Varieties of language

. . . language, while existing to serve a social function (communication)

is nevertheless seated in the minds of individuals. (Guy 1980)
The reader should know, however, that this position is controversial.
Unfortunately it appears to be opposed to the view of William Labov, who (as
we shall see) is the most influential of all sociolinguists. Labov takes a very
clear position on this issue, as witness these remarks in a discussion of the
English spoken in the American city of Philadelphia:
. . . the English language is a property of the English speech community,
which is in turn composed of many nested subcommunities. There is no
doubt that Philadelphia speakers of English are members of the larger
community of American English speakers, and the even larger
community of all speakers of English. It might also be said that
Philadelphia is in turn composed of many smaller subcommunities. But
the data presented here show that the linguistic world is not indefinitely
complicated. (1989: 2)

I began this paper with a question about the possible objects of linguistic
description. As far as I can see, the individual speaker is not such an
object. This essay, like other studies of sociolinguistic variation, shows
that individual behavior can be understood only as a reflection of the
grammar of the speech community. Language is not a property of the
individual, but of the community. Any description of a language must
take the speech community as its object if it is to do justice to the
elegance and regularity of linguistic structure. (1989: 52)
The context of these remarks is a long (and impressive) discussion of varia-
tions in a single complicated feature of the English spoken in Philadelphia, in
which he shows that a representative sample of speakers hardly varies at all
even on the finest details. His data are beyond dispute, but they only seem to
show that individuals in Philadelphia are very similar as far as this one feature
is concerned. It does not follow that Philadelphians agree on all features, nor
that every human being belongs clearly to a single community, nor that every
community (however defined) will show the same amount of internal agree-
ment. Moreover, the existence of agreement among speakers does not show
that 'language is not a property of the individual', any more than similarities of
height or income in some population show that height and income are not really
properties of the individual.

2.2 Languages
2.2.1 'Language' and 'dialect'
We shall spend the rest of this chapter looking at the most widely
recognised types of language variety: 'language', 'dialect' and 'register'. We
30
 22 Languages

shall see that all three types are extremely problematic, both from the point of
view of finding a general definition for each one which will distinguish it from
the others, and also from the point of view of finding criteria for delimiting
varieties.
We first need to consider the concept 'language'. What does it mean to say
that some variety is a language? This is first of all a question about popular
usage: what do ordinary people mean when they say that some variety is a
language? Having answered the question in this form, we may or may not wish
to take 'language' as a technical term, and say how we propose to use it in socio-
linguistics. We shall want to do so if we find that popular usage reflects some
kind of reality to which we should like to refer in sociolinguistics, but if we
come to the conclusion that popular usage reflects no such reality, then there
will be no point in defining 'language' more explicitly in order to use it as a
technical term.
One thing that is not in question is the importance of studying popular
usage of the term 'language' simply as part of English vocabulary, along with
'well-spoken', 'chat' and other vocabulary which reflects the parts of our cul-
ture which are related to language and speech. It is part of our culture to
make a distinction between 'languages' and 'dialects' - in fact, we make two
separate, distinctions using these terms, and we may draw conclusions from
this fact about our culturally inherited view of language (in the same way
that we can use vocabulary as evidence for other aspects of culture - see
3.2.1).
We may contrast our culture in this respect with others where no such distinc-
tion is made. For example, according to Einar Haugen (1966), this was the
case in England until the term dialect was borrowed in the Renaissance, as a
learned word from Greek. In fact, we may see our distinction between
'language' and 'dialect' as due to the influence of Greek culture, since the dis-
tinction was developed in Greek because of the existence of a number of clearly
distinct written varieties in use in Classical Greece, each associated with a differ-
ent area and used for a different kind of literature. Thus the meanings of the
Greek terms which were translated as 'language' and 'dialect' were in fact
quite different from the meanings these words have in English now. Their
equivalents in French are perhaps more similar, since the French word dialecte
refers only to regional varieties which are written and have a literature, in con-
trast with regional varieties which are not written, which are called patois. The
point of this discussion is to show that there is nothing absolute about the dis-
tinction which English happens to make between 'languages' and 'dialects'
(and for readers familiar with some language other than English, this discussion
will hardly have been necessary).
31
Varieties of language

What then is the difference, for English speakers, between a language and a
dialect? There are two separate ways of distinguishing them, and this ambiguity
is a source of great confusion. (Haugen (1966) argues that the reason for the
ambiguity, and the resulting confusion, is precisely the fact that 'dialect' was
borrowed from Greek, where the same ambiguity existed.) On the one hand,
there is a difference of size, because a language is larger than a dialect. That is,
a variety called a language contains more items than one called a dialect. This
is the sense in which we may refer to English as a language, containing the sum
total of all the terms in all its dialects, with 'Standard English' as one dialect
among many others (Yorkshire English, Indian English, etc.). Hence the greater
'size' of the language English.
The other contrast between 'language' and 'dialect' is a question of prestige, a
language having prestige which a dialect lacks. If we apply the terms in this
sense, Standard English (for example, the kind of English used in this book) is
not a dialect at all, but a language, whereas the varieties which are not used in
formal writing are dialects. Whether some variety is called a language or a dia-
lect depends on how much prestige one thinks it has, and for most people this
is a clear-cut matter, which depends on whether it is used in formal writing.
Accordingly, people in Britain habitually refer to languages which are unwritten
(or which they think are unwritten) as dialects, or 'mere dialects', irrespective
of whether there is a (proper) language to which they are related. (It would be
nonsense to use 'dialect' in this way intending its 'size' sense, of course.) The
fact that we put so much weight on whether or not it is written in distinguishing
between 'language' and 'dialect' is one of the interesting things that the terms
show us about British culture, and we shall return to the importance of writing
in 2.2.2.

2.2.2 Standard languages

It is probably fair to say that the only kind of variety which would
count as a 'proper language' (in the second sense of 'language') is a standard
language. Standard languages are interesting in as much as they have a rather
special relation to society - one which is quite abnormal when seen against the
context of the tens (or hundreds?) of thousands of years during which language
has been used. Whereas one thinks of normal language development as taking
place in a rather haphazard way, largely below the threshold of consciousness
of the speakers, standard languages are the result of a direct and deliberate
intervention by society. This intervention, called 'standardisation', produces a
standard language where before there were just 'dialects' (in the second sense,
i.e. non-standard varieties).

32
 2.2 Languages

The notion 'standard language' is somewhat imprecise, but a typical standard

language will have passed through the following processes (Haugen 1966; for a
somewhat different list, see Garvin and Mathiot 1956 and Garvin 1959).
(1) Selection - somehow or other a particular variety must have been selected
as the one to be developed into a standard language. It may be an existing vari-
ety, such as the one used in an important political or commercial centre, but it
could be an amalgam of various varieties. The choice is a matter of great social
and political importance, as the chosen variety necessarily gains prestige and
so the people who already speak it share in this prestige. However, in some
cases the chosen variety has been one with no native speakers at all - for
instance, Classical Hebrew in Israel and the two modern standards for
Norwegian (Haugen 1994).

(2) Codification - some agency such as an academy must have written diction-
aries and grammar books to 'fix' the variety, so that everyone agrees on what is
correct. Once codification has taken place, it becomes necessary for any ambi-
tious citizen to learn the correct forms and not to use in writing any 'incorrect'
forms that may exist in their native variety.
(3) Elaboration of function - it must be possible to use the selected variety in all
the functions associated with central government and with writing: for example,
in parliament and law courts, in bureaucratic, educational and scientific docu-
ments of all kinds and, of course, in various forms of literature. This may require
extra linguistic items to be added to the variety, especially technical words, but
it is also necessary to develop new conventions for using existing forms - how
to formulate examination questions, how to write formal letters and so on.
(4) Acceptance - the variety has to be accepted by the relevant population as
the variety of the community - usually, in fact, as the national language. Once
this has happened, the standard language serves as a strong unifying force for
the state, as a symbol of its independence of other states (assuming that its stan-
dard is unique and not shared with others), and as a marker of its difference
from other states. It is precisely this symbolic function that makes states go to
some lengths to develop one.
This analysis of the factors typically involved in standardisation has been
quite widely accepted by sociolinguists (for more details and examples, see
Fasold 1984, Milroy and Milroy 1985, Haugen 1994). However, there is ample
scope for debate and disagreement about the desirability of certain aspects of
standardisation. For instance, it is not essential either that standardisation
should involve matters of pronunciation as well as of writing (Macaulay 1973),

33