Scott Sumner had a recent post about how people get so hung up on how to define words, or what labels to attach to certain ideas, that they lose sight of actual issue at hand. I largely agree, and I think a very important skill to develop is the ability to “taboo your words,” as Eliezer Yudkowsky once said. 

Very briefly, Yudkowsky imagines two people who seem to disagree with each other over whether the proverbial tree that falls with nobody around to hear it still makes a noise. Two people argue with each other about it, one saying yes and the other saying no. But the one saying “yes” is defining noise to mean “acoustic vibrations,” while the one saying “no” is defining noise to mean “auditory experience.” Both parties actually agree about the state of the world. They both would agree that the falling tree generated acoustic vibrations and did not generate an auditory experience – but because they’re using the same word (“noise”) to describe these different phenomena, they are going in circles over a disagreement that doesn’t actually exist. 

Now, you’d expect this dispute to be easily rectified once they realized the source of the confusion. And in this specific case, it might even work out that way – once they realized they’re just using the same word to mean different things, they would also realize their dispute has been dissolved and go on with their day. Wouldn’t it be odd if instead, they ended up starting another argument insisting the other person’s definition of “noise” was objectively incorrect? 

That’s pretty unlikely to happen because nobody feels personally invested in how “noise” is defined. But it happens all the time in political discussions, because so many terms, once politicized, carry an emotional valence with people. Sumner’s post focused on whether or not “addiction” is properly categorized as a “disease” as opposed to a “character flaw.” Someone with a disease is automatically coded sympathetically, whereas someone whose problems arise from character flaws is coded unsympathetically. Thus, people fight each other tooth and nail over the semantic issue of what constitutes a “disease” because they want to support a more (or less) sympathetic view towards people addicted to drugs. 

We see the same thing play out in political disputes. For example, racism is a very charged, highly valanced word. Everyone agrees that racism is bad – which is why there is so much furious dispute over what is or isn’t racist. To successfully brand some idea or action as racist is an automatic victory in the debate over whether or not it’s good or bad.  

For example, suppose you think equal outcomes is intrinsically good, and you wanted to convince other people to think the same way. One way to do that is to engage and refute arguments made by the Princeton philosopher Henry Frankfurt that equality of outcome has no intrinsic value, or the arguments by Michael Huemer in support of the same conclusion. Or you can skip all that and insist that unequal outcomes are racist. That is, you can claim not merely that unequal outcomes can be a result of racism – the claim is that unequal outcomes just are racist, by definition. And if unequal outcomes are racist, that means they must be bad, because anything racist is bad. You can simply define your side of the debate into victory. 

This appears in the worldview of Ibram Kendi. By Kendi’s lights, equality (treating people equally without regard to race) isn’t the goal. His goal is equity (treating people differently according to race to get equal outcomes). As a result, he thinks discriminating on the basis of race can be a good thing – and if racial discrimination can be good, that means racial discrimination can’t be what it means to be racist. He says this very plainly, stating “racial discrimination is not inherently racist” because what actually matters is the outcome of discriminating on the basis of race – i.e., “whether the discrimination is creating equity or inequity. If discrimination is creating equity, then it is antiracist. If discrimination is creating inequity, then it is racist.” If discriminating against people on the basis of their race leads to more equal outcomes, and unequal outcomes are racist, then you’d have to be a racist to be opposed to racial discrimination. And we all agree that racism is bad, right? Hence the push to define “racism” not in terms of processes, but of outcomes.

Or say you’re someone who is opposed to an “either/or” mindset. You can try to convince people to abandon such a mindset through arguments and reason. Or, you can take a shortcut and just declare that using an “either/or” mindset is actually a way of promoting white supremacy culture – and since we all agree that white supremacism is a bad thing, we don’t need to do any more work. Framing something as either/or promotes white supremacism, and white supremacism is bad, therefore the either/or framing is bad. Case closed. 

Of course, such tactics can lead movements to eat their own tails. After all, one key point of Kendi’s worldview is that there is an all encompassing and mutually exclusive binary. You are either an antiracist, or you are a racist – there are no other options and no middle ground. Again, he says this plainly in his book, writing “One endorses either the idea of a racial hierarchy as a racist, or racial equality as an antiracist. One either believes problems are rooted in groups of people, as a racist, or locates the roots of problems in power and policies, as an antiracist. One either allows racial inequities to persevere, as a racist, or confronts racial inequities, as an antiracist.” So, by framing the issue of racism as an either/or dichotomy, Kendi is also promoting tenants of white supremacy culture. To be clear, that’s not actually my view – I don’t believe Kendi promotes white supremacy culture. My point is simply to show the absurdities people can trap themselves in when trying to play these definitional games. 

Unfortunately, I don’t know of a good way to get people to drop the baggage and association attached to words and phrases and focus on the underlying issues instead. But I’m told recognizing a problem exists is at least a necessary step to fixing it – and this is definitely a problem that needs a fix.