Over at his Evolution Blog, Professor Jason Rosenhouse has written an article entitled Multiverses in which he argues that popular objections to the multiverse by advocates of cosmological fine-tuning are irrational. In particular, Rosenhouse argues that the multiverse does not violate Occam’s razor, because the proper way to measure the simplicity of a theory is not by counting the number of entities it postulates, but by the elegance and simplicity of the mathematical equations underlying the theory. Rosenhouse also quotes from a short article entitled The Multiverse Strikes Back in Scientific American by Max Tegmark, who observes:
[I]t’s quite striking to me that the mathematically simplest theories tend to give us multiverses. It’s proven remarkably hard to write down a theory which produces exactly the universe we see and nothing more.
In his reply to scientific criticisms of the multiverse concept in the same issue of in Scientific American by George F. R. Ellis, Tegmark points out that the multiverse is an inevitable consequence of other prominent theories in physics (eternal inflation, string theory and quantum mechanics), all of which are testable. But then he goes on to argue for the existence of an even larger and crazier multiverse than the ones entailed by these theories. Tegmark, unlike most scientists, believes in an absolutely unrestricted multiverse in which “anything that could happen does happen” in some parallel universe. Tegmark evidently thinks that this kind of multiverse would demolish the fine-tuning argument for God’s existence at one stroke. While Professor Jason Rosenhouse is not as sold on the multiverse as Tegmark, he strongly agrees that a multiverse would undercut fine-tuning arguments.
Well, I hate to disappoint them both, but Professor William Dembski anticipated this line of argument back in 2003, in an article entitled Infinite Universe or Intelligent Design? Dembski uses the example of Arthur Rubinstein to show that design inferences are possible even in the unrestricted multiverse contemplated by Tegmark, where anything that can happen does happen, in some parallel universe. Dembski’s argument is at once engaging and persuasive, so I hope he will forgive me for quoting him at length:
Consider the following possibility: Was Arthur Rubinstein a great pianist or was it just that whenever he sat at the piano, he happened by chance to put his fingers on the right keys to produce beautiful music? It could happen by chance, and there is some corner of an infinite universe where everything is exactly as it is on planet earth except that the counterpart to Arthur Rubinstein cannot read or even appreciate music and happens to be incredibly lucky whenever he sits at the piano. Examples like this can be multiplied. There are corners of an infinite universe where counterparts to me cannot do arithmetic and yet sit down at a computer and write probabilistic tracts about intelligent design. There are even extremely remote pockets of an infinite universe where my Chicago Cubs win the world series. Perhaps Shakespeare was a genius. Perhaps Shakespeare was an imbecile who just by chance happened to string together a long sequence of apt phrases. An infinite universe, in virtue of its unlimited probabilistic resources, ensures not only that we will never know but also that we have no rational basis for preferring one to the other.
It might appear at this point that Professor Dembski is arguing that an in unrestricted multiverse, inductive inferences (which form the basis of science) would be unreliable, but that is not his intention here. Indeed, Dembski even proposes a sensible way of keeping bizarre scenarios – such as a Rubinstein who cannot read music, but gets lucky whenever he sits at the piano – out of our own universe:
Not so fast. Given unlimited probabilistic resources, there does appear to be one way to rebut such anti-inductive skepticism, and that is to admit that while unlimited probabilistic resources allow bizarre possibilities like this, these possibilities are nonetheless highly improbable in the little patch of reality that we inhabit. Unlimited probabilistic resources make bizarre possibilities unavoidable on a grand scale. The problem is how to mitigate the craziness entailed by them, and the only way to do this, once such bizarre possibilities are conceded, is to render them improbable on a local scale. Thus, in the case of Arthur Rubinstein, there are portions of an infinite universe where someone named Arthur Rubinstein is a world famous pianist and does not know the first thing about music. But it is vastly more probable that in portions of the universe where someone named Arthur Rubinstein is a world famous pianist, that person is a consummate musician. What’s more, induction tells us that ours is such a portion.
Professor Dembski realizes that not everyone will be convinced by this line of argument, so he fleshes out the reasoning that justifies our conviction that when Rubinstein performs in a concert, he is indeed a consummate musician. What is critical is that Rubinstein’s performance exhibits specified complexity.
But can induction really tell us that? How do we know that we are not in one of those bizarre portions of an infinite universe where things happen by chance that we ordinarily attribute to design? Consider further the case of Arthur Rubinstein. Imagine it is January 1971 and you are at Orchestra Hall in Chicago listening to Rubinstein. As you listen to him perform Liszt’s Hungarian Rhapsody No. 2 in C sharp minor, you think to yourself, “I know the man I’m listening to right now is a wonderful musician. But there’s an outside possibility that he doesn’t know the first thing about music and is just banging away at the piano haphazardly. The fact that Liszt’s Hungarian Rhapsody is cascading from his fingers would thus merely be a happy accident.”
The idea that Rubinstein is just banging away at a keyboard and getting lucky seems to you absurd. But if you take seriously the existence of an infinite universe, then you need to take seriously some counterpart to you pondering these same thoughts, only this time listening to the performance of someone named Arthur Rubinstein who is a complete musical ignoramus. How, then, do you know that you are not that counterpart?”
To answer this question, let us ask a prior question: What leads you to think that the man called Rubinstein performing in Orchestra Hall is a consummate musician? Reputation, formal attire, and famous concert hall are certainly giveaways, but they are neither necessary nor sufficient. Even so, a necessary condition for recognizing Rubinstein’s musical skill (and therefore the design in his performance) is that he was playing a complicated arrangement of musical notes and that this arrangement was also specified (in this instance, the concert program specified that he was to play Liszt’s Hungarian Rhapsody No. 2 in C sharp minor).
In other words, you recognized that Rubinstein’s performance exhibited specified complexity. Moreover, its degree of complexity enabled you to assess just how improbable it was that someone named Rubinstein was playing the Hungarian Rhapsody with apparent proficiency but did not have a clue about music. Granted, you may have lacked the probabilistic and information-theoretic apparatus to describe the performance in these terms, but the implicit recognition of specified complexity was there nonetheless, and without that recognition there would have been no way to attribute Rubinstein’s playing to design rather than chance.
The same line of thinking that warrants our conclusion that the specified complexity exhibited in Rubinstein’s playing is the result of design also warrants the theist’s conclusion that the specified complexity we find in the cosmos and in living things is the product of an Intelligent Designer of nature.
In the theory of intelligent design, specified complexity is a reliable empirical marker for design. It is how we preclude the interplay of chance and necessity and properly detect the agency of an intelligence. Granting this use of specified complexity (and we certainly use it this way for human artifacts), on what basis could we attribute natural phenomena that exhibit specified complexity to material mechanisms, which by definition operate purely through the interplay of chance and necessity? Note that we are not just talking about an analogy here (as in classical design arguments that depend on finding similarities between artifacts and biological systems, say). Rather, we are talking about an isomorphism – the specified complexity in artifacts is identical with the specified complexity in natural systems (be they cosmological or biological).
Dembski’s argument shows how design inferences can be legitimate, even in an unrestricted multiverse of the kind proposed by Tegmark:
It follows that the challenge of an infinite universe to intelligent design fails. It fails because there is no principled way to discriminate between using the unlimited probabilistic resources from an infinite universe to preclude design and using specified complexity to infer design. You can have one or the other, but you cannot have both. And the fact is, we already use specified complexity to infer design. Moreover, unlike an infinite universe, which is inherently beyond the reach of empirical inquiry, specified complexity is an empirically determinable feature of objects, events, and structures. Bottom line: Regardless whether the universe is finite or infinite, it is possible for empirical evidence to confirm intelligent design in nature.
Now, at this point, some readers may still be inclined to argue: “Yes, but in Rubinstein’s case, we know the identity of the design-maker: he’s the man performing on stage. In the universe’s case, we don’t know the identity of the designer.”
To refiute this argument, let’s consider a different case. Suppose that our planet is visited by an invisible alien who announces in advance, over a loudspeaker, that he will be performing Liszt’s Hungarian Rhapsody No. 2 in C sharp minor at a certain concert hall, on a certain day, at a certain time. At the appointed time, the invisible alien sits down at the piano and does as he promised. (Don’t forget that in Tegmark’s unrestricted multiverse, there are invisible aliens.) The element of specification is here: the alien did as he promised. Wouldn’t it be rational to conclude that a skilled musician was at work here?
Now let’s suppose that the alien does not announce himself, but that at a certain concert hall, on a certain day, at a certain time, a piano starts playing. It plays music that you’ve never heard before, but the harmony is exquisite and perfect. Each piece of music that you hear is a tour de force, worthy of a professional musician. It exhibits the element of specificity, because there are no mistakes, and each of the pieces exhibits a depth and harmonic richness that makes them worthy of a concert performance. Wouldn’t it be rational to conclude that a skilled musician was at work here?
I submit that the design inference has nothing to fear from a multiverse. What do readers think?