Astral Codex NaN: Dialoguing with Scott Alexander about the Culture War Frame
Alexander responds to my last post
After my last post, my second critiquing Scott Alexander’s Astral Codex Ten, I posted the article to r/slatestarcodex. There, Alexander actually took the time to respond to me! This was actually very cool and unexpected. We definitely disagree, but he did his typical thing of writing out his thinking in a way that was engaging and clear.
I posted a lengthy reply on reddit, but I thought I’d repeat it here for others who might be following along at home. It’s reproduced below.
The Main Point of Contention
Thanks for getting back to me. Given your response here, and some of the comments posted below, it seems like you and others in this group take umbrage at being lumped into the culture war narratives. The argument, as I understand it, is that the community around your blog explicitly rejects this type of culture war and is just trying to focus on reason and logic. This is a great goal, and I do think the quality of discussion overall is high. However, I think that this goal isn't always met in practice. As someone who does identify as somewhat left, I do find that when I read your work consistently I find little barbs and asides that feel like uncharitable or unreasonable attacks on positions that I agree with. It's worth pointing these out for a few reasons.
First, if the goal of the rationalist community is to give all ideas a fair shot in discussions, then it's fair for someone to point out when they think the arguments aren't meeting that goal. It sounds like you and I disagree on whether the points you made in the Pinker / Gardner piece meet that threshold, but I definitely stand behind my raising questions about them.
Second, I don't believe that anyone can really argue from a purely objective, "view from nowhere" perspective. We all bring specific biases and alliances to our writing, and those perspectives--like all models--simplify or mischaracterize elements of the real world. In your comment, you say I'm trying to "smuggle in" culture war terms. But the article is framed as an argument over rationality! And you say explicitly that you're on Pinker's side! It's set up as a conflict between opposing sides, and to understand who those sides are we need the broader context of the culture they came from. That culture is U.S. internet culture, which has been strongly influenced by bitter argumentation and flame wars between right-aligned and left-aligned factions. (You've written nicely about this in previous posts, where you say that you're part of the "grey tribe" and that it's hard for you to be truly critical of your ingroup. Which is a very honest and authentic thing to say! I feel the same way about my groups.) I don’t think we can ignore that broader context.
Just as you argue that I'm smuggling in the culture war, I feel like you're smuggling in a stance of objective neutrality. That's the major thesis of my article! And I completely stand behind that. I think that you were not thoughtful or gracious in how you addressed Howard Gardner's essay. From my reading, Gardner's essay isn't even a real opposition to Pinker's book. I think he liked the book and had some minor notes about how he wanted the discussion contextualized (you might need to emphasize factors other than rational calculation in decision-making). That's pretty standard, boilerplate academic discourse–including the fact that he linger longer on the negative than the positive. By trying to "rebut" this commentary, Pinker was turning it into a tribal disagreement--and by using the same team framing, you were reinforcing that.
To me, that's the big, general point. You think I'm unfairly smuggling this into a culture war frame, and I think you're unfairly smuggling in an objective, neutral position (one that is actually missing here). You stand by your hostility towards Gardner, I stand by my frustration with this essay.
Specific Points
My other major response is that it feels like you're also missing the points I'm trying to make here. One rule of thumb I have for figuring out if someone is giving a fair characterization of someone else's argument is if the other person, on hearing that argument, would say "yes, that's more or less right about what I think." (This comes from Daniel Dennett.) Reading your response here, I do not feel like you've met that standard. A few places where I feel like you're not quite getting what I'm arguing, and where I think discussing this disagreement could still be fruitful:
You write: "You seem to think that I am accusing Gardner (and the various culture war teams you're trying to smuggle in here) of being 'irrational'": No, I don't think that you believe Gardner is irrational, but instead that he is against "Team Reason" or some rationalist point of view. That's the core framing of the article, right? He's not irrational--in the sense of never thinking rationally--but he's not on the rationalist side, so what does that really mean? I get that, and I think a lot of the steps you walk through are valuable ways of thinking that issue through. I just think that I'd also reject the dichotomy you end up at.
You write: "My conclusion was that Gardner (mistakenly) believes that Pinker means we should never use heuristics, and so even though both of them are actually doing rational things, Gardner claims to be anti-rationalism." I disagree! I don’t think that Gardner thinks of himself as being on a different team from Pinker at all, and I certainly don’t think he’d argue that Pinker’s books says “we should never use heuristics.” I think he sees himself as commenting on a book that he generally liked. Maybe there’s a place where Gardner says he’s against rationality or the rationalist community, but it’s not in that essay. He ends the essay writing “Cheers for the four REs!” Reason is one of those four things he’s cheering! Addressing this point reinforces the main point I made above: I think that Pinker, and then ACT, were adding an argumentative, us-vs-them narrative to a discussion that did not have to be framed that way.
You quote tweet my point 1 (everyone uses a mix of reason and heuristics) in my concluding argument, and then write: “Aaaaah! That was my entire point!” Yes! And I agree with that point! In the last section of my article, I was trying to build up the logic structure leading to my own conclusion. Not every bullet I wrote is intended as a rebuttal. Instead, I’m trying to say how, even starting from a shared premise, we ended up at a different conclusion. I’m sorry if that wasn’t clear. Let me be clear here: I like a lot of what you write, both in general, and in this essay. The extended discussion on what an argument over rationality could be about was great, and why the piece grabbed me in the first place. (But there were clearly some elements of it that rubbed me the wrong way enough to want to write a rebuttal.)
Digging Deeper into the Rationality Conflict
The following quote was most helpful for me in further clarifying what this back-and-forth is about: “Most people more or less agree on how to use rationality and heuristics, but one real difference between them is that some people are interested in formalizing and studying this process (for comparison: most people are biological organisms who live biological lives, but some people in addition to this are biologists). I think when someone tries to formally study the thing, they sometimes identify as ‘a rationalist’, and that confuses other people into thinking they're asserting something scary and different from and opposed to normal reasoning.” This is getting at what I consider the most exposed, fleshy part of my own argument. I agree that there are some people who want to build systematic knowledge through careful reasoning, and others who don’t want to do that work and just choose to do whatever seems to work (or what they see others doing). However, I believe that in almost all cases even those people who are just doing what seems to work believe that there’s some underlying system or science to the approach, even if they didn’t derive it. In that sense, they feel like they are part of the same kind of epistemic community that rationalists believe they are part of.
To illustrate: say I’m interested in effective altruism because I am a rationalist. I may not go out and collect data, or do novel statistical analyses, or even just try to reproduce (i.e., re-run and check) the code that generated that analysis. Yet I am happy to use the results that come out of analyses of the effective altruism community, and will prefer them over heartstrings-pulling commercials, because I believe they’ve gone through the kind of epistemic vetting that makes them valid. So a fundamental part of my being a rationalist is my reliance on the community of reasoner’s I’m a part of, and using the collected wisdom of that community.
Herbal healers, in my estimation, have the same sort of formal structure to their decision making. They don’t go out and test every herb to see how it works, but they rely on the collected knowledge of others around them. They believe that their knowledge has gone through valid, epistemic confirmation, given the rules they’ve set up for themselves about what counts as knowledge.
Now, I would argue that the methods that Medieval herbalists, for example, were using wouldn’t be very good at separating true from false. But once you make the move to decouple “rationalist” from “specifically uses computational, mathematical, or algorithmic approaches to come to conclusions,” I think you end up in a position where the defining feature is something more vague related to “formally studying” and “systematizing.” And I would argue that formalizing and systematizing exists along a spectrum and includes lots of different methods. Among them are writing things up in big 100+ page books of plants and roots with pictures and notes on how they’ve been able to help (or not help) people. So a Medieval Saint doing that may not have developed the degree of sophistication of modern science, but it wasn’t something qualitatively different from what lots of other mainstream, traditional Western thinkers were doing. Biology before Darwin was heavy into just collecting lots of artifacts and trying to come up with taxonomies to organize them. Darwin gave a theory that attempted to explain what you could see, and it revolutionized the discipline, but I would argue that biologists before Darwin were working in the same discipline, and it was just as scientific–or rationalist–despite being much less sophisticated.
My “gut check” for this line of reasoning would be asking a bunch of people who hold beliefs that I think are unscientific or non-rational why they believe what they do. If I’m right, I would expect them to talk about a larger body of knowledge that they believe, particular teachers or sages that they follow, and particular rules or principles that they follow (even if they’re a bit sketchy). If I’m wrong, I’d expect people with these views to say “it just sounds right to me” or “that’s just always the way it’s been.” It could probably go both ways, depending on the person. But if the formalizing and systematizing is something that uniquely divides people who identify as rationalist from non- or anti-rationalist, then I’d expect to see a big, significant difference if a representative sample of both groups was coded for how they justify their beliefs.
Conclusion
To summarize, I would say that the main point of my article was that there really *isn’t* a principled way of dividing rationalist from anti-rationalist people that relies on “reason” as it’s commonly understood. Instead, what divides people is their belief that one particular set of rules for vetting knowledge is better than another. All those rules are model-based, and so simplifications. Ultimately that means that they have certain values baked in–values about what it’s important to focus on and what can be ignored. So choosing the rationalist or the anti-rationalist side is really about choosing which set of values makes the most sense to you. Which is fine! That’s how knowledge has always worked. But I do think it’s a different point from the one you were attributing to me above.