Testing the Multiverse: Beyond the Limits of Science?
“I don’t see them,” said Frank Wilczek, who won the Nobel Prize in Physics for deciphering the inner structure of protons and neutrons (they’re almost all empty space). “The application of science can give insight into any question that makes sense. It may not give an answer. It may advise that the question is ill-posed, or it may provide only partial insight — but I don’t think anything should be ruled out as inaccessible to science.”
A multiverse, as we shall see, is a test case.
The essence of science — what science is actually doing — may seem obvious to scientists, but not to philosophers of science.
Bas van Fraassen is just such a philosopher of science, at San Francisco State University, and a professor of philosophy emeritus at Princeton University in New Jersey, and he is a strict empiricist, meaning that he believes knowledge qua knowledge must come directly from observation and data. “When I think about science,” he told me, “I think of it as a large human enterprise that has certain criteria of success, and, as an empiricist, I say that all such success relates to what’s observable. When science is successful, it gives the best possible descriptions and explanations of what we find in the observable realm.” (All quotes are from “Closer to Truth.”) [5 Reasons We May Live in a Multiverse ]
Van Fraassen said he is not a “scientific realist,” meaning that he does not accept that the scientific criterion of success is “truth in every respect” or “truth, period.” He said he rejects the prevailing notion that science can penetrate deeper than “just what’s observable” and “postulate all things needed in order to explain observable things.”
He has argued that we must decouple what we observe from whatever underlying reality may be generating those observations. In other words, there are (at least) two levels of reality: One consists of the rules and regularities of the physical world, which science can access and measure. But the other level, the ultimate source of those rules and regulations, science can never even access, much less come to know.
“To say you accept a scientific theory,” van Fraassen said, “[only] means that you believe that it is empirically adequate with respect to the observable phenomena, what we find via inspection and measurement. I do not see the scientist as Columbus discovering America, but rather as constructing models and theories in order to represent observable phenomena.”Strict empiricism offers a more modest mission for science. So does its philosophical cousin, anti-scientific realism, which rejects, for various reasons, the common-sense idea that what we perceive is what really exists (for example, doubting that reality is mind-independent). Both strict empiricism and anti-scientific realism see an unbridgeable gap between deep reality and human cognition, because, like in a relay race, the “baton of information” is passed from one medium to another: from the actual source in deep reality, to electromagnetic radiation, to man-made instruments of observation, to biological sense organs, to neural processing in brains.
At each stage of that “information relay,” argues the strict empiricist (and the anti-scientific realist), information is lost. There are several gaps, which combine to constitute an overall, unbridgeable gap.
As a strict empiricist, van Fraassen affirms only that which human beings can know for sure. He allows only observations and models into his realm of certitude, but not the underlying realities that generate them. “The scientific realist must have theories referring to real things,” he stressed, “but to an empiricist, that is not a scientific explanation — it’s a metaphysical explanation — and not the thing to do.”
Princeton physics professor J. Richard Gott described the boundaries of science in terms of what science can and cannot know. “We’ve learned a great deal about the universe — age, structure, initial conditions, how it started, how it’s developing. But a theologian might say, ‘Well, have you really answered the question of why is there a universe, as opposed to no universe at all?’ It’s easy to imagine no universe at all. Science is not prepared to answer this question, at least not at the present time,” he said.
Is there anything in the middle, so to speak, between what science can and cannot know?
The multiverse, stretching the limits of understanding
In answering that question, Gott pointed to a multiverse. “Scientists now speak of a multiverse ,” he said, “many different universes beyond our observational capability. But we talk seriously about them.”
South African cosmologist George Ellis sharply asserted that view in the journal Nature. Ellis, who is well known for challenging the multiverse, argued that “attempts to exempt speculative theories of the universe from experimental verification undermine science.
“This is a very powerful argument,” Linde noted in response. “You cannot prove anything about things that you cannot see. Fortunately, this argument is wrong. Here’s what’s often missed in the discussion of the multiverse: If we have many experimental or observational facts that can be explained only in the context of one particular theory (e.g., multiverse), these facts constitute experimental or observational evidence in favor of this theory.” [Note: “…in favor of this theory” is precise language; it does not mean “…evidence that proves this theory.”]
“Thus, anybody disliking the theory of the multiverse,” Linde continued, “should be asked to present an explanation of these observational or experimental facts in any other context, not involving a multiverse. Many people tried. Nobody succeeded. That’s why we take it so seriously.”
To make sure I understood what he meant by “experimental or observational facts that can be explained only in the context of one particular theory,” or in other words, “that which cannot be explained otherwise,” Linde invoked the “anthropic principle”. This oddly profound principle constrains theories of physics and cosmology by human existence, because humans must obviously exist if they are to observe the facts that they try to explain. This seemingly trivial, tautological or even nonsensical claim is, depending on one’s views, either a surprisingly powerful predictive and explanatory tool or an embarrassing abandonment of the scientific method and an excuse for failure.
Two Nobel Laureates in physics have offered me their diametrically opposing views on the anthropic principle.
Steven Weinberg at the University of Texas at Austin, one of the pioneers of this new way of thinking about science, said that assuming “different versions of the multiverse idea, the anthropic principle is just common sense.” The reason, he said, was that “if there is a vast number of universes — in which the various constants [of physics], including the energy in empty space [known as the cosmological constant], vary from universe to universe — it is natural that we will be only in the kind of universe that could support life.” In other words, if the cosmological constant, which is a kind of repulsive gravity, would be greater than a certain maximum amount, then its anti-gravity force would resist and overpower gravity, preventing gravity from forming agglomerated bodies (galaxies, stars, planets).
On the other hand, when asked his view of the anthropic principle, physicist David Gross at the University of California at Santa Barbara said, “I hate it.”
“Anthropic considerations acquire real physical meaning only if one has many potential options,” Linde explained to me, “but only if some of them are compatible with the existence of observers.
“The multiverse provides these options”, he asserted. “The most famous problem addressed by anthropic considerations is the size of the cosmological constant.” In other words, in studying the energy density of empty space, the vacuum, why is the cosmological constant so astonishingly small but still not zero?
Linde said there are multiple problems in physics and cosmology, and only a multiverse can solve them, thus validating the theory of the multiverse as “science,” even if justified in this new kind of way.
“I couldn’t make any sense of this ‘anthropic principle’,” Linde explained, “until I proposed a model of an inflationary universe consisting of many different parts with different properties [different laws of physics]. The same picture appears in a much more convincing way in [cosmic] eternal chaotic inflation and finally became even more convincing after the discovery of 10^500 vacua in string theory.” (See Linde’s “A brief history of the multiverse.”)
This theoretical finding was that there are (very roughly) 10^500 different, theoretically feasible configurations or ways that string theory can generate different, theoretically feasible laws of physics (based on all the possible stable, geometrical or topological configurations of an infinitesimally small, higher-dimensional “manifold” that string theory proposes as the foundation of space and time and particles and forces). Each one of this vast number of specific configurations or ways characterizes its own kind of universe — and perhaps doing so in reality innumerable times to define innumerable universes — all of which taken together compose the multiverse. (This does not mean that the multiverse is inextricably tied to string theory, however elegant; there are other mechanisms that could generate multiple universes.)
Moreover, Linde claimed, this also means that the same facts that can serve as observational or experimental evidence in favor of a multiverse can simultaneously serve as observational or experimental evidence in favor of string theory. In defense of his position [in recent correspondence following first publication of this essay], Linde stressed that no one has developed a better, more robust theory to explain the cosmological constant than the multiverse and string theory (not even after 18 years — the cosmological constant was discovered in 1998 with the stunning observation that the expansion of the universe was accelerating, not slowing down due to gravity as everyone had expected).
Linde admitted that this is an unusual way to think about science, and he cited Weinberg: “Now we may be at a new turning point,” Weinberg said famously, “a radical change in what we accept as a legitimate foundation for a physical theory.”
“New theory is born in pain,” Linde reflected, “but so far, the parents are quite happy, or at least cautiously optimistic,” and he again quoted Weinberg. “I found a report of a discussion at a conference at Stanford,” Weinberg said, “at which [U.K. Astronomer Royal] Martin Rees said that he was sufficiently confident about the multiverse to bet his dog’s life on it, while Linde said he would bet his own life. As for me, I have just enough confidence about the multiverse to bet the lives of both Linde and Rees’ dog.”
The failures of multiverse theory
With Linde’s permission, I circulated his remarks to others quoted in my “Confronting the Multiverse” essay, and I received everyone’s permission to post the remarks that follow.
In response, Ellis stated that Linde’s argument requires three parts: experimental or observational facts to be explained, a viable theory that can explain these facts and no other theory that can work as well. And Ellis claimed that for the multiverse, “there are problems with each part.”
“First, the anthropic puzzle” — i.e., how human observation seems to select or “determine” those laws of physics that are consistent with human existence — “is not a problem in physics,” Ellis said. “It’s a philosophical issue. By that I mean that there is no requirement of a new theory because of some experiment that contradicts the standard model of particle physics plus the standard model of cosmology. Rather, the aim is to explain values of fundamental constants involved in those theories. But that is always in terms of new theories involving other constants, which then in turn need to be explained in terms of further theories with further constants, and so on. It is quite unclear how many constants there are to be explained, when to stop the chain of explanation, or indeed what constitutes a valid ‘explanation’ — for this is a philosophical (or perhaps psychological?) aim.”
Second, Ellis challenged each of what he labeled “the three pillars” of Linde’s argument for multiple universes: (a) cosmic eternal chaotic inflation plus (b) the string-theory mechanism that generates vast numbers of different kinds of universes to give (c) an anthropic explanation of the value of the cosmological constant. Ellis claimed that competing theories and observational evidence disfavor Linde’s theory. In any case, Ellis said, “using the multiverse to anthropically explain all the constants of physics does not work, because so many constants will have wrong values.” In other words, according to Ellis, anthropic reasoning (i.e., where human existence must constrain physical conditions) in the context of a putative multiverse cannot predict the actual values of physical constants.
In summary, Ellis said arguments for the multiverse fail “because their requirements are not fulfilled.” Pointedly, Ellis asserted that Linde’s multiverse-generating theories of cosmic inflation are ruled out or disfavored by recent observational data.
In response, Linde denied that “eternal chaotic inflation” is ruled out or disfavored by recent observational data, stressing differences among eternal chaotic inflation, chaotic inflation in general and its simplest version (all of which Linde discovered, or invented, depending on one’s views). Only inflation’s simplest version, he said, is ruled out by recent observations.
“The main idea is that in a broad class of theories, inflation can start even if initial conditions in the universe were chaotic, which explains the name ‘chaotic inflation,'” he said. “High-temperature phase transitions, which were the basis of the first versions of inflationary theory, are no longer required. Almost all inflationary models studied now belong to this general class.”
Linde stressed that just because he rejects Ellis’ arguments against the multiverse does not mean that he thinks the theory of the multiverse is complete.
Personally, in my epistemology of a multiverse, as it were, I am influenced by Weinberg’s simple statement: “I quite agree with Linde” — which is consistent with Weinberg’s own position permitting anthropic arguments to address scientific problems.
Challenging current belief
Although in questioning the multiverse, Ellis is in the minority — certainly among contemporary cosmologists — I admire his challenge to current belief. After all, that is what MIT physicist Alan Guth (the originator of cosmic inflation theory), Linde and others did in postulating cosmic inflation theory and multiple universes in the first place. Ellis has made people think hard about the assumptions underlying the purported reality of multiple universes, which, if real, would radically expand the scope of existence.
Paul Davies, physicist and director of the Beyond: Center for Fundamental Concepts in Science at Arizona State University, stated that the multiverse theory can be indirectly tested in two ways.
“First, if a multiverse is one of many predictions of a theory, and if its other predictions can be directly tested,” he said, “then (assuming it survives those tests) we increase our confidence that there is a multiverse. Secondly, we can use statistical analysis related to anthropic selection, as Linde points out, to make testable predictions. So, for example, the multiverse explanation of the weakness of [repulsive] dark energy predicts that as we learn more about the formation of galaxies, the measured value of dark energy should come closer and closer to the maximum permitted that is consistent with galaxy formation. If some other explanation is correct, then there is no such constraint.”
In other words, Davies made a provocative point about the fit between the actual, measured value of (repulsive) dark energy and the theoretical maximum value that such (repulsive) dark energy could take, but still allow for human existence by not overpowering, and thus not nullifying, the gravitational attraction needed to form galaxies, stars and planets. The tighter this fit, Davies said, the greater the consistency with a multiverse.
But, dare I say it, the same would also be true for consistency with a Design/Designer, would it not? (“The Multiverse and God” will be the topic of my third and final essay on the multiverse.)
Yet Davies challenged the multiverse by combining the implications of a multiverse with the so-called “Simulation Argument.” This draws on two premises or assumptions, that (i) alien civilizations exist throughout the universe and (ii) continue to increase their computational capabilities. Given those two premises, then randomly chosen observers, like human beings, would most likely find themselves living in a simulated or “fake” universe (unless there are cosmic or technological “showstoppers” preventing whole-world simulations).
Davies argued that in a multiverse, “‘fake universes’ would be even more likely to vastly outnumber real universes [than in our one universe without a multiverse], so that if we live in multiverse, we would be overwhelmingly more likely to be living in a simulated reality. But that would imply that the laws of physics in our universe are also overwhelmingly more likely to be simulations and therefore cannot be used to conclude that there is a multiverse! So there is an inconsistency at the heart of the multiverse concept.”
Let’s unpack the argument and lay it out. To Davies, “If you take seriously the possibility of a multiverse of all possible universes, including all possible variations, then there would have to be at least some of those universes where sentient civilizations would advance to the point where they would have sufficient computing power to simulate entire fake worlds (like in the ‘Matrix’ movies). Simulated universes are much cheaper to make than the real thing. So once you’ve got civilizations throughout the multiverse that can simulate universes,” Davies stressed, then this is what they will do, and do increasingly.
As a result, “the number of fake universes in a multiverse will proliferate greatly and very soon outnumber real ones.” [Is Our Universe a Fake?]
Fake universes, Davies said, “undermine all the arguments for a multiverse,” because arguments for a multiverse are based on the physics of this universe. “But if ours is a simulated universe, then our laws are simulated, too, which means that the whole of physics is a fake.” And if the whole of physics is a fake, Davies said, then the whole argument for a multiverse collapses. The reason is that while the multiverse argument proceeds from the physics humans have discovered in this universe, people cannot use this argument because it then leads, surprisingly, to the conclusion that this is a fake universe, with fake physics.
That, Davies claimed, is an internal contradiction of the multiverse theory. “You’re hoist by your own petard if you want to claim that there is a true multiverse with all possibilities, because it leads you to accept the fact that it’s just full of fakes,” he stressed, which then undermines the reality of physics, which in turn undermines standard arguments for the multiverse.
The best way to avoid this circular contradiction, in my opinion, is by asserting that simulating whole worlds, especially the simulation of consciousness, is impossible — even impossible in principle, even with an infinite number of universes, a claim that causes its own set of problems. [The Singularity, Virtual Immortality and the Trouble with Consciousness]
Another way might be to suppose that any simulated laws of physics in simulated universes would likely be based on “real” laws of physics in the simulators’ “real” universe. (But for this counterargument to work, there would need be no exceptions, none whatsoever — which in such a vast universe, I think, is a bar too high.)
At first, the Simulation Argument (the possibility that this universe is a fake) seems to have nothing to do with arguments for the multiverse, yet each, curiously, threatens the other. Just as Davies showed how simulations would undermine the multiverse, a multiverse of infinite numbers of universes would play havoc with the Simulation Argument of this universe. (The reason is that infinities mess up measurements and statistics, because when all things occur an infinite number of times, the relative occurrences of all things acquire a kind of strange equality.)
In support of a multiverse, philosopher John Leslie posed a novel argument. “If there is any ‘Creative Factor’ responsible for there being a universe at all, something rather than nothing, anything not a blank,” he argued, “then one could well expect that this cosmos-creating Creative Factor would operate more than once.” That Creative Factor would be a principle, force, God — it doesn’t matter what, he said. “Why on Earth would it have operated once only? Wouldn’t it make more sense to suppose that it had operated countless times? If so, then there would be grounds for believing in a multiverse that didn’t depend on accepting that the laws of physics we observe are genuine instead of ‘fakes.'”
Human beings seem hardwired to want to know things, all things; people want to push knowledge to its ultimate limits, plumb the depths of ultimate reality. Science or metaphysics? Science and metaphysics? Which is the multiverse?
Kuhn is co-editor, with John Leslie, of “The Mystery of Existence: Why Is There Anything at All?” (Wiley-Blackwell, 2013). Read more of Kuhn’s essays on Kuhn’s Space.com Expert Voices landing page.
Follow all of the Expert Voices issues and debates — and become part of the discussion — on Facebook, Twitter and Google+. The views expressed are those of the author and do not necessarily reflect the views of the publisher. This version of the article was originally published on Space.com.