Social decisions are sometimes made by single individuals or small groups and sometimes by a widely encompassing set of traditional rules for making the social choice in any given situation, e.g., a religious code.1
The last two methods of making social choices are in a sense extreme opposites, developments of conflicting tendencies in a democracy. The rule of a single individual is the extreme of administrative discretion, the rule of a sacred code the extreme of rule by law. But in dynamic situations the rule of a sacred code leads by insensible steps to dictatorship. The code needs interpretation, for conditions change, and, no matter how explicit the code may have been in the first place in determining how society shall act in different circumstances, its meaning becomes ambigous with the passage of time. It might conceivably happen that the job of interpretation passes to society as a whole, acting through some democratic process–“vox populi, vox dei.” Or it can happen that interpretation passes to the hands of the people individually and not collectively; in this case, as soon as differences of opinions arise, the religious code loses all its force as a guide to social action....But more likely, in view of the authoritarian character of the sacred code, the interpretation will pass into the hands of a single individual or a small group alone deemed qualified....The classification of methods of social choice given here corresponds to Professor Knight's distinction among custom, authority, and consensus, except that I have subdivided consensus into two categories of voting and the market. (F.H. Knight "Human Nature and World Democracy" in Freedom and Reform,...1947, pp. 308-310.)---K. Arrow (1951) 'Introduction" Social Choice and Individual Values, p. 1.
As regular readers know, I have been researching Arrow's reflections on uncertainty and his impact(s) on and exchange(s) with Rawls for half a decade now; I was alerted by one of my mentors in the history and philosophy of economics, David M. Levy, to the treasure trove(s) of philosophical issues in the vicinity. I never met Arrow, but I did email him once, in 2010, to learn about his readings in the history of physics (due to a comment by Koopmans suggesting that Arrow was Koopmans' source for some of his remarks on the methodology of physics in his exchange with Vining). But I never get a response (perhaps my note ended up in his Spam-box). I did not try very hard to meet or contact Arrow (for an overview of his life and influence see the New York Times obituary) because I wanted to sort out my own (critical) views on his work on uncertainty before I exposed myself to his influence.
In A Theory of Justice, in a footnote (recall my post here), Rawls, a careful reader, calls attention to the significance of some footnotes in Arrow's Social Choice. The first footnote (quoted above) articulates a political interpretation of religion: religion is understood as (i) the storehouse of law-like "rules" for (ii) decision making in particular circumstances. So, religion provides both historical memory -- a way to codify experience -- and a means (iii) to help small groups decide on behalf of the collective. That the rules are supposed to be (in their ideal form) law-like is made clear at the start of the footnote.
The rest of the footnote is a comment on the instability of sacred codes as species of the rule of law. According to Arrow this rule of law is not a stable equilibrium and devolves (on a slippery slope with insensible steps -- one wonders what measure Arrow is assuming here -- to serfdom) into dictatorship. Arrow insists that the cause of such instability is the inability to freeze meaning ("meaning becomes ambiguous with the passage of time"). [Quine, who knew his Arrow (see his autobiography), would appeal to a similar argument in one of his attacks on Carnap's defense of analyticity by way of stipulative definitions]. The reason for this is, as Arrow indicates, that (the science of) interpretation cannot be turned into a consensus-generating device.
My aim here is not to expose the somewhat shaky foundations of Arrow's sociology of religion nor to explore why he decides to engage with Frank Knight from the very start of Social Choice. (As he indicates in the same footnote, he is relying on Tawney's Religion and the Rise of Protestantism.) But it is worth noting that the way Arrow has articulated the point, the very same argument applies, it seems. to a secular rule of law with authoritative judges as secular elite, priesthood. On Arrow's view the rule of law is inherently unstable and always runs the risk of devolving because interpretations cannot be stabilized. The rule of law need not devolve in dictatorship, of course (it need not be intrinsically authoritarian), after all interpretation can conceivably "pass" to "society as a whole."
What is notable, however, is the claim that "as soon as differences of opinions arise, the religious code loses all its force as a guide to social action." Arrow is a very precise writer, and this is no slip of the pen. On his account without consensus (or full agreement) the rule of law (religious or secular) becomes impotent. That is to say, the instability of interpretation must undermine the authority of the religious code/rule of law which apparently must speak with a single voice. Once this is in place Arrow can turn to his formal exploration of social choice. One can respect Arrow's formal skills, his mentorship of many influential students, and his many contributions to economics, not the least general equilibrium theory, and also recognize that despite Arrow's view, there may be versions of authority (even religious authority) that can co-exist in creative tension with discordant interpretation(s) at a given time. In fact, one may well understand the nature of civilization as accepting this cardinal fact.