23 internautes sur 25 ont trouvé ce commentaire utile
Kevin M. Iga
- Publié sur Amazon.com
There are many books that purport to prove the existence of God, or prove the non-existence of God. This is a book that purports to calculate the probability of God existing.
More accurately, it formulates the existence of God as something that we may not be entirely certain of, and therefore depends on vague estimates of probabilities, which are subjective measures of one's predilections to believe arguments of one type or another. Thus, this book acknowledges the subjective character of the question, and instead of focusing too much on the AUTHOR'S probability of the existence of God, it provides the reader with the tools to calculate the READER'S probabiliity of the existence of God.
This is a perfect example of a situation where the usual notion of probability often taught (when you flip the coin a thousand times it comes up heads as often as it comes up tails, for instance) comes up short. This notion makes no sense when it comes to matters of fact, which cannot be repeated in experiment. For this, Unwin describes the Bayesian interpretation of probability, pioneered in the early 1900s, where probabilities measure rational belief.
Unwin's work here applies the Bayesian notion of probability to five classical arguments used in the debate over the existence of God. As such, he has added something new and interesting to the debate.
In execution, however, there are problems and fallacious arguments--a feature that can be beneficial in a classroom where these deficiencies can be debated and discussed. First, Unwin's choice of five arguments determines the resulting probability more than he would like to admit. If some arguments had been subdivided into subarguments, the probability would change, giving that argument more weight. Similarly, there are fairly abstract philosophical arguments that were not brought in at all, which may be good for most people who are suspicious of such abstract arguments (perhaps with good reason) but it is important to recognize that not all arguments for or against the existence of God are considered.
Furthermore, the most difficult issue of all problems with the bayesian approach to probability, the initial a priori probability, he skips over facilely by declaring it to be 1/2. This may perhaps be better defended than any other number, but the explanation here is lacking.
Unwin also has a rating system to deal with the effect of each evidence area on the overall probability of the existence of God, that is very coarse, as it must be in such situations (can you imagine anyone arguing that the existence of evil in the world, given that God exists, is 23% as opposed to 24%, for instance?) The representative percent probabilities (1/11, 1/3, 1/2, 2/3, 10/11) he gets is fairly influential over the resulting answer, more than Unwin would like to admit. He certanly doesn't get two decimal places of accuracy, as he claims.
Now there is structure to bayesian analysis that Unwin does not discuss. In a chain of evidence, evidence that favors a hypothesis cancels evidence that works against that hypothesis in a very precise way, so that given his choices of probabilities, he is bound to get fairly moderate numbers (especially since he made sure to have arguments that favor the existence of God and arguments against).
Now, beyond the math, I should also mention a few other criticisms:
1. Philosophers and theologians have brought a great deal of nuance to these arguments that Unwin does not acknowledge. To take one example, the problem of evil in the world is not necessarily stacked against a traditional theist as it at first appears, depending on to what extent you accept various explanations offered by theists.
2. As in any philosophical work, there are arguments Unwin makes that would not work for everyone. For instance, his use of the anthropic principle to explain the fruitfulness of our universe depends on the many-worlds interpretation of quantum mechanics, which from a scientific methodological viewpoint, is as metaphysical and unmotivated as the existence of God in the first place. Still, the fact that Unwin does not focus on the ANSWER but on the PROCESS the reader can use to find his or her own answer, ameliorates this consideration considerably, as long as the reader goes through this exercise.
3. Unwin compares two positions: the traditional monotheist position of the Abrahamic faiths against the materialistic version of modern atheism. Comparing many positions at once would make this story much more complicated, so it's easy to see why he didn't bother. But for many people, the competing theories are of a different kind.
5. The most compelling reasons for belief, for many people, often involve issues beyond rationality, such as personal encounters with the divine, or the influence of a community of belief. And it is not at all clear that these reasons are "bad" reasons that should be shunned in these considerations. But they are excluded from the start in Unwin's work.
Still, the fact that Unwin seeks to provide tools, not answer questions, makes this book valuable as a *beginning* of a conversation.
His description of faith as the extra-rational piece that goes beyond reason falls short of traditional understandings of faith. Faith is not the magical extra extent to which one believes something that one would not believe otherwise. It is a trust that one chooses to make when reason is there: "faith is a leap into the light, not a leap into darkness". Even if this were the case, it is hardly acceptable to posit faith and rationality as adding together--as if a particularly gullible individual who has very good reason to believe something could end up with a probability of higher than 100%! Yet this is the model he argues for, somewhat unconvincingly.
With all these disagreements aside, I recommend the book, not as a source of answers, but a way to start thinking about these questions for yourself, perhaps leading you to write another book that reflects your perspective on this question. After reading this book, I was upset at the problems I mentioned above, but decided to organize a group of students to read this together, not in spite of the problems, but perhaps because of them. It is a positive thing for discussions to be deeper, especially when one does not agree with the other side.
Beyond all of this, Unwin is a very clear writer, and explains concepts of probability, risk, and Bayesian analysis so that anyone can understand it. He is also chatty and pleasant to read, with a sense of dry wit that can be delightful or annoying, depending on your personal taste in humor.
Overall, I recommend this book highly, as a beginning, not an end, to discussion.