Speech community (Dell Hymes 1972)
et Sprechbund 'speech area'

Dell Hymes, Models of the Interaction
of Language and Social Life [1972],
repris dans Dell H. Hymes, Foundations in Sociolinguistics.
An Ethnographic Approach
, Philadelphia,
University of Pennsylvania Press, 1974;
reprint London, Tavistock, 1977, pp. 29–66.

(47) Speech community is a necessary, primary concept in that, if taken seriously, it postulates the unit of description as a social, rather than linguistic, entity. One starts with a social group and considers the entire organization of linguistic means within it, rather than start with some one partial, named organization of linguistic means, called a "language." This is vital because the notion of "a language" can carry with it a confusion of several notions and attributes that in fact have to be sorted out.

The first confusion is between the notions of a speech community and a language. Bloomfield (1933). Chomsky (1965) and others have in effect reduced the notion of speech community to that of a language, by equating the two. The result is to make "speech community" itself a redundant concept, having no part to play in research, beyond honoring its definitional foundations with its nominal presence. Definition of a speech community in terms of a language is inadequate to the bounding of communities, either externally or internally. Externally, the linguistic and communicative boundaries between communities cannot be defined by linguistic features alone. Forms of speech of the same degree of linguistic difference may be counted as dialects of the same language in one region, and as distinct languages in another, depending on the political, not linguistic, history of the regions. This is so in parts of Africa, for example, and lies beneath the appearance of linguistic neatness in Europe. Were the standard languages removed from above them, a mapping of Europe's linguistic units would look much more like native North America. With regard to internal bounding of a community, two different conceptions have been advanced. Many have implicitly assumed a "natural" unity among members of a community, in virtue solely of identity, or commonality, of linguistic knowledge […]. Bloomfield. and some others following him, have postulated a quantitative measure of frequency of interaction as defining a community. It is clear from work of Barth (1969), Gumperz, Labov, Le Page, and others that definition of situations in which, and identities through which, interaction occurs is decisive […].

(49) Neustupný has coined the term Sprechbund 'speech area' (parallel to Sprachbund 'language area') for the phenomenon of shared features of speaking across language boundaries. Thus, Czechoslovakia, Hungary, Austria and southern Germany may be found to share norms of greeting, acceptable topics of inquiry, what is said next, etc. If sharing of grammatical knowledge of a form of speech is not sufficient, neither is sharing of knowledge of rules of speaking. A Czech who knows no German may belong to the same Sprechbund (and get by in it with a certain tolerance and goodwill), but not the same speech community, as an Austrian. The notions of Sprachbund and Sprechbund are oriented /50/ toward cultural patterns as attributes of communities. When we consider corresponding phenomena from the standpoint of persons, we see even more clearly that provenance and the dimension of use are quite distinct. Persons often command more than a single form of speech, of course, and may command knowledge of more than one set of norms as to speaking. The range of languages within which a person's knowledge of forms of speech enables him to move may be called his language field. The range of communities within which a person's knowledge of ways of speaking enables him to move communicatively may be called his speech field. Notice that the two are distinct. A scholar's language field may not entail communicative participation for some of his languages (the great French linguist Meillet is said to have spoken and written no language other than French, although he read many); the kind of participation might be said to be in cultural worlds, rather than, or only indirectly, in communities. Again, one's command of a certain language (so identified by provenance of its resources) may be particular to one's local community, so that the command does not permit easy access to other communities in which the same "language" is known. On the other hand, the knowledge of speaking rules required to move within a field larger than one's own community may be complemented by quite minimal command of another form of speech, far less than would be required for normal participation in the other communities.

(50) To participate in a speech community is not quite the same as to be a member of it. Here we encounter the limitation of any conception of speech community in terms of knowledge alone, /51/ even knowledge of patterns of speaking as well as of grammar, and of course, of any definition in terms of interaction alone. Just the matter of accent may erect a barrier between participation and membership in one case, although be ignored in another. […]

(51) A speech community is defined, then, tautologically but radically, as a community sharing knowledge of rules for the conduct and interpretation of speech. Such sharing comprises knowledge of at least one form of speech, and knowledge also of its patterns of use. Both conditions are necessary.