This comment is without significant impact except for a more rigorous theory of criminal procedure. Laudan (2007?) makes the following comment in footnote 4:
It is sometimes suggested that this potential loss of liberty [the convicted innocent] represents a risk that the citizen runs under the rule of law that he did not run while in a state of nature. That is dubious since it seems plausible to suppose that, even in a stateless society, there would be plenty of surrogate and unofficial forms of vigilante-ism that would mete out a primitive form of justice that would frequently result in persons being punished for actions that they did not commit.
My contention is that vigilante-ism does not occur in the State of Nature. A cursory examination makes it appear to be the State of Nature, since there is not a government “met[ing] out the primitive form of justice” but there is a supposed social code on behalf of which the vigilante is acting. The risks in the State of Nature are a passionate violence and preemptive violence, not justified violence acting In The Name Of. It is possible, likely even, that the punisher and the punished have not discussed this code and do not share this code, but the vigilante presumes the code to apply nonetheless. This problem of a lack of shared code is not necessarily resolved by the Social Contract.
So, the contention Laudan is responding to is accurate: in the State of Nature there is no risk of a wrongful conviction, because there are no (juridical) convictions in the State of Nature. This is not to say, however, that the State of Nature is preferable to a regime that can wrongly convict.
Laudan, Larry. (2007). The social contract and the rules of trial: Re-thinking procedural rights. While this a working draft and ‘not for quotation” I have determined this to be an acceptable use of reference. If the author disagrees and wishes the discussion to be removed I will gladly do so upon request.