Cognitive decoupling

From In-Groups, Out-Groups, and the IDW by Jacob Falkovich.

Over the past year or so, Sam Harris and Ezra Klein spent several tweets, a dozen emails, and a two-hour podcast vehemently disagreeing with one another. The ostensible cause of this disagreement was a dispute about whether or not there’s a genetic component to the black-white IQ gap in the US. However, neither was willing to commit to a concrete position on the issue. Both danced around the actual claim while deferring to various experts who may or may not suggest that a genetic component is more or less probable.

Did they disagree about Charles Murray? In the podcast, Klein says that he opposes Murray’s social policies but allows that Murray is “a lovely guy interpersonally” who should not be silenced. Sam Harris agrees that Murray is a good guy who shouldn’t be silenced but caveats that “his social policies are not social policies I’m advocating.”

So, what are these men actually disagreeing about?

In a thorough analysis of the Harris-Klein controversy, John Nerst suggests that what is actually at issue is whether the discussion of the racial IQ gap is a matter of science or of politics. Harris sees Murray as a scientist whose arguments and conclusions fall well within the academic consensus of genetic and cognitive science. Klein sees Murray as a ‘policy entrepreneur’ who advocates for reprehensible conservative policies. The notion that racial differences are genetic (and, in Klein’s understanding, therefore immutable), Klein argues, has historically been used to support destructive policies of bigotry and discrimination.

This isn’t merely a difference in emphasis. Underlying this debate is a divergence in modes of thought so wide that the two men were left practically at cross-purposes.

Harris is a scientist by training, and scientists depend on what rationality researcher Keith Stanovich1 calls “cognitive decoupling.” Decoupling separates an idea from context and personal experience and considers it in the abstract. It is the approach used in the scientific method, when performing thought experiments, and when generalizing principles from individual examples. In decoupling thought, arguments follow one another according to the formal rules of deductive logic.

The contrary mode of thinking sees every argument embedded in a particular context. The context of an idea includes its associations, implications, and the motivations and identities of those who advance it. In a lengthy written response to Harris published before their podcast, Klein makes use of all of the above:

The belief in black deficiency has been instrumental. It was used to justify slavery and to quiet moral qualms over unyielding oppression and violence.

(Association)

If the disparities we see in American life are the result of an intrinsic inferiority on the part of black Americans, then that diminishes the responsibility white Americans have to correct those disparities.

(Implications)

This also feeds into Murray’s proposal to devolve the entirety of the government’s social supports into a more regressive version of a universal basic income.

(Motivation)

For two white men to spend a few hours discussing why black Americans are, as a group, less intelligent than whites isn’t a courageous stand in the context of American history; it’s a common one.

(Identity)

To put this simply: You cannot discuss this topic without discussing its toxic past and the way that shapes our present.

(Context)

Klein is a political journalist. In politics, one rarely encounters an argument posed as a neutral thought experiment or research query. If a subject is under discussion, the arguments marshaled by both sides are usually advocacy in the service of one group or another. A political thinker always sees ideas in context, and the most important context is: who does this argument benefit and who does it disadvantage?

Very interesting to see this discussion. I have long been aware of the fact of two different conversations that can occur and the need to recognize and separate them.

One conversation is the scientific one. Is X real and if it is real, why is it real, and to what degree? The purpose of this conversation is discovery. What are the facts and how confident can we be in those facts? The world is dynamic, evolving, and the most interesting systems are often the most complex. Beyond the natural sciences, there is little in which we can have absolute certainty and even the natural sciences are hedged about by terra incognita. What can we know?

Which is well and good among like-minded inquirers. Regrettably, not everyone is so like-minded. For others, every conversation is a political conversation or an emotional conversation and the questions you ask are taken to be markers for your opinions. Many people's conversations are not scientific. They are conversations about power and opinion and emotion. What you feel about something takes priority over the facts themselves. Invocation of implication, motivation, association, identity, and context are wielded as rhetorical tools. Not to discover truth but to shape a desired outcome to the conversation.

I have always used the phrase compartmentalizing to describe keeping these different conversations separate from one another. The feelers cannot comprehend a truth seeking conversation and can become exercised by the questions being asked because they over-interpret those questions. For them, a question is never a question, it is a sign of an opinion that is right or wrong in their pantheon of opinions.

Compartmentalizing is a good term but cognitive decoupling is better. Compartmentalizing is simply keeping discovery conversations apart from opinion conversations. Cognitive decoupling describes the functional difference between the two conversations.

The important point is that the capacity to abstract the general from the specific and then to isolate the general from context in order to test hypotheses dispassionately and without implication, or impugned motivation, or association, or identity, or context is critical. Not many people do it at all much less do it well. Yet it is an easily comprehendible approach that lends itself to training. And a necessary predicate to progress. If we cannot engage with reality, then reality engage will eventually engage with us to our detriment.

All this is not to imply that implication, motivation, association, identity, and context are irrelevant. Once you have a testable hypotheses of some rigor from the bald logic and analysis of data, it is important to put that hypothesis into context, question the motivation, etc. Too often, though, the process is inverted and questions are never asked because someone wields the rhetorical aspect of implication and context, etc. in order to prematurely shut down inquiry in the first place. "You are not allowed to ask that!" is the sentiment and it is a poisonous sentiment for anyone committed to progress. We have to be able to test all ideas in all ways and not just hide them away because we feel uncomfortable about them.


UPDATE: Related - Toolbox-thinking and Law-thinking by Eliezer_Yudkowsky.