The Twelve Virtues of Rationality

Yesterday, Raymond Arnold sent out an email announcing that those at Highgarden had come up with a really cool art project and invited everyone to come down and participate. The idea was to paint panels depicting the Twelve Virtues of Rationality, to put on the upper panels of the wall in the living room area, and they had headed off to buy paint. This did in fact sound like quite the cool art project and a chance to influence what would be on the walls for potentially quite a while, so Laura, Alexander and I got our things together and headed down to check it out. While we were sad that one of those who came up with the idea had left, things otherwise went pretty great, complete with fresh baked goods.

Brainstorming for the panels offered a few obvious good ideas, some of which were even practical for people at our skill level to paint. Mostly, you had your classics. One does not want to take things too far and end up with virtual clip art, but books scattered all around is hard to beat for scholarship, people arguing is hard to beat for argument, and things like telescopes and microscopes are hard to beat for Great Visual Moments in Empiricism, as much as I was hoping for the Leaning Tower of Pisa. Others were more difficult to get a handle on.

I had an instinct that we should take the sword metaphor and have them running through the panels as a linking motif, and it sounded really cool, but on reflection we all agreed that was both more martial than we would like and an awkward fit in too many places.

By our skill level, of course, I mean Raymond and Grace’s skill levels. I haven’t painted since grade school, and from my brief experiment attempting to do so, my skills are at about the level where I last left them. Those with higher skill tried to explain that paint does not work the way I instinctively think about art working, and suggested started with a rich colorful background. It is quite likely I should have heeded this advice, but I felt like the chance of me creating something viable was pretty low anyway so I would experiment with smaller things, avoid wasting massive amounts of time and paint, and if I got somewhere then hey, maybe I would start over with a real background. The old instinct of one mustn’t waste things is almost always wrong for inexpensive values of things, given the value of time and life being short, but it was drilled into me pretty hard and having too little of that instinct has its own problems, so I am reluctant to push on it too hard.

I will, of course, never be a painter at that rate, since as Picasso pointed out, the way one becomes a painter is to paint. It seems likely that the test of whether you are a painter is that painters are completely unafraid to waste paint. For now, I am comfortable with that, since I am hard at work on lots of other skills at once, but damn it seemed like it would be a lot of fun if I were to become even semi-competent.

In the end, both Raymond and Grace had awesome looking backgrounds going and I am eager to see how things turn out, but right now I am more curious to continue thinking about the virtues themselves. It is a strange list with several serious problems. The first, as Allison put it, is that there are too many syllables. Your virtues should be concise, simple things. One could respond that this is a problem with English rather than the fault of the virtues, but it seems likely that it also represents a failure to properly identify and name the core elements of what you are attempting to promote. Similarly, twelve seems like a lot of virtues, and a lot of them seem related, especially if you read their explanations. Could we get most of the punch (and therefore, more than all the punch) by invoking simplicity and combining virtues and/​or leaving some of them out? There are also several that have names that do not, to at least many people, mean what their name implies. Perfectionism, for example, invokes a different set of ideas than the virtue is aiming for, and the invoked set involves a lot of things people consider bad.

They also need an organizing principle, especially if there are going to be so many. I loved the virtues of the Ultima series for many reasons, but a lot of it was that it was so elegant on so many levels to organize them around the venn diagram of the three principles: Truth (blue), Love (red) and Courage (yellow). Those are some top quality principles. One could argue there are other principles that deserve equal billing, or that one of those selected does not quite make the top three, but if you think truth, love or courage is a bad idea, then I am pretty sure something has gone horribly wrong.

Could we do a similar thing with the virtues of rationality? Truth is kind of the central thing we are looking for, so it can’t be one of the principles. We instead are looking inside of truth to an extent, so we can think of these as Truth of Truth, Love of Truth, and Courage of Truth. Truth of Truth is the pure data: So let’s call that principle Data. Love of Truth is shown by working towards the goal, which is something Eliezer talks about all the time: it is no good to say that you should seek truth, or that you lack truth, unless you actually do seek truth. So let’s call that principle Effort. Finally, there is Courage, which is more complicated. A lot of what makes rationality hard is constantly being torn between the need to stand on your own and the need to take in the wisdom (and efforts) of others and stand both toe to toe with friends and on the shoulders of giants.

Thus, Courage of Truth can be thought of as having two sides. There is the Courage to seek Truth Together, or Collaboration. Share your knowledge, publish your findings, study hard, debate what it means, listen to others and seek Aumaman Agreement. There is also the Courage to seek Truth Alone. Speak the truth, even if your voice trembles. If everyone else says the sky is pink, or the new baggage inspections will make us safer, or all three line segments are the same size, but the evidence says it isn’t true, you need to stand alone. Rationality will fail unless you have the courage to stand alone. Until we another better way of putting that, let’s call that Independence.

This leads us to a structure of three principles, one of which has two aspects. By combining them into a venn diagram, we now can either have or not have Data, we can have or not have Effort, and we have either Collaboration, Independence, or neither. That’s two times two times three, or twelve virtues. Perfect!

Can we make them fit? Yes we can! For extra fun, stop here and try the exercise yourself, then see how your answers compare to mine. There is certainly a non-zero amount of shoehorning that needs to go on, since unlike Lord British we are given a fixed set of pegs to put into exactly enough holes, but it fits a lot better than one might think. The majority of the explanations below are my own words, but I also quote liberally from Eliezer Yudkowsky’s original document:

Independence and Effort without Data is Curiosity. No one else can make you curious, you must become curious yourself; the burning itch to know comes from within. Only then can you gather data and/​or work with others. One might be tempted to say that curiosity proceeds effort, but this would be a misinterpretation, as Eliezer Yudkowsky explicitly points out: Be wary of those who speak of being open-minded and modestly confess their ignorance. Curiosity seeks to annihilate itself, so curiosity without effort is not the true curiosity.

Independence and Data without Effort is Relinquishment. That which can be destroyed by the truth should be, and it does not matter what others think.

Data and Effort without Independence or Collaboration are Lightness. Let the winds of evidence blow you around like a leaf, and surrender to the truth as quickly as you can.

Data, Collaboration and Effort together are Evenness. Only with great effort can one properly account for data and for the input of others. One must be constantly vigilant to treat all sides fairly. If you attend only to favorable evidence, picking and choosing from your gathered data, then the more data you gather, the less you know.

Independence, Data and Effort together are Empiricism. The roots of knowledge are in observation and its fruit is prediction. To find the truth, one must ultimately test with one’s own eyes.

Collaboration alone is Argument. This is the essence of collaboration to seek the truth, and cannot be done alone. You cannot move forward on factual questions by fighting with fists or insults. Seek a test that lets reality judge between you.

Collaboration and Effort together without Data are Simplicity. Simplicity is the avoidance of unnecessary data; Antoine de Saint-Exupéry said: “Perfection is achieved not when there is nothing left to add, but when there is nothing left to take away.” When combining the knowledge of many, the goal is not a mass of “big data” that none understand; the goal is to take from each only that which is valuable, and combine them into a whole that is both smaller than, and greater than, the sum of its parts.

Effort alone is Humility. To be humble is to take specific actions in anticipation of your own errors. The core essence of true humility is the effort to improve!

Data lone is Precision. The strongest data is the most precise data. The narrowest statements slice deepest, the cutting edge of the blade.

Collaboration and Data together without effort are Scholarship. Study many sciences and absorb their power as your own. Each field that you consume makes you larger. Share your works with others, so that they may do likewise. The real effort begins, however, not with learning the words of others, but in figuring out what the consequences may be.

The absence of Independence, Data, Collaboration and Effort is, of course, The Void. The important thing is not any abstract principle, but to cut the enemy!

See how neatly it all fits together?

Untitled