In a recent paper for a philosophy supervision, following a lead in Kraut (2007, 39n), I used a quote from Ian Hacking which seemed to quote Michel Foucault, “Objects constitute themselves in discourse (Hacking 1979, 51)”. When my supervisor read this he remarked on the unusual construction saying “What can he mean by the active mood in “Objects constitute themselves in discourse”? Is that different to saying that objects are discursively constituted? (Maybe we should check the French)”. Which is where the ‘fun’ started. Hacking does not give a precise reference for this ‘quotation’. He does give three works in a references section at the end of the article, viz., Michel Foucault, Les Mots et les choses, (Paris: Gallimard, 1966), translated as The Order of Things, (London: Tavistock, 1970), The Birth of the Clinic, (London: Tavistock, 1973): 199 and Archaeology of the Human Science; A Scetch of a History, [sic.] forthcoming.
Now a Google search of the precise terms “Objects constitute themselves in discourse” leads only to results which contain Hacking’s original material (1979, 51) or derivations of it or references to it. A search without the quotation marks gives further references, one of which is to a book chapter by Richard Rorty (“Foucault and Epistemology”, in Hoy, D. (ed.) 1996 (1986), 41-49) Rorty’s text is worth quoting in full, because it leads to a dichotomy.
One might try, however, for looking for a theory of how ‘objects constitute themselves in discourse’. When Foucault uses phrases such as this, I think he is offering the following account of his relation to the epistemological tradition: ‘Whereas Descartes and Locke and Kant and the positivists and the phenomenologists have assumed that the job of signs was to represent pre-existent reality (even if only phenomenal reality, constituted by consciousness), I will show you a new way to look at what people say. From this new perspective, you will not see words as linked to things by relations like “impression”, or “symbolization” or “synthesis” or “reference” or “truth”. Instead you will see them as nodes in a network of texts, and this network as making up “practices that systematically form the objects of which they speak”. (Rorty 1996, 42)
Rorty’s first quote in this paragraph is a reference to Hacking’s original article and its derivatives, but his final quote is from Foucault himself, “A task that consists of not – of no longer – treating discourses as groups of signs (signifying elements referring to contents or representations) but as practices that systematically form the objects of which they speak (1987, 49)”. Foucault’s original has, “… mais commes des pratiques qui forment systématiquement les objets dont ils parlent (1969, 67)”.
And the dichotomy is this: we have Hacking’s original sense of ‘objects constituting themselves in discourse’, a sense which has been perpetuated by, inter alia, Rorty, Kraut and others and for which no direct text by Foucault can be found (by me) and then the sense quoted by Rorty later on, for which the Foucauldian text is available, of objects being systematically formed by the practices of discourse.
As a final aside, Rorty’s text “One might try, however, for looking for a theory of how ‘objects constitute themselves in discourse’”, when translated into French, is given as, “On pourrait essayer malgré tout, en cherchant une théorie de “la manière don’t les objets se constituent eux-mêmes en discours (Hoy (éd.) 1989, 55-56)”. Now this construction ‘la manière don’t les objets se constituent eux-mêmes en discours’ contains a pronominal verb which, according to my sources in the Languages Department at the Open University, is best translated as something like, ‘the manner in which objects themselves are constituted in discourse’. Perhaps Hacking (and the rest) have inadvertently committed a calque in virtue of someone’s mistranslation. If only we had the text.
I have asked Professor Hacking what his original source was, but have not yet had a reply. Unfortunately, we cannot ask Rorty or Foucault. If anybody can throw any light on where Hacking might have got his unusual construction from, I (and my supervisor) would be pleased to hear.
Foucault, M. 1969. L’Archéologie du Savoir. Paris. Gallimard
Foucault, M. 1973. The Birth of the Clinic: An Archaeology of Medical Perception. London. Taylor & Francis.
Hacking, I. 1979. “Michel Foucault’s Immature Science” Noûs, 13:1. 39-51.
Hacking, I. 2002. Historical Ontology. Cambridge MA. Harvard University Press.
Hoy, D.C. (Éd.) 1989. Michel Foucault: Lectures Critiques (traduit de l’anglais par Jacques Colson). Bruxelles. De Boeck-Wesmael.
Hoy, D.C (Ed.) 1996 (1986). Foucault: A Critical Reader. Oxford. Blackwell Publishers Ltd..
Kraut, R. 2007. Artworld Metaphysics. Oxford. Oxford University Press.
 In Historical Ontology, 2004, Cambridge MA, Harvard University Press (p.98), Hacking uses the ‘quote’ from Foucault originally appearing in the Nôus article (1979, 51), but he also goes on to use the same wording in his own right, “Since most, if not all, knowledge is “immature” in this way, attempting to understand how objects constitute themselves in discourse must be a central topic, not exactly of theory of knowledge, but of what I would now call historical ontology (2004, 98)”.
 See, e.g., this thesis by Jacqualine J Heywood, p.37 https://www120.secure.griffith.edu.au/rch/file/15aeb56b-02e7-8963-ac4f-bc743325b59b/1/02Whole.pdf