In the pre-modern world, law existed in multiple configurations, sometimes in overlapping configurations, that we might recognize as both 'governmental' and 'non-governmental'. For instance, to use the example I know best, Islamic 'law'- al-shari'ah- in its medieval forms was generally not directly attached to governmental forms at all, or only peripherally. In fact, the translation of 'shari'ah' as law is problematic, given the very narrowness the semantic range of 'law' has acquired in modern English. Shari'ah literally means 'way' or 'path,' more indicative translations of its actual scope and manner of application, at least in the pre-modern, pre-modern-state world (things have changed dramatically in the past couple hundred years). Shari'ah, in the first place, was primarily produced in the 'non-governmental' sphere, and was often applied in that same sphere, without recourse to coercive force (though it could also be applied using some form of coercive force, social if not governmental). For instance, in the event of a dispute between two parties over, say, property arrangements, a common recourse would be to go to the local mufti (someone trained in fiqh, Islamic jurisprudence) and seek a legal opinion. The mufti would review the case and then offer a fatwa, a legal opinion if you will. This fatwa was non-binding: no coercive force, indeed no physical force lay behind it. It might be employed in a formal court (the court of the qadi), or it might not (usually, it seems, the latter). Its legal force was more socially based: the contending parties both recognized the mufti as a viable arbitrator, and social convention (and social force in a somewhat Foucaltian sense) aided in this agreement and cohesion.
It would take far too long to go into further details about the interworkings of medieval shari'ah, but suffice to say, its rendering- both theoretically and in practice- of 'law' was considerably more complex and nuanced than the way we usually think of 'law' in the modern, state-dominated world. I would not argue that shari'ah was ever entirely 'non-coercive,' but no society can ever be entirely non-coercive in some sense. Nor am I suggesting that contemporary radicals would want to implement, or could implement, shari'ah, which is obviously both religiously and historically contingent in the manner I have described. However, I do think that if anarchists want to rethink law, rules, and public order, serious appraisal and analysis of pre-modern, 'traditional' modes of law and order are in order.