Farewell to Reality

by Jim Baggott

  To a large extent, we have already examined the rationale that Weinberg used in our earlier discussion of fine-tuning. The effect of a non-zero cosmological constant is to accelerate the expansion of spacetime. If the cosmological constant is too large, spacetime expands too quickly and galaxies and clusters of galaxies cannot form. In 1987, the cosmological constant was assumed to be zero, but the possibility of vacuum fluctuations of the kind predicted by quantum theory suggest something rather larger. A factor 10120 larger, in fact.

  Weinberg came up with an alternative approach: ‘Perhaps Λ must be small enough to allow the universe to evolve to its present nearly empty and flat state, because otherwise there would be no scientists to worry about it.’17

  Whereas Hoyle’s prediction was turned into an example of anthropic reasoning post hoc (Hoyle did not invoke observer self-selection at the time he made his prediction in 1953, although he was subsequently comfortable with this logic), Weinberg’s paper was overtly anthropic in nature. It is titled ‘Anthropic Bound on the Cosmological Constant’, and he referred readers interested in learning more about anthropic reasoning to Barrow and Tipler’s book, Paul Davies’s 1982 book The Accidental Universe and a 1983 paper by Carter.

  Weinberg refined his prediction in a paper published two years later. At the time, there was no evidence for a small positive cosmological constant, but he speculated that the anthropic upper bound was large enough to show up eventually in astronomical observations. Nine years later, observations of Type Ia supernovae confirmed that his instincts were correct.

  The carbon coincidence and the anthropic upper bound on the cosmological constant appear on the surface to be good examples of anthropic reasoning leading to genuine predictions that were subsequently tested and verified through observation and experiment. But before we rush to embrace the weak anthropic principle as science, let’s dig a little deeper.

  As I mentioned, Hoyle did not use anthropic reasoning in making his prediction. What, then, did he do? He took a perfectly good scientific observation — the relative abundance of carbon in the universe — and argued that there must be a valid scientific reason for this. If his ideas about stellar nucleosynthesis were right, then there had to be a physical mechanism by which carbon is formed in the interiors of stars. The most logical physical mechanism involves the existence of a resonance at an energy that will enhance the rate of carbon formation over the rate of disintegration of the beryllium nuclei. The fact that we are intelligent, carbon-based life forms is neither here nor there.

  A study of the history of Hoyle’s prediction led Kragh to conclude:

  Only in the 1980s, after the emergence of the anthropic principle, did it become common to see Hoyle’s prediction as anthropically significant. At about the same time mythical accounts of the prediction and its history began to abound. Not only has the anthropic myth no basis in historical fact, it is also doubtful if the excited levels in carbon-12 and other atomic nuclei can be used as an argument for the predictive power of the anthropic principle.18

  Similarly, even though Weinberg’s logic was overtly anthropoid, he could have reached the same conclusions without referring to the anthropic principle. The question again concerns physical mechanism, not the fact of human existence. The mathematical formulae in Weinberg’s 1987 paper concern only physical parameters, such as the density of mass-energy in the universe, the gravitational constant, and so on. Weinberg took some perfectly good scientific observations — the relative abundance of galaxies in the universe — and argued that there must be a valid scientific reason for this. The most logical physical mechanism demands a practical upper bound on the cosmological constant.

  Writing in 2008, the theorist Lee Smolin observed: ‘Just as Hoyle’s argument has nothing to do with life, but is only based on the observed fact that carbon is plentiful, Weinberg’s first argument has to do only with the observed fact that galaxies are plentiful.’19

  These are examples of relatively straightforward scientific deduction from observed facts. Yes, I know ‘observed’ must mean we’re here to make the observations. However, whilst it is certainly important to be aware of selection effects in scientific observation (and of course, we run head-on into such effects in quantum theory), we don’t tend to be in the habit of putting every observation in a human context, even though no observation of any kind is possible without humans.

  All scientific knowledge is, after all, intrinsically human knowledge.

  At issue, then, is not whether the weak anthropic principle is scientific. Rather, it is whether it actually adds any real value to scientific reasoning. To answer this question we would need to find an example of a scientific (and therefore testable) prediction that is unambiguously driven by anthropic reasoning. To my mind, this would have to be a prediction that depends on a selection bias stronger than simply observers observing the facts of nature, since it seems we should always be able to lift consideration of selection effects out of the logic without greatly impairing our ability to reach a valid conclusion.

  I know of no such anthropic prediction. And, I suspect, given that science as it is widely practised is based on detached observation and experiment, the circularity of the logic means that any such anthropic prediction wouldn’t actually be scientific. In other words, I suspect that the Copernican Principle is too deeply embedded in the very fabric and meaning of science that to try to eliminate it, even through the agency of a relatively mild form of privilege, renders the result unscientific.

  Science is just not set up this way.

  All things being equal

  Finally, we must now turn our attention back to the relationship between the anthropic principle and the multiverse, which is the principal reason for the recent surge of interest in anthropic reasoning.

  In an interesting twist, it has been argued that the application of the weak anthropic principle to the notion of the multiverse actually ‘saves’ science from the threat posed by intelligent design. The subtitle of Susskind’s The Cosmic Landscape reads: ‘String Theory and the Illusion of Intelligent Design’. We’re encouraged to accept the anthropic multiverse because this is the best solution to the fine-tuning problem. Reject it and we’re stuck with intelligent design as the only alternative explanation for the universe we inhabit.

  In Anthropic Bias, Bostrom argues that, all things being equal, a multiverse theory is indeed more probable:

  … consider a single-universe theory hU on which our universe is fine-tuned, so that conditional on hU there was only a very small probability that an observer-containing universe should exist. If we compare hU with a multiverse theory hM, on which it was quite likely that an observer-containing universe should exist, we find that if hU and hM had similar prior probabilities, then there are prima facie grounds for thinking hM to be more probable than hU given the evidence we have.20

  Bostrom uses Bayesian decision theory to make his point. Similar arguments have been made by superstring theorists, so it’s worth a short excursion to explain this logic.

  Thomas Bayes was an eighteenth-century mathematician and Presbyterian minister who developed a theorem concerning probability, published after his death in 1761. In its modern form, Bayes’s theory is sometimes used to evaluate and compare scientific theories in relation to observational and experimental evidence. A scientific theory is allocated a prior probability, which we can think of as expressing a scientist’s ‘degree of belief’ in it. After reflecting on some relevant observational or experimental data, we compute a posterior probability. If this is greater than the prior probability then the evidence is clearly in favour of the theory and our degree of belief in it increases.

  If we are confronted by two distinct theories with equal prior probabilities, and the evidence increases the posterior probability of one of them above the other, then we clearly have a means to choose between them. We keep the theory with the higher posterior probability and reject the other (at least, until we get more evidence).

  The
point that Bostrom is making is that if our two theoretical models under consideration (which he calls hU and hM) both have equal prior probabilities, then simple statistical probability favours the multiverse model. The fine-tuned single universe hU is simply much less probable than an observer-selected universe existing among a great multiplicity of possibilities, hM.

  I have no argument with this, as it stands. But if I take hU to be the current authorized version of reality — in the form of the standard models of particle physics and big bang cosmology — and hM to be any contemporary multiverse theory, then I have a big problem. These two theories do not have equal prior probabilities. By virtue of all the observational and experimental facts that support it, my degree of belief in hU is high.

  In case we’ve forgotten, let’s quickly remind ourselves of the status of the multiverse described by superstring/M-theory. We first assume that elementary particles can be represented as vibrations in filaments of energy. We assume a supersymmetric relationship between fermions and bosons. We assume that superstring theory’s extra spatial dimensions are compactified in a Calabi—Yau space. We accept the M-theory conjecture. We assume that our universe is but one of a large number (possibly an infinite number) of inflating spacetime regions in a multiverse. We assume that the 10500 different possible Calabi—Yau shapes are physically realized in different universes, resulting in universes with different physical parameters — different particle spectra, different physical constants and laws.

  There is no observational or experimental evidence for any of these assumptions. So, my degree of belief in hM is virtually non-existent. Applying Bayesian logic at this point doesn’t change the picture much, if it all.

  Of course, the superstring theorists will argue that this isn’t a valid comparison. The multiverse theory hM should subsume the single-universe theory hU, such that hU is seen to be an approximation of the larger theory. Fine. When the theorists are able to use the multiverse theory to calculate everything that the standard models of particle physics and big bang cosmology can calculate, I’ll be ready to reconsider.

  Where does this leave us? In The Cosmic Landscape, Susskind explained that he could find only two kinds of arguments against anthropic reasoning:

  As much as I would very much like to balance things by explaining the opposing side, I simply can’t find the other side. Opposing arguments boil down to a visceral dislike of the anthropic principle (I hate it) or an ideological complaint against it (it’s giving up).21

  I’d like to offer a third argument. I don’t hate the weak anthropic principle. I don’t think its adoption by theorists means they’ve given up. I reject the weak anthropic principle because it is simply empty of scientific content. It adds absolutely nothing to the debate. And yet it is used by some contemporary theorists to provide a rather facile logic, a veneer to deflect the fact that multiverse theories themselves are not scientific. Anthropic reasoning is the last refuge of theorists desperate to find a way to justify and defend their positions.

  So, what do I think is going on? How do I explain the fine-tuning of the universe? My hands are in the air. It’s a fair cop. I have no explanation because science does not yet have an explanation. We may be here because, by happy accident or the operation of some complex natural physical mechanisms we have yet to fathom, the parameters of the universe just happen to be compatible with our existence.

  And this is indeed the point. Scientists (even theoretical physicists) should not be afraid to say that they don’t know. Nobody is expecting them to have all the answers to human existence. We want them to speculate, to push the frontiers of their science. But when their ambition to give answers drives them to tell fairy tales, smothered in a sugar-coating of anthropic logic, let us all be clear that we’ve left science far behind.

  * This is because the helium nucleus, consisting of two protons and two neutrons, is also an alpha particle. The process involves three alpha particles combining together to form a carbon nucleus.

  * If this seems rather generous (the Nobel Prize is valued at about £750,000, which is shared between recipients), then we should probably stop to reflect on the new Fundamental Physics Prize, established on 31 July 2012 by Russian entrepreneur Yuri Milner. Milner’s new Foundation is ‘dedicated to advancing our knowledge of the Universe at the deepest level by awarding annual prizes for scientific breakthroughs, as well as communicating the excitement of fundamental physics to the public’ Prizes are valued at $3 million (almost £2 million), and many notable string theorists were among the nine to receive its first awards. In December 2012, Stephen Hawking was awarded a special prize of $3 million and a second $3 million prize was shared by a select group of experimental physicists working at CERN’s LHC. For more information, see www.fundamentalphysicsprize.org.

  * The Templeton Foundation acquired the services of leading PR company Bell Pottinger to handle the media buzz that was provoked by the announcement.

  ** For example, the Templeton foundation publishes Big Questions Online (www.bigquestionsonline.com), on which can be found a recent (10 July 2012) posting by Steven M. Barr, Professor of Physics at Delaware University, titled: ‘Does Quantum Physics Make it Easier to Believe in God?’.

  12

  Just Six Questions

  Defining the Destination at the End of a Hopeful Journey

  As far as the laws of mathematics refer to reality, they are not certain; and as far as they are certain, they do not refer to reality.

  Albert Einstein1

  I’ve found it hard to reach this point without my head filling with all kinds of homilies about journeys and destinations. I set out with the intention to provide you with at least an entertaining journey, one that I trust you have travelled hopefully. These are serious issues, and I still hope to persuade you that there’s quite a lot at stake. But I also hope I’ve been able to make my points in a way that you’ve found enjoyable, with a minimum of po-faced proselytizing.

  We’ve now arrived at our destination, and it’s incumbent on me to try to wrap things up as best I can and give you some sort of definition of precisely where we are. In thinking about how best to do this, I considered providing you with some kind of summary, but this seemed to involve simply repeating all my main points, and I’m conscious that I’ve done that a couple of times already.

  Instead, I figured it would be more interesting to anticipate some of the questions that might now be lurking at the back of your mind, and try to answer them as best I can.

  I’ve thought of six.

  If fairy-tale physics isn’t science, what is it?

  This is a very good question (and thank you for asking). As I’ve tried to explain, when there is no longer an attempt to ground theoretical developments by forcing them to relate to the hard facts of empirical reality, the result is metaphysics. Therefore, the easy answer to this question is to declare that fairy-tale physics is metaphysics.

  But this won’t do. I went to great lengths in Chapter 1 to explain that reality (in its broader, non-empirical sense) is a metaphysical concept and that it is extremely difficult, if not impossible, to eliminate all the metaphysical elements from what we would have no difficulty in accepting as perfectly legitimate science. Philosophers of the logical positivist persuasion had a go at this in the 1920s and 1930s, and failed.

  I’m not a professional philosopher, but it seems to me that the root of the problem lies in what we tend to demand from a description or explanation of the physical world if it is to be acknowledged as satisfying and meaningful.

  Just what does it mean to comprehend or understand something? We might argue that we understand something when we have grasped the rules governing its properties and behaviour and can use these rules to manipulate the present or predict the future.

  But there are rules, and then there are rules. We demand descriptions based on rules of the kind that convey insight and understanding. If we strip all the metaphysics from science — all the stuff we can’t prove by reference to
the facts but nevertheless believe to be contingently true (in the spirit of the Veracity Principle) — then what we are left with is a rather vacuous instrumentalism. Theories then serve a simple purpose. They are instruments that we use to relate one set of empirical facts to another, but in a way that conveys no real understanding.

  I want to illustrate why I don’t think this is very satisfactory by reference to a famous allegory devised by American philosopher John Searle. Searle formulated his ‘Chinese room’ thought experiment to attack some assumptions regarding artificial intelligence, but it will serve our purposes here to help us recognize what we mean by ‘understanding’.

  Searle sits on a chair in an empty room. We slip him pieces of paper beneath the door carrying questions written in Chinese which require simple, single-word answers. Searle doesn’t speak Chinese, but he has a handbook titled Empirical Rules for the Interpretation of Chinese Characters. The rules described in the handbook tell him how to interpret the characters without providing a translation. They are rules of the type ‘if then. By blindly following the rules, Searle can select quite sensible answers in a way that does not contradict previous answers.

  He slips his answers back to us beneath the door. From our perspective, he is having an intelligent conversation with us, in Chinese.

  But Searle doesn’t understand Chinese. He is just a computer, following an algorithm. To have an intelligent conversation with the universe, we need more than mechanical algorithms or instrumental rules that allow us to manipulate the present or predict the future. We need a structure: we need to understand the letters, the vowels, the words, the grammar and syntax. We need to understand the basic elements of reality — light, matter and force, space and time — and how these elements combine together to construct the poetry of the universe.

  Metaphysics is an inherent and perfectly natural part of the language we use in our dialogue with nature. Eliminate it completely and the language becomes devoid of real meaning. We find we can no longer hold a sensible conversation in it.