Roger Penrose has a theory. He believes that the Big Bang was not the unique moment of creation that is commonly believed, but rather just one instance in a series of regularly recurring births and deaths that the universe cycles through over unimaginatively long time scales.
Well, that’s interesting, if a little mind-boggling, I can imagine you saying. But what concerns us now isn’t so much why this eminent mathematical physicist is convinced of this particular bouncing picture of our universe or what his specific theory implies for physics (or anything else). As fascinating as such questions are, they can be pretty technically daunting, even for professional physicists.
But one thing we can meaningfully discuss are the key questions of scientific verification and evidence. After all, surely what distinguishes a scientific theory from philosophical speculation is its testability, and it’s pretty hard to fathom how we might develop objective, persuasive evidence of the very early or very late universe: we certainly can’t run the cosmos back in time to check what was happening billions of years ago any more than we can wait around billions of years to check if and when another Big Bang somehow manages to occur.
So is it hopeless? Surprisingly, no. For Professor Penrose claims that concrete proof of his theory can be found by closely examining the radiation produced right after the Big Bang – the so-called cosmic microwave background (CMB). If his theory is correct, he maintains, evidence of the past universe would be found in the first light of our present universe, lurking as circular patterns in the CMB.
The good news is that the CMB is perhaps the most rigorously studied phenomenon in physics over the past 25 years. Once thought to be the veritable epitome of smoothness, increasingly detailed international experiments have enabled scientists to identify its minuscule fluctuations with remarkable levels of sophistication. Which is much more than simply an impressive technical accomplishment, because it is these very fluctuations from the very early universe which in time give rise to the bountiful numbers of galaxies that we see today. Indeed, our hugely heightened understanding of the CMB has done nothing less than transform fundamental physics, giving rise to an entire generation of “quantitative cosmologists” and singlehandedly elevating cosmology from its once dubious reputation of “hand-waving metaphysics” to its present status as one of the most dynamic fields in science today.
Given all of that, then, you might be forgiven for thinking that testing Professor Penrose’s theory should be pretty straightforward: simply take a close look at the CMB and see if circular patterns can be found. If they’re there, then the theory is, if not proved, at least strongly supported, while if they aren’t there, then the theory must surely be disproved, or at least severely weakened.
But a closer look reveals that it’s not quite that easy.
The first problem is a statistical one. Of course, if you look closely enough, you would expect to find some circular patterns somewhere in any collection of data. The key question is, how much? How do we know that a given pattern is confirming evidence of our theory or simply what we would expect in the normal course? In other words, To what extent can we distinguish a random signal from observed evidence?
At first that, too, might sound quite straightforward: if we see a lot more circular patterns than the universe would naturally produce, then we have some proof that Penrose’s theory is correct. But how can we tell how many circles the universe would “naturally produce”? Or, as Penrose puts it, we have to find some way of assessing what we mean by “a random sky”.
“A key question is, What do you mean by a ‘random sky’? Many people who do these sorts of things mean a sky by which you take information taken from the observed sky, which you then feed into your model of ‘the random sky’.
“But from our perspective, this is a circular argument, because if these effects are really happening then it makes no sense to have them being so integral to your definition of randomness.”
In other words, if our sense of “a random sky” is in some way directly based upon what we actually see, and if what we actually see is in turn at least partially a product of factors in accordance with Penrose’s theory, then any attempt to try to determine whether our observations reflect something “out of the ordinary” are necessarily bound to fail.
What we need, then, is to find some way of assessing randomness by some separate theoretical framework that has nothing directly to do with what we’ve actually observed. But in physics, such a framework often doesn’t exist – particularly in a field like cosmology where it’s typically impossible to repeat experiments (i.e. run the universe over again). Which means that if we’re not careful, we might find that what we observe will become inextricably bundled into our underlying theoretical understanding, making it impossible even in principle to distinguish what might have happened from what has actually happened.
The core issue, then, doesn’t really have much to do with physics at all (at least at first), but is instead a bread-and-butter TOK knowledge question: To what extent do our observations provide evidence for our theories? And if we want to make progress in physics or any other field, the answer to that question better not be: None at all.
Howard Burton, firstname.lastname@example.org
To best explore the topics raised in this post for teaching and learning, please see below a number of related expert resources that are part of Ideas Roadshow’s IBDP Portal which offers a dynamic IB-specific database of 500+ authoritative video and print resources featuring more than 80 world-leading researchers, including two Nobel Laureates, explicitly created to meet the needs of both teachers and students in the Diploma Programme – please visit our new website for further information.