Monday, October 15, 2007

On the environment: Pascal's wager

In an attempt to convince non-believers, the great thinker Blaise Pascal argued roughly as follows:
if God exists, and one doesn't believe, the loss is infinite,
if God exists, and one does believe, the gain is infinite, the cost finite
if God doesn't exist and one doesn't believe, the gain is nothing
if God doesn't exist and one does believe, the gain is nothing, the cost finite
He then argues that because the infinite trumps the finite every time, if there is a small probability that God exists then the right choice is to believe.
He is absolutely correct in this simple game, although as many others have pointed out, there are other possibilities which complicate the game, and hence the analysis.
In particular, his assessment of the existence of God was based on faith, not on evidence, and as such, it gives no reason to suppose that there is a positive probability of God existing (there is also, of course, the more powerful objection that it supposes belief in the correct God: there are several religions dominant among humans alone: how do we choose which one?) Wikipedia has a rather nice discussion of the issues raised, and the arguments for and against Pascals ideas.

However, on environmental issues, I think that there is another version of Pascal's wager which is, while perhaps less compelling to those of a religious bent, rather important.

Consider the following: there are those who argue that we are experiencing global warming (G), and those who insist that we are not:
if G exists, and we do nothing, we face catastrophe
if G exists, and we act, we may avert catastrophe, and the cost is moderate
if G doesn't exist and we do nothing, we are fine
if G doesn't exist and we act, the cost is moderate, and we are fine.
Since catastrophe is awful, we ought to act.

The difference for me, as a non-theist, between Pascal's wager and the Global Warming wager, is that Pascal based his argument that the probability of the existence of the all-powerful outcome (in his case, the existence of God) on faith, and not on evidence. The faith was conclusive for him, and hence sufficient to give (for him) a positive probability of the outcome.

The situation with respect to Global Warming is different: there is evidence that Global Warming is happening, and that it may lead to a catastrophic breakdown in the environment. This is the all-powerful outcome: the question is whether there is a positive probability that the scientists who argue for its truth are correct.
As a scientist, I understand that there is no proof: that is a notion reserved to mathematics and logic, in which we can have absolute certainty. The theory of gravitation is a scientific theory: it holds that certain physical situations can be modeled accurately in certain ways: it produces testable predictions, and if a prediction fails, it needs explaining: this leads to modified, improved, or outright new theories: general relativity, for example. These new theories then get tested, modified, etc.

Global warming is somewhat different: it is a post-hoc explanation for an observed phenomenon, namely that average temperatures have increased by a significant amount over the past few years. It is difficult to compare average temperatures to periods before they were measured, and this makes it difficult to judge whether similar situations have existed in the past, or to what extent we are responsible (or should I say irresponsible?) for causing this warming.
Global warming nay-sayers tend to argue that the warming we have seen isn't happening, is natural, has happened before and isn't really all that significant.

I think that the evidence that things have warmed recently is significant: if you study the statistics of records, you can how likely it is that in 100 years of measurements, what is the likelihood that a new record will be set in a given year, and what is the likelihood that new records will be set in, say, three consecutive years. The likelihood that we would have set as many record highs over the past ten years, given a stable weather pattern, is miniscule. Take it over thirty years, and the evidence points not just to things being warmer, but to the trend being continued warming.

I'm not going to argue that the warming is not natural, or that it hasn't happened before: if global warming exists, whether or not it is natural, or has happened before, if it is likely to cause problems, we ought to try to do something about it.

On the fourth point, that global warming is not that significant, I invoke Pascal's wager. There are serious arguments, based on serious evidence, made by serious people, that global warming is likely to be catastrophic. There are scientists, a few, who discount this. But they are few, and they don't argue that to act would be catastrophic, just expensive, or ineffective.

So, the G(lobal) W(arming) version of Pascal's wager is laid.
GW: are you listening?

Yours, just getting warmed up,
N.

No comments: