Legal Physics, Cosmic State Machines, and Elegant Rules (Part 1)
After leaving Prof. Isbell’s machine learning course today, I was hungry. In my mind, I’d settled on Chipotle for an early dinner and set off from Clough to the TSRB parking lot to find my car. While I walked, I ruminated on something that Prof. Isbell had shown the class in lecture: he’d, essentially, derived a computational form of Occam’s razor for machine learning in demonstrating that the shortest hypothesis for a maximum (or minimum, depending on what you’re searching for)-a-posteriori hypothesis, when represented in binary, was also the most preferable, since it’s the simplest (in that it contains less information than other hypotheses that fit the data while also minimizing the information about noise that must be transmitted along with the hypothesis).
I recalled that, in the most bizarre and overreaching formulations of the “legal physics” that I’d considered in law school, one could say that certain molecules might be, for instance, trespassing. This isn’t a new formulation in the law: I recall reading cases from the early 20th century, if not the 19th, in which a court found trespass when a defendant caused an aerosol (or some other particulate) to enter the zone extending vertically from the edge of another person’s land. Construed in the broadest terms, the law would recognize a trespass if a person skirted the edge of another person’s land and some of their dandruff were to cross the boundary. There are controls for these silly consequences of extending the law too far — like de minimis harm and, therefore, no consequences to the defendant — but the violation is still present, so far as the law’s cosmic state machine is concerned.
And this, really, is what the law is: a cosmic state machine. It’s a human attempt to circumscribe the actions that humans — and even animals and inanimate materials — can take. A type of animal might be protected, for instance, from anthropogenic harm while on federal land, but subject to hunting if it crosses into land with a different legal status.
In its more enlightened moments, the judiciary recognizes that the law operates best when it circumscribes those actions that social norms, policymaking, and regulatory capture — circumscribed by meta-norms (i.e. human rights) — have deemed harmful to the collective welfare and leaves all other actions to individual discretion. The law, when considered in this narrow scope, is defined by its complement; even the most informed members of a population can’t know what’s best for the individual or the general social good, so we limit ourselves to prohibiting only those actions that we’re certain are deleterious when applied equally to all isomorphic conditions. Or, at least, we try to limit ourselves.
But the law, when defined in the broadest sense possible — as Lessig did in Code, using his four-part framework — continues to both expand and constrain the scope of human possibility. It expands possibility when it creates new tools that permit new actions, but constrains when it finds new applications of these tools to be harmful. There’s absolutely no agreement among legal scholars that this is so, and I have no desire to represent this hypothesis as if it’s commonly accepted. But this has always, to me, seemed to be the hypothesis that best describes the practice of law.
Much of these laws seem to be driven by a drive to remedy “flaws” in the fundamental architecture of the world. It’s possible for a human to beat another human to death because we’re able to produce the necessary velocity in a sufficiently dense object to harm and, eventually, kill another person. No law of the universe prevents a fist from making contact with another person’s face, so we attempt to create disincentives, hoping to reduce the likelihood that a person’s fist will make unwanted contact with another person’s body. But this is a pretty inelegant solution: during the moment in which a person deliberates whether to strike another person, we’re only able to intervene in the most attenuated way, fostering a culture that will disapprove of the violent act and engendering a sense of the legal consequences that might accompany the action.
So, if we could intervene more directly, should we? If we could change the laws of physics to recognize people as objects, determine a threshold of force that would be harmful to another person-object, and physically prohibit the object from making contact, should we?