The Philosophy of Similarity

Some good ideas sound like bad ones.

Writers will often note that one proposition looks like a different one. Next, they will poke holes in the latter proposition in order to refute the former. Tom Scocca provides one such example in his review of Kathryn Paige Harden’s The Genetic Lottery. Harden’s book argues that some of the variation in intelligence stems from our genes, and this variation can explain some of our economic inequality. The book also does not advocate for racism or any sort of racial differences in intelligence. Rather, she says that no such evidence supports the belief that racial IQ differences stem from genetics. In short, she supports the stuff that has evidence, and doesn’t support the stuff that lacks evidence. Weirdo. Harden also believes that we can maintain a commitment to social equality in light of these genetic differences. Maybe errs in her analysis of the genetic data and political implications thereof. If so, I’d like to see them refuted. I want to hold accurate beliefs. Instead, arguments against her work (and similar ones, like that of E.O. Wilson) debate the views she doesn’t hold. Scocca, for instance, notes the similarity between Harden’s work and the work of earlier phrenologists, who used differences in skull shapes to advocate for a racist social structure. He also compares it to Charles Murray’s The Bell Curve, a book which, according to Scocca, poses a genetics-driven explanation for racial IQ differences. 

Introduction to Similarity

Is The Genetic Lottery similar to phrenology and race science? Of course! Both argue that some social inequality reflects genetic differences. Harden’s book is also similar to books that argue against genetic determinism. It’s also similar to Stephen Hawking’s A Brief History of Time, since both discuss scientific research. It’s also similar to Hamlet; both were written in English. A lot of things are similar in a lot of ways, but that’s not how arguments work. We can’t reach a conclusion through mere similarity. Consider the argument below.

Premise 1: Babe Ruth played baseball.

Premise 2: Baseball is similar to cricket

Conclusion: Babe Ruth played cricket. 

Premise 2 doesn’t prove the conclusion. Babe Ruth did play baseball, and we can prove that through a variety of historical evidence. Meanwhile, no such evidence exists for his cricket career. Furthemore, our disbelief of Ruth’s cricket career does not stem from differences between baseball and cricket. Imagine a game called “bazeball,” which mirrors the rules of baseball, except that balls that bounce off the top of wall before entering stands count as doubles rather than home runs. That sport would be even more similar to baseball than cricket is, but our argument still fails. 

Premise 1: Babe Ruth played baseball

Premise 2: Baseball is super duper similar to bazeball

Conclusion: Babe Ruth played bazeball.

Similarly alone doesn’t help an argument. While we often employ comparisons in moral arguments, we use these comparisons to exemplify morally relevant premises. Ultimately, the It’s that morally relevant premise that matters, not superficial similarity. I know that sounds confusing, so let me run through an example.

Abortion

Pro-choice adcovates will argue that a fetus is, in some important ways, not like a post-birth human. Pro-life advocates will argue that a fetus is, in some important ways, like a post-birth human. This debate includes other arguments involving bodily autonomy, but I’ll ignore that piece for now. To keep it simple, I’ll examine the question of whether or not the first-trimester fetus constitutes a human life. I intend to present arguments that both sides would consider satisfactory (though simplified) summaries of their beliefs. 

Pro-life side:

Premise 1: Killing a biological human life is immoral

Premise 2: A first-trimester fetus is a biological human life

Conclusion: Killing a first trimester fetus is immoral

Pro-choice side:

Premise 1: Killing a living being that lacks mental activity is morally permissible

Premise 2: A first-trimester fetus is a living being that lacks mental mental activity

Conclusion: Killing a first-trimester fetus is permissible.

On the pro-life side, the interlocter can note the common intuition against killing a biological human. Babies aren’t the smartest, but we still oppose killing them, as we do with adults that suffer from mental impairments. Meanwhile, the pro-choice side can argue that biological human-ness doesn’t seem to hold any moral importance. We don’t condemn men for killing millions of their sperm. We can also imagine a lab that grows and kills artificial organs for research purposes. Here, scientists would regularly kill biologically human entities, but I don’t think most of us would mind. 

Pro-choice advocates can point toward the regular killing we commit every day. We kill bacteria on our hands and weeds in our yard, and we don’t seem to mind that. However, we condemn the killing of other humans, and at least acknowledge that some level of harm occurs when we do the same to certain animals. Why is it okay to kill bacteria, pretty bad to kill cats, and really awful to kill humans? The key difference here involves mental activity. One cannot deny the biological human-ness of the fetus, but it lacks the necessary neurological infrastructure of the post-birth human. On the flip side, pro-life debaters could draw attention people in comas and people with certain neurological disabilities. 

I could continue, but I don’t intend to solve the abortion debate here. Rather, I think this issue helps exemplify why we use similarity in moral arguments. Each side creates a comparison in order to highlight a morally relevant factor. For the pro-life, they cites the biological humanness of a baby and a fetus. Meanwhile, the pro-choice side notes the non-mental-ness of a fetus and various other bioligical entities. The disagreement arises from the correct-ness of each argument’s first premise. Does the biological category of human matter, or does the mental infrastructure matter? That’s part matters. It doesn’t matter whether a human fetus is more similar to an adult human or a cactus, however we’d determine that. 

Returning to Scocca, the level of similarity between Harden’s book and racist ideas doesn’t matter. If people want to argue against Harden, they’d need to use these similarities to find some underlying moral premise. In other words, we’d have to find what makes attempts at race science incorrect, and then see if Harden’s book shares those traits. To start, I’ll make a few arguments below that I think hold water. 

Premise 1: Phrenology is a methodology that produces inaccurate scientific results.

Premise 2: We should ignore the conclusions of methodologies that produce inaccurate scientific results.

Conclusion: We should ignore phrenology.

In this case, can we swap phrenology for Harden’s genetic research? If so, then we can ignore her findings. However, our decision to ignore her findings would stem from a belief in the science being wrong. It wouldn’t depend on her research being similar to phrenology. . If Harden gets the science wrong, that’s a case worth making, but superficial comparisons to junk science don’t help us. I’d change my mind if someone refuted her data or methodology, but none of the arguments I’ve seen against her work provide such a refutation.  Note that, In the argument above, we can substitute homeopathy and p-hacking, even though they don’t seem similar to phrenology.

Here’s another argument against a hypothetical viewpoint that I will call Theory X:

Premise 1: Things that rank people into racial hierarchies are immoral.

Premise 2: Theory X ranks people in racial hierarchies.

Conclusion: Theory X is immoral

This seems like a compelling case. However, Harden rebuts inequality and race science in the book, so it wouldn’t apply. Maybe the case looks like this instead:

Premise 1: Things that could be used to create racial hierarchies are immoral.

Premise 2: Harden’s research could be used to create racial hierarchies.

Conclusion: Harden’s research is immoral.

Of course, I think any good-faith reader of The Genetic Lottery would reject the second premise. Still, even if we grant it, premise one doesn’t seem right. Racists use all kinds of ideas to support their conclusions, like inequalities in income and educational attainment. Yet, these are often the same inequalities that progressives highlight to support redistributional policies. Thus, “could be used” is too weak, it would force us to ignore the pivotal problems that we want to solve. I don’t think anyone can find a relevant point of similarity between Harden’s work and the junk science of racists.

Bayes, I guess

I understand that most of us don’t think in terms of discrete premises and conclusions. More likely, most of us employ something similar to the Bayesian reasoning discussed in the rationalist community. Basically, we hold some extant belief on a topic (a prior), and then we update that prior based on new evidence. When I see an ad for a new diet supplement, I don’t examine the evidence for that individual supplement. Instead, I have a prior that none of that crap works, so I conclude that this new one probably doesn’t work either. I don’t have time to examine the scientific literature on each one, so use my prior. Maybe, one day, a pill will cause significant weight loss, and a series of randomized controlled trials should update my prior. Until then, however, I’ll assume that these supplements suck.

Furthemore, I think this kind of reasoning helps us avoid absurd conclusions. For example, one study found that a brief look at the Thinking Man statue reduced the viewer’s religiosity. When exposed to this finding, one should examine it against a few priors. First, people’s religious beliefs run too deep to be changed by a single artwork. Second, I don’t see much connection between thinking and rejecting religion. St. Thomas Aqunas remains one of the Western canon’s greatest thinkers, and there’s no shortage of atheist dummies. With that in mind, it shouldn’t surprise us that later psychologists haven’t been able to reproduce the study. Thus, people who stuck with their priors would have held a more accurate belief, while those who accepted the study’s finding would have held a less accurate belief.

Since I still like my philosophy 101 premises and conclusions, let’s make Bayes-esque arguments with them. Imagine I watch a Youtube video and see an ad for New Weight Loss Supplement X. I might think something like this:

Premise 1: A weight loss supplement is extremely unlikely to work

Premise 2: New Weight Loss Supplement X is a weight loss supplement.

Conclusion: New Weight Loss Supplement X is extremely unlikely to work.

This is a valid argument that hinges, and the cogency hinges on premise 1. For now, I feel safe in accepting this premise. Again, future research could find some successful supplements. If it does, I’ll reject premise 1 or change it to something like “A weight loss supplement other than [thing the research supports] is unlikely to work.”

But This Thing Isn’t That Thing

I suspect something similar occurs when people balk at discussions of genetic differences. I share some of this squeamishness, and I think we should remain skeptical of those who argue that nature justifies the status quo. I remain extra skeptical of claims that justify racism. I imagine Scocca, Harden, and my readers maintain similar priors. That’s great, because none of these beliefs contradict The Genetic Lottery! Haden doesn’t think there’s evidence for racial differences in intelligence, and she doesn’t think genetics justifies current inequalities. She also doesn’t think that genetics alone determines our intelligence.

In that case, what prior does her book contradict? Is it the belief that genetic differences have nothing to do with inequality? I don’t think many of us accept that. We understand that some people suffer from disabilities, like blindness, that can hurt their potential. Scocca himself mentions the (likely genetics-based) height difference between LeBron James and Chris Paul. Does he have a prior that genetics never impacts behavioral traits? How likely is that? We know that dogs act differently from cats regardless of how humans train them. Siblings act differently despite undergoing the same parenting. Also, how plausible is it that evolution has impacted every organ except the brain? Maybe I’m missing something, but I can’t find a reasonable prior that would cause someone to reject Harden’s book. 

Ultimately, some good ideas sound like bad ones. Unfortunately, similarity alone can’t tell us which ones are right and which ones are wrong. We must maintain priors to avoid accepting an  outlandish idea, but we have to consider whether our priors are reasonable. We also need to engage in some self-reflection, and ask ourselves why an idea feels right or wrong to us. It’s easy to reject ideas that seem to come from the wrong “team,” but that’s just not how truth works. As I will continue to reiterate on this blog, we can’t shortcut reason.