The Laboratorium (3d ser.)

A blog by James Grimmelmann

Soyez réglé dans votre vie et ordinaire afin
d'être violent et original dans vos oeuvres.

Programming Property Law

One of the things that sent me to law school was finding a used copy of Jesse Dukeminier and James Krier’s casebook Property in a Seattle bookstore. It convinced me – at the time a professional programmer – that the actual, doctrinal study of law might be intellectually rewarding. (I wonder how my life might have turned out if I had picked up a less entertaining casebook.)

Although everything it was interesting, one section in particular stood out. Like every Property book, Dukeminier and Krier’s book devoted several central chapters to the intricate set of rules that govern how ownership of real estate can be divided up among different owners across time. If the owner of a parcel of land executes a deed transferring it “to Alice for life, and then to Bob and his heirs,” then Alice will have the right to enter on the land, live on it, grow crops, and so on for as long as she lives. At her death, automatically and without any need of action on anyone’s part, Alice’s rights will terminate and Bob will be entitled to enter on the land, live on it, grow crops, and so on. As Dukeminer and Krier patiently explain, Alice is said to have an interest called a “life estate” and Bob a “remainder.” There are more of these rules – many, many more of them – and they can be remarkably complicated, to the point that law students sometimes buy study aids to help them specifically with this part of the Property course.

There is something else striking about this system of “estates in land and future interests.” Other parts of the Property course are like most of law school; they involve the analogical process identifying factual similarities and differences between one case and another. But the doctrines of future interests are different. At least in the form that law students learn them, they are about the mechanical application of precise, interlocking rules. Other students find it notoriously frustrating. As a computer scientist, I found it fascinating. I loved all of law school, but I loved this part more than anything else. It was familiar.

The stylized language of conveyances like “to Alice for life, and then to Bob and his heirs” is a special kind of legal language – it is a programming language. The phrase “for life” is a keyword; the phrase “to Alice for life” is an expression. Each of them has a precise meaning. We can combine these expressions in specific, well-defined ways to build up larger expressions like “to Alice for life, then to Bob for life, then to Carol for life.” Centuries of judicial standardization have honed the rules of the system of future of estates; decades of legal education has given those rules clear and simple form.

In 2016, after I came to Cornell, I talked through some of these ideas with a creative and thoughtful computer-science graduate student, Shrutarshi Basu (currently faculty at Middlebury). If future interests were a programming language, we realized, we could define that language, such that “programs” like O conveys to Alice for life, then to Bob and his heirs. Alice dies. would have a well-defined meaning – in this case Bob owns a fee simple. We brought on board Basu’s advisor Nate Foster, who worked with us to nail down the formal semantics of this new programming language. Three more of Nate’s students – Anshuman Mohan, Shan Parikh, and Ryan Richardson – joined to help with programming the implementation.

I am happy to announce, that following six years of work, we have developed a new way of understanding future interests. We have defined a new programming language, which we call Orlando (for Orlando Bridgeman), that expresses the formalized language of Anglo-American real-property conveyances. And we have implemented that language in a system we call Littleton (for Thomas de Littleton of the Treatise on Tenures), a program that parses and analyzes property conveyances, translating them from stylized legal text like

O conveys Blackacre to A for life, then to B for life if B is married, then to C, but if D marries to D.

into simple graphs like

Then, if something else happens, such as

A dies.

Littleton updates the state of title, in this case to:

Littleton and Orlando model almost everything in the estates-in-land-and-future-interests portion of the first-year Property curriculum, including life estates, fees tail, terms of years, remainders and reversions, conditions subsequent and precedent, executory limitations and interests, class gifts, the Doctrine of Worthier Title and the Rule in Shelley’s case, merger, basic Rule Against Perpetuities problems, joint tenancies and tenancies in common, and the basics of wills and intestacy. They are no substitute for a trained and experienced lawyer, but they can do most of what we we expect law students to learn in their first course in the subject.

We have put a version of Littleton online, where anyone can use it. The Littleton interpreter includes a rich set of examples that anyone can modify and experiment with. It comes with includes a online textbook, Interactive Future Interests, that covers the standard elements of a Property curriculum. Property teachers and students can use it to explain, visualize, and explore the different kinds of conveyances and interests that first-year Property students learn. We hope that these tools will be useful for teaching and learning the rules that generations of law students have struggled with.

To enable others to build on our work and implement their own property-law tools, we have put online the Littleton source code and released it under an open-source license. Anyone can install it on their own computer or create their own online resources. And anyone can modify it to develop their own computerized interpretation of property law.

To share the lessons we have learned from developing Orlando and Littleton, we have published two articles describing them in detail. The first, Property Conveyances as a Programming Language, published in the conference proceedings of the Onward! programming-language symposium, is directed at computer scientists. It gives formal, mathematical semantics for the core of Orlando, and shows that it captures essential features of property law, such as nemo dat quot non habet. The second, A Programming Language for Future Interests, published in the Yale Journal of Law and Technology, is directed at legal scholars. It explains how Orlando and Littleton work, and describes the advantages of formalizing property law as a programming language. As we say in the conclusion:

To quote the computer scientist Donald Knuth, “Science is what we understand well enough to explain to a computer. Art is everything else we do.” For centuries, future interests have been an arcane art. Now they are a science.