Halliday describes how, for many years, writings on linguistic “behaviour” opened by criticizing linguists who limited themselves to the “code.” This limitation, he notes, had been “accepted as a fact of life,” even though it resulted from a specific historical phase in which code and behaviour were “rigorously held apart.” Only recently had the two begun “quietly merging again,” restoring a view of language “simultaneously as system and as process.”
He insists that this renewed interest is not a return to a linguistique de la parole or theory of performance. The two notions are “in fact opposed,” because a performance model accepts the separation of code and behaviour and then studies behaviour “as if it was unrelated to any code.” This misunderstanding stemmed from the dominance of competence–performance frameworks, where actual utterances fall “outside the scope of linguistics.” Halliday cites Lyons’ 1968 claim that linguistic theory “cannot be concerned with the production and understanding of utterances in their actual situations of use.”
This stance, he argues, reflects the long tradition of philosophical grammar, which studies the code “in isolation from behaviour.” Its lineage runs from the Modistae through rationalist grammars to contemporary formal approaches such as transformational theory. Philosophical grammar depends on a polarity between the ideal and the actual, allowing the linguist to devote attention only to the ideal, since anything too distant from it fails to meet the “rigorous demands” of the theory.
Such an ideology, Halliday notes, inevitably provoked reactions. Movements demanding a study of the actual appeared revolutionary but were “reactionary,” because they accepted the same dichotomy between the “pure” and the “contaminated,” merely shifting interest to the latter.
Ethnographic and descriptive grammars, by contrast, do not rely on such a concept of purity. They “attempt to handle code and behaviour under a single rubric.” Halliday connects this orientation to classical linguistic theory—attention to auctores rather than artes—and to later figures such as Boas, Sapir, the Prague School, Malinowski, Firth, and the glossematic linguists Hjelmslev and Uldall.
The two traditions are not inherently incompatible, yet in the mid-twentieth century they drifted so far apart that communication “almost” broke down. Halliday attributes this partly to Chomsky’s formalization of American structuralism, which reaffirmed the search for “a formal theory of language as code.” This emphasis placed structuralists at odds with European linguists attempting to account for system and process together. Halliday cites Hjelmslev as a key example of a theory in which code and behaviour are inseparable “as system and process,” though this remains “widely misunderstood.”
For Halliday, the central problem is that formal models produce an excessively reduced “idealized picture of language.” When this limitation became evident, new attempts emerged to bring behaviour back into the scope of theory, such as speech act theory and notions of “communicative competence.” Yet a model that excludes behaviour risks reducing language to an abstraction that “ends up by painting itself into a corner.” Only by acknowledging the interdependence of code and behaviour can linguistic theory adequately describe human communication.
Bibliographic reference:
M. A. K. Halliday, “Language as code and language as behaviour: a systemic-functional interpretation of the nature and ontogenesis of dialogue”, in Semiotics of Culture and Language. Volume 1: Language as Social Semiotic, edited by Robin P. Fawcett, M. A. K. Halliday, Sydney M. Lamb and Adam Makkai, Bloomsbury Academic Collections.