I've just been reading Jeremy Bentham's rather good refutation of natural law/rights. A major problem for the whole 'natural law' approach to morality & politics is that there doesn't seem to be the slightest bit of evidence to suggest that any such laws exist. They're mere inventions, made up to support the prejudices of their advocate. Despite their supposed 'self-evidence', such 'laws' end up conflicting with those proposed by others. As Bentham notes, "the systems are as numerous as authors".
The whole approach seems to involve a confusion of physical (natural) and human laws. Natural laws cannot be broken, they simply describe the way our universe works - there are no normative notions involved, for they do not allow themselves to be broken. "If there were a law of nature which directed all men toward their common good, [human] laws would be useless... it would be kindling a torch to add light to the sun."
The notion of natural rights is similarly incoherent: "nonsense on stilts", according to Bentham. "Right is with me the child of law... A natural right is a son that never had a father".
One point I particularly liked was that "those who cite [natural law] with so much confidence... must suppose that nature had some reasons for her law. Would it not be surer, shorter and more persuasive, to give us those reasons directly?"
Quite apart from being false, natural law should also be opposed because of its negative consequences. These fictions also encourage dogmatism: "There is no reasoning with fanatics, armed with natural rights, which each one understands as he pleases". Further, there would be anarchy if each individual rebelled whenever the state offended his perception of Divine or Natural law (as apparently recommended by Blackstone, a contemporary of Bentham).
Social Contact Theory
I don't mind social contract theory as much as natural law, since (at least in my experience) it tends to support liberalism or libertarianism rather than conservativism. But while we're on the topic of political fictions (and since I've been revising Hume a bit too), it's surely worth a mention...
Original Contract: Some of the early social contract theorists wrote as if the social contract were a historical fact - as if men in the 'state of nature' had come together and explicitly consented to submit to a government for their mutual benefit. This consent was supposed to be the basis of state legitimacy.
The problem, as Hume pointed out, was that no such events ever took place - and even if they had, this would be insufficient for modern legitimacy. As a historical fact, the origin of states tended to be a violent seizure of power (whether through conquest or usurpation); the consent of the populace was never a serious consideration. In any case, such consent would only bind those who were party to it. New generations never got any say in the matter. We surely would not consider them bound by their ancestors' choices.
Tacit Consent: A more promising move is to suggest that citizens have somehow given their tacit consent by living in society. As Socrates argued in the Crito, he had lived in Athens his whole life, despite numerous opportunities to leave, and so he had given his implicit consent to be bound by the laws of the city.
The problem here, Hume suggested, is that a poor peasant doesn't have any real choice about what land he lives in. We might as well say a sleeping man put on board a ship has freely consented to obey the captain, for he could jump overboard (and drown) if he didn't like it.
Further, Social Contract theory analyses political obligation in terms of the obligations arising from contractual promises. But this begs the question of the basis of contracts - why are we obligated to fulfill our promises?
Hume thought there were two kinds of moral duty: those arising from natural instinct (e.g. sympathy), and artificial ones arising from "a sense of obligation, when we consider the necessities of human society". Examples falling in the second class include "fidelity" (keeping promises), and also "allegiance" to the state.
But then, Hume asks, what point is there in attempting to reduce allegiance to fidelity, when both rest on the same foundation? There is no need for this elaborate fiction, simple utility is enough: "The general interests or necessities of society are sufficient to establish both".
I don't find this objection decisive, however. It seems vulnerable to a Kantian view of humans as free agents, subject to the categorical imperative of always treating other people as ends, and never as a means only. Such a view would seem to provide a deontological (rather than consequentialist) justification of promise-keeping, but not government. On such a view, some sort of consent would be necessary to legitimise the State after all.
The first objection does seem decisive though: it simply isn't true that everyone can choose their country of residence. So this cannot be considered grounds for tacit consent.
Hypothetical Consent: More sophisticated accounts are based not on real consent (whether explicit or tacit), but rather, mere hypothetical consent. That is, we merely ask whether people would consent to the government over anarchy, if asked.
As commentators here have noticed, Hume's arguments simply don't apply to Rawls' style of social contract theory. [Well, that's not entirely accurate. The 'same foundations' objection discussed above still seems relevant. But as I said, it's not decisive, at least not for deontologists.] Instead, we may return to Bentham and his distaste for fictions.
In response to those who suggested contractual theorising was justified on the basis of being a useful fiction, Bentham made the obvious rejoinder that if you're going to assert false premises, why not cut to the chase and just assert the conclusion? "Indulge yourself in the license of supposing that to be true which is not, and as well may you suppose that proposition itself to be true, which you wish to prove, as that other whereby you hope to prove it."
The wittiest objection I've heard is that "a hypothetical contract isn't worth the paper it's not written on"! I forget who said that though (leave a comment if you know).
Hypothetical consent cannot create a real obligation (if one results, it's because of reasons other than the 'consent'). There's also the problem of anarchists who wouldn't consent anyway.
A more promising avenue might be to ask whether a government is worthy of our consent. This is closer to utilitarianism. The whole notion of the social contract would then be nothing more than a heuristic, or mental shortcut. It cannot serve as the real foundation of state legitimacy.
Update: Siris defends natural law from Bentham's criticisms.