Zen and the Art of Rationality


29


Eliezer_Yudkowsky

Followup toEffortless Technique

No one would mistake my writings for ancient Eastern wisdom.  Successfully or not, I aspire to clearly set forth the reasoning, antecedent assumptions, and pragmatic use of my conclusions.  Successfully or not, I aspire to cut my proposals into modular pieces, so that a user can reject one mistake without destroying the whole.  This standard of writing is inherited from the ancient traditions of technical thinking, not the ancient traditions of Zen.

No one would mistake my writings for ancient Eastern wisdom.  My goals are not the goals of Buddha or Lao Tse.  Feeling Rational suggested that emotions should follow from beliefs but not beliefs follow from emotions:  the ideal is to free yourself of all attachment to preferred conclusions about reality, arrive at your beliefs of fact by weighing the evidence without prejudice, and then feel fully whatever emotions follow from these beliefs-of-fact.  In stereotypical Eastern philosophy, you are supposed to free yourself of all attachments, not just attachment to beliefs-of-fact apart from evidence; you are supposed to relinquish all desire.  Yes, I know it's more complicated than that - but still, their goals are not mine.

And yet it oftimes seems to me that my thoughts are expressed in conceptual language that owes a great deal to the inspiration of Eastern philosophy.  "Free yourself of attachments to thinking that the universe is one way or another:  Arrive at your picture of the world without prejudice, and then feel fully whichever feelings arise from this picture.  Let your emotions flow from your beliefs, not the other way around."  It's not a Buddhist conclusion, but the language owes a nod in the direction of Buddhism.  Even if a Buddhist teacher would vehemently disagree, they might still grasp immediately what was being proposed.  Grasp it more clearly, perhaps, than an old-school (i.e. pre-Bayesian) Western rationalist.

No one would mistake my writings for ancient Eastern wisdom.  And this is well, because I can't stand people who try to pass off their ideas as ancient wisdom.  As if that were a recommendation!  The fifth-century Chinese philosopher Xiaoguang Li observed that ancient civilizations are revered, and yet ancient civilizations are not wise like venerable human elders are wise.  A civilization further back in time is younger, not older.  The current civilization is always the senior, because the present enjoys a longer history than the past.  Incidentally, does it change your opinion if I tell you that Xiaoguang "Mike" Li is actually a friend of mine who lives in the Bay Area?

So be it far from me to spray-paint my work with a patina of venerability.  And yet in too many ways to list here, my work owes a nod in the direction of Buddhism, Taoism, Zen - and even Bushido.  Yes, Bushido! See e.g. the Musashi quotes in the Twelve Virtues of Rationality. Whatever their other flaws, samurai had a deep grasp of the virtue of perfectionism as a life-principle.  To Westerners, "perfectionism" refers to something that seems like work, makes people unhappy, and causes software to ship late.

Of the virtue of curiosity, I said:  "A burning itch to know is higher than a solemn vow to pursue truth."  Here is the conceptual language - but not the propositional statements - of Lao Tse admonishing, "Stop talking about morality and righteousness, and people will regain the love of their fellows."  People are not naturally rational - but you sure can trip over your own feet by thinking too much about "rationality" instead of paying attention to the obvious evidence.  Learned virtues are powerful but dangerous; they have many degrees of freedom for error.

Western religions demand submission to God, bended knee and bowed neck.  Many Christian saints achieved their canonization by going to great lengths of voluntary suffering.  You obey God's precepts out of dutiful morality and reverence, on penalty of judgment and damnation.  Such concepts have contaminated Eastern street religions as well, of course.  But so far as Eastern religious philosophy is concerned, one speaks of harmony with the Tao, rather than submitting to the Tao.

When I ask myself whether rationality seems more like submitting to the commands of Bayes, or moving in harmony with the Bayes, the latter seems far closer to the mark.  By placing yourself in correspondence with the Bayes, you wield the power of the Bayes.  If you misstep in the dance (accidentally or deliberately), there is no judge who damns you, or any divine watcher disappointed in you:  You have failed yourself.  The laws of probability theory still govern you, entirely indifferent to your submission or defiance.  The consequences of your disharmony will occur to you according to the natural order of things: the Bayes does not condemn you for your disobedience, but reality will not go according to your hopeful plans.  Neither guilt nor repentance will save you, since the Bayes cares nothing for your allegiance.  Worshipping the Bayes will not gain its favor, for the Bayes has no ego-desire to demand your praise.  Probability theory is there to be used, not believed-in.  There is no ancient Taoist manuscript that agrees with such Bayesianity, but the language...

The axioms of Bayesian probability theory make no mention of clothing, and therefore a valid derivation is valid whether you wear a lab coat or a clown suit.  The Bayes makes no mention of solemnity or silliness, and therefore lecture on rationality is just the same whether spoken in deep portentous tones or a high squeaky voice from inhaling helium.  Understanding what probability theory constrains and does not constrain, we are free to be spontaneous in all other respects.  This purity and freedom is preached in no Buddhist tract, but there is something of an Eastern aesthetic about it - and a mathematical aesthetic also, but math knows no East or West, and is simply math.

Miyamoto Musashi said:

"The primary thing when you take a sword in your hands is your intention to cut the enemy, whatever the means. Whenever you parry, hit, spring, strike or touch the enemy's cutting sword, you must cut the enemy in the same movement. It is essential to attain this. If you think only of hitting, springing, striking or touching the enemy, you will not be able actually to cut him."

Likewise in rationality.  Every step cuts through to the truth in the same movement.  Every step carries the map through to reflect the territory.  If you fail to achieve a correct answer, it is futile to protest that you acted with propriety.  Whether you wear a lab coat or a clown suit, however much it might naively seem to associate with science, does not affect whether you cut through to the correct answer.  (This is why I'm not afraid to borrow the language of Taoism, or verse-form when the whim takes me; mere style makes no difference to probability theory.)  You might think that such focus, such purposefulness, is more Western than Eastern - but where is the equivalent declaration of Musashi's by a Western philosopher?

Lest I seem to give the East too much praise, I note a well-known truism to the effect that Westerners overestimate the average quality of Eastern philosophy because only the good stuff gets imported.  Buddhism seems "atheistic" because you don't read about the ten thousand minor deities unabashedly worshipped on the street.  Such selectivity is right and proper, and I make no apology for it.  I am not trying for authenticity, that is not my purpose.

Likewise, I don't spend much time pondering my "Western influences" because they are as natural to me as breathing, as unseen to me as air.  If I had grown up in Taiwan, my writing would probably sound far more Buddhist and Taoistic; and perhaps I would talk of the inspiration (though not advice) I had received from reading some Taiwanese book about Greek philosophers, and how I often felt closer to Judaism than my forgotten childhood Buddhism.

Nonetheless, I think it a wise thing for an aspiring rationalist to read at least one book of Buddhist or Taoist or Zen philosophy - preferably a book in its original English, recommended to you by some mathematician or programmer or scientist.