In the analogy, the prize represents the benefits of cooperation (i.e., of not having to fight over everything). The reason why the no-communication case relates to real-world negotiation is elaborated on by Schelling in The Strategy of Conflict (see pp. 77-84 of the PDF (67-74 by the printed numbers)). Basically, it comes down to the need to establish an ultimatum, as well as the fact that communication is impaired by mistrust.
The reply to both objections is essentially “If not here, where?” Only one option has a salient property that distinguishes it from the others, and so we choose it because we expect that they will expect that we will expect that they… [etc.] …will choose it. This is the same mechanism by which any legal code acquires its respect, despite not being regarded as “an end in itself.” In the same way that seemingly pointless math can help achieve coordination in the example, seemingly pointless deontology can help achieve coordination in human affairs.
Alternative legal codes won’t acquire near-universal respect just because people think, “well we’ve got to use some standard, and this standard is ‘salient’, so we might as well abide by that.” With regard to any given conflict of interests, there might be any number of “salient” standards one can choose, and individuals will always be inclined to support the salient standard that is pursuant to their own interest. Which brings us to the point that the only kind of saliency that can command near-universal assent to a rule is an ex ante tendency for the establishment of the rule to benefit any given person.
I’m not certain about this matter one way or another, since I haven’t done the kind of extensive empirical research that would be needed to determine it. However, I would like to know what your supporting evidence for this is
As Mises argued, you can see it in the clear state of affairs that political parties nigh invariably promise higher living standards to their flock. This in turn stems from the clear state of affairs that people generally vote in favor (according to their often flawed judgment regarding means) of their own pocketbooks and their own comfort.
De Jasay suggests that the influence of the state over people’s preferences is not just a result of its influence over their understanding of the efficacy of means: e.g. “Once begun, the addictive nature of redistribution sets in motion unintended changes in individual character and the family and group structure of society.”
Individual character has as much to with means as does any other product of human custom.
A number of studies have shown that increased material wealth does not lead to increased happiness, at least after a certain level (1, 2).
In fact, it appears that relative status and dominance are a major influence over behavior (e.g., this; the anecdotal evidence that terms like “tall poppy” and “crab bucket” exist).
I think Gilles Saint-Paul makes an important point here regarding the dubious (and I think feather-bedding oriented) “happiness economics”:
In fact, people always make choices that are not supposed to make them happy but instead are linked to how they view themselves and what goals they have in life: career, status, the transmission of properties, abiding by some community values, etc. For example, an individual could report a low level of happiness because he is concerned about his children. That does not mean that he made the wrong choice by having children. In other words, happiness is only one part of individual welfare, not the whole of it.
Also, as so often occurs with statistical research, econometricians have also gleaned from the data conclusions that directly oppose the “Easterlin Paradox” and “hedonic treadmill” 1, 2, 3, 4, 5.