Your article is interesting, and a lot of the points you make are valid. In practice, LW-style rationality might well have some of the effects you describe, especially in the hands of those who use it or understand it in a limited way. However, I don't think your point is valid as a general argument. For example:
If you win too much, deprive those close to you the chance of winning, effectively causing them anxiety.
To me, this seems to be based on a fallacious understanding of LW-style "winning." Winning here means accomplishing your goals, and using a "rationality" toolkit to win means that you accomplish more of your goals, or accomplish them better, than you would have without those tools.
For some people, being right about everything is a goal. For some people, harmonious social relationships are a goal. For a lot of people, these are both goals, although they may be prioritized differently, i.e. a different weight may be placed on each. If the goal of being right conflicts with the goal of harmonious relationships, and harmonious relationships are prioritary, then according to the toolkit of "rationality", it is rational to lay off a bit and av...
[A]re there times when it should be desired to lose[?]
When you should "lose", "losing" is the objective, and instrumental rationality is the art of successfully attaining this goal. When you do "lose", you win. On the other hand, if you "win", you lose. It's very simple.
Some of the things (like eating healthier and exercising more) I did not let go, because I felt the damages of my role reversal were less than the damages of their habits; however, other ideas, arguments, beliefs, I did let go because they did not seem worth the pain I was causing my parents.
Why call this losing instead of winning-by-choosing-your-battles? I don't think members of this community would endorse always telling others "I know a better way to do that" whenever one thinks this is true. At the very least, always saying that risks being wrong because (1) you were instrumentally incorrect about what works better or (2) you did not correctly understand the other person's goals.
More generally, the thing you are labeling rationality is what we might call straw vulcan rationality. We don't aspire to be emotionless computrons. We aspire to be better at achieving our goals.
Eliezer wrote a cute piece about how pathetic Spock was to repeatedly predict things had <1% of succeeding when those sorts of things always worked. As outsiders, we can understand why the character said that, but from inside Spock-the-person, being repeated wrong like that shows something ...
You've confused goal-winning (LW sense) with social-winning.
Rationality is the optimal tool for goal-winning, which is always what is desirable. This relation is established by definition, so don't bother criticizing it.
You can show that our current understanding of rationality or winning does not live up to the definition, but that is not a criticism of the definition. Usually when people debate the above definition, they are taking it to be an empirical claim about spock or some specific goal, which is not how we mean it.
EDIT: Also, "air on the side". It's "err" as in "error". Read Orwell's "politics and the english langauge".
So, sometimes actions that are generally considered rational lead to bad results in certain situations. I agree with this.
However, how are we to identify and anticipate these situations? If you have a tool other than rationality, present it. If you have a means of showing its validity other than the rationalist methods we use here, present that as well.
To say that rationality itself is a problem leaves us completely unable to act.
If the action you are engaging in is not helping you achieve your goals, than it is not rational.
You are describing a failure of rationality rather than rationality itself.
I find it interesting, even telling, that nobody has yet challenged the assumptions behind the proposition "Rationality is a tool for accuracy," which would be that "rationality is the best tool for accuracy" and/or that "rationality is the sole tool that can be used to achieve accuracy."
I guess to me it seems automatic to 'unpack' a general goal like that into short-term specific goals.
Sure, I can see that. Were it similarly automatic for me, I'd probably share your intuitions here.
I understand rationality to be related to a set of cognitive tools rather than a certain personality or genetic type. Like any other tool it can be misused. You can kill a person with a spoon, but that is a misuse of its intended function. You cut a pound of raw meat with a chainsaw, but that is a misuse of its intended function. Tools are designed with both intended purposes and functional limitations. Intended purposes serve to provide the user with an understanding of how to achieve optimal impact. For example, some intended uses of a sword would be killing, disabling, acting, or training (and many more). Tools can be used outside of their intended purposes. The use might not result in optimal output, it might even damage the tool, but it is possible. A sword can be used to cut wood, clear shrubbery, as a decoration, a sword could even be used as a door stop. Doorstop has long departed from the intended function for a sword upon its design, but nevertheless it exists as possibility given the structure of a sword. Functional limitations are desired uses that a tool cannot meet given its structure. A sword alone cannot allow you to fly or breathe underwater, at least not without making significant alterations to its structure, rendering it no longer a sword.
Every tool exists with both intended functions and functional limitations. From reading some essays on this website I get the impression that many members of this community view rationality as a universal tool. That no matter what the conflict a certain degree of rationality would provide the appropriate remedy. I would like to question this idea. I think there are both functional limitations to rationality and ways to misuse one's powers of reasoning. To address these, it is first necessary to identify what the primary function of rationality is.
From reading various articles on this website I would suggest that rationality is seen as a tool for accuracy in obtaining desired results, or as Eliezer puts it, for “winning.” I agree with this analysis. Rationality is a tool for accuracy; increased accuracy leads to successfully obtainment of some desired result; obtainment of some desired result can broadly be described as “winning.” If rationality is a tool for increasing accuracy, then the questions becomes “are there ever times when it is more beneficial to be inaccurate,” or in other words, are there times when it should be desired to lose.
I can think of two situations where increased accuracy is detrimental: 1.) In maintaining moderation; 2.) In maintaining respectful social relations.
1.) *It is better to air on the side of caution*: The more accurate you become the faster you obtain your goals. The faster you obtain your goals the quicker you progress down a projected course. In some sense this is a good thing, but I do not think it is universally good. **The pleasure winning may deter the player from the fundamental question “Is this a game I should be playing?”** A person who grew up playing the violin from an early age could easily find themselves barreling along a trajectory that leads them to a conservatory without addressing the fundamental question “is becoming a violinist what is going to most benefit my life? It is easy to do something you are good at, but it is fallacious to think that just because you are good at something it is what you should be doing. If Wille E. Coyote has taught us anything it is that progressing along a course too fast can result in unexpected pitfalls. Our confidence in an idea, job, a projected course, has no real bearing on its ultimate benefit to us (see my comment here for more on how being wrong feels right). While we might not literally run three meters off a cliff and then fall into the horizon, is it not possible for things to be moving too fast?
2.) *”Wining” all the time causes other people narrative dissonance*: People don’t like it when someone is right about everything. It is suffocating. Why is that? I am sure that a community of dedicated rationalists will have experienced this phenomenon, where relationships with family, friends, and other personal networks are threatened/damaged by you having an answer for everything, every causal debate, every trivial discussion; where you being extremely good at “winning” has had a negative effect on those close to you. I have a theory for why this is, is rather extensive, but I will try to abridge it as much as possible. First, it is based in the sociological field of symbolic interactionism, where individuals are constantly working to achieve some role confirmation in social situations. My idea is that there are archetypes of desired roles, and that every person needs the psychological satisfaction of being cast into those roles some of the time. I call these roles “persons of interest.” The wise one, the smart one, the caring one, the cool one, the funny one, these are all roles of interest that I believe all people need the chance to act out. If in a relationship you monopolize one of these roles to the point that your relations are unable to take it on, than I believe you are hurting your relationship. If you win too much, deprive those close to you the chance of winning, effectively causing them anxiety.
For example, I know when I was younger my extreme rationality placed a huge burden on my relationship with my parents. After going to college I began to have a critique of almost everything they did. I saw a more efficient, more productive way of doing things than my parents who had received outdated educations. For a while I was so mad that they did not trust me enough to change their lives, especially when I knew I was right. Eventually, What I realized was that it is psychologically damaging for a parent’s 20 something year old kid to feel that it is their job to show you how to live. Some of the things (like eating healthier and exercising more) I did not let go, because I felt the damages of my role reversal were less than the damages of their habits; however, other ideas, arguments, beliefs, I did let go because they did not seem worth the pain I was causing my parents. I have experienced the need to not win as much in many other relationships. Be they friends, teachers, lovers, peers, colleagues, in general if one person monopolizes the social role of imparter of knowledge it can be psychologically damaging to those they interact with. I believe positive coexistence is more important than achieving some desired impact (winning). Therefore I think it is important to ease up on one’s accuracy for the sake of one’s relationships.
- Honestly I have more limitation and some misuses I to address, but decided to hold off and see what the initial reception of my essay was. I realize this is a rationalist community and I am not trying to pick a fight. I just strongly believe in moderation and wanted to share my idea. Please don't hate me too much for that.
- HungryTurtle