There's a possibly relevant passage in H. L. A. Hart's 1949 paper called 'The Ascription of Responsibility and Rights' (Proceedings of the Aristotelian Society, vol. 49 [1949], pp. 171-194). On p. 188, Hart considers a behaviourist view, on which we analyze a description of a choice partly in terms of "a general hypothetical proposition or propositions to the effect that X would have -responded in various ways to various stimuli, or that his body would not have moved as it did or some physical consequence would have been avoided, had he chosen differently" (p. 188). Hart suggests (?) that the concept of action is normative and holistic. He charges behaviourism with "the common error of supposing that an adequate analysis can be given of the concept of a human action in any combination of the descriptive sentences, categorical or hypothetical, or any sentences concerned wholly with a single individual" (p. 189). He draws an analogy between (on the one hand) the difference between concepts of action and concepts of bodily motion and (on the other hand) the difference between classifying some patch of ground as a piece of earth and as a piece of property (where the latter characterisation relies on legal norms). In both cases, he says, the relevant concepts (of actions and of property) depend on "accepted rules of conduct" (p. 189)

Hart might have been influenced by Friedrich Waismann's paper, 'Language Strata', which was presented to the Jowett Society at Oxford in 1946, but wasn't published until 1953 (when it appeared in Logic and Language [Second Series], ed. A. G. N. Flew [Basil Blackwell] -- Flew says the paper had exercised "considerable influence" after its presentation at Oxford). Waismann doesn't mention behaviourism till the final (and very long) paragraph of his paper (p. 29). He then adds, "An action may be viewed as a series of movements caused by some physiological stimuli in the 'Only rats, no men' [behaviourists'] sense; or as something that has a purpose or a meaning irrespective of the way its single links are produced" (p. 30). Waismann closes with the example of a spoken sentence, which can be regarded as "a series of noises" (akin to mere bodily motions) or as "a vehicle of thought" (which, since they're meaningful, are akin to actions) (p. 31). [I don't think Waismann makes anything like Hart's point about rules, but I'm not sure.]

There's also this from a 1963 paper by May Brodbeck, where she's summarizing an old objection to behaviourism: "We observe only manifest behavior, like a hand going up or pulling a lever, not the internal meaning of an action, like hand-raising or voting. This meaning lies in the logical connections the action has with the complex of desires, intentions, choices, reasons, conventions and moral rules that are all inextricably involved together in social life" (May Brodbeck, 'Meaning and Action' Philosophy of Science, Vol. 30, No. 4 [Oct., 1963], pp. 309-324 (at.p. 310). She attributes this objection to R. S. Peters (The Concept of Motivation [London, 1958]) and Peter Winch (The Idea of a Social Science [London, 1958]).