LANGUAGE AND

SOCIAL CLASS
LANGUAGE AND SOCIAL CLASS

• If you are an English-speaker or know something about English­speaking societies, you will be able to estimate the relative
social status of the following speakers solely on the basis of the linguis­tic evidence given here:

• Speaker A
• I done it yesterday.
He ain 't got it.
It was her what said it.
• Speaker B
• I did it yesterday.
He hasn 't got it.
It was her that said it.
WHY DOES SOCIAL DIFFERENTIATION HAVE THIS EFFECT ON
LANGUAGE?

• If you heard these speakers say these things, you would guess that B was of higher social status
than A, and you would almost certainly be right. How is it that we are able to do this sort of thing?
• The answer lies in the existence of varieties of language which have come to be called social-class
dialects or, by some writers, sociolects. There are grammatical differences between the speech of
these two speakers which give us clues about their social backgrounds. It is also probable, although
this is not indicated on the printed page, that these differences will be accompanied by phonetic
and phonological differences - that is to say, there are also different social-class accents. The
internal differentiation of human societies is reflected in their languages. Different social groups
use different linguistic varieties, and as experienced members of a speech community we
WE MAY NOTE PARALLELS BETWEEN THE DEVELOPMENT OF THESE
SOCIAL

• varieties and the development of regional varieties: in both cases barriers and distance
appear to be relevant. Dialectologists have found that regional-dialect boundaries often
coincide with geographical barriers, such as mountains, swamps or rivers: for example,
all Traditional Dialect speakers in the areas of Britain north of the River Humber
(between Lincolnshire and Yorkshire) still have a monophthong in words like house
('hoose' [hu:s]), whereas speakers south of the river have had some kind of [haus]-type
diphthong for several hundred years; and in the USA the border between Northern and
Midland dialE�cts
• at some points runs along the Ohio river. It also seems to be the case that the greater the
geographical distance between two dialects the more dissimilar they are linguistically: for
instance, those regional varieties of British English which are most unlike the speech of
London are undoubtedly those of the north-east of Scotland - Buchan, for example; while
in North America the biggest linguistic differ­ences between regional varieties of English
would be found by comparing the speech of Newfoundland with that of, say, Mississippi.
• The development of social varieties can perhaps be explained in the same sort of way - in
terms of social barriers and social distance. The diffusion of a linguistic feature through a
society may be halted by barriers of social class, age, race, religfon or other factors. And
social distance may have the same sort of effect as geographical distance: for example, a
linguistic inno­vation that begins amongst the highest social group will affect the lowest
social group last, if at all.
SOCIAL STRATIFICATION

• Of the many forms of social differentiation, for example by class, age, sex, race or religion, we
shall concentrate in this week on the particular type of social differentiation illustrated in the
example of speakers A and B - social stratification.
• Social stratification
• is a term used to refer to any hierarchical ordering of groups within a society especially in terms of
power, wealth and status. In the industrialized societies of the West this takes the form of
stratification into social classes, and gives rise linguis­tically to social-class dialects.
• Social-class stratification is not universal, however. In India, for example, traditional society is
stratified into different castes.
THE CLASS SOCIETIES OF THE ENGLISH-
SPEAKING WORLD THE SOCIAL SITUATION
• In the class societies of the English-speaking world the social situation is much more
fluid, and the linguistic situation is therefore rather more complex, at least in certain
respects. Social classes are not clearly defined or labelled entities but simply aggregates
of people with similar social and economic character­istics; and social mobility -
movement up or down the social hierarchy - is perfectly possible. This makes things
much more difficult for any linguist who wishes to describe a particular variety - the
more heterogeneous a society is, the more hetero­geneous is its language.
IDIOLECT

• For many years the linguist's reaction to this complexity was generally to ignore it - in
two rather different ways. Many linguists concentrated their studies on the idiolect - the
speech of one person at one time in one style - which was thought (largely
erroneously, as it happens - see p. 29) to be more regular than the speech of the
community as a whole
• Dialectologists, on the other hand, concentrated on the speech of rural informants, and in
particular on that of people of little education in small isolated villages,
• It is only fair to say, however, that there are two additional explanations for why dialectologists
concentrated on rural areas in this way;
• First, they were concerned to record many dialect features which were dying out before they were
lost for ever.
• Secondly, there was a feeling that hidden somewhere in the speech of ·older, uneducated people
were the 'real' _or pure' dialects which were steadily being corrupted by the standard variety, but
which the dialectologists could discover and describe if they were clever enough.
• Therefore, Dialectologists then began to incorporate social as well as geographical
information into their dialect surveys.
• For example, workers on the Linguistic Atlas of the United States and Canada, which was begun in
the 1930s, divided their informants into three categories largely according to the education they
had received, and thereby added a social dimension to their linguistic information.
• They also began, in a rather tentative kind of way, to investigate the speech of urban areas. It was
not really until after the Second World War, however, that linguists also began to realize that in
confining dialect studies to mainly rural areas they were remaining singularly ignorant about the
speech of the vast majority of the population - those who lived in towns. A large amount of
linguistic data that was both interesting in itself and potentially valuable to linguistic theory was
being ignored or lost.
THE DIFFERENCES BETWEEN LINGUISTICS
DIALECTOLOGISTS
LINGUISTS AFTER THE SECOND WORLD WAR

• It is only after the Second World War, linguists realized that:

• 1. they are obtaining an imperfect and inaccurate picture of the speech of different areas.
• 2. They are confining dialect studies to rural areas, they missed important information
about the majority of people who live in towns.
• 3. Urban dialectologist faced the problem of describing fully and accurately the speech of
large towns and cities with heterogeneous populations such as in a city like New York
with a population of 8+ millions?
URBAN
DIALECTOLOGISTS
• Those urban dialectologists who
recognized that this was the case
were therefore forced to work
out how they were to describe,
fitlly and accurately, the speech
of large towns and cities, and it
was in response to this problem
that urban dialectology eventu­
ally became sociolinguistic.
WILLIAM LABOV LABOV’S WORK:

•
His informants were randomly selected (not friends or relatives).His informants’ speech was representative
of New York city (esp. the Lower East Side) since these informants were a representative sample.
• Therefore his work was an accurate description of all the varieties in this area.
• Labov showed that variation is not free in the speech of New Yorkers as was claimed by traditional
linguists, e.g ‘guard’, ‘beard’, and ‘bad’.
• Variation is not random, but determined by extra-linguistic factors in a predictable way (if they were of a
certain class, age, and sex).
WILLIAM LABOV LABOV’S WORK WAS A REVOLUTION BECAUSE
HE:

• 1. developed techniques to draw out normal speech from people in spite of the recorder.
• 2. developed methods for quantitative measurements of linguistic data.
• 3. The problem of heterogeneity of speech communities has been overcome because we
are able to correlate linguistic features with social class accurately.
 SOCIAL AND REGIONAL DIALECT VARIATION
IN BRITAIN
• In Britain, at the other end of the
social scale, however, the situation
is very different. Speakers of the
highest social class employ the
dialect we have called Standard
English, which, as
LEXICAL EXAMPLE

• To take a lexical example, there is in the Standard English

dialect a single word scarecrow signifying the humanoid
object farmers place in fields to scare off birds.

• At the other end of the pyramid, on the other hand, we find

a far greater degree of regional variation in the most
localized regional English dialects. Corresponding to
scarecrow we have bogle, flay-crow, mawpin, mawkin,
bird-scarer, moggy, shay, guy, bogeyman, shuft, rook-
scarer, and several others
GRAMMATICAL DIFFERENCES

In Standard English, for example, we find both:

•He's a man who likes his beer
•and
•But regional variation in nonstandard British English varieties is much greater. All the following are possible:
•He's a man who likes his beer.
•He's a man that likes his beer.
•He's a man at likes his beer.
•He's a man as likes his beer.
•He's a man what likes his beer.
•He's a man he likes his beer.
•He's a man likes his beer.
• If we are to obtain a correct picture of the relation shipbetween language and social
stratification we must be able to measure both linguistic and social phenomena so that we
can correlate the two accurately.
• As far as social class is concerned this can be done relatively easily (it is still far from
simple) by the socio­logical method of assigning individuals a numerical index score on
the basis of their occupational, income, educational and/or other characteristics, and then
grouping them together with others with similar indexes (although the justification for
differ­ent groupings may be controversial).
• Measuring language is more difficult. The solution developed by Labov and since used
by others is to take linguistic features which are known, either from previous study or
intuitively by the linguist as a native speaker, to vary within the community being
studied, and which are also easily countable in some way. For instance, in two separate
surveys, one in Detroit, USA, and one in Norwich, England, the same grammatical
feature appeared to be suitable in this way.
• In Standard English the third-person present-tense singular form of verbs has an affix,
orthographic -s, which distin­guishes it from other persons: I know, we know, they know,
but she knows. In East Anglia, the area of England in which Norwich is situated, and in
Detroit this -s is often not present, at least in the speech of some people. This means that
the following sorts of forms occur:
• She like him very much.
He don't know a lot, do he? Itgo ever so fast.
MEASURING

• Since Standard English has the -s, and since the standard variety is generally most closely
associated with higher social groups,
• it was suspected that there might be a direct correlation between social-class position and
usage of -s. To investigate this po.ssibil­ity was relatively easy, since there was no
difficulty in measuring this linguistic feature: it was simply a matter of listening to tape-
recordings made during the surveys and counting the number of times a speaker did or
did not use -s.
INHERENT VARIABILITY.

• the situation portrayed in both the above cases could be regarded as being a case of
dialect mixture. We could say, that is, that in the first case what we are really faced with is
two different dialects, one with and one without the -s.
• It is better to describe the present-day situation as a case of inherent variability.

• Inherent variability means that the variation is not due to the mixture of two or more
varieties but is an integral part of the variety itself.
• Thus according to the 'dialect-mixture view' Detroit speakers vary their verb forms
because they mix Detroit Black English (which in its 'pure' form does not have -s) with
Standard English (which does). According to the 'inherent-variability view', on the other
hand, this variation is simply one characteristic of Detroit African American English.
• The evidence for this second view is that this kind of variation takes place on a very wide
scale, involving all speakers and a very large number of other linguistic features.
Moretellingly, this kind of variability is found even in the speech of very young children
who have not been exposed to other dialects. Linguistic varieties appear to be inherently
variable as a rule rather than as an exception, and inherent variability is probably the
linguistic counterpart of social heterogeneity.
• A number of other, rather more complex grammatical features have also been shown to correlate with social
class in the same kind of patterned manner - to characterize, by their frequency of occurrence, different (but
non-discrete) class dialects. Consider, for example, what happens if we wish to negate the following sentence:
• I can eat anything.
• There are two possibilities in the standard variety of English. We can either negate the verb, to produce:
• I can 't eat anything.
• or we can negate the pronoun:
• I can eat nothing.
SOCIAL-CLASS ACCENTS,

Social-class accents, in contrast to grammatical features, appear rather more difficult to handle.
We know, from our experience as native speakers, that there are a whole range of socially
determined accents, but how exactly are we to correlate these phonetic and phonological features
with sociological para­meters? The usual method is to investigate, singly, the pronunci­ation of
individual vowels and consonants. It is, for example, relatively simple to count the presence or
absence of particular consonants in any stretch of speech. In Norwich the following three features
were studied:
• the percentage of n' as opposed to ng in walking, running,
• etc. - [w:J:kl).] versus [W:J:k1?] .
CONCLUSION

• At this point it is important to ask: what exactly is the value

• of this kind of information? One reason why this question is important is that a number of
writers on sociolinguistics have derided this kind of work as 'correlational
sociolinguistics', as if the whole point of the exercise were simply the correlation rather
than what we can learn from it. The answer is that, first,
• t shows precisely what sort of information we are working with when we assign a social
status to a speaker on the.basis of linguistic evidence. Through our linguistic experiences
we have become sensitive, normally subconsciously, . to correlations of this type between
social class and standard or local linguistic forms. Secondly, it tells us something about
the social structure of the particular communities.
• Thirdly, it illustrates the point made above about the idiolect. Although individuals will
sometimes use one verb form, and sometimes another, the average percentage for each
group falls into a quite predictable pattern.
• Fourthly, it tells us a lot about social-class dialects. Even though we are concentrating on
only one feature rather than on a variety as a whole, it is still apparent that, like regional
dialects, social-class dialects are not distinct entities. They merge into each other to form
a continuum. We can if we like talk of 'the middle-working-class Norwich dialect', but if
we do we must be very clear (a) that our division into five social classes may be arbitrary;
(b) that the linguistic differences involved are simply relative and involve the frequency
of occurrence of particular features; and (c) there may be differing results if other
linguistic variables are taken.
• Finally, and most importantly, it gives us a great deal of information about, and insights
into, the processes involved in linguistic change - one of the biggest mysteries there is
involving human languages, and one which sociolinguistics has done much to help us to
understand better in the last forty years. As we shall see in the next chapter, the
correlation of linguistic variables with sociolinguistic variables is not the goal of this type
of research; in the study of linguistic change, this type of correlation is not where we
finish but where we start.