There’s a pretty clear cut list of states that tend to be universally known by Colombians: New York, Texas, California, etc. Alabama, where I went to school, isn’t usually high on that list, and when it is, it’s either because of the Lynyrd Skynyrd song or for a history littered with some of the worst of America’s past. Texas makes for a much better (and more accurate, in my case) state to claim; it only invokes John Wayne-esque finger pistols and questions about horses and cowboys.

Yet one thing I’ve found myself increasingly enjoying is explaining to foreigners and New Yorkers here alike (there’s a lot) some of the incredible work that people are doing in Alabama and across the South. The Times, as I say to earn some credibility with this NY audience, recently sang the praises of the construction of a memorial honoring victims of lynching in Montgomery. I opine it’s a truly moving memorial that is, quite literally, unlike anything I’ve seen in the US. It’s something to help Americans heal and reckon with our past; I’ve lived in 5 counties throughout the South, and each one, I discovered, has had a lynching take place there. This is one example of such work.

Here is another. After laying eyes on the memorial, a friend from the Equal Justice Initiative, the organization behind the memorial, gave me a copy of Just Mercy, chronicling the work founder Bryan Stevenson has done as a lawyer in Alabama. It’s a poignant and tragic look at a broken justice system and its victims, and a celebration of those who have received both justice and mercy from the work EJI is doing. It’s on the NYT best-selling nonfiction for four years for a reason: read it.

A side note on the book: it’s a curious title that leaves a kind of linguistic suspense. Does the title mean only mercy, or that mercy is just? Should I guess and color my reading with that interpretation, or ignore the title until its meaning is more clear? The first is a tough question, unanswered until the last chapter, but it’s a fair assumption that a title from a Harvard JD isn’t a poor use of language, but rather intentionally ambiguous. What is law, anyway, without a clever use of words.

One thing I’ve been chewing over after reading is a thought on policy-making, namely, that good policy accounts for future mistakes as inevitable, as integral parts of human-based systems rather than growing pains able to be eradicated. Good policy-making asks about false positives (Type I error) and false negatives (Type II error): what happens when things go wrong?

In the case of Stevenson, it’s when things go wrong in the context of court. What happens when someone is wrongly convicted of a crime? Let’s ignore, for this context, the humanistic elements in the discussion on policy around sentencing, if that’s even possible (as Stevenson puts it, “Each of us is more than the worst thing we have ever done”).

You could make a substantial list of the number of things that go into judicial error, but it’s well documented that convictions and sentencing is heavily influenced by who the suspect is. Are they wealthy, educated, from a respected family? Poor and uneducated? White? Black? The book is colored with stories of this, of inmates without fair representation, facing juries whose minds are all but made up pre-trial and judges unwilling to admit any sort of uncertainty.

The laws that Stevenson challenges in the book are predicated on complete accuracy, an unwavering faith in the success of institutions to accurately and objectively determine, 100% of the time, fair allocations of right and wrong. Any amount of time spent thinking about human nature renders this a fairy tale. Here’s two of such laws:

  • The death penalty for homicides
  • Mandatory life imprisonment for juveniles convicted of homicide

It’s easy to show how flawed these laws are, as they don’t account for any kind of institutional error. There are two assumptions at work here: (1) a judicial ruling is always correct, and (2) a defendant given the death penalty or life imprisonment is forever a danger to society, incapable of offering something else.

Any mistake in either one yields harsh consequences, especially the first. A false conviction that leads to an execution is, quite ostensibly, irreversible. The premise to mandating sentencing thirteen year-olds to life in prison is that they’re always innately incapable of becoming productive members of society. Just imagine, for a second, the horror of getting either of these wrong. What kind of a society is that?

There are of course two sides to this coin: Type I errors and Type II errors, which could occur in either the conviction or sentencing stage. In each case, an analysis of potential cost yields laws based on complete certainty a tragedy. If someone is sentenced to the death penalty or a kid is handed life imprisonment and either the conviction is unjust (he’s innocent) or the sentencing (he isn’t a threat to society), that’s a far more egregious Type I error than a Type II Error, where someone having committed homicide is spared the death penalty and given life imprisonment, or a kid avoids spending the rest of his life in prison for a crime he committed when his primary concerns were trying to fit in in the 7th grade. Comparing the consequence of each, Type I appears far more disastrous.

One point of clarification here is that evaluating the potential consequences of error is a function of both how bad the error would be and the likelihood of error – essentially a calculation of expected value. Yet, in these instances, the probability of committing each error would be unlikely to differ enough to overcome how significantly worse executing an innocent man would be.

In short, policy debates often exist in a silo, where the assumption is that institutions will always function how they’re supposed to. They’d be far more effective and representative if we shake this assumption and incorporate a little more error into our models.