• Elliot Sober’s Objection to the Fine-Tuning Argument

    It is a rare and wonderful moment in critical thinking when you approach an idea, think that it’s crazy and implausible on first glance, but later realize is dead right.

    That’s happened to me. The idea was the weak anthropic principle.

    To review, some people have argued that the apparent improbability of our universe being able to support life is evidence for theism. To this, some have countered that we couldn’t observe anything except life supporting universe (we’re here, duh!) and so therefore the so called ‘fine-tuning’ of the universe is meaningless. I thought this was trite and had trouble understanding how any human being could ever think was a good response. And then I read Elliot Sober.

    In one of Sober’s papers on the subject of fine-tuning, he says:

    “Suppose you use a net to fish in a lake and observe that all the fish in the net are over 10 inches long. At first, this observation seems to favor the hypothesis that all the fish in the lake are more than 10 inches long over the hypothesis that only 50% of them are. But then you learn that the net has holes that are 10 inches across. This makes you realize that you were bound to obtain this observation, regardless of which hypothesis about the lake is true.”

    With the fine-tuning argument, we are fishing with a net that cannot catch anything except universes that allow for us to exist. Just like Sober’s ten-inch fish no longer counted as evidence in favor of either hypothesis once the holes in the net were taken into account, so the same principle applies for us.

    Meditate on that long and hard. Reread the above a few times. It’s really difficult to grasp this, but once the lightbulb goes of you’ll see that it cannot possibly be false.

    A piece of evidence only counts in favor of one theory over another if that evidence is more likely to be observed if the first theory is true than if the second theory is true. Whether or not a god exists, we must observe a fine-tuned universe out of necessity, we cannot possibly observe anything else.

    To put it in bayesian term (see my post on Bayes’ Theorem) a , if fine-tuning is the “evidence” we are trying to explain, and our background knowledge contains the proposition that we exist, alongside the weak anthropic principle which states that all of the conditions necessary for our existence must be so because we exist, then it logically follows that there is a 100% chance we will observe fine-tuning even if there is not a god.

    Now there are alternative ways to look at this argument under Bayes’ Theorem. Perhaps we might not make “fine-tuning” the object of our explanation. Maybe we’ll put it into our background knowledge and instead make the object of explanation “a life-permitting universe exists.” Forget it, says Richard Carrier:

    “But then [our background knowledge] still contains ‘observers exist’ which still entails ‘a life-bearing universe exists’… If [we] tried putting ‘observers exist in [the evidence we’re trying to explain], [our background knowledge] would then contain the Cartesian fact ‘I think, therefore I am,’ which then entails [the evidence we’re trying to explain]. So we’re back at 100 percent again. If [we] tried putting “I think, therefore I am” in [the evidence box], [the] conclusion would only be true for people who aren’t observers (since [our background knowledge] then has no observers), and since the probability of there being people who aren’t observers is zero, his calculation would be irrelevant (it would be true only for people who don’t exist)…”   (p.410, The End of Christianity)

    That’s quite a lot to think about! If this doesn’t make sense to you, I’d suggest reading my blog post on Bayes’ Theorem carefully, because of course much of this quote only makes sense if you understand Bayes’ Theorem, and even then it is still difficult to grasp.

    The logic here, I think, is mostly simple. What we observe is 100% likely under the no-god theory, no matter how we try to frame the fine-tuning problem in Bayesian terms. That’s Carrier’s contention, anyway.

    What I have trouble grasping is the last bit: If we put “I think, therefore I am” in the evidence box, our background knowledge won’t contain the knowledge that conscious agents, or observers, exist. Carrier hasn’t done terribly well explaining this, which is regrettable. That said, our background knowledge must include the logical fact that “either conscious agents exist or they do not.” If they do, then that entails our evidence “I think therefore I am” with 100 percent certainty, no god required. If conscious agents do not exist, then we could never even possibly have the evidence “I think therefore I am.” Now taking basic logic out of background knowledge is not something we can responsibly do if we want to use Bayes’ Theorem to reach a logical conclusion.

    This appears to have exhausted the possible ways of rehabilitating the fine-tuning argument. It has collapsed. However, I think there is one more way in which it might be rehabilitated, but luckily, Carrier has beat me to the punch in rebutting it, and it’ll be what we look at next.

    EXTRA!! EXTRA!! Read all about it. Bill Jeffries and Michael Ikeda wrote a wonderful article on fine-tuning that makes similar points to Sober, and here it is for FREE. I will probably discuss it in greater detail in a future post.

    Category: Uncategorized


    Article by: Nicholas Covington

    I am an armchair philosopher with interests in Ethics, Epistemology (that's philosophy of knowledge), Philosophy of Religion, Politics and what I call "Optimal Lifestyle Habits."