Two Problems with the Nonaggression Principle
Libertarians typically favor reducing government to a minimal or night-watchman state, one limited to police, courts, and national defense. (For example, the Libertarian Party platform says that “the protection of individual rights is the only proper purpose of government.”) Some libertarians argue for this minimal state by appealing to the nonaggression principle. This is a mistake.
The nonaggression principle (NAP) prohibits initiatory force. Adapting a formulation by philosopher Roderick Long, it can be cast as follows:
NAP: Any act of forcible interference (or threat thereof) with another individual’s person or property is a violation of rights, unless the act is a response to forcible interference (or threat thereof) by that person.
I will argue that the NAP has unacceptable consequences. We can modify the principle to avoid those consequences, but this comes at a severe cost: the principle will then no longer do any real work in the argument for the minimal state.
First Problem for NAP: Past Injustices
Suppose I steal a priceless painting from you and then give it to Bob, a friend of mine. Even if Bob knows the painting was stolen, he did not use or threaten force. According to the NAP, this means that forcing Bob to give the painting back would be a violation of his rights. Putting this point more generally: the NAP prevents rectification of past crimes or injustices, so long as the original criminal has transferred the proceeds of his iniquity to someone else.
We could avoid this unpalatable consequence by revising the NAP:
NAP: any act of forcible interference (or threat thereof) with another individual’s person or legitimately acquired property is a violation of rights, unless the act is a response to forcible interference (or threat thereof) by that person.
The libertarian could then claim that Bob did not legitimately acquire the painting, since it was stolen; thus we are allowed to use forcible interference to get it back.
However, adding “legitimately acquired” to the NAP opens up a huge can of worms. Which properties count as legitimately acquired? A stringent, Nozick-inspired answer to that question: my property was legitimately acquired if I got it through a sequence of transfers each of which was free and uncoerced, and where this sequence goes back to an original acquisition of property that was just and fair.
Even if we suppose, rather improbably, that there was a just initial acquisition of property, it beggars belief to claim that the current distribution resulted from a sequence of free and uncoerced transfers through the subsequent centuries. To see this, we need merely point out that the distribution of wealth in the United States was hugely affected by the slavery and blatant oppression of black Americans, not to mention the slaughter of Native Americans and theft of the land they occupied in the early days of the republic.
On the stringent view of legitimate acquisition, even a democratic socialist who wants a big welfare state could simply agree with the NAP: the socialist will just add that, since almost no currently owned property was legitimately acquired, and since the NAP only applies to legitimately acquired property, the NAP imposes no real restrictions.
Perhaps libertarians can come up with a less stringent theory of the legitimate acquisition of property. Perhaps, as suggested by David Gordon in a related context, the notion of convention will play a role. Perhaps this theory, when added to the NAP, will imply that most acts of initiatory force are still forbidden; and, finally, perhaps this can in turn be used to argue that only the minimal night-watchman state is justifiable.
But observe: even with such a theory in place, the dispute between the libertarian and the socialist would then have nothing to do with the revised NAP, which they can both accept. The NAP, which was originally put forward as the basis for the libertarian position, no longer plays any role in the argument. All the real work would be done by the yet-to-be-provided ancillary theory about what counts as legitimately acquired property. (Sandy Ikeda made a related point about the NAP.)
Second Problem for NAP: Taxation
When the government sends me a tax bill, it comes with a threat: they can garnish my wages or even throw me in jail. This threat is not in response to any act of forcible interference on my part; I was just sitting here not paying taxes. But this implies that the tax bill is, according to NAP, a violation of my rights.
However, if taxes themselves are a violation of rights, then even the minimal state is forbidden, insofar as the minimal state funds courts, law enforcement, and national defense by levying taxes.
Some libertarians might accept this and conclude that even the minimal state is too much state. These libertarians, like the 2020 vice presidential candidate Spike Cohen, may become anarchists.
But this goes beyond what most libertarians advocate. The Libertarian Party platform calls for extremely limited government but does not demand the end of compulsory taxation. Similarly, Libertarian presidential candidate Jo Jorgensen advocated eliminating the federal income tax but did not suggest getting rid of all taxes.
Libertarians might try to avoid the anarchist suggestions of NAP by modifying the principle so as to allow just enough taxation to support the minimal state:
NAP: any act of forcible interference (or threat thereof) with another individual’s person or property is a violation of rights, unless either the act is a response to forcible interference (or threat thereof) by that person or the act is an instance of the government collecting taxes for a legitimate purpose.
But the libertarian will now need another ancillary theory explaining which governmental purposes are legitimate and why. After all, the socialist could fully agree with the revised NAP, so long as the socialist also claims that a large welfare state is a legitimate purpose for government taxation.
The libertarian will no doubt dispute the socialist claim and propose instead that anything beyond the minimal state is not legitimate. However, once again, the dispute between the libertarian and the proponent of big government will have nothing to do with the revised NAP, which both sides can accept. The nonaggression principle, which was supposed to ground the claim that only the minimal state is legitimate, now does no work at all.
When faced with the extreme consequences of NAP, one cannot reply by simply saying, “Well, I don’t want to push it that far.” Principles are not like buses that you can take as far as you want and then get off; they imply what they imply. If you don’t accept the implications of the principle, you don’t really accept the principle; you must reject it or at least qualify it. However, the obvious attempts to qualify or hedge the NAP to avoid such consequences would mean that the principle no longer does any real work in supporting the minimal state.
Of course, none of this implies that libertarians are wrong in advocating only the minimal state; it only means that they would need to support this view on other grounds. Libertarians might, for example, follow Mises in saying that “social utility is the only standard of justice.” In other words, instead of the rights-based talk of NAP, one might try to defend the night-watchman state on broadly utilitarian grounds, claiming that such a state will lead to greater happiness and human flourishing.