I.

INTRODUCTION

Variation in language, as a sociolinguistic phenomenon, may depend on

different factors. On the one hand, language can vary according to the
situation in which the speech act takes place and according to the relation
between the speakers (style 1). Varieties may also be associated with specific
functions in particular situations (register 2). On the other hand, language
varieties may be characterized by the geographical and the social
background of the speaker. A variety associated with the geographical
location in which it is used is called regional variety or regional
dialect , whereas variation in language due to social factors is referred to
as social variation or social dialect.

Social variation studies developed from traditional dialectology when

scientists understood the complexity of language variation. In the late 18th
century, dialectologists treated variation in language as a result of the
geographical origin of the speaker. In the 1950s, sociolinguists started to
concentrate on social factors‘ relevance to language variation. They agreed
that the dialectologists´ point of view was too restricted, and that
geographical location was not enough to account for linguistic variation.
Firstly they pointed out that a language is subject to constant change, i.e.
the mobility of the speakers of different dialects of one specific language,
and the resulting interaction between these dialects cause modification or
substitution of linguistic features. Secondly they started to investigate
differences in society which proved to be relevant to variation in language. It
becomes obvious that the current sociolinguistic approach to language
variation is in two ways more complex than the dialectologists' view, as not
only the aspect of constant change is taken into consideration but also the
influence social factors have on language varieties.

In the first chapter, tise paper will begin by explaining what is meant by
a variety of a language . In addition, the paper will give the reader a general
idea of what is the focus in this field of scientific investigation. The first
subsection will be dealing with the concept of the linguistic
variable , describing its function in sociolinguistic studies, followed by a
range of examples illustrating variation on all linguistic levels. After having
introduced the term social variable in the second subsection, the paper will
give a few concrete examples, in order to illustrate social factors' relation to
linguistic variation and to shortly discuss the problem of social stratification.
The paper will also touch upon the relation between linguistic variation and
the individual's act of identification.

The second chapter will describe the sociolinguistic research work. Here, the
different stages of the procedure will be explained, discussing the main
problems and difficulties that may emerge during sociolinguistic
investigation.

II. THE SOCIOLINGUIST'S INTEREST IN THE FIELD OF

SOCIAL VARIATION

Every single language has a repertoire  3 of varieties, including a standard

variety which is the result of deliberate intervention by society. A language
undergoes standardization4 in order to create a standard language which
serves as an orientation for linguistic norms. A non- standard variety of a
particular language may differ from the standard language on all linguistic
levels. It may be characterized by differences in pronunciation, grammar and
in vocabulary. A variety differing from the standard variety in pronunciation
only is often called accent, whereas variation in grammar and vocabulary
may be referred to as dialect. A linguistic variety differs from the standard
variety on at least one of these levels. It is shared by a speech
community which is defined by the use of certain linguistic features and by a
common attitude towards the variety 5. The members of a particular speech
community may not all know nor use the entire repertoire of "their
language", but they are aware of the norms about the selection of varieties,

" [...] so we may define a variety of a language as a set of linguistic items

with similar [regional or] social distribution." (Hudson 21996, p.22)

The sociolinguist concerned with the relation between social and linguistic
features confines himself to the investigation of a restricted number of
variables. He concentrates on a certain social variable and identifies its
variants which appear to promote the usage of a certain variant of a
linguistic variable instead of another. "The choices among the variants of a
linguistic variable are influenced by both social and linguistic forces" (Fasold
1990, p.272). The fieldworker has to deliminate the speech community which
he will be focusing on from other communities, and it is necessary to know
who is using the relevant features in which context. He will then prepare a
procedure in order to elicit relevant data confirming his hypothesis about the
relation between linguistic and social variables in this particular speech
community.

A. Linguistic Variables
In the study of language variation linguistic, variables function as scientists´
tools, enabling them to investigate, recognize and analyse particular speech
patterns. A variable can be seen as a set of alternative features,
called variants, which can be substituted for one another without changing
the meaning of the word. It was William Labov who introduced this concept
to sociolinguistic studies:

"[...] a linguistic variable is a [linguistic] item [...] that has alternate

realizations, as one speaker realizes it one way and another a different way
[...]" (Wardhaugh 1997, p.140). It is the scientist’s aim to find a
sociolinguistic explanation of the preferred use of one variant instead of
another, and to relate the variation in language to certain social factors. A
linguistic variable can have a number of variants which differ from one
another on a phono- logical level. There is, for example, the present
progressive form of verbs where the -ing- suffix is realized differently by
different speakers. The variable (-ing) has the two identifiable variants [-ing]
and [-in´]. Some linguists might argue that the difference between the two
realizations of the -ing-variable is a morphological difference. (Hudson 1996,
p.43) There are other phonological features like the (h) variable at the
beginning of words which may be realized in two clearly distinct ways. The
word home, for example, does either begin with an audible [h] or it is
pronounced [h=Ø]. The non-occurrence of this phoneme is called zero-
pronunciation. (Wardhaugh 1997, p.138) Other phonological features may
show quantative variation (ibid.), i.e. the scientist has to distinguish between
different degrees of, for example, frontness or backness in the realization of
a vowel. A clear-cut distinction between the variants is not possible and the
identification of the relatively differing variants is much more complicated.
This is also the case when a variable shows multi-dimensional variation, that
is when more than one characteristic of the pronunciation of a vowel has to
be taken into consideration. The variants of a particular vowel-variable may
not only differ in their degree of frontness or backness but at the same time
in their degree of lip-rounding or -unrounding and tenseness or laxness. The
scientist has to decide which pronunciation features are taken into account
when graduating the variants.

As I have already indicated, morphological variation can also be relevant to

studies of linguistic variation. Investigators have looked at the presence or
absence of morphemes, for instance, at the third person singular -s in the
present tense form of verbs. The variable (thinks) ,for example, may be
realized as either [thinks] or [think] . The non-occurence of the -s-suffix
would be regarded as a non-standard realization of the variable. Again one
could argue that this is rather a phonological difference. "[...] differences in
either pronunciation or in morphology [...] are in any case hard to keep
seperate [...]" (Hudson 21996, p.43). Variation on the level of syntax has
been investigated for negated sentences, among other things. The sentence
(He hasn´t got any money either) is a syntactic variable which has the
possible variants [He hasn´t got no money either] (double negation 6) and
[He hasn´t got no money neither] (multiple negation  7). It is rare for
differences in syntax to be investigated by sociolinguists because syntactical
features seem rather unsensitive to variation and are difficult to recognize in
ordinary speech.

Sociolinguists have also looked at the varying usage of lexical items.

Investigating the social distribution of certain synonyms, however, is very
rare as significant data is most difficult to elicit. Lexical variation is rather
relevant in investigating different registers. As has been elaborated
throughout this section, linguistic variation is investigated on all linguistic
levels. Though scientists mainly concentrate on the realization of phonemes
as, for one thing, pronunciation features do occur most frequently when
natural language is being investigated. Secondly, pronunciation is most
sensitive to variation; an individual speaker never pronounces one word
twice in exactly the same way. Thirdly, pronunciation is, in contrast to the
grammar and the vocabulary of a language, less liable to standardization
and at the same time individually marked. Moreover, phonological features
are more quickly adopted than changes in grammar or vocabulary 8.

B. Social Variables
A social variable can be defined as a social factor with an influence on
language variation which, analogous to the linguistic variable, can occur in
various ways. If the choice of one particular variant of a linguistic variable
instead of another is not attributed to regional differences or differences in
style or register, sociolinguists try to explain the variation by quantifiable
factors in society which are known or expected to be influencing language.
There is a wide range of social differences between speakers which have
been found to relate to linguistic variation. The following remarks will be
confined to three very influential social variables: socio-economic status, sex
and race.

1. THE SPEAKER'S SOCIO-ECONOMIC STATUS

The social variable of socio-economic status was found to influence the
realization of, for example, the (-ing) variable at the end of present
progressive verb-forms. In his 1974- Norwich-study, Peter Trudgill
investigated the occurrence of either [-ing] or [-in‘] with a number of
speakers. Dividing the group of speakers into social classes, he found out
that the higher the socio-economic status of the speaker (and the more
formal the style), the more probable the occurrence of the standard variant
[-ing]. In casual style, the least formal speech style, the speakers who were
related to the lowest class did use the non-standard variant [- in‘] in virtually
one hundred percent of the cases.

Trudgill's procedure was designed to find evidence for how speakers' social
status and the degree of formality in their speech relate to the use of the (-
ing) variable. Before starting to conduct this investigation, Trudgill had to
divide his selection of speakers into groups which were to represent five
differerent social classes. According to their occupation, income, education,
place of residence and their father's occupation he assigned the participants
either to the middlemiddleclass, the lowermiddleclass, the
upperworkingclass, the middleworkingclass or to the lowerworkingclass. The
question often raised in this context is, if it is appropriate to stratify society
in this way, reducing various social factors to a single scale and assuming
that the resulting concept of socio-economic status is universally valid. It is
obviously difficult to define groups of people on the basis of their social
background in general. William Labov who established social stratification in
his New York study argued that different social factors are relevant to
different linguistic variables.9 Scientists agree that the criteria taken into
consideration to provide a representative stratification of society must be
chosen with regard to the variable under investigation.10 The second complex
of Trudgill's study was the aspect of speech style. As he was looking for
evidence for the relation between formality in speech and the use of the (-
ing) variable, he had to elicit different degrees of formality. The participants
used the most formal style while reading prepared wordlists paying most
attention to their pronunciation. The most casual speech was produced in
natural conversation, talking to, for example, family members or friends. The
two intermediate speech styles were produced while reading out a reading-
passage respectively while talking to the interviewer. This style stratification,
too is one of the methods developed by William Labov.

2. THE SPEAKER'S SEX

One of the first quantative studies of social variation which was carried out
by John Fischer in 1958 in New England, is concerned with the (-ing) variable,
relating the choice of variants not only to social class but also to the
speaker’s sex. He carried out this investigation among school children,
forming two seperate groups, one group of girls and another of boys. These
two groups were again divided according to social status. Fischer found out
that the lower class boys used the most the non-standard variant [-in‘] while
the middle class girls clearly preferred the RP- standard variant [-ing]. The
choice between the two variants did not only depend on the speaker's sex
and their location on the social scale, but also on the degree of formality of
the situation. In other words, the children adjusted their pronunciation of the
variable in question to the situation in which the speech act took place. They
payed least attention to the use of the RP-variant in casual speech while
talking to their class mates, for instance. One could speculate that girls
rather aspire to be associated with higher social status than boys do. What is
obvious, though, is that the (-ing) variable has social significace not only for
the fieldworker, but for the speakers as well.
The (-ing) variable is generally regarded as a social marker in the English
speaking world, i.e. it carries social information about the speaker. People
are aware of markers which show class- and style stratification. Using the
variant [-in'], which is stigmatized in parts of the English speaking world,
associates one's pronunciation with lower-class speech. A social marker can
become a stereotype as soon as it is conciously varied in order to identify
oneself with a certain group of speakers.

3. THE FACTOR OF RACE

William Labov’s New York City study, which was carried out in 1972,
illustrates the factor of race and its relation to the use of a certain linguistic
features. Together with other scientists, Labov investigated the speech of
Afro-American adolescents which proved to be different from the speech of
"white" Americans and others by various features. Labov found out that the
variable (is), i.e. any present tense form of the verb to be, with its variants
[is] and [is=∅] is the most distinctive feature, and that "white" speakers
prefer [is] in most of the cases whatever their social status. For the "black"
speakers it appeared that the more they identified with the "black
community", the more frequent was the use of the variant [is=∅]. This
manner of identification with a certain group of people by means of a specific
linguistic feature was investigated by Labov among Afro-American teenagers
in Harlem. He divided his selection of speakers into four groups, identifying
"core members" of a gang called "the Jets", "secondary members",
"peripheral members" and "non-members" (Hudson ²1996, p.185). He found
out that the closer the speaker's relation to the gang, the more often
occurred the socially marked variant [is=∅]. Even the non-members
occasionally used this stereotype identifying themselves with the "black
community", but at the same time they distanced themselves from the gang,
using it less often than any of the members. This study illustrated that
linguistic variables may be employed for the purpose of identification with a
particular group or speech community which is defined by a particular social
variable.

III. THE COLLECTION OF DATA

Before beginning with the actual research work and the collection of data
about a particular sociolinguistic phenomenon, the fieldworker propounds a
hypothesis, i.e. he defines an assumption of the relation between a particular
social and a linguistic variable. He has to choose a social variable which he
expects to be relevant to the variation of a linguistic feature. Usually the
linguist chooses a linguistic variable which has been previously observed by
other scientists and which is therefore likely to be socially significant. Some
scientists even fall back upon predetermined lists of variables which have
been made by dialectologists or other linguists. The fieldworker also has to
decide in what way he expects the two variables to relate to each other. This
initial decision already contains the problematic aspect of social stratification
as the variable needs to be appropriately graduated. He then has to settle
which variants of the linguistic variable carry social information at all, and to
which social variant they will probably relate to.

The first step in collecting sociolinguistic data is the preparation of a

questionnaire which is necessary for the performance of the interviews. It is
designed with regard to the hypothesis. It is important to elicit sociolinguistic
information as definite as possible, either confirming or refuting the
scientist's assumptions. Creating this questionnaire, the scientist has to
deliberate on the best way of bringing about representative data without
which the interview becomes subject to unnatural conditions. This might turn
out to fatally influence the results of the interviews. Formulating the
questions, the fieldworker has to think over which words or grammatical
constructions are suitable for the interviews. He has to bear in mind that
the linguistic environment around the varied feature "[...] systematically
influences the frequency with which each variant could be expected to
appear" (Trudgill, 1984, p.246). It was again William Labov who found out
that neighbouring phonemes have an effect on the pronunciation of the
varying feature. If the linguist is interested in variation of linguistic features
according to different degrees of formality, he might use William Labov's
concept of style stratification which is generally regarded as a sensible
method to elicit different speech styles.

The second step in this procedure is the selection of speakers that will be
participating in the interviews. The sociolinguist has to formulate a set of
qualifications with regard to the expected results. In other words, the
participants have to fit the social variable(s) under investigation. This rather
subjective method of selecting speakers is called judgement sample and it
has been criticized for leading to misinterpretation. The alternative method,
the random sample, provides an equal chance to be selected for everyone in
the population to be sampled. The choice of participants is arbitrary and, on
the one hand, this makes the method much more objective. On the other
hand, though, a randomly selected group of speakers will have to consist of
a much greater number of individuals as only a few of them will be relevant
to the study. Consequently, most of the recorded material will be identified
irrelevant. Therefore, judgement sample is generally preferred in
sociolinguistic studies and scientists try to avoid misleading subjectivity at
this early stage of investigation. Obviously, this part of the research work is
rather lenghthy as a large number of speakers is required to guarantee a
representative amount of data. Moreover, it is important to ensure that the
selected speakers are all willing to be interviewed and recorded. This is the
basic requirement for consistency in speech. As soon as the fieldworker is
satisfied with the sample of speakers and their grouping, and as soon as he
knows what he will be looking for during the interviews, he will start to make
recordings. The presence of a tape-recorder during the interview, however,
represents a problem which is referred to as the observer's paradox . It is
obvious that the situation of a recorded interview with a linguist is a rather
formal one: "Won't speakers become self- conscious, and try to make their
speech more standard in the presence of a tape-recorder?" (Spolsky, 1998,
p.10). Although people's speech is more formal under such conditions than it
would be if they were talking to friends or family members, the apparatus is
the most important tool in the sociolinguistic interview. Some scientists have
tried to alleviate this problem. Lesley Milroy, for example, carried out a study
in Belfast being part of the social network of the speakers. She was accepted
as a friend, consequently her presence did not increase the formality of the
speech. This is, of course, an exceptional case as usually the fieldworker
comes into the participant's homes as a stranger, and the problem of the
observer's paradox cannot be overcome because everyday language cannot
be analysed without making tape-recordings first.

After the data collection has been completed, the investigator starts with the
identification of the variants by listening to the recordings. This is a rather
difficult stage of the study so far as the linguist's subjective expectations
might influence the recognition of the occurrences. This problem mainly
applies to studies concerned with phonological variation. Pronunciation
features may differ in several phonetic dimensions. It is not always possible
to reduce the analysis of a linguistic variable to one distinct feature. The
fieldworker has to count the occurrences of each variant for each text
separately to be able to compare different figures for different texts. He
starts, for instance, by counting the occurrences of variant a) in the casual
speech of group 1), he continues by counting the occurrences of variant b) in
casual speech of group 1) and so on, until he reaches the last group 4) to
count the occurrences of variant c) in most formal style. To make the
comparison easier, the occurring figures are reduced to percentages. It
matters then to discover the differences between the texts and their
relevance to the hypothesis. Some differences might not be statistically
relevant at all, but if they are they require sociolinguistic explanation.

The scientist can then start to interpret the data, bearing in mind that the
phonetic environment can make speakers favour one variant instead of
another and that his hypothesis is not fully confirmed. The stage of
interpreting the data is certainly the most difficult one, as it involves some
important decisions. First of all, the speech patterns need to be precisely
described and explained. The study is regarded as succesful when the
findings can be generalized and applied to a level outside the actual
investigation.

IV. CONCLUSION
In the course of my work at this paper, I realized that the field of
sociolinguistics is very extensive and difficult to survey. When I was looking
for relevant literature on my topic, I found it hard to decide which aspects I
should take into consideration and which to leave out. I noticed that many
different topics in this field of linguistics are closely related to one another
and it was difficult to keep them separate. As this field seems to be subject
to constant controversial discussion among sociolinguists, it is difficult to find
a central theory which would help to find one's way through the range of
opinions, experiences and approaches of the experts.

I decided to shortly discuss language variation in general in my introduction

in order to help the reader locate the level of Social Variation. There are of
course many aspects of this topic which, in this paper, have only been
touched upon as I have confined myself to a general introduction. There is,
for example, a wide range of investigations which were carried out under
very different circumstances. I think it would be interesting to directly
compare a selection of studies and to evaluate the different proceedings. I
used a few sociolinguistic studies to illustrate the general procedure but did
not go into detail about one particular research work as this would have been
beyond the scope of an "Einführungsseminar"- termpaper.

LIST OF REFERENCES

- Cheshire, Jenny. 1982. Variation in an English Dialect. Cambridge:

University Press.

- Fasold, Ralph. 1990. The Sociolinguistics of Language. Cambridge:

Blackwell.

- Hudson, R.A. ²1996. Sociolinguistics. Cambridge: University Press.

- Hughes, Arthur et al. ³1996. English Accents and Dialects. New York: Oxford

University Press.

- Labov, William. 1966. The Social Stratification of English in New York

City. Washington D.C.: -

- Milroy, James. 1992. Linguistic Variation and Change. Oxford: Blackwell.

- Spolsky, Bernard. 1998. Sociolinguistics. New York City: Oxford University

Press.

- Stork, Hartmann. 1972. Dictionnary of Language and Linguistics. London: -

- Trudgill, Peter. 1974. The Social Differentiation of English in
Norwich. Cambridge: University Press.

- --- (ed.). 1984. Applied Sociolinguistics. London: Academic Press.