Saturday, February 6, 2010

Guys, I've solved it!

Yesterday my father brought me a new book.

The Big Questions - Steven E. Landsburg

It is a book written by a economist/mathematician about all kinds of interesting philosophical topics from religion to morality to the limits of human knowledge.   Trade agreements, costs, distribution of wealth and taxation were topics on which the book makes some good points but like many economists he makes the fundamental error of presuming people are perfect economic units maximizing their own happiness.  People are not perfect economic units and they make terrible decisions; any model that does not account for that is a waste of space.  Additionally when he tries to branch out to philosophical arguments he often gets caught up in the mathematics of the situation while totally failing to account for the psychology.  He tackles several classic philosophical problems, one of which follows below:

A train is speeding out of control down a track.  On the track ahead of it 5 people are tied up, unable to escape.  There is a switch that you can hit to send the train down an alternate track on which 1 person is tied up.  Should you hit that switch?

Most people answer yes.  Some people think it is okay to hit the switch and some people think you have an obligation to hit it, but the overwhelming majority feel that hitting the switch is the best choice.

A train is speeding out of control down a track.  On the track ahead of it 5 people are tied up, unable to escape.  There is a very large person standing near the track and you are confident that if you push that person onto the track it will stop the train, saving the 5 people.  Should you push that person onto the tracks?

Most people answer no.  The author of the book answers yes, and in fact is completely shocked that *anyone* would answer no.  His view on the matter is that it is simply a matter of numbers, saving 5 people is more important that saving 1, QED.  This is a classic example of someone trained to believe in people as perfect economic units, and not so much a salesman.  I can tell you from personal experience that being a salesman instead of a academic gives you a *drastically* different (more accurate) idea of how people actually behave.

Firstly, the world is always full of uncertainty.  In the examples above people are told that shoving the man will 100% stop the train and that the train will 100% not stop on its own, but in real life people doubt.  They don't know what is going on, they aren't sure how things will be, and their grasp of the physics of the world around them, particularly when a split second decision is required, is not very good.  Thus people are naturally cautious when performing a violent, dangerous act to effect a future good.  Everyone is sure that pushing that man onto the tracks is really, really bad and the only way to make up for it is being completely certain that it will stop the train.  Given that in the real world the train may stop on its own, someone else might be near a switch that will divert it and the man might not stop it anyway it is extremely reasonable to not interfere.  In the first example though you are not adding any additional harm to the situation if there is something you don't know about going on.  Barring something very strange happening there is no reason to think you made things worse.

The second issue here is that there is no accounting made for personal guilt.  This should not be news, but people care more about their own suffering than someone else's.  Pushing a man onto the tracks will cause tremendous personal guilt, particularly if it doesn't work!  Also, it is a known fact that people are tremendously more willing to inflict pain, suffering and death on others as long as they are physically removed from the situation.  Hitting a switch to drop a bomb is hugely different than beating someone to death with a rock even if the net result is one death.  People intuitively know this and know that their personal guilt from shoving someone to their death would be far greater than their personal guilt from hitting a switch to divert a train.  This may not be the 'correct' way to reason from an abstract, moral point of view but it is exactly how people think.

I don't take umbrage with the conclusion that in a world of perfect information we should choose 1 death over 5.  If somehow these situations were clear and certain it is the correct moral choice.  However, the assumption that people will actually make choices like moral logicians in a world of perfect knowledge is completely irrational.  When you ask people questions they respond by thinking about them in terms of the world they live in and what they know.  When you forget that you are setting yourself up for some really rude surprises.

Check this out:


  1. The idea that morality is a body count is something that could only be cooked up in university graduate level economics (or first year philosophy). I suppose the author would argue that it is less morally acceptable to pass a homeless person with flimsy clothing in -40 weather, doing nothing, than it is to randomly assault a stranger with and axe and cut off their leg. The correct counter-argument to this is, "Oh, I didn't know you were an idiot."

  2. I have a ton of respect for learning and research in general, but I have to say that the academic lifestyle really lends itself to coming up with conclusions that don't mesh well with real life. The detachment from other people often seems to lead academics to propose ideas that most people would disagree with, though often those people cannot really say with certainty why they disagree.

    It is valuable to think about things and speak about theory but it is very important to keep in mind that your theory may not match the world. The ability to argue convincingly for a theory is not strongly correlated with the correctness of the theory.