Preserving Legal Corpus Callosa

Introduction

Meet Stanley. He is the winner of the 2005 Defense Advance Research Projects Agency's Grand Challenge (DARPA), a 212 km off-road obstacle course in the Mojave Desert. It was not close. Out of 23 finalists, only Stanley finished the race in under seven hours. What is Stanley's secret?

Stanley's Secret

Right and Left Brains

Of all the other autonomous (driverless) vehicles in the competition, Stanley appears to be the most “right-brained.” The right brain is responsible for imagination, intuition, and holistic thinking while the left brain is responsible factual analysis, detail, and ordering. Generally, the right brain drives creativity while the left manages logical functions. The two are connected by another brain structure, the corpus callosum. Some tasks are better served by one hemisphere over the other, but all complex tasks require some degree of cooperation between the two. The amount of cooperation needed may be overlooked. I would postulate that Wiles' proof of Fermat's last theorem required just as much creativity as Beethoven's Moonlight Sonata.

All of the vehicles in the competition, such as Kat-5, had a relatively common driving algorithm: external sensors fed low-level modules that controlled the vehicle's speed, direction and decision making. Each robot had a predetermined set of instructions and made logical choices based on the environment. Stanley operated similarly, but also had an adaptive, holistic learning program that kept a log of "human reactions and decisions" when a human drove the vehicle in test runs. In this respect, Stanley did not operate on a discrete set of rules since his rules could adapt. Stanley succeeded because he was the most human.

Constrasting Stanley and Kat-5 may be instructive for approaching law. The DARPA contest suggests that there is an advantage to using human, evaluative processes over mechanical, logical ones in navigating one's way through future, unforeseeable obstacles. Kat-5, or any robot, cannot tell the difference between a speed bump and a pedestrian who falls in the street—unless the human programmer foresees this problem ahead of time. In Jerome Frank's words, “no one can foresee all the future permutations and combinations of events.” So, are human balancing tests preferrable? Do we have ways of determining when to use balancing tests and when to use bright line rules?

Unask the question

The hitch here is that these two modes are not distinct. The human creation of a bright line rule is an inherently non-robotic, evaluative process. And, the human application of bright line rules elicits other human, illogical right-brained processes that may be far more disturbing than what was eaten for breakfast. On the other hand, balancing tests seem to always incorporate robotic subroutines. “Undue burdens” and “substantial effects,” are really lists of robotic subprocesses, such as determining whether certain factors are absent or present. The two processes are inseparable to the point that previous questions have the zen equivalent of a mu value.

Then, the real difference between “right-brained” Stanley and “left brained” Kat-5 is not straightforward. The programs of Stanley and Kat-5 both involve left and right brain processes. Since all robots have initial human programming, the logical robot rules have an inherent hybrid right-left process imbued in their code. But, beyond this, they all follow a discrete set of rules and procedures. Except Stanley. He refines his initial ruleset using adaptive “open to interpretation” left-brained programming and refines the rules based on subsequent hybrid processes. Stanley's first real advantage over Kat-5 and the rest of the DARPA field is that his right and left brains are more frequent communicators; he has a more robust corpus callosum. Second, he has better debugging software programmed into his brain. He still operates within rules and procedures but his rules have been refined through many iterations of integrated left-right brained processes.

Stanley's software operates like U.S. statutory and constitutional law should operate. The legislature creates initial principles needed to govern, reflecting a set of desireable societal goals. The principles are tested out in the course of everyday life and flaws inevitably arise from their rigid application to circumstances unforeseen by the intial programmers. So, the judiciary steps in, applies law to fact, and refines the intial rules to better serve positive goals. After numerous iterations of applyng law to fact, precedent develops that helps the system navigate through similar future obstacles. Each iteration is a debugging exercise that uses previous errors to avoid future ones.

Initial Takeaways

First, mushiness in law is fine, if not desirable, so long as there are adequate procedures built into the system to hone and polish the law later. This process of error correction or "debugging" in the legal system will be expounded upon in another essay.

Second, no system functions optimally on extreme principles. In the present metaphor, the best qualified robot was the robot that struck a balance between the formalistic extreme and the adaptive extreme. A similar argument against "extreme operation" can and will be made about other systems.

Third, courts should restrain from judicial restraint. The essential tenet of judicial restraint is that “the legislature is supreme.” There may be positive value in conservatism, knowing that the law will be the same every day. Nevertheless, increased judicial restraint moves us closer toward the maladaptive behavior of a Kat-5. It threatens to divorce communication between the left and right hemispheres of the legal system. Here, I agree with Justice Stevens—rational basis review “is tantamount to no review at all.”

The reality is that judicial restraint and its dopplegangers function less as judicial posture and more as a legally acceptable presentations for everyday life—stomachable alternatives to human commitments to unjust outcomes. Even so, the theater has tangible effects. The audience starts to frame arguments and painfully unfunny jokes in legal terms completely void of any substantive meaning. And, through grotesque osmosis, we enter robotic states of mind that threatens to shut down our corpus callosa. In time, we start to lose the main advantage that we have over even the most adaptives of Stanleys.