LogicMaps

 Although the rules of law and the rules of science were first conceived in natural language, the scientific laws have become decidedly more formal and mathematical in modern times, where the rules of law are still expressed in the ambiguities of natural language. The legal profession exists because we require human experts (lawyers and judges) to make logical sense out of these informal, written rules, deciding – often as a matter of logic – when they do and do not apply. Science no longer needs such experts. It has calculators and computer programs.

Scientists had an inherent advantage in getting to automated reasoning tools because they began writing their rules in formal mathematics many centuries ago. The formal logic required to bring automated reasoning tools to legal professionals is a little younger, but it exists. It has been around for about a century and a half now. It is used by engineers to reason about hardware and software designs with theorem provers, SAT solvers, model checkers, and other automated decision procedures. But as with the mathematical tools, your rules must be stated in a formal language before you can begin, putting such tools out of reach for most lawyers. Until now. Generative AI has reached the point where it can be used to translate natural language into this kind of formal logic, not unlike how it translates say, English to French. This is the kind of task LLMs are natively good at.

This now allows legal professionals to apply automated reasoning directly to the natural language rules, skipping the formalism entirely. The most useful of the logical reasoning tools for legal professionals is the decision procedure — a method for mechanically proving whether or not a given factual scenario is covered by the rules. This decision procedure can now be automatically derived from the natural language rules and externalized as a natural language diagram — the LogicMap — which allows any competent speaker of the language to determine the applicability of rules, even with little or no legal training.

But don’t worry, this is one of those benign applications of technology to professional occupations that augments, rather than replaces human skills — like calculators for scientists, or spreadsheets for accountants, or spelling and grammar checkers for writers. These LogicMaps have uses all across the life cycle of legal rules:

• at the very beginning of the rule making process, by legislators and statute drafters, or by lawyers drafting contracts, to better understand the consequences and correctness (or incorrectness) of rules as they are being drafted

• when laws are being refined by regulatory agencies, to ensure the fidelity of regulations to the original statutes

• when regulations are being amended, or contracts are being renegotiated, to understand the net difference between the old and new versions of the rules

• when rules are being applied, by judges or regulators, to decide conformance

• when clients are being advised about eligibility, liability, or safe harbor provisions of rules

• by software developers, when coding the logic for public-facing applications of rules, or coding automated conformance applications for regulated businesses

• even by ordinary citizens to decide whether to hit ‘I accept’ before agreeing to the use a software service.

You can view some recent real-world examples of LogicMaps here: