Wednesday, April 1, 2015

FTA part 4: Why fine-tuning means the universe is improbable (part of the Aron series)

Greg says:
OK, great, all that is good to know for me. Hopefully, it will help us avoid talking past each other.

I'll go ahead now and engage with your points from your previous post. First, I am not sure if we want to go down the path of discussing Spinoza. I don’t have much interest in it, and you didn’t seem to push it too hard, so we’ll put that out of mind, unless you want to bring it back up at some point.

But perhaps what I could say about it is to give you a bible verse that is germane to the topic of P(FT | G):

Isaiah 45:18 - he who created the heavens,
he is God;
he who fashioned and made the earth,
he founded it;
he did not create it to be empty,
but formed it to be inhabited

OK, on to the topic of P(FT | ~G). In vernacular, what is the probability that the finely-tuned aspects of the universe would arise naturalistically? This is the term in the fine-tuning argument that typically takes on values like 10^-60 or 10^-120, etc. It’s the term that I called “epsilon.”

In particular, you said that I needed to be careful with that term, and you are absolutely right. Like you said, if you really want to think deeply about this, it’s not really accurate to just take the 10^-60 number and say that’s the probability. Again, you are 100% correct. So why do so many people (including myself) do that? TBH, I think for most people, they probably don’t realize the subtlety. For me, it’s just so much easier to communicate the idea that way, and in the end, if you want to go deeper, the conclusion is the same anyway, because you are just pushing back the fine tuning one step. But the route is indeed more subtle.

BTW, please read the following as if I were discussing an exciting topic that I like to think about and on which I am interested in hearing your feedback, rather than some guns-blazing attack on the non-theistic worldview. Because the former is the way I mean it, rather than the latter.

To make matters concrete, let’s unpack the discussion by focusing on the example of big G, which is the constant that is often said to be finely-tuned to one part in 10^60. It is not necessarily the best example to take, since it has its flaws, but it’s certainly an easy one to discuss. As well, anything I say here can be easily generalized to discussion of other constants.

Now, without loss of generality (and for the sake of discussion), we are free to pick units such that G = 1, so that the life-permitting range for values of G is 1 - 10^-60 < G < 1 + 10^-60. This is a narrow range for sure, but it does not necessarily translate into a low probability, because that depends on the probability distribution from which we are randomly drawing values of G.

(And of course, we are now ranging into philosophical/metaphysical speculation...we have no known mechanism by which we may suppose the existence of a probability distribution, nor one from which a value of G could be “drawn”, but in the end I don’t think it matters. I think the argument makes intuitive sense and will apply to just about any theoretical/hypothetical mechanism that one could come up with.)

Let’s imagine for a second that the probability distribution from which we are drawing the value of G is uniform from zero to two. In that case, p(G) = 0.5 uniformly on that interval. Then P(1 - 10^-60 < G < 1 + 10^-60) does indeed equal 10^-60. In this “special” case, the range of life-permitting values really does equal the probability of getting a value within that range.

But what if the probability distribution were a normal distribution with mu = 1 and sigma = 10^-61? In other words, a really, really tight distribution around the desired value of G = 1. In that case, we are *virtually guaranteed* to have a value of G within the life permitting range.

However…(do you see where I’m going with this?)...the only way you may legitimately assume we have such a “special,” atypical probability distribution for G is if you admit there is fine-tuning in the probability distribution itself. How in the world would one, apart from an intelligent creator with a purpose in mind, possibly justify having a probability distribution that forces this otherwise seemingly serendipitous, life-permitting value of G? (And that’s just one finely-tuned parameter.)

In other words, in my estimation, by correctly pointing out that the narrow fine-tuning range does not equate directly to probability, you escape the fine-tuning at that level, only to encounter it at a deeper level. With the fine-tuning argument, not only do you have to answer the age-old question of, “Why is there something rather than nothing,” but also, “Why is the something (that be, rather than nothing) the way that it is?”

[See summary page of this discussion, with links to all the posts, here.]