stanford: CSLI.
Sinclair, J. (1991). Corpus, Concordance, Collocation. Oxford: Oxford University
Press.
Skehan, P. (1998). A Cognitive Approach to Language Learning. Oxford: Oxford
University Press.
Weiner, J. and Labov, W. (1983). Constraints on the agentless passive. Journal of
Linguistics 19, 29-58.
Wolfram, W. (1993). Identifying and interpreting variables. In D. R. Preston (Ed.)
American Dialect Research (pp. 193-222). Amsterdam: John Benjamins.
Winford, D. (1996). The problem of syntactic variation, In Arnold, J, et al (Eds.)
Sociolinguistic Variation: Data, Theory and Analysis. Selected Papers from
NWAVE 23 (pp. 177-192). Stanford: CSLI.
25
Table 1. Story openers in the corpus of narratives
All-female conversations
All-male conversations
Total
Markers of a shared reminiscence
Remember when clause
0
1
1
What about that time when
0
2
2
You know when clause
What about X
Pivotal lone when-clause
0
1
0
1
0
5
1
1
5
I can’t forget that time when
0
1
1
_____________________________________________________________________
Total
1
10
11
Temporal subordinate clauses introduced by:
once
6
0
6
when
5
2
7
the other day
5
0
5
one time
2
0
2
one day
3
0
3
last time
1
0
1
yesterday
1
0
1
_____________________________________________________________________
Total
23
2
25
Zero
9
34
43
Miscellaneous
clause right
left dislocation
there was X
you know X
you see
see
you should have seen X mate
fuck me
he’s a bastard mate
oh it’s horrible
it wasn’t half fun
1
9
6
2
3
2
0
0
0
1
1
7
3
3
2
0
0
1
1
1
0
0
8
12
9
4
3
1
1
1
1
1
1
_____________________________________________________________________
Total all story openers
58
64
12
_____________________________________________________________________
26
Transcription Conventions
..
=
?
short pause (not timed)
utterance latched on to previous turn (with no discernible pause)
question marks show the end of a stretch of talk interpreted as an
question
<LAUGHTER> angled brackets give additional information
[
extended square brackets show the beginning of an overlap
→
an arrow indicates that the line to the right is the one where a
given example occurs
Notes
i
ii
Thanks to Lesley Milroy for noting the function in turn-taking
Thanks to David Adger for this insight and many other helpful comments
27