Abstract A growing theoretical and research literature is challenging the presupposition, common to both ‘descriptive’ and ‘prescriptive’ linguistics, that rules are constitutive of language. This paper briefly surveys some of the arguments, and adopts a position that excludes rule-governedness as an inherent property of language. Nevertheless, rules exist, and the large and continuing demand for ‘authorities’ such as grammars and dictionaries attests to the fact that they play an important role in the ecology of language. The paper explores some ways in which this role can be appropriately understood. A starting-point is the distinction between first-order languaging and second-order language: rules, as a socio-cultural imposition on language, relate to the latter. They are not defining characteristics of language, but are manifestations of a human need to create them. The rules (or ‘formulae’) are, in fact, the outcomes of a continual process of formulation. Equally important to understanding the ecological role of formulae is implementation, which is also an on-going process. Implementation, however, has received limited attention from linguists. Three areas in which the societal and social-psychological impetus for the implementation of formulae are examined. They are: the confidence in predictability that underpins human attempts to communicate; the belief in a standard language; and the desire to intervene and enhance linguistic behaviour in language pedagogy and other fields of applied linguistics. The paper presents the starting-point of a systematic treatment of the topic, and is intended to stimulate further academic discussion and research.