prior probability

As we mentioned in our previous post, we will be presenting our work in progress on “probabilistic interpretation” at this year’s Federalist Society Annual Faculty Conference. Since the conference is in San Francisco, our talk will make reference to the fictional Starfleet Command (from the Star Trek sci fi series): imagine going warp drive and beaming down to “Planet Newgarth” during the tumult over the Case of the Spelencean Explorers, a make-believe murder prosecution professor Lon Fuller wrote about in the late 1940s on Earth. It’s the year 4299 A.D., so why are there no “machine judges” on the Supreme Court of Newgarth? In our talk, we imagine just such a scenario and discuss how a machine might decide (not just predict the outcome of) a close case.

That is the title of our most recent work in progress; one of the research projects we were working on while we were on sabbatical. (Part 1 of the paper will be published in the University of LaVerne Law Review this spring and is available here.) Part 2 presents Lon Fuller’s “Case of the Speluncean Explorers,” one of the most famous thought experiments in legal studies. In our paper (part 2), we imagine an alternative system of voting by appellate judges, a bayesian or cardinal voting system in which judges assign a score to their preferred judicial outcome. Appellate courts generally use an ordinal system of voting (i.e. one judge, one vote) to decide cases. By contrast, we propose a simple cardinal voting system for deciding appellate cases, using Fuller’s hypothetical case to illustrate how our simple system of cardinal voting would work in practice.

Image credit: Wikimedia Commons

Via Chris Rodley: “A runaway trolley is about to create five trolley problems. Do you pull the lever and divert it, so it only creates one?”