Praise of some popular LW articles

by AllAmericanBreakfast7 min read20th Jul 20201 comment


RationalityWorld Modeling
My composition teacher in college told me that in some pottery schools, the teacher holds up your pot, examines it, comments on it, and then smashes it on the floor. They do this for your first 100 pots.

In that spirit, this post's epistemic status is SMASH THIS POT.

In line with my previous experiment in reading LW articles with a critical attitude, focusing on the weakest passages, I decided to try the opposite. I picked two popular LW articles suggested at the top of my feed, selected the passages I thought were strongest, and tried to express my agreement and build on them.


Strongest Passage in "Reason Isn't Magic"

by Benquo

Quote 1

[Manioc transplantation to Africa] is offered as a cautionary tale against innovating through reason, since there's a lot of information embedded in your culture (via hundreds of years of selection), even if people can't explain why...

First of all, it's not clear things got worse on net, just that a tradeoff was made. How many person-days per year were freed up by less labor-intensive manioc handling? Has anyone bothered to count the hours lost to laborious traditional manioc-processing, to compare them with the burden of consuming too much cyanide?

The concept of "cultural selection" implies that the survival of weird, costly traditions indicates that they must make a culture more fit, so breaking them is likely to make a culture worse off. Benquo is rightly skeptical of this little-tested idea.

What about slack, the idea that net negative traits might survive because competition isn't fierce enough to wipe them out?

What about the idea that weird, costly rules might improve group cohesion enough to have a net positive effect on cultural fitness, even though their direct effect is negative?

What if fitness tradeoffs differ in different environments, time periods, or cultural contexts? Maybe simpler preparation produced higher cultural fitness in Africa, while the avoidance of cyanide poisoning produced higher cultural fitness in South America.

What if  "reason" is being used inappropriately here to lump together any change made based on individual thought? Didn't cultures throughout history and around the world develop traditions for how to employ reason to produce good results?

Quote 2

Third, the actually existing Portuguese and Africans involved in this experiment weren't committed rationalists - they were just people trying to get by. 

That's true, and I'd also point out that being a "committed rationalist" can mean not only an individual striving for better reasoning, but a set of traditions and institutions that shape and improve individual thought. The Portuguese and Africans certainly had traditions of this kind, though perhaps not yet developed enough to avoid manioc poisoning.

Benquo is right to point out that insofar as tradition and reason stand in opposition, we should compare the ideal to the ideal and the actual to the actual. I'd build on that and say that tradition and reason often harmonize, and that if the cultural evolution hypothesis is correct, we should expect that over time, cultures that bring tradition and reason into better harmony should win out.

Impact on conclusion:

The problem with this argument [about manioc exemplifying the dangers of innovating through reason] is that it's a nonsense comparison... If this is the best we can do for how poorly reason can perform, reason seems pretty great... 

There's absolutely no reason to think that the sort of gradual iteration that accretes into tradition never enters a bad positive feedback loop. Even if you think modernity is an exceptional case of that kind of bad feedback loop, we had to have gotten there via the accretion of premodern tradition and iteration!

The only way out is through. But why did we have this exaggerated idea of what reason could do, in the first place?

Benquo argues that the manioc story wildly distorts the relative benefits and dangers of reason and tradition. This is true, but implicitly accepts the tradition vs. reason dichotomy that I find misguided. There are obvious ways in which tradition can serve to shape and promote innovation overall, even if they appear to slow it down in the short term.

The debate has political undercurrents. Why the urgency to defend reason or tradition when the conceptual architecture is so hazy, unless there's some coded signalling going on? Maybe this whole thing is being taken as a metaphor for the risk of an AI singularity?

Strongest Passage in "It's okay to be (at least a little) irrational"

by Kaj Sotala

If, on the other hand, the community makes it clear that it's okay to be irrational, for as long as you're trying to get rid of that, then you can actually become more rational. You don't need to rationalize reasons why you're not being irrational, you can accept that you are irrational and then change it.

School wants students who are hard workers with genuine curiosity, who don't just want to get an A but really want to learn. This year, I've allowed myself to optimize for ease and good grades, and to admit that I'm not curious about everything. This has been important for clarifying my priorities and making time for real learning.

I'd build on this and say that it's easy for a community to be mistaken about its own aims. Teachers, students, and administrators think school is about learning, but it might be mostly about credentialism. The less pressure there is for participants to signal the ostensible ideals, the more free they are to uncover the real purposes of the institution they're participating in and navigate it or improve it.

What are other institutions that might benefit from less pressure to meet the ideal? Would we endorse the following statements?

  • In a relationship, it's okay to be loveless, for as long as you're trying to get rid of that, then you can actually become more loving.
  • In a job, it's okay to be lazy, for as long as you're trying to get rid of that, then you can actually become hardworking.
  • As a parent, it's okay to be authoritarian, for as long as you're trying to get rid of that, then you can actually become more liberal with your child.
  • In society, it's okay to be prejudiced, for as long as you're trying to get rid of that, then you can actually become accepting.
  • In budgeting, it's okay to splurge, for as long as you're trying to get rid of that, then you can actually become frugal.

I do feel a sense of pressure to meet all these ideals, and that the unrelenting pressure often produces feelings of resentment, anxiety, confusion, and manipulation. I might open a dialogue with some loved ones about this idea, and see what their thoughts are.

Impact on conclusion:

And it's not just a community thing, it's also an individual thing. Don't simply make it clear to others that some irrationality is okay: make it also clear for yourself. It's okay to be irrational.

I'd build on Kaj's remarks by pointing out that if this acceptance is so clearly beneficial and superficially accessible, then there might be some powerful forces getting in the way of it. In general, it's a nice idea, but I cringe when I imagine having a conversation about the relaxing of specific ideals. "Hey honey, is it OK that I don't love you all the time?" "Sure, babe, as long as you're trying to love me!"

Maybe what Eliezer accomplished in the post that inspired Kaj is that he found a way for Kaj's particular form of irrationality to be interpreted as a signal of rationality. What strategy would that imply for dealing with pressures to be loving, hardworking, liberal, accepting, and frugal?

In a relationship, perhaps you'd find a way to frame time apart as a way to regenerate energy for reconnecting with them later. Taking time for myself is how I get ready to love you better.

In a job, you might speak to the difficulties of the work you're doing. Tell your boss how difficult the job is, but then emphasize how much effort you put into it, and the good sense of humor you have about it. Me being tired and jaded only illustrates what a good work ethic I must have to be showing up for this job.

As a parent, you might have a frank conversation with your kid, telling them that when they provoke you, it stresses you out and makes you more likely to be snappish and controlling. Me admitting my controlling tendencies is the first step toward a more open relationship with my child.

Dealing with prejudice, you might use a prejudiced mutual acquaintance as an example. Ask whether they might be able to let go of their prejudice if they didn't feel shamed for not having absorbed all the little details of how to signal non-prejudice. Entertaining the idea that relaxing our condemnation of a hopeless bigot's prejudice, so that it might diminish in the long term, implies that the same rule might apply to our lesser infractions.

In budgeting, you might really play up the times when you do spend money. Make a big, fun deal about when you spend money, really making sure to get the most enjoyment out of it. Money is to help me enjoy things, and by budgeting well, I empower myself to get more enjoyment out of my money.


Responding to Benquo's piece, I notice that committing to a mindset of praise still gives you opportunities to think beyond the piece you're reading. Examples include teasing out implications, adding further justifications for the author's argument, or filling in the hidden context. It's also unsatisfying to prohibit any criticism. Sometimes, the article you're reading points the right direction, but doesn't choose the best route. A mindset of praise helps you acknowledge the former without getting so hung up on the latter that you wind up writing it all off out of contrarianism.

After reading Kaj's piece, I also see that a praise mindset leads to the generation of new ideas, activates my imagination, and makes the article much more meaningful. I have ideas for things I want to talk about with my girlfriend tonight. An article that you merely read is like carrying around an empty cup. To criticize it is to check it for leaks and shatter it when you find them, so that you're not tempted to put some precious liquid in it. To praise it is to pour in liquid and see what it will hold.

If there is no method of "true thinking," only the production of praise and criticism, then it seems that practice in both modes is the only way to get better at this game we're playing.

On the other hand, perhaps "true thinking" is allowing yourself to have private reactions both of criticism and praise, and then checking them against other people's private reactions to see if they line up. None of the other commentators on Benquo's piece made quite the points I did.

Eliezer responded to Kaj's post differently from me. Instead of asking for relief from being policed, or trying to arrange a diminution of mutual policing, he took Kaj's post as a reason to stop policing others quite so much. A cursory glance over the comments makes it seem that most discussion was around whether it's true that the pressure to be rational is counterproductive, whereas my approach was to try and come up with lots of examples of where the idea might be true, in order to allow some practical tests in my own life.

If "true thinking" is shown by having your own reactions line up with those of other people, then this didn't happen. Maybe Benquo's and Kaj's posts weren't provocative enough to generate enough comments? Maybe they were open enough that the commentariat's thoughts were too diverse to converge on a particular reaction? Maybe I or they were not doing "true thinking," but merely rationalizing our reaction of praise or critique?

Or maybe a praise mindset tends to produce divergent reactions, while a critical mindset tends to produce convergence? There are many different things we could say are right about the truest parts of any given argument, but the problems with the most false parts of an argument will be specific and apparent to all.

One advantage of this approach to reading - committing either to a critical or praise mindset, and selecting a specific passage that you believe is the truest or most false portion of the argument - is that it acts as a commitment device. No automatic strong reaction to the text? Keep searching until you find one! Don't have something to say? Find something. It's there to be found. I think both the critical and the praise approaches are good techniques for activating your mind, your imagination, your curiosity. I hope to see future articles generated using these techniques, to see how they work out for other people.

Finally, I like this format of responding to multiple unrelated articles in a single post. It smashes together two ideas that you might not ordinarily think to consider together. It makes for a more generative conversation, when I don't normally post comments on old articles. It also allows me to generate enough content for a blog post, when my reaction to a single article might not be sufficient.