A nation ignorant and free, that never was and never shall be. Thomas Jefferson
George Rebane
[This is the addended transcript of my regular KVMR commentary broadcast on 8 November 2017.]
Numeracy is the name given to a set of math oriented skills that an educated lay person is supposed to have in order to hold a job, function as a householder, and understand the day’s news and issues. Such skills include the ability to do arithmetic, understand graphs and charts, be able think logically, comprehend the basics of chance and risk, and so on. For a numerate person, the mastery of these basic skills need not be at the level of a technician, but sufficient to comprehend today’s work-a-day world that will be ever more complex tomorrow.
Well, it appears more and more with each passing day that this is not the case. The innumeracy of our electeds is already legend among those who can and do pay attention. Combined with their other character shortcomings, our fiscal messes provide enough evidence of that. But now, warning bells are also starting to ring from the highest levels of our judicial branch. Recent cases involving gerrymandering reveal that our Supreme Court is proudly innumerate as they reject arguments that describe the basics of playing with congressional district boundary lines to ensure that one or the other party gains or remains in power - a very technical yet centuries old enterprise in our republic.
I was recently surprised to learn through a correspondent that “the Supreme Court is allergic to math”. Oliver Roeder of 538.com points out “the justices, the most powerful jurists in the land, seem to have a reluctance to taking math and statistics seriously.” Currently this is in evidence in the case of Gill v Whitford that “will determine the future of partisan gerrymandering”, and how that will infringe on voters’ rights. There is some serious math involved in the algorithms that are used to compute the gerrymandered district boundaries. And at least four justices have voiced that the math is “unwieldy, complicated, and newfangled”, with one justice calling the whole thing “baloney”. Meanwhile the reality of computing such boundaries to affect political outcomes continues, no matter how little of the matter is accepted or understood by the court. (more here and here)
Now there are some ways to minimize the court’s numeracy requirements, as taught in the system sciences, ways that have yet to be broached by the lawyers in the case, who seem to be equally limited. And many legal scholars are beginning to understand that the fault here lies with traditional legal education, as delivered by the nation’s top law schools. Such training ignores the promise of advancing technology that more and more cases the court hears will hinge on sophisticated notions that can only be described with statistics and algorithms, the understanding of which will bear on the already muddied legal definitions of what is ‘fair’ and ‘just’.
The real worry here is that if our Supreme Court is flummoxed by hoary disputes such as gerrymandering, how can it possibly rule in cases involving, say, subtle discrimination, equal opportunity, or complexly regulated markets, all the measures of which today involve available numbers and statistics. And we have yet to consider the expected tsunami of cases that hinge on the rampant implementation of intelligent machines, and advanced proprietary processes such as deep learning with big data. Without a numerate judiciary, our nation will go forward with an ensconced powerful cadre of mentally medieval gatekeepers making critical decisions in a technology-dominated world. I think it’s past time to broaden the requirements for a 21st century law degree, and also for some mandated remedial training for those who already rule how the rest of us should lead our lives.
My name is Rebane, and I also expand on this and related themes on Rebane’s Ruminations where the addended transcript of this commentary is posted with relevant links, and where such issues are debated extensively. However, my views are not necessarily shared by KVMR. Thank you for listening.
[Addendum] To reduce the numeracy acumen of the courts, specifically SCOTUS, in cases such as to determine the best national prescription for gerrymandering congressional districts, it would be more than useful if the contending sides would gather with the court and attempt to put in place some of the guidance developed by the Harvard Negotiation Project led by Roger Fisher and William Ury (here, here, and here).
The contending sides should first attempt to define a mutually acceptable ‘utility function’ (UF) or simply the utility that can be used to compute how acceptable or ‘good’ is a given solution – in this case the set of gerrymandered boundaries drawn by some algorithm. We should recall that ALL utility functions are subjective – there is no UF that comes down the mountain to us inscribed in stone; we make them up to suit our needs. Such a utility will input the boundaries and the socio-political demographics of the underlying geographical area, and output a number. The larger the number, the better the mutually acceptable solution.
All utility functions are subjective and can be first described in the King’s English without resorting to any fancy math, algorithmics, or how the utility is computed – the UF is just a black box with known in puts and a single number output. UFs may also be defined in terms of several attributes, which is the usual case. (Here, such things as the district’s ratio of area to perimeter length, and closeness to, say, the state’s proportion by party affiliation.) The main notion here is that the definition of acceptable solutions and their relative ratings will be known to all parties involved, who can then work supportively work together to find either the ‘best’ or a mutually acceptable solution. Such a process also identifies the winning algorithm that computed the solution which can then be codified into law or mandated for use in some other way.
If such a mutually acceptable UF cannot be agreed upon, then both parties present to the court the UF that suits each, and once approved by the court, the search for the best gerrymandering algorithm proceeds in an open and transparent process. In this case the final solution will probably be the one that discovers what in game theory is known as the Nash Equilibrium (q.v.) solution. An NE solution is one for which the computed UF for either party will be diminished if that party tries to unilaterally abandon the solution while the satisfied party maintains it. In short, it pays for all parties to accept the NE compromise.
This approach can be used to resolve many other cases of contention where a single outcome that will be imposed on both parties is required. And nowhere in the proceedings does it require the judge/justices to acquire any “newfangled” numeracy skills.
For more esoteric adjudications involving numerical consideration wherein the parties will be asymmetrically judged – one wins (more), the other loses commensurably – the court can select an appropriate approach from to the rich body of work known as Fair Division Methods (google ‘fair division methods’)
The overarching principle that I believe our legislative and judicial institutions should abandon is purveying the fiction that there is some universally acceptable, absolute, or gold standard definitions of ‘fair’ and ‘just’ that are either sought or appended to rulings rendered by our institutions of collective power. There are none, and the worst that can be done is for our mandarins to project the hubris of certainty and absoluteness from their high-level perches.
The problem that we in the nebbish seats have is that we are either ignorant or indifferent to what is going on and how things are resolved while our attention is elsewhere. The usual response of the masses has been to not care until things become so bad as to require the aggrieved to reach for the pitchforks and torches, and march to the town square. Although there have been exceptions, such attempted solutions seldom turn out well.
We are quietly heading into an AGI (artificial generalized intelligence) crisis as we busy ourselves rearranging deck chairs. Take note of the comment stream below.
Max Tegmark in Life 3.0: Being Human in the Age of Artificial Intelligence discusses the advantages of Robojudges and the implications.
. . . it’s impossible for human judges to master all technical knowledge required for every possible case, from thorny patent disputes to murder mysteries hinging on the latest forensic science, future robojudges may have essentially unlimited memory and learning capacity. One day, such robojudges may therefore be both more efficient and fairer, by virtue of being unbiased, competent and transparent. Their efficiency makes them fairer still: by speeding up the legal process and making it harder for savvy lawyers to skew the outcome, they could make it dramatically cheaper to get justice . . . Starts at Kindle Location 1941
At Kindle location 1974-1975:
Do we want human judges to have AI-based decision support systems, just like tomorrow’s medical doctors?
Yes! I am for providing our innumeracy challenged justices an AI Decision Support System to handle their math problems, including balancing their checkbooks.
Posted by: Russ | 08 November 2017 at 08:05 PM
"I think it’s past time to broaden the requirements for a 21st century law degree, and also for some mandated remedial training for those who already rule how the rest of us should lead our lives."
I'd rather they follow their instincts to not jump off the deep end and accept a theory of what is an acceptable partisan gerrymandering vs. an unacceptable partisan gerrymandering based on a mathematical analysis that the court is not equipped to judge AND THAT IS NOT EXPRESSLY EMBODIED IN LEGISLATION. If Wisconsin or California wants to enshrine a mathematical test for acceptable gerrymandering, they should BUT TO DATE, THEY (WE) HAVE NOT DONE SO.
The only reference to "baloney" I could find was Roberts' statement:"... the intelligent man on the street is going to say that’s a bunch of baloney" when the SCOTUS accepts or rejects a claim, that “It must be because the Supreme Court preferred the Democrats over the Republicans. … And that is going to cause very serious harm to the status and integrity of the decisions of this court in the eyes of the country.”
So, no, that was not Roberts saying the approach was baloney, only that the average Joe on the street would consider a ruling by them based on these theories as baloney. And I suspect the average Joe on the street who never grokked algebra would think exactly that.
Posted by: Gregory | 08 November 2017 at 08:41 PM
It is my understanding from several sources that all gerrymandering in the country has been done for some time now by machines implementing this or that algorithm. Before we have numerate judges, I believe the approach I outline relieves them of having to "accept any theory of what is an acceptable partisan gerrymandering vs. an unacceptable partisan gerrymandering based on a mathematical analysis".
I wonder who accused Roberts of saying the approach was baloney.
Posted by: George Rebane | 08 November 2017 at 10:19 PM
The US Constitution leaves the issue of how people vote to the states. I think it would be a usurpation of the tenth if the SCOTUS tries to take that away.
Posted by: Todd Juvinall | 09 November 2017 at 09:49 AM
"I wonder who accused Roberts of saying the approach was baloney".
George, you just threw out (it's unclear to me if it was a direct quote from 538.com) that one justice called it baloney, and the only "baloney" uttered at the court that I could find was the quote from Roberts that was not in his voice. If there was another, I'd like to see it.
Give the Toddster a cigar for his 9:49am
Posted by: Gregory | 09 November 2017 at 10:19 AM
Russ 805pm wed, whether the gerrymanderings were by neato algorithmics or not, they are not shoved down the people's throat by the computers but generally voted upon by legislators or other native criminal classes in each state, according to state law. The idea that we should make available AI devices to lead justices and judges by the nose when the math gets above where you are comfortable is abhorrent to me. A truly wretched idea. Bad. Really and truly bad.
Posted by: Gregory | 09 November 2017 at 08:34 PM
In this discussion we should not confuse the definition of the subjective utility function (that measures the goodness of a set of gerrymandered boundaries) with the algorithm that generates the boundaries which maximizes the utility function (provides a set of optimized boundaries). These are two distinct tasks, the first is based on human subjective inputs, the latter is based on some very sophisticated optimization techniques that are rendered into code. Humans cannot reasonably be expected to beat the optimization program to generate better boundaries as evaluated by their approved utility function.
Posted by: George Rebane | 09 November 2017 at 11:20 PM
In context, the 10th amendment was another state's rights law designed to appease and accommodate the ownership of slaves. Like the electoral college, the 10th was designed to give equal political power to the less populated slave states so that they could prevent the federal government from passing anti-slavery legislation or voting laws that would allow abolitionists to gain public office and overturn the practice of slavery. The Constitution is riddled with antiquated features that favored slavery in the name of 'state's rights.'
Posted by: Robert Cross | 10 November 2017 at 09:24 AM
Quite the imagination there BobbieC. That is a twist I have never heard of or considered at all.
Posted by: Todd Juvinall | 10 November 2017 at 10:23 AM
Yes, RC 924, I'm sure you see the Constitution as a relic that needs abolishment... the essence of the "Living Document" argument... that it means what you want it to say is in danger with every textualist appointed to the bench.
States do not have rights, they have powers, and the 10th restricts federal powers, no more, mo less. A good idea that the Congress used the Interstate Commerce Clause to drive around in much the same way the Panzerkampfwagens drove around the Maginot Line.
Posted by: Gregory | 10 November 2017 at 10:43 AM