The Law is a domain of rules, and many of those legal rules are embodied in statutes and regulations. Since rules can be expressed logically and computers can reason deductively, computationally modeling statutory reasoning should be easy. One simply inputs a fact situation to the computer program; the program identifies the relevant rules, determines whether or not the rules’ conditions are satisfied, and explains the answer in terms of the rules that applied or did not apply.
Building a computational model of statutory reasoning, however, presents challenges. As explained below, statutes routinely are vague, syntactically ambiguous as well as semantically ambiguous, and subject to structural indeterminacy. If a computer program is to apply a statutory rule, which logical interpretation should it apply, how can it deal with the vagueness and open-texture of the statute's terms, or determine if there is an exception?
The chapter draws a contrast between deductively applying a statute and the complex process of statutory interpretation, which frequently involves conflicting reasonable arguments. Classical logical models may break down in dealing with legal indeterminacy, a common feature of legal reasoning: even when advocates agree on the facts in issue and the rules for deciding a matter, they can still make legally reasonable arguments for and against a proposition.
Reasoning with statutes, however, remains a pressing necessity. The chapter examines various AI & Law approaches that address or finesse these issues: a normalization process for systematically elaborating a statute's multiple logical versions, a logical implementation for applying a statute deductively, and more recent models of business process compliance and network-based statutory modeling, both potentially useful for cognitive computing.
Questions addressed in this chapter include: How can statutory rules be ambiguous, both semantically and syntactically? How do lawyers deal with these ambiguities, and how can computer programs cope? What are normalized legal drafting, Prolog, and a Prolog program? What is depth-first search and how does it differ from breadth-first search? What is legal indeterminacy and why is it a problem for logical models of legal reasoning? How can logic programs assess business process compliance with regulations? What problems do isomorphic knowledge representations of statutes address? What are citation networks and statutory network diagrams, and how can they support cognitive computing?