Embedded within the laws of physics are roughly 30 numbers—including the masses of the elementary particles and the strengths of the fundamental forces—that must be specified to describe the universe as we know it. Why do these numbers take the values that they do? We have not been able to derive them from any other laws of physics. Yet, it’s plausible that changing just a few of these parameters would have resulted in a starkly different universe: one without stars or galaxies and even without a diversity of stable atoms to combine into the fantastically complex molecules that compose our bodies and our world. Put another way, if these fundamental parameters had been different from the time of the Big Bang onward, our universe would be a far less complex universe. This is called the “fine tuning observation.” The fine-tuning problem is to find out why this is.
As someone whose thought is both fueled and constrained by the scientific tradition, I am only interested in explanations that are scientific. This means a candidate explanation must be three things: confirmable, falsifiable, and unique. By “confirmable,” I mean that the hypothesis must lead to further consequences, otherwise unexpected or surprising, which could be confirmed by novel but possible experiments. “Falsifiable” means that it is possible to specify a novel but doable experiment that would invalidate the hypothesis if the experimental result contradicted the predictions of the hypothesis. “Unique” means that there are no other simpler or more plausible hypotheses that make the same prediction.
Any explanation that fails these tests should be abandoned. After all, it is possible to imagine a multitude of possible non-scientific explanations for almost any observation. Unless we accept the stricture that hypotheses must be confirmable, falsifiable, and unique, no rational debate is possible; the proponents of the various explanations will never change their minds.
Yet several of the most popular explanations for the fine-tuning problem fail these tests. One such hypothesis is that there is a god who made the world and chose the values of the parameters so that intelligent life would arise. This is widely believed, but it fails the test for a scientific explanation.
Another hypothesis that fails the test is what we may call the “anthropic multiverse.” Though there are many variations on this theme, the essential idea is that our universe is one of a large or infinite set of worlds that exist simultaneously, each with different, random values for those 30-some physical parameters. Hence, our universe has the very rare property of having parameters that give rise to sufficient complexity to make it hospitable to intelligent life. To connect this hypothesis with observations we have to limit ourselves to the study of the subpopulation of universes in which we could live. This is called using the anthropic principle.
The anthropic multiverse cannot make any falsifiable predictions, though. Here is one proof: We can divide all the parameters that define each universe into two classes. First, there are those that matter to the existence of life—change one of those, and your universe is no longer hospitable to life. But since we already know, or could deduce from our existence, the values of these parameters, they can’t be used to falsify predictions of the anthropic multiverse. Second, there are parameters that don’t matter to the development of intelligent life. Those parameters can take any value and still yield up a universe teeming with life. These parameters are distributed randomly, so they might take any value in our universe. Because any and every value is allowed, this second set of parameters can’t be used to falsify predictions of the anthropic multiverse either.
Even if an anthropic multiverse is fundamentally unscientific, though, that does not mean we need to throw out all multiverse theories. One way to make a multiverse theory scientific is to suggest that complex universes like ours must be typical in the population of universes. Now we can make predictions without invoking the anthropic principle. For instance, in my first book, “Life of the Cosmos,” I proposed the “theory of cosmological natural selection,” which predicts that the parameters of physics are fine-tuned to produce many black holes, which is the case in our universe, as we see by its great chemical and astrophysical complexity. It turns out that a universe that makes many stars, and hence many black holes, is also filled with the oxygen and carbon needed for life. This theory made a few falsifiable predictions that have so far held up, despite several opportunities to contradict them with real observations in the last two decades. One of these predicts that no neutron star can have more than twice the mass of the sun. These predictions involve properties that are otherwise very surprising; were they to be confirmed there would be no other explanation on the table.
There is one property of the inflationary multiverse which just plausibly could be confirmed—were certain parameters very delicately tuned—which is the observation of patterns in the cosmic microwave background that could be explained by other bubble universes having collided with ours. But if this is not observed it doesn’t falsify the hypothesis. It just means those parameters are not so finely tuned.
Readers of popular science may have encountered claims that some predictions of the anthropic principle have been confirmed. However, I argue that those claims are based on at least three different kinds of fallacies. First, there is the assumption that the properties we observe around us—for example, the fact that carbon and galaxies are plentiful in the universe—are essential for life. However, we can seek an explanation for why galaxies are plentiful without needing to assert that galaxies are helpful for life. We can see by observation that galaxies are plentiful without having to be in one. Second, many of these anthropic arguments make untestable claims about properties of hypothetical universes that will remain forever unobserved. Finally, there is the “inverse gambler’s fallacy”: Observing a single trial with an improbable outcome and deducing that it must be one of a large number of trials.
Defenders of theistic explanations assert that it might be the case that there is a god who made the universe and tuned its parameters so that we could exist. It might. Similarly, defenders of anthropic multiverse scenarios assert that it might just be the case that our universe is one of a vast collection of worlds with random laws and parameters. This also might be true. But science is not about what might be true, it is about what can convincingly be argued for by rational argument from public evidence. If we weaken this standard to admit the anthropic multiverse, we open the door to equally unscientific theistic explanations. The proponents of each can (and do) argue with each other, but they will never convince each other, for they have given up the method and criteria that are necessary to make a convincing case for a claim in science. Meanwhile, the fine-tuning observation is a challenge that requires a scientific explanation.
Editor's picks for further reading
FQXi: Our Not-So-Special Universe
In this blog post, Zeeya Merali reviews recent papers questioning whether our universe is really as fine-tuned as we thought.
TED: Lee Smolin on Science and Democracy
In this video, blogger Lee Smolin explores the similarities between science and democracy.
Discover Magazine: Science's Alternative to an Intelligent Creator: the Multiverse Theory
Tim Folger talks with proponents and critics of the anthropic multiverse.