Young African-American males experience a high incarceration rate. Progressives, conservatives, and libertarians each have a
desired cause for this.
Progressives: racism in the criminal justice system
Conservatives: high propensity of young African-American males to commit crimes
Libertarians; the war on drugs
Progressives prefer the oppressor-oppressed axis, which makes racism the desired cause. Conservatives are most comfortable with the civilization-barbarism axis, which makes criminal behavior the preferred cause. Libertarians prefer the freedom-coercion axis, which makes the war on drugs the preferred cause.
I claim that trying to argue that one of these is the cause is an unwinnable argument. Each of these causal forces has an element of truth, or at least plausibility. The chances are slim of coming up with an empirical analysis that decisively rules in favor of one cause and rules out all other causes.
In general, an unwinnable argument about causality is any argument in which one tries to affirm that X is the sole cause of Y or that X is not at all a cause of Y under circumstances of high causal density.
For example, arguments about the role of financial deregulation in the financial crisis of 2008 tend to be unwinnable. The case for seeing financial deregulation as the sole cause is compelling only to people who are inclined to espouse it. The case for seeing financial deregulation as not a factor at all is compelling only to those of us who are inclined to emphasize other causes.
Some further claims:
1. When there is a desired cause (meaning a cause that fits well with one’s political axis in the three-axes model, chances are the issue involves an unwinnable argument.
2. If your objective is to win an unwinnable argument, then you will tend to engage in normative sociology. To turn your desired cause into the cause, you have to filter out evidence that might support another causal factor and only discuss evidence that supports your desired cause.
It hardly requires saying that I think that it is counterproductive to try to win an unwinnable argument. It is almost as counterproductive to try to reason with someone who is convinced that they can win an unwinnable argument.
I am not saying that it is counterproductive to try to make an argument for or against something being a causal factor. However, I think that it does help to keep in mind that when a desired causal factor is involved it is challenging to remain objective in assessing the evidence.
There is a Cowen-Hanson paper Are Disagreements Honest? that you should read if you have not done so already. One of the reasons that disagreements can persist is because the protagonists engage in normative sociology.
As a libertarian I see the truth and fallacy within all three. I think I’m a pretty typical libertarianm
Maybe on some progressive or conservative site there is some version of me making a similarly even-handed claim on a post by their version of Arnold. Ah, I’m just kidding around.
Agreed — all are factors (the drug war, racism, and urban black culture). And I think that conservatives and progressives might agree as well. I’d expect the differences to be mostly in the amount of weight placed on the 3 factors.
Actually, I don’t see progressives and conservatives giving any ground.
I don’t remember seeing libertarians not giving ground. I know it is just an example but it wouldn’t even occur to me to claim the drug war is the sole reason for disparate black crime.
Maybe if I had to argue for my team every couple years I’d be just as biased, but I doubt it.
It is insufficiently appreciated, not least among libertarians, that diversity of opinion, including what Sir Isaiah Berlin calls incommensurables (absolutely incompatible values), is an inescapable cost of freedom. If had all the same opinions, there would be no need for freedom. By vastly expanding the range of expressible and actionable thoughts and convictions, freedom mass-produces dissent and political strife. Under these conditions, what matters is that we be able to manage this divisiveness in such a way as to maintain peaceful conditions (plus high levels of productivity and personal autonomy). That is largely the job of politics. It’s a difficult job, but it needs to be done. And, indeed, we have stumbled upon and designed quite successful forms of keeping peace in free societies, which is unappreciated by libertarians as is what I call “the spontaneous order of politics and the state,” which should be a natural attractor of the libertarian’s curiosity, but she is barred for ideological reasons to conceive that evolution happens everywhere, not only in the market, and that in the political sphere there are lots of moves made as if by an invisible hand (i.e. successful evolutionary peace- and-trust-building adaptations that people are not aware of).
For more see: sage://viewer/#feed/http%3A%2F%2Ffeeds.feedburner.com%2FRedstateeclecticCommentary
Sounds like begging the question on first blush. “THIS intervention is necessary to prevent the other interventions.”
I would also counter that libertarians by, for example, wanting to free the roughly half of people who are in prison on drug-related charges (and taxing policing, security, judicial resources, not to mention definitely causing corruption and barbarism in addition to disparate impacts- and not least noteworthy NOT solving the drug problem) are actually more apt to focus on what you suggest they under-appreciate.
I have had the exact argument with a conservative who assured me that marijuana was the gateway drug to hell.
Seems to me the real world is hardly ever “A causes X” but rather almost always “A, with B, and no C, unless there’s also D, 50% of the time causes X, and 24% of the time X and Y”
Or “A less than 100 units causes X, but A more than 100 units has no effect at all”
But I think Arnold may have cherry picked the example, as all three arguments, and the interactions between them, have merit here.
Perhaps worthwhile to find a topic with less well founded causes?
A corollary to Arnold’s Law (or maybe this IS Arnold’s Law) might be that high causal density gets funneled into the desired cause during two-sided debates. I have thought this came about because of the two-party system. Now I wonder if it is a more basic feature of human nature.
Politics is for persuasion. You want people to focus on one direction. This has to come at the expense of focusing on other directions or diluting effort. So, you argue that that one thing is the keystone problem in order to move the effort in that direction.
You don’t even have to really think that the problem is the solution. Maybe your side just doesn’t want the actual problem(s) to be focused on. Or you just want to hurt the other side. Your argument takes whatever form serves what you think are your sides’ interests.
Maybe it is what makes us smart that is what makes us stupid.
Boyle’s Law says there is an inverse relationship between volume and pressure. But no one ever claimed that was the only thing that influenced pressure, or that the relationship was perfect, or the same for all gases, or continued to apply outside a certain operational envelope. There is definitely ‘causal density’ with regards to gas pressure. And this was in the mid 17th century, before any modern theory of gas statistical behavior and certainly before quantum mechanics.
And yet the proposition of Boyle’s Law is something that can be investigated and argues about in a fairly straightforward and ‘winnable’ way. Criminal activity also has a tremendous causal density, and yet there are correlation relationships we can observe and chart that are at least as tight as Boyle’s data was.
I submit (I would say in concordance with Tullock’s The Organization of Inquiry, that what makes an argument unwinnable is not causal density but when it’s in the ‘not even wrong’, i.e., not even properly framed as a legitimate argument or any proposition posed in a testable, measurable, falsifiable way.
That is, what makes a proposition the basis of an ‘unwinnable’ argument is that the proponent is playing pretty fast and loose with their underlying model, and the empirical meaning of the vague terms they are using to describe nebulous phenomena, and playing ‘motte-and-bailey’ word games to preserve the option of moving the goal-posts when it comes to any attempt to contradict their factual assertions or the logic of their arguments.
In other words, they are preserving a kind of ‘unfalsifiability-of-convenience’ while denying the fact of that unfalsifiability. What makes an argument unwinnable is that your opponent refuses to give you enough information to allow you to prove them wrong, no matter what you do. They will not reduce their assertions to firm and tractable claims.
A good example is the recent econoblogosphere kerfuffle over ‘austerity’. Is this an unwinnable argument, and if so, why? Sure, there’s a lot of causal density. But I submit that it’s not unwinnable if you can force someone to define what they mean by ‘austerity’, what their model predicts, and what results would look bad for their theory’s claims. And what really makes the argument unwinnable is that you can’t get these people to put down their chips on any particular number on the roulette table in advance of the the ball drop.
So, what really enables ‘normative sociology’ is the tolerance of this kind of convenient hedging from which one is always able to salvage the reputation of their favorite explanation, no matter what happens.
“There is definitely ‘causal density’ with regards to gas pressure.”
Nice one.
The larger point is that I think our host creates a perception which amounts to a false equivalence between the sort of claims he lists.
The opposite of ‘winnable’ is ‘losable’. The ‘conservative’ proposition that criminal behavior it ‘the‘ sole cause of disparate black incarceration rates is losable. Indeed, do not many progressive scholars and critics of the criminal justice system make just this claim – that there is evidence that white and black detectable criminality rates are more similar than their incarceration rates, and that therefore anyone who makes the ‘unitary explanation’ case is wrong? I have seen conservative scholars admit that other factors are at play, and indeed, few people make the kind of extreme, ‘unitary’, claim Kling is posing here. They say something more reasonable, which is that the relationship is strong and has such strong predictive and explanatory power, that to the extent other factors influence incarceration rates, they are of secondary or minimal importance by comparison.
The ‘libertarian’ proposition that it is all due to the drug war is also losable, depending on how you measure it, but one gets into fuzzy territory when one tries to estimate the no-drug-war counterfactual scenario, or starts to weigh whether disparate incarceration rates for non-drug-related offenses may be second and third order effects of the drug war. I have also seen libertarians freely admit that other factors are at play, and make similarly reasonably moderated statement such as that above.
But the ‘progressive’ proposition that it is all due to racism is not losable. If one’s model of human behavior assumes human-subgroup statistical uniformity in the absence of racism, then any disparity is evidence of racism one way or another, because the cart is put before the horse. I honestly don’t think I have ever seen a progressive look at the data sets that would support the alternative propositions above and ever moderate their position.
So we have an odd man out here, which is the true origin of the unwinnability of these arguments, and the “a pox on both your houses” framing is inapposite.
Spot on Handle. Building further off of Arnold’s example, we can look around the world and observe the fact that everywhere we look black crime rates are higher than other ethnic crime rates. Now, maybe their incarceration rates are lower in Haiti, but that’s just because the overall criminal justice system is totally screwed up in Haiti (i.e. overall barbarism is higher.) So in fact, Arnold picked an almost perfect example of a case that completely proves his theory wrong. Crime has a strong correlation with gender, genetics and age — give me two populations and tell me their gender/age/ethnicity and I’ll make a bet on which has a higher crime rate (and win every time.)
Note. Nothing in this proves Arnold wrong.
In fact, I’m willing to go with nothing proves Arnold wrong.
Complex systems endure epistemological problems. For such systems a precise causal analysis mandates a trial and error discovery process. Identifying the exact premise and the ideal fix for the phenomenon in question is closely connected with the resolution of this involved knowledge problem.
Social sciences actually pose insurmountable epistemological and structural questions, the micro components and causal connections are infinite. Raw data is essentially unstructured and we employ various theories to look for patterns. With unstructured data, eventually those pure theories determine the relevance of the patterns we observe.
Hayek precisely stated in ‘The Sensory Order’ — “much that we believe to know about the external world is, in fact, knowledge about ourselves”
http://politick.me/2014/04/20/relativity/
Is this argument really not winnable? At least as between the progressive and conservative view on the one hand, and the libertarian view on the other?
It seems like you can pick any particular criminal law where libertarians agree that the criminal law is a legitimate and appropriate exercise of state authority (maybe murder), and then test whether that particular law shows a higher incarceration rate for some ethnic group than others?
There seems to be a bit of circularity in the libertarian argument. An ethnic group has higher incarceration rates because the government has enacted an overly broad criminal law, so the ethnic group is therefore caused to break that law more frequently?
The problem with both the Progressive and Conservative way of defining problems is that, even if (or to the extent) they are true, they do not lead to viable solutions.
What if young black males really are more likely to commit crimes? What is the solution? Increasingly draconian penalties? The more you push a population down (no matter how the target group is defined), the more insular and less invested they become in participating openly in civilized, voluntary, mutually-beneficial interactions. It devolves to a state of war pretty rapidly — more crime, not less. What then? Deport all blacks to Africa, the way Lincoln dreamed of doing? Mass extermination? In other words, if you define the problem by Conservative terms, the implied solution only makes things worse.
And what if its true, as Progressives contend, that there is systemic racism in the criminal justice system? When you want to prove that any particular case was improperly handled, it’s no evidence to show macro, global statistics. What do you do to identify the presence and effect of this systemic racism in any particular case? How do you solve that global, systemic problem? Give half of black male convicts a pass, randomly? That’s not a justice system. Who would invest any trust in a system of justice that overtly depends on spinning a wheel of fortune? Therefore, if you define the problem by Progressive terms, the implied solution only erodes the criminal justice system altogether.
Furthermore, the Conservative and Progressive positions feed and depend on each other — young black males really are more likely to commit crimes, because drugs are defined as crimes, and they naturally also commit more violent crimes surrounding the drug trade. But Progressives are correct to point out that the criminalization of drugs is itself an example of the systemic racism they are talking about, since the criminalization of drug use is very clearly driven (politically) by anti-black sentiment (often as a variant of middle class anti-poor sentiment, generally).
As a result, the Progressive and Conservative arguments are destined to go around and around each other, arguing in circles, like a chicken-and-egg debate, indefinitely. They both not only support each other’s fundamental premises, but neither of them leads to a rational solution.
The argument between Progressives and Conservatives is unwinnable because the problems, as they’ve each defined them, are intractable, and the solutions they imply are not viable.
The libertarian position cuts through both of them. Poor minorities would cease to commit as many crimes if fewer behaviors were defined as crimes, obviously. Decriminalization would also reduce the need for drug dealers to resort to extra-legal violent methods to solve their problems, since their disputes will be solvable through ordinary, legal means. I doubt that any legislation can eliminate racism, but decriminalization (of all but the most directly-harmful violent and property crimes) would at least remove the ability of the government to express its racism through police dogs, no-knock raids, random frisking, roadside strip-searches and felon-branding.
The larger problem is that, in a pseudo-democratic infotainment-based kleptocratic slave state, as we live in now, the State WANTS social issues to break down on intractable lines. As Orwell described in Nineteen Eighty-Four, the point of war was not to win, but to never stop fighting wars. They like it when battle lines never shift meaningfully. I see that as a metaphor for the intractable social issues that the State is called upon to solve. The State LOVES it when millions of people call on them to be an (ineffective) solution to an intractable problem. It excuses them from ever having to accomplish anything. It gives them job security. Statism loves a stalemate.
If genetics/biology cause violence and poverty, then we should work on learning the mechanisms, indicators, and invest in potential biological treatments. That is the long term view, but it’s really, really important.
If culture causes violence and poverty, our culture should shame it so that people make better choices and feel embarrassed to perpetuate the problem culture.
It seems that Blaise Pascal wrote on this a long time ago. Judith Curry made a reference to this today in “Pascal on the art of persuasion” in her article.
I think Pascal had some good suggestions on how to proceed.
I don’t agree with what you say about conservatives. I think their desired cause is the welfare state.
Bingo, Dave. The conservative cause is breakdown of black family caused by welfare state.
With all do respects for libertarians, the war on drugs argument is crap. If drugs were legalized (as they largely should be) you would still end up with largely same group in prison. Think, broken windows policing. Also, many so called drug offenders in jail are really in jail for something else, and pleaded guilty to a lesser drug charge. The liberal argument is so dumb I will not address it other than to say is there a different level of black prison population from areas controlled by blacks.
I would say the Welfare State argument is common among both libertarians and conservatives. The Drug Law argument is unique to libertarians, because Conservatives soundly reject the idea of repealing drug laws. But every libertarian I’ve ever heard address the issue would definitely agree that the Great Society expansion of the welfare state halted the decline of poverty and reversed it, corroded the family, and made crime vastly worse.
Arnold didn’t not say that.
On high incarceration rates for young African-American males:
The way I see it the “libertarian” explanation makes no sense. Sure, drug laws explain why incarceration rates are higher in general, but what exactly is being asserted here-that young black men are more likely to be involved either on the supply or demand side of the illegal drug trade? Why? Libertarians have no satisfying answer that I’ve heard, they usually fall back on racist enforcement: blacks don’t actually use or sell more illegal drugs than white people, but law enforcement is racist. So when you get back to the root of the actual problem, Libertarians and Progressives are more or less in agreement on the issue.
Now, I find Progressive explanation to be paranoid: it’s possible, but it sounds like a conspiracy theory that’s just a little too convenient and hard to believe.
To be fair, the Conservative explanation isn’t so much an explanation as it is a description. What is actually needed is an explanation of why young black men engage in criminal behavior more than whites do. I think most conservatives are a little uncomfortable speculating about that, but at least they don’t deny that it’s true.
Personally I think it’s a cultural issue, related to perverse effects of group identity amongst black people.
The libertarians are right about one thing though: if we legalized drugs, there would be fewer people incarcerated in general. That’s not really why we should do it (and we should do it), but it’s true.
The problem with the financial deregulation argument is that it attributes an effect to a non existent cause.
You forgot to mention the welfare state as a cause, based on the idea that idle hands are the workshop of the devil.
And BTW I think that the cause is not so important, the solution is. More police on the street might help so it seems to me whatever the cause it is worth a try.
Yes. And I’m also in favor of empowering citizens. Or at least reducing the disincentives for citizens to act more like cops on the street.
But look at what the the discussion was like on that front with the Zimmerman/Martin case. Even though citizens would always have less impunity than cops people still love elitism. Politically we are a long way off.
But they are also (finally) getting mad at cops behaving badly- with all the wrong lessons learned.
Maybe the answer is more entry-level cops that lose their jobs at first offenses.