Structural Normativism for Law and AI
The paper “No-Miracles Argument of Law: Jurisprudence Meets Structural Realism” has been published open access in the Journal for General Philosophy of Science. Its central thesis is that law should not be understood primarily as a collection of isolated rules, principles, or precedents, but as a structured normative field in which legal meaning emerges through relations: priority and exception, competence and authority, conflict and balance, possibility and constraint.
Inspired by scientific structural realism, the project develops structural normativism as a way to make these relations visible. What makes a norm legally meaningful is not only what it states, but where it stands, what it depends on, and how it interacts with the wider order.
The European Union makes this especially clear. It is treated as a legal subject, yet much of its reality lies in structure: treaties, competences, institutions, procedures, doctrines, and relations between legal orders. Its coherence does not come from a single legal source, but from the organization of many normative elements into a shared form.
Structural normativism studies this organization. It examines order relations in legal systems, the logic of balancing, and the architecture behind legal dogmatics. Normative entities emerge through modal structures: through what a legal order treats as necessary, possible, impossible, or contingent.
For AI, this matters deeply. Legal AI systems cannot responsibly work with law as a database of fragments. A legal answer is not valid simply because a rule can be retrieved or a precedent resembles a case. It depends on structure: on authority, hierarchy, interpretation, exceptions, and the space of legally possible conclusions.
Structural normativism offers a foundation for representing this depth. It points toward AI systems that do not merely predict or summarize legal outcomes, but make the relations behind legal reasoning explicit. Such systems would need to show how norms connect, where tensions arise, and why one path of reasoning is stronger than another.
In this sense, the project links legal theory to responsible AI. It asks how law can be modeled without flattening its normative complexity, and how AI might support legal reasoning while remaining transparent, constrained, and accountable.