Science Needs Rational Debate

I wrote Objective Judgment, Chess Competition and How Science Is Failing. Here are some further comments on those issues.

Skill at winning chess games is skill at chess. To a good approximation, there's just one thing, not two separate things. By contrast, skill at doing scientific research is different than skill at debating scientific issues.

So how could debate help science, in parallel to how chess competitions help chess? The best chess tournament performers are approximately the same as the best chess players, while the best debaters are not the best scientists. So there's a concern that emphasizing scientific debates more would favor people who are good at debate but less good at science.

I have two answers.

First, we don't just need more of the kind of debates we sometimes have today. We need rational debates that do objective truth seeking. We need a type of debate where skill at debate helps you figure out the best conclusion. Skill at debate should make debates more productive, not help biased people make "their" side win.

The type of debate where you use rhetorical skill to make "your" conclusion win the debate, regardless of the truth, is irrational. We need debates using truth-seeking methodology and debaters who are trying to organize the arguments for all sides to figure out an appropriate conclusion given current knowledge. Flawed, biased, adversarial people can participate in debates and bring up arguments, but for debate to be productive it's important that at least one person involved is being more objective. Someone has to organize the arguments and analyze things fairly instead of just choosing a side and then making arguments for "his" side.

Debate should be a somewhat cooperative process where people try to document, organize and untangle the arguments about a topic. Having experts in the arguments for each conclusion being considered is helpful. If no participants liked or advocated a conclusion, or had much expertise about it, then the participants might not know all the arguments it has or treat it fairly.

What's the difference between a debate and a critical discussion? In a debate, people have reached contradictory conclusions and they want to address this disagreement. So, if the debate reaches a conclusion about the issues, then at least one person will change their mind. The people can't both or all be fully right.

Second, anyone can debate. If a scientific researcher is bad at debate, other people can debate about the ideas he discovers. If no one agrees with him and makes the arguments he would, then he'll need to participate (or find or create better proxies). If zero people will participate who have anything good to say about his ideas, then there's a problem. To give his discoveries and arguments a fair hearing in debate, there needs to be someone with expertise about them who is willing and able to talk about them.

Participating in debate shouldn't usually mean getting on a stage in front of an audience or dealing with time limits. While that's one way to debate, it's not necessary and has various downsides. One can share arguments by writing essays, papers, books or forum posts. In general, if no one writes down arguments publicly, then they won't be taken into account in current debates on the topic. But if someone shares those arguments, then debaters can use them.

Direct participation in debate can just be saying that you wrote a few essays, writing one sentence summaries of your main arguments (with links or citations to the full essays), and specifying where they should go in a debate tree. Debating doesn't require talking in voice or interacting with people in real time; those are both optional. Nothing about reaching a true conclusion requires voice or real time interaction.

If you put arguments in a debate tree, then others may reply. They may share refutations of your arguments for the debate tree. Then you or anyone else who knows how should give rebuttals of those refutations – or reconsider your conclusions. (Changing your conclusion doesn't necessarily mean deciding that the other side is right. You could e.g. decide that you don't know enough and need to do more research. You could reduce your confidence and become more neutral. As you learn more, you might find that something similar to your original view was correct, or that the other side was correct, or some other conclusion entirely.)

As long as people make good, relevant arguments about the issues, a scientist who is able to think about the issues should be able to participate in debate reasonably well. He should be able to explain how and why he reached his conclusions. He should be able to explain his lines of reasoning.

In a good debate, just understanding your own reasoning and being able to explain it is enough to participate productively. Special skill at debate isn't necessary. Just being able to think conceptually about your expertise, and communicate about it, is good enough.

What if people argue in problematic ways that take debate skill to untangle and defuse? Then someone with such skills will need to do that job. Hopefully there are enough rational people in the world, with debate skill, to pay attention to all the important topics. If there aren't, we better encourage more people to develop those skills.

People often think they can recognize bad arguments that they can't refute – so they can just say those arguments are bad without giving counter-arguments. (Or they think they can refute the argument, but don't want to spend time actually doing it – and expose their refutation to criticism.) Viewing ideas as refuted without refuting them has a high false positive rate. People reject both good and bad arguments that way. Trying to refute an argument is typically the best way to figure out more accurately whether it's good or not. If you're able to refute it, you can look at the refutation in retrospect and learn something about the quality of the argument. But before you have a refutation, you can't reliably know if you will be able to refute it or what sort of refutation it will be. (This is an example of how you can't predict the future growth of knowledge. David Deutsch pointed out that if you could rationally predict that in the future we would know something, then you would know it now.)

In general, if an argument is easy to refute, then refuting it isn't much work. If refuting it is a lot of work, then it wasn't obviously wrong and you shouldn't have assumed it was wrong. If it takes you a lot of work to refute arguments that should be easy to refute, then your methods are incorrect, and you should improve them. Learn how to refute things more efficiently instead of assuming they're false without refuting them.

The world needs some thinkers, with skill at debate and philosophy, who can evaluate arguments including dealing with arguments that don't focus on science. Debate doesn't reliably stick to one field or one expertise. Some generalists with some expertise at epistemology are needed sometimes to help with debates. Most scientists can't do that very well, but science needs that, so they need to rely on other people (e.g. debate experts) or else more scientists need to learn more about debate, philosophy and more.

If some scientists want to participate in debate, that's great (as long as they learn to debate well instead of making excuses about how they are scientists not debaters). If other scientists don't want to be involved in debates and want to stick to research, that is fine, but then they should be happy if anyone else is willing to do the debating that they don't want to do, and they should respect the outcomes of the debates that they chose to leave to others. Scientists who don't want to debate should also have some willingness to answer some questions about their field so the debates can be more informed.

Relevant skills that someone needs to bring to debates, to make them more productive, include being able to criticize a pattern or category of bad arguments instead of responding to each argument one by one. Responding one by one sometimes doesn't work due to there being a steady stream of similar arguments. By the way, it doesn't require bad faith to keep making similar arguments. If you keep refuting similar arguments one by one but don't see and point out the pattern or similarity, then you should not be surprised if the person making the arguments is unaware of the pattern or similarity. Or they might know the pattern but not see what's wrong with it because you only gave refutations of individual things instead of criticizing the pattern.

Rational debate involves evaluating ideas which often requires some skills outside of a scientific field like chemistry. Whether chemists participate or not, someone with appropriate skills needs to in order to sort out controversies in chemistry. The sort of being "good at debate" that we should value and respect is skills like dealing with patterns of bad arguments and being good at writing clearly and organizing ideas. Those are some of the skills that help debates rationally reach conclusions instead of being inconclusive messes or biased fights between opposing sides. (It's also typically hard to have productive debates without some domain experts. E.g. debating chemistry issues without any chemists involved usually doesn't work. Some expertise at both debate and at chemistry should be present in order to have a productive debate. One person who knows about debate and one chemist could do a good job if they were reasonably objective instead of each being biased for a side. If people are being pretty biased, then it'd be important to have both types of expertise on each side.)