Traditionally, for liberals, the central problem in political philosophy has typically been the problem of political obligation. Because of their concern with liberty, and hence the proper limits of state power, liberals have tended to be particularly interested in to what extent citizens are obliged to obey the demands of the institutions of the state. And indeed, the question of political obligation ought to be central to political philosophy. A political philosophy which somehow evaded the questions of what legitimacy the coercive power of the state has, if indeed it has any, and of what conditions are placed on that legitimacy would be in serious difficulties. It would have to either admit its incompleteness or struggle to explain what made it distinctively political, since a central feature of politics, that of the claimed legitimacy of political arrangements, apparently fell outside its scope.
Such a philosophy would not be able to explain when, if ever, resistance to an established power is acceptable, and how far that resistance might go. It would make it unclear whether we should regard the taxman as any different from the thief, or the magistrate as any different from the kidnapper. Arguments about exactly what the state should do may be all very well, but they all rest on and indeed return to the assumption that the state should be in a position to rely, as it must, on its citizens to cooperate in any of its projects at all. That assumption requires justification.
Explaining obligation generally has of course been of significant importance to moral philosophy more generally. Promising, as a generally uncontested source of obligations, has been a typical example. Some explanations have focused on the evolution of the practice of promising. It is clearly useful for members of a group to be able to reliably bind themselves to completing some future action: it builds bonds of trust and allows the completion of cooperative tasks, otherwise impossible because of the uneven distribution of rewards and hence incentives across time and individuals.
Because of this utility of the practice of promising, there is a temptation to regard the obligations which arise from promising as derived from the benefits of the practice as a whole. That, however, seems to be a mistake: to promise is to be bound by something more than the usually fairly small possibility of the undoubtedly significant benefits derived from the practice as a whole disappearing. There is, after all, a sense in which to promise is to deny oneself the possibility of calculating the general moral benefits and costs of fulfilling or not fulfilling the promise. Someone who I had promised some money would typically not be impressed if I had given the money to charity instead, where, let us assume, it would do more good, and rightly so.
Utilitarianism's account of political obligation seems to make the same mistake as its account of promising. It dismisses the liberal concern with a justification of the coercive power of the state by sidestepping it, refusing to see any problem as distinctively political, since the demand to act so as to create the greatest happiness of the greatest number applies across all actions, from both the highly personal to the deeply political. As with promises, the utilitarian explanation of any obligation relies upon the utility, in the strict sense, of that obligation, and so long as political obligation is useful, in that strict sense, the citizens are appropriately obliged.
Likewise, conservatism, by downplaying the importance of individual freedom and stressing that of social heirarchies, sees less to worry about in the state's coercive power than liberals do, as long as that power is exercised in favour of those heirarchies. Libertarians, both of left and right, tend to veer in the other direction, denying the state any legitimacy whatsoever, beyond the protection of sets of prior property rights, either to the full products of one's labour or to whatever one can appropriate. Liberals, though, however they may periodically beg, steal or borrow from these other traditions, tend to regard the question of what grounds the state's authority as central. One might even describe it as one of the marks of liberalism, at least as a political philosophy, even if that does produce odd outliers like Mill, as it certainly captures the centrality of social contract theory, like that of Locke, to liberalism.
The social contract assimilates the relationship of state and citizen to mutual promisers: as long as the state does not violate certain guarantees it made, the citizens are obliged to obey it, and likewise. The problem for such theories is at least two-fold. Firstly, they need to locate the promise in time and space, and secondly, they need to both explain and justify its terms. The first can be got round through implied consent - by failing to remove oneself from the state, or by accepting its benefits, consent to the coercion of the state is given - but the second then looms all the greater. Hume I think once compared inplied consent to demanding that a drowning man promise to be your slave before rescuing him, and then expecting that he regard that promise as binding.
Promises extracted with the threat of force, which of course includes instances of failure to aid, are not binding as promises, as Hume's argument explicitly acknowledges. They might be binding for other reasons - duties to minimise harm, for example - but they are not binding as promises. Whether or not the gunman holding my family hostage extracts a promise from me to rob the nearest bank is surely irrelevant to whether or not I should rob the bank. If I should rob the bank, it is because I have a duty to prevent harm to my family, not because someone was able to exploit their threat advantage to coerce a promise from me.
Indeed, in the absence of a suitable alternative, depreciating the role of voluntariness in creating the obligations of a promise seems to make it hard to understand the importance of the actual act of promising in creating the obligations which we seem to believe flow from it. It makes it hard to understand why, for example, we would regard those who uttered the words 'I promise' whenever an electrode was touched to a particular part of their brain by a torturer as free from any obligations which might, under other circumstances, result from saying that phrase.
The objection that can be derived from that observation is, for a pure social contract theory, fatal: the burdens of refusing to promise are so high as to make not promising nigh-on impossible, and so the idea that consent, in any meaningful sense, to the state's coercive power is granted in such a situation becomes ridiculous. As with promising, which consent may well be a subset of, the problem is, fairly obviously, that consent requires that the consent be given voluntarily - involuntary consent is, after all, something of a non sequiter - and when the result of not consenting is probable death, it is difficult to understand the decision to consent as voluntary. Contemporary liberal political theory tends, where it confronts this problem, to deal with it by arguing for whatever arrangements would be consented to under idealised conditions designed to remove the possibility of coercion. Thus the tagline of this blog: to each according to his threat advantage is not a principle of justice because to each according to his threat advantage is coercive.
Of course, to each according to his threat advantage may be a perfectly acceptable principle of prudence. Prudence, however, is not the same thing as justice. It may well, for example, be prudent to avoid making a song and dance of trying genocidal ex-tyrants, and indeed we may have an obligation, perhaps of justice, to those whose suffering would weigh heavy in that calculation of prudence. I, though, would be incredibly wary of making such judgements: history, after all, has a nasty habit of not running entirely according to plan, and discerning the effects of any one action may well be a mug's game. It is, let us not forget, too early to judge whether the French Revolution was a good thing or not.
However the calculation were to come out though, it would tell us nothing about any obligations we might have to those tyrants in particular. To narrow the focus even further, it would tell us nothing about any obligations we might have to those tyrants in particular as the result of promises they extracted through the threat of the continuation of their reign, quite literally, of terror. Those obligations would seem to fall under the scope of justice rather than of prudence, thus, of course, ruling out appeal to to each according to his threat advantage as an appropriate distributive maxim.