Justice Holmes, Meet Dr. Turing: Law Is Computation

05/07/2015 04:03 pm ET Updated May 07, 2016

In his famous work of legal theory, The Common Law, Oliver Wendell Holmes, Jr. memorably declared: "The life of the law has not been logic." This pronouncement sums up a deeply held tenant of American legal theory. Most lawyers in the United States believe that there is an element of irreducible human uncertainty that is somehow an indispensable ingredient in law. This assumption is running squarely into the developments of legal technology and computational law. Justice Holmes, meet Dr. Alan Turing.

Turing was a British mathematician famous for a number of foundational accomplishments in computation theory, only hinted at in the recent bio-pic The Imitation Game. Computation theory, as expounded by Turing and other math wizards, is not just about computers. Rather, a "computation" is any rule governed, step-wise process. These processes are surprisingly common in both the natural and human-constructed worlds, including such diverse examples as many biological functions, the manufacturing assembly line, and games ranging in complexity from Chutes and Ladders to poker or chess. Computation theory provides a means for specifying such processes in a formal way, a bit like how the alphabet and writing allow the specifying of language formulations in print for current storage and later reconstruction.

Using a computational approach, a step-wise process like a board game can be fully described in its stages, inputs and transitions. This careful specification both clarifies and records its elements. The terms of that description can be in words, in computer code, in the gears of a mechanical calculator, or even in pictures, like a flow chart. And here comes the really potent next step: a computation specified in one mode of description (including in a "natural language" statement like the written rules for poker) can also frequently be specified in some other mode, like the binary code used by our digital computers. Put a few programming steps in between, and you get internet poker.

The Turing Machine, a formal proposition rather than an actual physical construct, is posited as having the capacity to describe any computable process. Digital code and computer processing may not have quite this reach, but they do a lot of computational work. They have enabled us to take many of the computational processes in the world and embody them into the on/off descriptions of binary code which our actual machines can then read and implement. From self-driving cars to cell phone apps, the astonishing tech-driven advances of our time all depend on this transformation.

Which brings us back to law. Legal rules, whether set out in contracts, regulations or judicial decisions, often look like a Chutes and Ladders game writ large. They have the same kind of stepwise branching logic, although the Internal Revenue Code has orders of magnitude more chutes and ladders than the board game. Complexity matters, but it doesn't contradict the core point: law is computation too. And because it is computation, we can represent it in software as well. Not emulate it in software, but represent its logic and process directly in the code.

The legal technology industry has already picked some of the low hanging fruit of legal sub-processes, like company formation, will drafting, and legal research. Once we express legal formulations themselves in computational terms, figuring out the implications for a particular set of facts becomes a quick exercise in technology. If the credit card companies had to file a computer code version of their contracts, you could press a couple of keys and find out what happens when you miss a few payments and the rates triple. Want to contract to sell your house? There'll be an app for that. That nasty Internal Revenue Code? TurboTax, H&R Block, and the IRS itself are on portions of it already.

An even more comprehensive approach to "computational law" has begun to embed the rules directly into the objects they govern. I recently tried to sync up my phone to the Bluetooth in a rental car -- the technology made me stop the car so that I could proceed more safely with the process. Imagine a "smart security," like a share of stock that knows who owns it and will implement any legal rules about selling it directly into the proposed transaction. If you imagine it carefully enough, computer engineers can make it happen.

Not all of these developments will be welcome -- the cellphone cop in my car may have been correct, but it was also annoying. Furthermore, none of this will squeeze all of the uncertainties out of law. There will still be places for human interpretation, where good software will stop and ask a person for guidance. On the whole, however, an explicitly computational approach to legal design will serve to make the law more available, transparent, and predictable, and put a bit more logic back into the system. Even Justice Holmes should approve of that.