John Searle, answered from the texts and cited to the page.
Social reality is constructed out of three resources: assignment of function (we use a log as a bench; we use the heart to pump blood), collective intentionality (the irreducibly we-mode in which two or more agents share an intention, as when several people push a car together), and constitutive rules of the form X counts as Y in context C.1
Institutional facts — money, marriage, governments, professorships, declarations of war, criminal convictions — are layered iterations of the counts-as formula. They are ontologically subjective (they depend on collective acceptance for their existence) but epistemologically objective (we can know that Smith was promoted as we can know that the mountain is tall).2
Collective intentionality is a primitive — it cannot be reduced to individual intentions plus mutual belief without losing the phenomenon. Language is the fundamental social institution, because it is constitutive of the others.345
We still need a more thorough account of rules and institutions. And we need to answer a lot of questions. Are all social facts institutional facts? Are there constitutive rules of, for example, wars and cocktail parties? What makes something into a "constitutive rule" anyway?The Construction of Social Reality, pp. 29–32
Notice that in 1b and 3b the observer-relative statement is epistemically objective; in 2b it is subjective. 'These points illustrate the ways in which all three distinctions cut across each other: the distinction between the intrinsic and the observer relative, the distinction between ontological objectivity and subjectivity, and the di between epistemic objectivity and subjectivity.The Construction of Social Reality, pp. 13–14
There is an enormous difference in the physical appearance of a bonfire and a rusty shovel, but the underlying principle in each case is exactly the same: oxidization. Similarly, there are enormous differences between baseball games, $20 bills, and national elections, but the underlying logical structure is the same.Philosophy in a New Century, p. 31
Not all freestanding Y terms are of this form. When the bank creates money simply by lending money that it does not have, the Declaration looks like this: We make it the case by Declaration that a status entity Y exists with the functions F. Thus in lending Jones the money, the bank makes it the case by Declaration that a certain percentage of the money loaned now exists that did not previously exist.Making the Social World, pp. 100–101
Constitutive rules also regulate, but they do more than regulate; they constitute the very activity that they regulate in the way that I have suggested. The distinction between brute facts and institutional facts, I have argued and will continue to argue here, can only be fully explained in terms of constitutive rules, because institutional facts only exist within systems of such rules.Mind, Language and Society, pp. 136–138
Developed in The Construction of Social Reality (1995), consolidated in Making the Social World (2010). The irreducibility of we-intentions is the central technical dispute with Bratman, Tuomela, and Gilbert — each of whom proposes a different way of reducing or supplementing the we-mode. I hold that the reductions miss the phenomenon: a we-intention is not a special complex of I-intentions plus mutual recognition; it is its own primitive form of intending. The counts-as formula — X counts as Y in C — has been criticized as too narrow (some collective practices emerge without a clear status declaration) and too broad (almost any social fact can be forced into the form). I hold it captures the structural feature that distinguishes institutional facts from brute facts.