We want to know what to do. Should we save the tiger, permit more roads or believe in nuclear power? More exactly, we want to know what most to do when there is a range of options, all of them attractive to some, but not all of which can be satisfied: safer or cheaper power, easy access to the countryside or its preservation, tea or tigers. We want to know how policies should be arrived at, and how decisions should be made. Two procedures have been considered, but in their actual forms both democratic and market forces are imperfect mechanisms, delivering results that often seem unacceptable. Of course, it is possible simply to stipulate that the policy that gets the most votes, or the one that wins out on a cost-benefit analysis, is then straightforwardly the one that should be put in place, but this is a view adopted only by a handful of die-hards. Most people recognize that these institutions as we have them are deeply flawed. But what of ideal forms? If we could perfect our voting or pricing mechanisms, then we could have much more confidence in the results they deliver. But, I argued, there is an important sense in which we are here in danger of getting things the wrong way round. For there is no independent notion of what the ideal form of democracy or the market would be, which might then be seen to generate, as a matter of routine, an unbroken series of acceptable verdicts on policy decisions.