Ubi Lex, Ibi Poena: Designing Norm Enforcement in E-Institutions

Specifying and Reasoning About Multiple Institutions

“…Grossi et al [11] explore a classification of norms into three kinds: substantive norms that are generally high-level in nature and not directly enforceable, check norms that specify how to verify substantive norms and reactive norms that define how violations are to be sanctioned. Particularly relevant to this paper are the check norms that the authors view as so-called sub-institutions that carry out regular or continuous monitoring in respect of the upholding of the substantive norm.…”

Model Checking Norms and Sanctions in Institutions

“…[12,16,1]) can be reduced to particular patterns of specification of our language; finally, we contribute to the ongoing debate about regimentation and enforcement of norms [10,9,6,8], showing that when human institutions impose a specific interpretation of norms, it may be the case that properties that reflect them cannot be satisfied by artificial institutions under the assumption that agents are autonomous. As a consequence, regimentation of norms is not always a viable solution.…”

“…Without proper sanction mechanisms, the introduction of norms typically does not change the set of properties satisfied by an institution, given that autonomous agents may not comply with such norms [5,2,9,18,7]: as a consequence certain properties may not hold in an institution even if its rules are correctly stated. For instance, properties (P.1) and (P.2) do not hold in the new model of the Resource Management institution obtained by adding norm h1, despite this correctly requires that a holder gives a resource after it has positively answered to an agent.…”

“…Institutions define two kinds of rules [17]: norms (also named regulative rules [17]), which regulate existing activities, and constitutive rules, which create the very possibility of certain institutional actions. Artificial institutions are often designed to reflect constitutive and regulative rules defined by human institutions in artificial systems [10,9,7], and model checking can play an important role to evaluate the compliance of artificial institutions with rules of human institutions and to compare design alternatives arising from different interpretations of such rules. In particular, when we map human rules only onto constitutive rules of artificial institutions, we obtain systems where violations cannot occur (they are regimented [10,9]).…”

“…Artificial institutions are often designed to reflect constitutive and regulative rules defined by human institutions in artificial systems [10,9,7], and model checking can play an important role to evaluate the compliance of artificial institutions with rules of human institutions and to compare design alternatives arising from different interpretations of such rules. In particular, when we map human rules only onto constitutive rules of artificial institutions, we obtain systems where violations cannot occur (they are regimented [10,9]). Instead, when we introduce regulative rules into artificial institutions, we obtain systems where violations may occur due, for instance, to the agents' autonomy.…”