Friday, December 7, 2012

Towards a moral calculus

Notorious Atheist Sam Harris makes the case for applying scientific methods to questions of morality. So far as I can make out, his core arguments are:

  1. human wellbeing is a meaningful, objective concept
  2. if you pick any two responses to a particular moral dimension, you can in theory (if not in practice) scientifically assess which is going to lead to more human wellbeing.
  3. we have a duty to maximise human wellbeing throughout the world
  4. therefore we should challenge suboptimal choices in our own and in other societies
  5. in particular, we should not abdicate the concept of morality to religion, but should consider it scientifically.

All good stuff and hard to argue. He illustrates his talk with side swipes at a number of practices which he considers religiously motivated. These are less convincing, but should not distract from his core message.

He implies that we can determine a single metric for human well-being which can be assessed. This is where I lose him. It seems to me that there are a number of dimensions to consider. We can generally agree on which of each of the following is better:

  • freedom – slavery
  • sickness – health
  • life – death
  • a life of passive acceptance – a life of meaningful contribution
  • respect – contempt
  • safety – danger
  • scarcity – abundance
  • construction – destruction
  • pain – pleasure

We can probably resolve these into a limited number of independent dimensions. If there are is more than one dimension, then the best we can hope for from science is that it can show us how to get to the envelope where increasing one dimension requires a trade-off in another. At that point, any further change requires a value judgement as to which dimensions we consider most important. Science cannot help us with this.

It is also not as obvious as it may appear that any society is actually far from the envelope already. In any case, to assess this, we need to identify a robust set of rigorously defined set of dimensions of well-being and models showing how they are constrained. Then we can sensibly discuss individual cases and value judgements.

This is so obvious that it must have been done, or at least worked on, already. Presumably Mr Harris can point me in the right direction?