Romaine (1994)
The term 'sociolinguistics' was coined in the 1950s to try to bring together the perspectives of linguists and sociologists to bear on issues concerning the place of language in society, and to address, in particular, the social context of linguistic diversity.

Since the late 1950s mainstream linguistics has been conceived as a largely formal enterprise increasingly divorced from the study of languages as they are actually used in everyday life.

Sociolinguistics has close connections with the social sciences, in particular, sociology, anthrology, social psychology, and education. It encompasses the study of multilingualism, social dialects, conversational interaction, attitudes to language, language change, and much more.

Theoretical sociolinguistics is concerned with formal models and methods for analysing the structure of speech communities and speech varieties, and providing a general account of communicative competence. Applied sociolinguistics deals with the social and political implications of fundamental inequalities in language use in various areas of public life, for example school, court, etc.

Macro-sociolinguistics, sometimes also referred to as the 'sociology of language', takes society as its starting-point and deals with language as a pivotal factor in the organization of communities. Micro-sociolinguistics begins with language and treats social forces as essential factors influencing the structure of languages.

The book is called 'language in society' rather than 'language and society' so as to emphasize the fact that the study of society must accord a place to language within it at the same time as the study of language must take account of society.

CONTENTS
1. Language in Society/Society in Language
2. Language Choice
3. Sociolinguistic Patterns
4. Language and Gender
5. Linguistic Change in Social Perspective
6. Pidgin and Creole Languages
7. Linguistic Problems as Societal Problems
8. Conclusions

Chapter 1: Language in Society/Society in Language

The narrowing of modern linguistics to the study of grammar has ruled out investigation of many interesting questions about how language functions in society.

The purpose of this chapter is to explain why the notions of language and dialect are fundamentally social and not linguistic constructs.

Language versus dialect
The term 'dialect' has generally been used to refer to a subordinate variety of a language. For example, we are accustomed to saying the English language has many dialects. These dialects may be of different kinds. A 'regional dialect' is a variety associated with a place. Dialects of a language tend to differ more from one another the more remote they are from one another geographically. In this respect the study of dialects or dialectology has to do with boundaries, which often coincide with geographical features such as rivers and mountains. Boundaries are, however, often of a social nature, e.g. between different social class groups. In this case we may speak of 'social dialects'. Social dialects say who we are, and regional dialects where we come from. The term 'dialect' also has historical connontations. Historical linguists, for instance, speak of the Germanic dialects, by which they mean the ancestors of language varieties now recognized as modern Germanic languages. The entities we label as the 'English language' or 'Flemish dialects' are not, however, discrete. Any variety is part of a continuum in social and geographical space and time. The discontinuities that do occur, however, often reflect geographical and social boundaries and weaknesses in communication networks.

Language and dialect in Papua New Guinea
A preliminary example from north-west New Britain in the Pacific region will illustrate the problems in applying purely linguistic criteria in deciding what counts as a language or dialect.

'Lexicostatistics': the method relies on the counting of percentages of apparent cognates, i.e. related forms meaning the same thing, in a wordlist of one or two hundred items. Those who use this method generally regard varieties sharing between 81 and 100 per cent cognates as dialects within a language. If there are between 28 and 81 per cent cognates, then the varieties count as languages within a family. Fewer cognates indicate a more distant relationship.

Any estimate of the number of languages spoken in an area like Papua New Guinea is fraught with difficulties due to the problems inherent in defining terms such as 'language' and 'dialect'. The very concept of discrete languages is probably a European cultural artefact fostered by processes such as literacy and standardization. Any attempt to count distinct languages will be an artefact of classificatory procedures rather than a reflection of communicative practices. Lexicostatistics will not yield any non-arbitrary technical definition of terms such as 'language', 'dialect', 'family', etc.

Language and dialect in Europe
Other examples from Europe can be taken to illustrate the arbitrariness of linguistic criteria, and the importance of social factors in deciding what counts as a language or dialect.

The West Romance dialect continuum

The Germanic dialect continuum

Danish, Norwegian, Swedish

Max Weinriech's often quoted dictum, 'a language is a dialect with an army and a navy', attests the importance of political power and the sovereignty of a nation-state in the recognition of a variety as a language rather than a dialect. Situations in which there is widespread agreement as to what constitutes a language arise through the interaction of social, political, psychological, and historical factors, and are not due to any inherent properties of the varieties concerned.

The process of standardization is connected with a number of socio-historical factors such as literacy, nationalism, and cultural and ethnic identity. It results in the selection and fixing of a uniform norm of usage, which is promoted in dictionaries, grammars, and teaching. A standard language is a variety that has been deliberately codified so that it varies minimally in linguistic form but is maximally elaborated in function.

Some linguists have found the terms 'autonomous' and 'heteronymous' speech varieties useful as alternative labels to language and dialect.

The term 'language' is employed for a variety that is autonomous, together with all those varieties that are heteronymous upon it. Because heteronymy and autonomy reflect political and cultural rather than purely linguistic factors, they can change.

English: language and dialect

Accent versus dialect
Some linguists make a further distinction between 'accent' and 'dialect'. An accent consists of a way of pronouncing a variety. A dialect, however, varies from other dialects of the same language simultaneously on at least three levels of organization: pronunciation, grammar or syntax, and vocabulary.

Register and style
Varieties can be regarded as a clustering of features. In addition to regional and social dialects, two other varieties often discussed by sociolinguists are register and style.

While regional dialect reveals where we come from and social dialect what our status is, register gives a clue about what we are doing. The concept of register is typically concerned with variation in language conditioned by uses rather than users and involves consideration of the situation or context of use, the purpose, subject-matter, and content of the message, and the relationship between participants. Vocabulary differences - either a special vocabulary or special meanings for ordinary words - are most important in distinguishing different registers.

Examples of register variation: Javanese; 'mother-in-law' language found in most of the Aboriginal languages of Australia

A notion related to register is that of 'style', which can range from formal to informal depending on social context, relationship of the participants, social class, sex, age, physical environment, and topic. Stylistic differences can also be reflected in vocabulary, syntax and pronunciation.

Speech communities and communicative competence
The very existence of languages critically depends on the availability of a social group who claim a variety as their own and maintain its distinctiveness from the varieties spoken by their neighbors. Such a group can be called a 'speech community' and the conventions they share about their speech variety can be called 'communicative competence'.

A speech community is not necessarily coextensive with a language community. A speech community is a group of people who do not necessarily share the same language, but share a set of norms and rules for the use of language. The boundaries between speech communities are essentially social rather than linguistic. This means that terms such as 'language' and 'dialect' are, from a linguistic point of view, non-technical notions since there is no objective way to determine when two varieties will be seen by their speakers as sufficiently similar to warrant calling them the 'same' language. Any attempt to count distinct languages will be an artefact of classificatory procedures rather than a reflection of communicative procedures. Degree of mutual intelligibility is itself a function of the extent of social and other contact between the groups concerned and does not necessarily have much to do with actual linguistic differences.

Patterns of social interaction often transcend language boundaries. The so-called Prague School linguists introduced the notions of 'Sprechbund' (speech bond) and 'Sprachbund' (language bond), with a Sprechbund involving shared ways of speaking which go beyond language boundaries, and Sprachbund, relatedness at the level of linguistic form. A Sprachbund and a Sprechbund may not necessarily coincide. Membership in a community may be established and maintained primarily in terms of interactional rather than language norms.

The term 'communicative competence' is used by sociolinguists to refer to a speaker's underlying knowledge of the rules of grammar