Research Ideas

James Grimmelmann
Current through December 2019

I sometimes get questions from students and younger scholars about research topics, and I do my best to help them find specific projects that mach their skills and interests. This page is my attempt to do the same for everyone I haven’t had the chance to talk to individually. I got the idea from my operating systems professor, Margo Seltzer, who also taught me how to read and critique systems papers. That gift I’m not in a position to pass along, but this one I am.

These are all ideas that I have thought about writing up someday, but do not expect to get to for years. I would love nothing better than for someone else to come along and preempt the topic by doing it with their own creative twist. If any of these strike you as interesting, please take the ball and run with it. All I ask is that you send me a link to your paper when you finish, so that I can enjoy it (and cross that item off the list).

Law and Computer Science

The international law of borders is heavily geometric: it is always drawing lines and measuring distances. Someone with an appropriate mathematical grounding should analyze its doctrines using the techniques of computational geometry. There have been a few promising starts in this direction, but I’m not aware of any systemic treatment.

Contract law has an interesting relational structure, in which parties can take actions that affect each others’ sets of permissible actions. A formal logic or calculus could model contractual structure in an interesting way. There is a reasonable body of literature in the field of law-and-AI on this problem, which has been further invigorated by recent interest in smart contracts, but there is still much to be done.

Lawyers debate the precedent a judicial decision sets for future cases: trying to break an opinion down into its binding “holding” and non-binding “dicta.” There is significant disagreement about many of the details, and many approaches run into paradoxes, e.g., nontransitivity. The legal literature here has borrowed from game theory, but there may also be some room to use decision trees or other models with rich internal structure to map out what an opinion does and does not establish.

Property law has a set of algorithmic rules for resolving conflicting transfers of property, e.g., that a later transfer can prevail over an earlier one if it is recorded in a central index first. These rules have several interesting variations and in some theoretically possible cases appear to run into circularity problems. A good mathematical treatment might provide significant insight.

Some choice of law methodologies (e.g., the Neumeier rules) are amenable to crisp formalization. Build a system that takes a set of contacts with different jurisdictions and displays the resulting choice of law under different approaches. This could be useful as a teaching tool and also for building insight as to what drives the different approaches.

A good computational history of legal informatics remains to be written. Legal publishers began using reverse indexes in the 1800s to keep track of what later cases cite an earlier one, and numerous innovations in how laws are written and promulgated were driven by embarassing failures of previous systems to organize the flood of legal materials in a usable way. An account that uses computer-science concepts like “diff,” “log,” and “namespace” might provide a fresh perspective on the history.

The rise of computational photography could have massively disruptive effects for evidence, free-speech, and privacy law. Do a deep dive on how modern devices construct composite images in-camera and on the typical post-processing toolchain, add some Susan Sontag and a dash of Errol Morris, and you could shake people out of complacency in time for the legal system to adjust without disaster.

Digital copyright law rests on a series of unfounded technical distinctions: between memory and long-term storage, between program and data, between software and hardware, between reproduction and distribution, between reproduction and performance, between source code and object code, and between a product and a service. Explain why these distinctions collapse if forced to bear too much weight, and how to salvage as much of copyright as possible from the wreckage.

Apply Carol Rose’s Crystals and Mud in Property Law to talk about the fragility of blockchain-based property.

Legal scholars know about software and computer science. But they don’t know enough, and they don’t always understand why technologists worry about the things they worry about and not others, or which tasks are easy and which are hard. It would be nice to have a good top-ten list of things every computer scientist knows (and every lawyer should).

Intellectual Property

The most important change in IP law over the past decade has been the rise of Internet platforms’ IP policies. For most purposes, the rules they set — whether enforcing third party take-down notices, creating their own content-filtering systems, adjudicating disputes over names and addresses, or creating mechanisms to make some kinds of copying easy and other kinds hard — are the face of the IP system for most Internet users. There are some good analyses of parts of this shift, and some good related work on content moderation in the context of harmful speech, but there is plenty of room for a good IP-focused analysis of the whole sweep of the change.

The “copy” has had a good 300-year run as the central concept of analysis in copyright law. But the ease of digital copying, the “end of ownership,” and the remarkable rise of streaming and time-based access models mean that it is time to imagine copyright after the copy. A modern reconstruction of copyright would focus on who has control over a copy of a work (in the sense of European data protection law), rather than on the production of copies themselves. The exclusive rights, first sale, fair use, and paracopyright regimes like the DMCA’s anticircumvention and copyright management information sections will all need to be rethought.

It is sometimes said that copyright is “strict liability,” because even subconscious copying can be infringement. But that’s a misleading picture. Copyright takes account of the alleged infringer’s mental state when finding volitional infringement, when assessing contributory and inducement liability, in Section 512, when reducing liability for innocent infringement and enhancing it for willful infringement, when finding willful blindness, when finding criminal infringement, when imposing liability on corporate officers, and when finding implied licenses or copyright estoppel. Synthesize the existing literature on these various strands and explain whether copyright could or should prune back some of the excess.

Copyright caselaw is littered with the carcasses of startups that thought they had found safe patches in the middle of its system of exclusive rights: Aereo, ReDigi, Zediva, and many more before them. The world could use a good retrospective on them, one that maps out the difficulties with the different lines the courts have attempted to draw.

Content Moderation

All platform content moderation systems inevitably reinvent the legal system, including precedent, appellate review, dockets, choice of law, written decisions, and more. They will do so not necessarily because the actual legal system will compel them to, but simply because they must convergently adopt particular institutions simply to keep up with the volume of business. A good account would draw on history (e.g. the rise of the common law and the bureaucratization of compensations systems), on theory (e.g. Lon Fuller’s “internal morality of law”), and on a detailed study of platform content moderation itself.

I have made some sketches towards an economic model of section 230 in terms of the error costs of over-blocking good content and under-blocking bad content. Someone with a firmer grounding in microeconomics and game theory could probably do a better and more rigorous job of it.

Legal Theory

The Calabresi/Melamed distinction between property rules and liability rules is a staple of legal theory. But the conceptual distinction it makes is only appropriate for a narrow range of torts: basically, those that look like nuisance. Applying it to other kinds of conduct, and treating “property rule” as shorthand for injunctive relief, as modern scholarship often does, is the source of much mischief and confusion.

Occasionally, an otherwise sober-minded judge will write a dissent issuing dire warnings about a type of speech that strikes many of their colleagues as innocuous. Whether it’s Justice Jackson on Jehovah’s Witnesses, Justice Douglas on playing the radio on streetcars, or Justice Alito on video games, there’s a story to be told about these First Amendment freakouts.

Software design theory, drawing on architectural theory, has made trenchant observations about modular design principles, dealing with complexity, recurring patterns, and more. Bring some of this literature over into legal scholarship and describe how it can help with legal drafting.

Hot Takes

In 2010, Michael Sippey coined the phrase “even if it’s fake, it’s real” to describe the complicated relationship between ontological truth and functional reality on the Internet. A decade later, this is one of the most useful ways to understand everything from deep “fakes” to “fake” news.

The Internet is a source of Lovecraftian horrors, from unseeable images to human flesh search engines. Put your finger on what makes the Internet such a potent engine of madness and terror and you will do the world a service.

Service to the Profession

The world could use a better law review article template. On the one hand, it should have better typefaces and typography than the usual Times New Roman with one-inch margins that a majority of drafts use. On the other, it should be so easy-to-use that authors will feel comfortable adopting it.

Leveraging one law-review acceptance to improve an article’s chances with another “better” law review is an unethical practice that wastes editors’ time, lowers the quality of review, and makes the legal scholarly publication system worse for everyone. Organize a pledge not to expedite by law professors and you will make thousands of students’ and authors’ lives better.

Law is doing okay on the open access front — most law reviews now post their articles online and allow authors to post them as well — but it could do better. Get your favorite law school to adopt an open access policy on the Harvard model, which allows the university to distribute all faculty articles non-commercially.