Interesting, gave me food for thought🙂 I find the comparison to LegalRuleML interesting. It's a logic framework for description of legal systems for computers, i've read about in the last weeks. So because i find it interesting and i've already thought about it, i think i can share a short version of my thoughts: LegaLRuleML just can handle legal actions, so no interpretation inside that framework too. LRML doesn't know about commands and only uses norms as building blocks. Validity is generally not really a thing described inside LegalRuleML, instead the source of the norm must always be traceable. So one can check validity only outside of this framework. But to compare it to your point what is the difference in state-violence and nonstate-violance. There is no inherent differentiation between from whom a rule is originated inside LRML. There is only a way to prioritize different norms to one another, if they are conflicting. But that is again outside of the logic itself and can only be imposed from outside the framework. And to the norms themselves; how norms work in LRML seems to match the description here. Norms are also just described as rule and the consequence for breaking the rule.
