System, Norm and Speech
In 1952, Coseriu published the essay Sistema, norma y habla (“System, Norm and Speech”), and soon this distinction entered canonical knowledge of linguistics. In the essay, Coseriu deals with Saussure’s pair of terms, langue and parole, and comes to the conclusion that these are insufficient to cover the entire reality of language.
His main point of criticism maintains that the mere distinction between the functional system on the one side, the langue, and its realization on the other, the parole, inevitably leaves various facts unaccounted for. Already the Prague phonologists had pointed out what their forerunners in the School of Kasan had proven: that a phoneme, i.e. the notion of a sound, can have varying realizations in a specific language, e.g. because of varying phonetic contexts. One of the examples on which Coseriu bases his argumentation is the Spanish vowel system. In the language system, the langue, there are only five vowel phonemes: /a/, /e/, /i/, /o/, /u/. But on the level of the parole, endless realizations correspond to the phonemes. Furthermore, the phonemes /e/ and /o/ have a regular, traditional distribution depending on their phonetic context, according to which they are articulated as open or closed. As this open or closed articulation is the normal and expected one, and a different pronunciation may be possible, but not common, this realization can be identified as the Spanish Norm. Thus, the norm is the common, traditional realization of the system in a linguistic community. It not only exists in the phonetic, but in all fields of language. One mustn’t confuse it with the prescriptive norm, the exemplary norm in a community (cf. Das Korrekte und das Exemplarische), a short interview with Coseriu on this subject in German). The difference between system and norm is of great relevance for all issues of system linguistics. On the one hand, the norm goes beyond the system, as it contains more information; e.g. in the case of the Spanish vowel phonemes the information whether they are to be pronounced open or closed. On the other hand, the system also goes beyond the norm, as the system not only contains traditional realizations, but being a “system of possibilities” also such realizations that are not (or not yet) traditional. So for example the system of a language will only contain the rules for generating words, but not the specific realizations. A word such as “shavable”, for example, follows an English word generating pattern, but probably not the English norm. It is a possible word by means of the system (and as such it virtually already exists and is understood), but it is not common in English, not “normal”.
In the history of linguistics, there have often been tendencies to reduce language to usage. Especially in the past few years – notably in the context of cognitive linguistics – it has been attempted to define language exclusively via usage.
Structural linguistics is repeatedly accused of not recognizing the importance of usage as the core dimension of language.
Coseriu’s distinction between system and norm takes into account that linguistic systems in their nature as systems are organized by functionally opposite elements, and yet that a system can only be mediated by speech. The process of mediation, however, not only imparts the system, but also its common realization.
Sistema, norma y habla (con un resumen en alemán), Montevideo; also in RFHC 9, p. 113-181; (List of publications No. 8) Reprinted in Teoría del lenguaje y lingüística general. Cinco estudios, Madrid; 2nd ed. Madrid 1967.