Thursday, June 5, 2014

Cargo Cult Arguments - Reparations edition

It is quite interesting, from an argument perspective, to see how commentators respond to Ta-Nehisi Coates’ Atlantic article, The Case for Reparations. Much of the response has been to not directly engage with Coates’ arguments but rather to engage with the reader’s own concept of reparations. This probably has to do in part with the fact that it is a 16,000 word article. Rather a daunting hill to climb, particularly if you have any concerns about the premises to Coates’ argument. In fact, Coates alludes in an interview that perhaps many people are only acting like they have read the 16,000 words.

The criticisms of the concept of reparations center around a lot of operational and philosophical issues. Here, in no particular order, are the ones I see most commonly being raised:
• We don’t compensate for general crimes, we compensate for specific crimes
• Many actions that are condemned today were legal at the time
• Many actions that are interpreted today as repressive and detrimental were seen as progressive and beneficial at the time
• We don’t accept group guilt for crimes committed by individuals
• We don’t believe in blood guilt (assignment of responsibility for a crime to those who share an attribute such as religion or race)
• We don’t believe in intergenerational guilt (guilt of the fathers being visited upon the sons)
• We don’t retroactively impose new values on old actions
• All groups have suffered at the hands of individuals (and the government) to a greater or lesser extent. There is no basis for determining what makes one class of victims superior to another.
• The loss of 600,000 lives in the Civil War and the expenditure of a trillion dollars on social welfare programs (believed to disproportionately benefit African Americans) is held to expiate any conceived group guilt
• The negative socio-economic indicators of African-Americans today are not evidence of systemic discrimination or of history but are the consequences of present-day individual decision-making
• The capacity of African, Caribbean and Asian refugees and immigrants to thrive In America today is taken as evidence that there is no color based discrimination
• The fact that slavery and judicial discrimination were prevalent in some parts of the country but absent in others is held to limit group responsibility to only those directly involved
• Slavery and its consequences are held to have been a net diminishment of national productivity so there are no beneficiaries of any color
• Descendants of those who emigrated after the Civil War (or 1900 or 1945 or 1965) are held to have no potential liability
• A claim for African-American reparations must necessarily open the door for a claim by Native Americans and all other immigrant groups who have been disadvantaged and exploited in the past
• Policies have trade-offs, prioritizing one goal above another doesn’t mean that the second is disavowed
• Disparate impact is no basis for concluding that there is or was any nefarious intent
• Many negative consequences in the past were the unintended outcomes of pursuing other goals, i.e. there was no intent
• Total absence of limiting principle to the proposal of reparations (how far back, how small a crime, who is considered a victim, etc.)
• Belief that any transfer of wealth would have no long term impact on African-American socio-economic outcomes
• Rejection of race based decision making

All good points. But these are criticisms of the general concept of reparations, they don’t necessarily relate to Coates’ argument. If he took 16,000 words to make his case for reparations, it would be courteous to at least engage with his actual argument.

I ignored Coates’ article for a while but then began to get interested in looking at it as an argument. How well does Coates lay out a defined problem, the root causes of the problem, the measured manifestations of the problem, the direction and evidence of causation, a solution to the problem, and a plan for implementation of the solution? Everyone is critiquing the idea of reparations but no one seems to be critiquing the quality of Coates’ argument. So how good is it?

Not good. This is primarily 16,0000 words of rhetoric with very little in the way of a structured argument seeking to illuminate or discover. In fact, there is virtually no argument at all, merely assertion dressed up as argument. Paraphrasing Richard Feynman, this is Cargo Cult Argument. It has the form and feel of an argument but none of the rigor and structure. It is a Petri dish of logical fallacies and evidentiary errors.

Essays of this sort follow one of two paths (or sometimes a hybrid). One path is to make an assertion that such-and-such is true under so-and-so circumstances and usually includes some sort of recommendation based on the asserted truth. These arguments are then supported with evidence and logic in an effort to convince a reader of the validity of the argument. A distinguishing feature of truth seeking arguments is that they usually include robust data from credible sources, evidence that is accepted by mainstream practitioners, a clear structuring of the argument including definitions and assumptions, and most especially an engagement with counterfactuals and disconfirming evidence. In other words, they not only seek to prove the positive value of their argument but also seek to show why alternative explanations are not true.

The alternate form of argument is much more common because it involves much less work. This is to make a rhetorical argument rather than an empirical argument. With a rhetorical argument, you are seeking to persuade people of the truth and value of your position. It doesn’t matter if the position is empirically or logically supported so long as it is believed.

And then you have the blended versions. As mentioned, these are Cargo Cult Arguments. They look like the pursuit of truth, they might include some smattering of data and evidence, there is the pro forma of a scientific method, there might even be some maths and statistics. But it is strictly for appearances. There is no effort to lay out a coherent sequence of necessary statements connecting the predicate assumptions with the stated conclusions, critical concepts are left undefined, assumptions are made without being stated, strawmen arguments advanced, ad hominem attacks made, questions begged, circular logic indulged, etc. All these defects are in evidence in Coates’ article along with other biases, errors and fallacies including correlation pursued without causation, red herrings, a reliance placed on the affect heuristic, use of affirming the consequent, indulgence of amphiboly, fundamental attribution error, appeals to belief, appeal to emotions, unintentional appeals to ignorance, attentional bias, avoiding the issue, base rate fallacy, bias blind spots, biased sampling, clustering illusion, complex cause fallacy, composition fallacy, conditional probability, confirmation bias, congruence bias, conservatism bias, converse accident, Disregarding Occam’s Razor, expectation bias, failure to elucidate, fallacy of exclusion, false analogy, false assumptions, false continuum, false dichotomies, reliance on fluency heuristic, focusing effect, framing effect, goal confusion, group attribution error, hasty generalizations, hindsight bias, identifiable victim effect, ingroup bias, insensitivity to sample size, ipse dixit, genuine but insignificant cause, joint effect, limited depth fallacy, limited scope fallacy, loaded language, missing context, ignoring moral hazard, negativity bias, non sequitur, obfuscation, outcome bias, one sided argument, passive voice, post hoc ergo propter hoc, prejudicial language, priority confusion, reactive devaluation bias, selection bias, Semmelweis Reflex, shared information bias, missing situational awareness, special [pleading, spurious correlations, stereotyping, survivorship bias, tautology, too broad definition fallacy, tu quoque, unacknowledged trade-offs, the undistributed middle fallacy, unstated assumptions, weasel words, wrong direction fallacy, and zero sum heuristic.

Now it is not unusual for even short arguments to demonstrate multiple biases , errors and fallacies. We assume the audience has the same assumptions, knowledge, goals and priorities as we do and we thereby fail to make the case adequately. It is also true that the longer the argument and the more complex the argument, the more likely it is that there will be a greater variety of biases, errors and fallacies. There’s just more scope. Finally, no matter how many biases, errors and fallacies there are, it does not make the argument untrue. It just makes it untrue for the reasons offered. But as a matter of experience, the more biases, errors and fallacies, the more likely it is that the argument is not true.

Coates has a lot of biases, errors and fallacies in his argument. He also has an unusually wide variety of biases, errors, and fallacies. But he also has a notable intensity of biases, errors, and fallacies. They aren’t random. He relies heavily on the fundamental attribution error, appeals to emotions, avoiding the issue, biased sampling, circular reasoning, compositional fallacy, confirmation bias, disregard of Occam’s Razor, failure to elucidate, fallacy of exclusion, ipse dixit, framing effect, identifiable victim effect, limited depth fallacy, loaded language, missing context, post hoc ergo propter hoc, stereotyping, and tautological arguments. Each of these argument defects show up multiple times over the entirety of the article.

Because Coates’ is a rhetorical argument, you have to disambiguate the article to get at what might be the argument on which the rhetoric is built. Doing this requires a lot of work and charitable interpretation. Where the author does not specify the problem, the causes, the measured consequences, the proposed solution, how the solution is to be implemented, and the benefits of the solution, then you have to make a lot of inferences. Hybrid and rhetorical arguments are advantageous to an ideological advocate because they off-load the work of constructing the argument on to the reader. By having to disambiguate the unstated or obfuscated argument, the reader has to make inferences that they author can then reject, but usually still without actually clarifying their position. This process is already underway with Coates rebuking a number of critiques essentially by claiming to be misunderstood but without clarifying.

The first thing that comes in to focus as you sift through the words is that Coates has at best only a narrow targeted proposal. He wants the US House of Representatives to pass HR 40, the Commission to Study Reparation Proposals for African Americans Act, a bill introduced every year for twenty-five years by Congressman John Conyers, Jr. and defeated every year for twenty-five years. Coates’ is arguing that we should pass HR 40 in order to understand how reparations might work. But that begs the question. Why examine the how if there is no agreement on the whether. In other words, if there is no agreement whether reparations are owed, then there is no need to examine how they might be paid. Coates does this time and again, assuming into existence that which he is supposed to be proving.

Getting your opponent to accept an action that is predicated on your argument being correct is one of the oldest rhetorical devices around and is representative of the whole article. Coates appears not interested in establishing the moral or legal grounds for reparations. Instead, what he wants is non-black America to accept his interpretation of American history without argument and pay some undetermined amount of money simply on the basis of racial identity and with no identified improvement in the national welfare arising from that action. In doing so he prioritizes African-American interests as uniquely deserving above all other comparably positioned claimants such as Native Americans, lower class citizens, impoverished groups, Women, etc. In fact, it is critical for his argument that no other claimants exist.

In order to advance support for his direct proposition that HR 40 be passed and for the broader proposition that reparations are owed to all black Americans by all other non-black Americans, Coates spends probably 14,000 of the 16,000 words documenting particular episodes of egregious events in the past that disproportionately affected African-Americans. The problems with this approach are legion. Frequently, Coates advances interpretations of events that may be received in the cloakrooms of grievance studies programs but are not widely shared in mainstream academia. Often, Coates is describing negative events that were shared by all citizens. For example, local operators fleecing people moving from the country to the city are a staple of all immigrant experiences. Irish, Chinese, Eastern European Jews were all defrauded of possessions and in commercial transactions. Did it happen disproportionately to African Americans moving from the Mississippi Delta to Chicago? Perhaps, but that is not on the face of it obvious and Coates makes no effort to make the case that the real suffering of such internal migrants was different in kind or degree from the suffering of foreign immigrants. Coates routinely ignores the trade-off complexities of past decisions. For example, across the South, particularly in the Tennessee Valley and Appalachians, electrification in 1920s-1940s was seen as a critical predicate to development. But in the process of bringing that benefit to the region, innumerable citizens were displaced from their land and not infrequently defrauded by federal land purchase agents. Bad things happened in pursuit of beneficial ends – how is that calculus worked in creation of a heritable claim?

Coates relies heavily on local and narrowly constrained circumstances to make the case for an overall indictment of the nation. He spends a lot of time documenting the hardships of African Americans in Chicago. Was Chicago representative of the whole nation? Certainly not. Close? Outlier? No way to tell. Coates only documents the local conditions and then indicts the whole nation.

Looking through the entire article, there are about 150 major statements of fact or opinion that are either demonstrably wrong, arguable (matter of interpretation), unsupported, or asserted based on faulty logic. Rather than go through those individually, better to summarize.

Fundamentally the problem is that Coates advances no basis for race-based reparations. It isn’t that he advances a bad argument. It is that he advances no argument whatsoever. He dips his hand into a lot of kettles and pulls out whatever might look good in support of a raw assertion of the need for reparations but there is no effort to build a coherent, robust and convincing argument. There is no argument to be answered because there is no argument made.

People agree that bad things have happened in the past. People agree that some groups have suffered more than others in the past. People agree that some people today suffer disproportionately in terms of particular socio-economic measures of well-being such as morbidity, life expectancy, education attainment, income, wealth accumulation, etc. People agree that past circumstances influence in some degree current choices.

That is a platform on which to build. But from there you have to build some sort of consensus around numerous issues including:
• That the past determines (not just influences) the present
• That the role of personal current actions are minimal in terms of life outcomes compared to the consequences of past events
• That blood guilt, group guilt and generational guilt are valid bases for moral decision-making
• That new values can be the basis for exacting payment for old actions
• That the success of numerous immigrant groups does not refute the proposition that there is systemic racism today
• That there is any measurable positive national good that might arise from the payment of reparations
• That the claim by African-Americans is distinctly different from the claims of all other possible groups who have suffered in the past and continue to suffer today
• That there was some beneficial wealth creation that occurred out of the treatment of African-Americans that can and has been passed down to the non-African-American community
• That there are both just and feasible limiting principles that can guide reparations calculation and apportionment
Absent agreement on any of these issues, there is no basis for reparations other than raw assertion of personal preference. More fundamentally, as long as people are broadly interested in evidence, causation, personal responsibility, equal treatment of people in the face of the law, the moral foundation for claims, etc. the argument for reparations cannot advance simply by assertion. Coates waves all these issues away, repeats stories of suffering, and concludes that something is owed. There is no there.

Coates believes the current conditions of African Americans in the US are almost entirely attributable to institutional and societal racism which has been and is being exercised in a deliberate fashion.

In disambiguating Coates’ argument from the clutter and rhetoric, the most charitable construction seems to be something along the following lines. I have followed each necessary statement (necessary in order to make sense of Coates’ argument) with two numbers in brackets. The first is the percentage degree that I am confident that Coates would agree with the statement. The second number is my estimation of the degree to which the American public (an averaging of the general public and domain specialists) would accept the truth of the statement. For example, Necessary Statement 6 (deliberate white supremacy) is rated (100, 25). The first number, 100, indicates that this is almost a direct quote from Coates. The second number,25, is my estimate of what percentage of the population (including historians) believe that all the identified exploitations were undertaken specifically to support white supremacy (as opposed, for example, for personal gain, fear, or as unintended consequences in the pursuit of other goals).
1. Blacks in America have suffered repeated and grievous exploitations and financial setbacks at the hands of particular individuals and particular sets of municipal, state and federal laws. (100, 95)
2. Similar exploitations and financial setbacks were suffered by other specific groups (Native Americans, Women, the Lower Classes, Immigrants, Religious Minorities, Ethnic Minorities). (70, 100)
3. These exploitations benefited particular individuals and classes of individuals who were overwhelmingly non-black (including other specific groups as above). (100, 90)
4. The benefits accruing to individuals have passed down through multiple generations and have had a material benefit to present day non-blacks. (85, 20)
5. The legal/institutional disadvantages (such as redlining) created benefits to non-blacks which continue to accrue only to non-blacks to the present day. (80, 15)
6. The exploitations were undertaken consciously and deliberately to support white supremacy. (100, 25)
7. Slavery and continuing exploitation of African-Americans was net beneficial to white society above any alternatives. (95, 5)
8. The financial setbacks arising from past exploitation have prevented blacks from benefitting to the same extent as other groups. (100, 20)
9. Racism and white supremacy continue to disadvantage blacks today. (100, 35)
10. Non-white immigrant success (Nigerians, Haitians, Cubans, Chinese, Koreans, etc.) does not negate (Necessary Statement 9). (100, 20)
11. All negative disparate impacts are the result of design and not of unintended consequences. (90, 10)
12. The improvement in socio-economic measures (and narrowing of gaps) between 1900 and 1965 (and the subsequent decline) does not negate the hypothesis that all current negative outcomes are attributable to past exploitation. (90, 20)
13. No benefit of past circumstances have accrued to current day blacks. (100, 10)
14. No alternate explanations than white supremacy and active racism explain current black negative socio-economic numbers. (90, 10)
15. The black experience is sufficiently different in kind and degree to set it apart from comparable experiences of all other claimant groups. (100, 15)
16. Past exploitation of Native Americans, Women, Lower Classes, Immigrants, Religious Minorities, Ethnic Minorities have not had comparable detrimental present day consequences as suffered by African-Americans. (100, 10)
17. Racial reallocation of wealth from non-blacks to blacks is the only way to make blacks whole for the accumulated past sufferings. (100, 5)
18. Such reallocation has to be accompanied by a sincere endorsement of the Coates interpretation of history. (85, 5)
19. That financial reallocation will allow blacks to become successful members of the national community with comparable socio-economic outcomes going forward. (60, 0)
20. Such a reallocation will be of worth to non-blacks by allowing atonement for guilt over the past actions of other people. (95, 5)
I believe this is the actual disambiguated case for reparations as advanced by Coates in disjointed and scattered fragments throughout his long article. All the Necessary Statement need to be true in order for Coates argument to work. Presented in logical sequence and with clarity in this way makes it possible to go back to the article to find where Coates actually provides clear evidence to support each Necessary Statement. When you do so, all you get is clear evidence for Necessary Statement 1. All the other Necessary Statements are unsupported assertions. Could the other Necessary Statements be better supported? Marginally but not much. It is not that Coates did not provide the evidence that exists. It is that the evidence does not exist, is mixed and contradictory or is plagued by definitional problems and/or subjective interpretation. Absent robust evidence, all you can do is assert the statements to be true. Hence the reliance on rhetoric versus logic and evidence.

The Atlantic trumpets this as the article that makes The Case for Reparations. It does no such thing. It lays out Coates’ interpretation of reality. When you pay attention to his argument it becomes clear that this interpretation is not shared and will unlikely ever be shared by the public at large.


No comments:

Post a Comment