Can we rely on our cosmological picture of the universe, or might we have got it all wrong? Although complicated math seems to support theories like dark matter and dark energy, we should be ready to interrogate the reasoning behind them, writes Bjørn Ekeberg.
"It is not anti-scientific to question established beliefs, but central to science itself. At the creative heart of science is a spirit of open-minded inquiry," writes Rupert Sheldrake in The Science Delusion.
It's hard to imagine a scientific subject matter that would call for more open-mindedness than the ambitious project of understanding the entire cosmos from our own extremely limited standpoint within it.
Cutting edge science at HowTheLightGetsIn Global Read more And yet few scientific fields are more established in their core beliefs than cosmology. Since the late 1960s, most scientists have settled on a 'standard cosmology' as the only serious and viable research program in astrophysics. Its fundamentals are considered facts and rarely, if ever, questioned in public. The Big Bang story is entrenched as the most scientific world-picture we have.
Its proponents may be tempted to take this remarkable conformity as proof the theory rests on a solid foundation. But there are many reasons to be skeptical both of this claim, and the conviction with which it is often posited.
When the problem is particularly hard, people tend to follow the crowd and not to question authority.
In the spirit of scientific self-questioning, I'd like to briefly outline 8 potential problem areas for the credibility of the ruling theory.
Think of it as a guide to a 'stress test' of cosmology: these are dimensions where the theory looks particularly weak or unconvincing from a critical perspective and where it could easily face more serious scrutiny than from a lonesome philosopher of science.
My role is not to pass judgment, to propose alternative theories, or to attack science as such - quite the contrary. Like Sheldrake, I believe good science comes from properly understanding the limitations of our knowledge. And while it is remarkable what we can know about the universe, it is perhaps even more remarkable how little we actually do know.
1. Conformity Dynamics
Like any science, cosmology happens in a social context. The scientific ideal, of course, is that private beliefs do not matter, scientists only work with knowledge.
But there are many obvious ways in which beliefs, assumptions and non-empirical preferences shape the scientific picture. In cosmology, this includes a belief in universal mathematical laws of nature (Ekeberg, 2019), or a distinct preference for theories that are beautiful, elegant and based on simplicity of laws (Hossenfelder, 2018).
In a more general sense, science involves belief in the credibility of others, other scientists, institutions, authority and tradition - and the importance of this kind of belief becomes more apparent at a collective level.
Behavioral economist Cass Sunstein shows how the tendency to conformity in large groups can harden an emerging consensus on a problem into a dogma (Sunstein, 2019).
Social experiments in a multitude of contexts, he writes, demonstrate that when we are faced with difficult questions and limited knowledge, we tend to defer to those we perceive as authorities on the matter. Typically these are recognized experts in the field.
But experts, as Sunstein demonstrates, can be led astray and this can happen at a collective level despite the rational intention of every single one of its actors. When the problem is particularly hard, people tend to follow the crowd and not to question authority.
Is there any problem harder than the nature and extent of the entire universe?
2. Inferential Extent
Historically, Newton's physical laws made up a theoretical framework that worked for our own solar system with remarkable precision. The planet Neptune, for example, was discovered not through telescopes but via predictions based on Newton's model.
But as the scales grew larger, Newton’s theory proved limited. Einstein's General Relativity framework offered an extended and more precise model beyond the furthest reaches of our own galaxy. How far could it go - beyond our own tiny neighborhood that we are able to sample?
The Big Bang 'standard cosmology' research program that emerged in the mid-20th century effectively stretches the model's validity to a kind of infinity - either defined as the boundary of the universe radius (now calculated at 46 billion light-years) or in terms of the beginning of time.
That is a cosmic leap. As far as we can be confident that our empirical observations agree with both Newtonian gravity and General Relativity makes up about 0.1% of the estimated size of the universe under the Big Bang model. In other words, 99.9% of its estimated extent is an inference from an expansive General Relativity model.
There is no extrinsic way to test the validity of the model itself, which is involved in many steps of the way from determining what to observe to synthesizing observations and selecting data into supporting evidence for the model itself.
3. Observational Evidence
How do we see the universe? In any astrophysical observation, the theoretical framework is alpha and omega. Output from space telescopes and other experimental probes detect variables according to parameters determined by the theoretical framework, which in turn is the only guiding map scientists have to make sense of the chaos in the sky.
All the data in the world cannot help you if there are any fundamental problems with the framework you used to render, correlate and interpret the data.
Any advanced billion dollar space exploration technology faces nebulous conditions, extreme distances, and a scale that requires constant observational selection. This yields a vast array of data with a very high noise-to-signal ratio. In turn, these data have to undergo a major 'clean-up' in which assumptions from the theoretical framework are applied.
This process is analogous to how detectors in particle accelerators rely on extreme filtering and selection to isolate the events they are designed to search for. But whereas terrestrial physics can compensate by endlessly repeating experiments, astronomy cannot. Any sufficiently clean measurement will only be a map of a moment in time and space, and only by appeal to one of cosmology's many simplifying assumptions can it be considered representative of the cosmos as a whole.
In the end, astrophysicists can obtain satisfying numbers in agreement with their theoretical calculations. These results are then inferred as evidence that the hypothetical initial conditions necessary to perform the measurement are indeed correct. But how scientific is this rather cyclical course of inquiry?
4. Validation Criteria
What counts as evidence? Notably, cosmology has also stretched the goalposts by developing a unique criterion for validation. It's called 'concordance.' The argument goes that if you can show consistent patterns across multiple data sets, this can be inferred as evidence for the validity of the theory employed).
A similar criterion was invented in the early 20th century, and this was how atoms and Planck's 'quantum' came to be accepted as physical facts, despite not being demonstrable by any conventional standard. But as astrophysicist David Merritt points out in a critical review, "it was the agreement of the measured value of a single parameter, in multiple experiments, that lent credence to the reality of atoms and energy quantization" (Merritt, 2017).
Cosmologists' appeal to 'concordance,' on the other hand, is substantially weaker: "They mean that it is possible to find a single set of parameters that provides an acceptable fit to the conjunction of observational data sets, and not that there is independent confirmation of the value of any single parameter."
In this branch of physics, not only the theories but also validation of evidence have become increasingly mathematical and statistical in scope.
5. Data Over-reliance
But what about all the data? Yes, today we have more observational data sets than ever before and it will only keep rising in the next decade with new satellites and space telescopes and even bigger supercomputers.
But all the data in the world cannot help you if there are any fundamental problems with the framework you used to render, correlate and interpret the data. The more data you have, the more you rely on the theory to guide you, and the more the chances increase - however unlikely it may seem from the inside - that the research program could be led astray by a self-reinforcing process.
As astrophysicist David Lindley puts it, science is "the exploration of the unknown. But researchers in fundamental physics, knowingly or not, have adopted entirely the opposite strategy: they have declared in advance what they are looking for and are toiling to create a theory that matches their expectations" (Lindley, 2020)
And what exactly are these theoretical expectations based on?
6. 'Dark' Hypotheses
'Standard cosmology' is a many-headed Hydra, made up of many levels of auxiliary hypotheses and adjustable parameters.
Today's prevailing model is called Lambda CDM (for 'Cold Dark Matter') and it controversially relies on concepts like dark matter and dark energy. These dark elements are not empirical phenomena but rather mathematical inventions that work to uphold the validity of the framework itself. They are 'dark' insofar as they must exist, assuming the framework is right.
Because this completely unknown dimension, necessary to make the theory work, is estimated at 95% of the entire universe, many remain unconvinced. The so-called MOND theory (Modified Newtonian Dynamics), for example, is vying to replace the dark matter hypothesis without undermining the framework as a whole.
As for dark energy, its properties ("equations of state") are explicitly adjusted in order to keep the data consistent with the mathematical framework. This a clear example of what philosopher of science Imre Lakatos described as a key function of auxiliary hypotheses: to protect the “hard core” of the research program.
7. Metaphysical Core
The crux of standard cosmology is to combine a certain version of General Relativity with particle physics to hypothesize a set of initial conditions for an invisible 'early universe'.
The Big Bang itself has always been a metaphysical hypothesis. It is not based on astronomical observation - rather, the event marking the beginning of time is a consequence of a mathematical solution to General Relativity. But beyond this event, all the initial conditions that would have to exist in order to explain our current visible universe are also metaphysical in this sense.
It is at least not unthinkable that we have gotten our story of the universe wrong.
As I argue in my book, the Big Bang hypothesis has become a given, not because it is necessarily correct but because it is productive. It works because it provides an operational framework for conducting astrophysical research.
Initial conditions have to be assumed, and in turn, they can be inferred back from interpreting the data in which they were already presupposed. Over time, this repetition creates the semblance of verification even if the foundational premise remains untestable in obscurity.
But this circularity is also the closest the theory comes to a solid foundation.
8. Problematic Pillars
But as I pointed out in the wake of last year's debate in Scientific American, these are not really four independent cornerstones but rather co-dependent variables derived from observations that were interpreted within the context of the first premise.
For example, the pillar called "predicted abundances of the light elements," is based on the hot, dense, early Universe as premised by the Big Bang theory. (It is also a pillar with its share of observational anomalies.)
In this sense, the central ground 'pillar', which first set about the sea change in our view of the universe, was based on Hubble's telescope observations in the late 1920s. Measuring the redshift of certain known stars seemed to show that the Universe was expanding. At first, Hubble himself and other astronomers were unconvinced by this interpretation.
But when a new General Relativity solution for an expanding universe was developed in the 1930s, along with the enormous rise in nuclear physics research, a hard consensus began to form around the 'expanding universe' hypothesis. This interpretation of the redshift developed into a new crucial cosmological variable called the Hubble Constant (for which recent measurements, incidentally, have shown huge inconsistencies).
However, for the theory to galvanize into the leading research program, the most decisive 'pillar' was the Cosmic Microwave Background radiation (CMB), first detected in the 1960s. The Big Bang theory proposed to explain this mysterious radiation phenomenon as a leftover glow from the Big Bang. This may seem plausible enough but it is really to explain one unknown in terms of another unknown. It is also far from the only possible explanation.
The Big Bang theory got an early edge because it could make measurable predictions. Of course, if you work on this theory, you would be inclined to interpret the validation of a prediction of one parameter as proof of the theory as a whole. But that is yet another inference leap of cosmic proportions.
Leap of Faith
Boiled down, underneath data sets and mathematical inventions, the hard empirical pillars for the story of our universe come down to this: a consensus interpretation of an astronomical measurement, combined with a mathematical model capable of explaining a newly observed radiation phenomenon.
Whether that is convincing enough to maintain credibility beyond the vocal support of the scientific establishment is for others to judge.
Perhaps it seems implausible that such an advanced science could be incorrect. But from a careful look at the logic and foundations of the prevailing theory, it is at least not unthinkable that we have gotten our story of the universe wrong.
It is also not unthinkable that we will never know in a scientific sense. Because to reach the end of the universe with conviction requires a certain kind of faith.
Ekeberg, Bjørn. Metaphysical Experiments: Physics and the invention of the universe, 2019
Hossenfelder, Sabine. Lost in Math: How beauty leads physics astray, 2018.
Lindley, David. The Dream Universe: How fundamental physics lost its way, 2020
Merritt, David. Cosmology and convention, 2017
Merritt, David. A Philosophical Approach to MOND, 2020
Sunstein, Cass R.. Conformity: The power of social influences, 2019