Reply to Julia Galef on Tradition

Julia Galef, a prominent member of the "Rationality" community, responded in one of her videos a while ago to an argument very like my Structured Argument for Traditions. The video was posted years ago, so I don't think she's read my version, but the argument isn't exactly original; my argument is just a juiced up and clarified version of a common idea. Her video describes the argument for traditions, and her attempts at refuting it, in the context of "rationalists" being disproportionately likely to cast off traditional social technologies in favour of their own newly invented ones. I don't think the casual casting off of tradition is rational, and I don't think their attitudes are representative of rationalists in general, so I will hereafter refer to them more specifically as "LWians", for the website which stands at the centre of that community.

For context, the LWians tend to cast off these specific traditional social technologies:

  • Monogamy. LWians are disproportionately likely to engage in a behaviour called "Polyamory", which is the practice of openly having concurrent sexual relationships with multiple partners, sometimes within the context of a "primary" marriage or long term relationship. Some of them also engage in promiscuous sexual behaviour.

  • Traditional Gender Roles and Separations. LWians believe the life script for men and women ought to be roughly the same, with the only major difference being tendency of sexual preference and reproductive hardware. They mix men and women in spaces that are traditionally not mixed, consciously try to deprogram or flout gender-associated behaviours like masculinity and femininity, and even reject the characterization as men or women altogether.

  • Social and Communication Norms. LWians discuss and practice nonstandard modes of communication like nonviolent communication, ask, guess, and tell culture, and probably some other stuff I'm not aware of.

  • Assorted Taboos. LWians tend to engage in behaviours that are traditionally taboo, like mixing work and sexual relationships, cuddling outside of romance, etc.

These are some the deviations that they seem to have attempts at specific philosophical justification for. There are further subtler deviations from traditional behaviours which I have not heard argued for but which also seem to stem from a lack of respect for tradition. I have specific harsh criticisms for most of these deviations, and further meta-level criticisms of the state of mind that produces them, but that's not for today.

On the other hand, I do respect what the LWians are doing. Taking ideas seriously and striving to have a strong philosophical foundation for our beliefs and actions is something we try to do around here as well. I just disagree with them about the proper methods and results of such a strategy. We'll stay out of specifics for now and just use the above as context for the meta-level claims Galef makes in her video:

The video is in two parts. In the first part, she briefly introduces the topic and outlines the epistemic argument for tradition:

  • Traditions are ideas that have faced significant selection pressure over time for being beneficial to their hosts. Further, they are known solutions to a very complex social organization problem which one can't approach naively and expect good results. Therefore a tradition is decent evidence that that way of doing things is a good one.

This is roughly the argument I described in more detail in my post about a structured argument for tradition. Galef expresses some sympathy for it, but goes on, in the second part, to explain why she thinks this argument isn't very strong after all. Here are my best renditions of her arguments, and my comments on them.

  1. Bad social customs are "really sticky", and people don't get rid of them. That is, my premise #5 is wrong: "Given experience with a specific bad cultural practice, humans are generally able to critique it and improve upon it."

    In my experience, which accords with what we would expect from common sense, when people encounter strong evidence that their cultural practices are flawed, or have unusual amounts of skin in the game, they do some soul searching and figure out how to fix them. I myself have done this multiple times, as have many of the normal non-rationalist people I know. Galef may be seeing a perspective effect, where what looks to her like irrational adherence to stupid ideas is in fact rational adherence to good ideas with which she currently disagrees. To say more, we would have to hear what evidence leads the LWians to believe that cultural practice is sticky for most people.

  2. Traditions don't undergo mutation as much as genes, so the search space is not adequately covered.

    In my original argument, I failed to address this, but the error rate in genomes gets the job done at only 10^-8 per base pair per generation. The error in cultural transmission is purposefully directed, and considerably larger. Our intellectual ancestors played with all kinds of funny ideas, and culture today is noticeably mutated from 50 or 200 years ago. We can be sure that most things have been tried. Even the LWians' "nontraditional" ideas themselves have a rather long historical tradition, coming up multiple times in history but never quite managing to achieve sustainability.

  3. We are working from a prior of "status quo bias", and LWians specifically correct for this to achieve balanced rationality.

In some ways, the validity of the status quo bias is exactly what is under examination here. The status quo bias, which is a preference for the current practice one is familiar with over a new practice, must be rational in some common circumstance or it would not have evolved. The common circumstance in which it is rational probably being exactly the big common sense questions like which life cultural practices to use. Further, the demonstrations of irrationality shown by the heuristics and biases people tend to be pretty weak when we've examined them critically; its as often the researcher oversimplifying an actually complex situation as it is the brain getting it wrong on some corner case.

But this is besides the point. Most of us loud traditionalists didn't grow up this way, we specifically overcame the status quo bias to get here. The status quo bias is not that strong in practice; people are cautious, but they change what they are doing all the time. Irrationality is not an adequate explanation for why literally almost every civilized person ever disagrees with the LWians about proper cultural practice.
  1. The traditions may be optimized for conditions that no longer hold.

I missed this one in my original argument as well, and it would be fatal. However, people have not changed, nor have the fundamental problems we face in life, besides an easing up of material conditions. Our traditions should probably change a bit as conditions change, but the sane approach is to salvage our traditional culture as much as possible, not throw it out and rewrite the whole codebase every time we hit a bump in history.

  1. Social experimentation is a public good, so we should appreciate other people taking risks on weird social ideas, so that we may all benefit when they turn out to be good, and not have to learn the hard way when they turn out to be bad.

    Indeed, we are watching closely and documenting the failures of the LWians' attempts at cultural reengineering for our own social technology research, but we also note the high costs we pay as a society when a generation of our brightest intellectuals decide to engage everybody in cultural experiments.

  2. Through all of this, questioning the value of tradition contains an implicit claim that you can do better, in contradiction of my lemma #4 that "cultural practices constructed ex nihilo are not likely to be good".

    I reaffirm that premise. Culture is very complex. It is probably possible to understand and even design in some cases, but a naive approach that does not borrow heavily from past empirical testing is unlikely to get much right.

  3. There is a difference between what is beneficial to your genetic potential, culture, and social group, and what is beneficial to a flourishing life for you as a human.

    This objection is the most substantial and hardest to address, resting on a values difference. Can a practice be good or bad for the individual while being the opposite for the long term interests of the social and genetic system? If the individual's goals do not align with the long term flourishing of the cultural and genetic system of which they are a part, then yes, traditions proven to serve the latter do not serve that individual. If the LWians choose to divorce themselves from their context in this way, there is no simple way in which they are objectively wrong. However, the position of this blog right from the beginning has been that we want to live in a way that is sustainable on that whole-system basis, so we unify the "individual" human desires, which contain much flexibility and room for interpretation, with that whole-system sustainable flourishing. We don't have room here for a proper treatment of this subject, so we'll leave it at that and address the details in later posts.

I think that sums up my responses to Galef on the a-priori argument for tradition in general. However, we agree with Galef that the epistemic argument for traditions is not very strong, and is soundly overwhelmed by the object-level evidence in specific cases, so further discussion at this level is pointless. We ought to take traditions somewhat seriously in general, but actual case studies and analysis of specific cultural practices is the more powerful source of evidence which really closes the case, so our future discussions of tradition will be on the object level.