The Evidence Crisis

Thanks to a kind invitation from the Simons and John Templeton Foundations and the World Science Festival, last Friday (30 May 2014) I participated in a public discussion on ‘Evidence in the Natural Sciences’ with Professors Brian Greene and Peter Galison.

This discussion was the final act in a one-day symposium of the same name, held at the Simons Foundation’s Gerald D. Fischbach Auditorium on 5th Avenue, in New York City. These were comfortable, well-appointed surroundings. But the overwhelming message from the symposium was actually quite discomfiting. In its 300-year maturity, it seems that science is confronted with nothing less than a crisis of evidence.

The crisis takes many forms. I learned that mathematicians are increasingly resorting to computer-based proofs that signal a loss of certainty and the ‘end of conviction.’ Efforts are underway to develop computer-based algorithms that will soon provide the only way to review such proofs, leading one audience member to wonder how long it will take to eliminate mathematicians entirely from the process.

Eliminating humans, and their biases and general lack of self-criticism, appears to be the only workable solution to a crisis of evidence in the bio-medical sciences as well. This is a field in which John P. Ioannidis (now at the Stanford School of Medicine) famously declared in 2005 that ‘most published research findings are false’ [1]. This was real sit-up-and-take-note stuff. The research findings in question are of the kind that can lead eventually to clinical trials of new drugs.

I’d been invited to address yet another type of evidence crisis. Last year I published a book, called Farewell to Reality, which challenges some of the prevailing opinions about contemporary theoretical physics of the kind which address our ‘big questions’ concerning the nature of the physical universe. In it I argue that some theorists have crossed a line. They are suffering a ‘grand delusion,’ a belief that they can describe physical reality using mathematics alone, with no foundation in scientific evidence. I call the result ‘fairy-tale’ physics.

My role in our public discussion was that of interlocutor and facilitator. Greene is of course widely known for his Pulitzer short-listed The Elegant Universe and follow-ups The Fabric of the Cosmos and The Hidden Reality, his many radio and TV appearances and his growing role as a popular science educator (he is co-founder of the World Science Festival with his wife, former ABC News producer Tracy Day). Galison is a Harvard science historian with a flair for popularization, author of Einstein’s Clocks, Poincaré’s Maps and Objectivity. He has developed a couple of TV documentaries, about the H-bomb and about national secrecy and democracy, and is currently working on a film about the long-term storage of nuclear waste.

And then there was me, sitting in the middle. An interlocutor with an agenda. What follows is not a transcript of our discussion (I’m hoping that the Simons Foundation will post a video of this online), but rather a summary of my position.

So What’s the Problem?

Wind the clock back. On 4 July 2012, I watched a live video feed from the CERN laboratory near Geneva, and celebrated the announcement that a particle that looked a lot like the Higgs boson had finally been discovered.

This was a triumph for a theoretical structure called the standard model of particle physics. This is the theory that describes physical reality at the level of elementary particles and the forces between them and which helps us to understand the nature of material substance.

But our joy at the discovery of the Higgs was tempered by concern. We know that the standard model can’t be the whole story. There are lots of things it can’t explain, such as the elementary particle masses and the nature of dark matter. And it is not a ‘theory of everything’: it takes no account of the force of gravity.

We build scientific theories in an attempt to describe and hopefully explain empirical data based on observations and measurements of the physical universe around us. But in the twenty-first century we’ve run into a major obstacle. We have evidence that tells us our theories are inadequate. But we have no data that provide meaningful clues about how our theories might be improved. Theorists have therefore been obliged to speculate.

But in their vaulting ambition to develop a ‘theory of everything’, some theorists have crossed a line without any real concern for how they might get back. The resulting theories, invoking superstrings, hidden dimensions and a ‘multiverse’, among other things, are not grounded in empirical evidence and produce no real predictions, so they can’t be tested. Arguably, they are not science.

Albert Einstein once warned:2

Time and again the passion for understanding has led to the illusion than man is able to comprehend the objective world rationally by pure thought without any empirical foundations – in short, by metaphysics.

What did Einstein mean? Quite simply, there can be no science without evidence or at least the promise of evidence to come.

How Should We Interpret ‘Reality’?

I believe that the root of the problem lies in the way we seek to interpret the word ‘reality’. Pick up any text on philosophy and you’ll find discussions of reality under the general heading ‘metaphysics’. How come? Physical reality seems really rather tangible and logical. It confronts us every morning when we wake up. Surely, despite what the philosophers might say, we can be pretty confident that reality continues to exist when there’s nobody looking. The science fiction writer Philip K. Dick once declared: ‘Reality is that which, when you stop believing in it, doesn’t go away.’3

But reality is curiously schizophrenic. There is an ‘empirical reality’ of things as we observe or measure them. This is the reality that scientists try to address. The purpose of science is to seek rational explanations and ultimately an understanding of empirical reality by establishing a correspondence between the predictions of scientific theories and the results of observations and measurements. Such a correspondence gives us grounds for believing that the theory may be ‘true’.

Here’s an example. In 1964, Peter Higgs, Francois Englert and Robert Brout speculated that there must exist a special kind of quantum field – which became known as the Higgs field – responsible for giving mass to elementary particles. In 1967 Steven Weinberg used this field to predict the masses of some exotic particles called W and Z bosons, which we can think of as ‘heavy photons’. These particles were discovered at CERN in 1983, with more-or-less the masses that Weinberg had predicted. Consequently, the Higgs field was incorporated into the standard model of particle physics.

But there could in principle have been other possible explanations for the masses of the W and Z particles. If the Higgs field really exists, then it should produce a tell-tale field quantum – the Higgs boson. In 2012, establishing a correspondence between the empirical data produced at CERN and theoretical predictions for the behaviour of the Higgs boson gave us grounds to believe that the Higgs field really does exist and that the standard model is ‘true’ within its domain of applicability.

We would perhaps not hesitate to declare that lying beneath this empirical reality must be an independent reality of things-in-themselves, a reality of things as they really are. But such an independent reality is entirely metaphysical. Kind of by definition, we cannot observe or measure a reality that exists independently of observation or measurement. We can only speculate about what it might be like. As Werner Heisenberg once said: ‘We have to remember that what we observe is not nature in itself, but nature exposed to our method of questioning.’4

It is this independent reality that philosophers try to address, which is why their speculations appear under the heading of ‘metaphysics’. Now, philosophers are not scientists. They don’t need evidence to establish a correspondence between their interpretation of an independent reality and our empirical world of observation and measurement. They’re more than satisfied if their interpretation is rationally and logically structured and coherent.

There is truth here, but of a subtly different kind.

Crossing the Line

Contemporary theorists find themselves caught in a bind. Without any clues from empirical data to guide theory development, and ever eager for answers to the ‘big questions’ of human existence, it seems that theorists have had no choice but to cross the line from physics to metaphysics.

There’s nothing wrong with this. Theorists have been doing this for hundreds of years. But, as scientists rather than philosophers, they have speculated about the nature of an independent reality of things-in-themselves with the aim of getting back across the line as quickly as possible. Einstein’s special and general theories of relativity were founded in arguably metaphysical speculations about the nature of space and time. But Einstein was at pains to get back across the line and show how this interpretation of space and time might manifest itself in our empirical reality of observation and measurement. The rest, as they say, is history.

Contemporary theorists have simply stopped trying to find their way back. Worse, they have built a structure so complex and convoluted and riddled with assumptions that it’s virtually impossible to get back.

What do I mean? As they have explored the metaphysical landscape of a mathematically-defined independent reality, the theorists have misappropriated and abused the word ‘discovery’. So, they ‘discovered’ that elementary particles are strings or membranes. They ‘discovered’ that there must be a supersymmetry between different types of particle. They ‘discovered’ that the theory demands six extra spatial dimensions which must be compactified into a space so small we can never experience them. They ‘discovered’ that the five different types of superstring theory are subsumed in an over-arching structure called ‘M-theory’. Then, because they ‘discovered’ that there are 10-to-the-power-500 different ways of compactifying the extra dimensions, then each of these must describe a different type of universe in a multiverse of possibilities. Finally, they ‘discovered’ that the universe is the way it is because this is the only universe in the landscape of 10-to-the-power-500 different kinds that is compatible with our existence.

I want you to be clear that these are not discoveries, at least in the sense of scientific discoveries. They are assumptions or conclusions that logically arise from the mathematics but for which there is absolutely no empirical evidence. It’s not really so surprising that the theory struggles to make any testable predictions. There is simply no way back to empirical reality from here.

Don’t be blinded by all the abstract mathematics, all the ‘dualities’ which connect one kind of mathematical description with another. These help to establish ‘coherence truths’, of the kind X = Y. But when neither X nor Y correspond to anything in the empirical world that even hints at the possibility of an observation or a measurement, then we can be clear that this all remains firmly metaphysical.

Alarm Bells

We have a problem. The theorists are stuck on the wrong side of the line, and most believe there is no viable alternative. As Nobel laureate Steven Weinberg remarked to me a little while ago:5

String theory still looks promising enough to be worth further effort. I wouldn’t say this if there were a more promising alternative available, but there isn’t. We are in the position of a gambler who is warned not to get into a poker game because it is appears to be crooked; he explains that he has no choice, because it is the only game in town.

Obviously, we sympathize. But what if, instead of being obliged to attend remedial therapy, those addicted to gambling were able somehow to influence the rules, to make gambling an acceptable pastime? No scientist likes to be stigmatized, to be accused of pseudo-science. This is why some in the theoretical physics community are seeking to change the way we think about science itself.

For example, string-theorist-turned philosopher Richard Dawid recently argued:6

… final theory claims introduce the new conception of a scientific process that is characterized by intra-theoretical progress instead of theory succession… The status of a merely theoretically confirmed theory will always differ from the status of an empirically well-tested one. However, in the light of the arguments presented, this difference in status should not be seen as a wide rigid chasm, but rather as a gap of variable and reducible width depending on the quality of the web of theoretical arguments.

The problem with this is that as soon as we accept the notions of ‘intra-theoretical progress’ and ‘theoretically confirmed theory’ we risk completely disconnecting from any sense of real scientific progress. We risk losing respect for evidence, deepening the crisis, unplugging from empirical reality and training – how many? one, two? – generations of theorists to believe that this is all okay, that this is science fit for our modern, post-empirical age. We ensure they inherit an addiction to gambling.

Some are already talking of these theorists as ‘lost generations’:7

It is easy to estimate the total number of active high-energy theorists. Every day hep-th and hep-ph bring us about thirty new papers. Assuming that on average an active theorist publishes 3-4 papers per year, we get 2500 to 3000 theorists. The majority of them are young theorists in their thirties or early forties. During their careers many of them never worked on any issues beyond supersymmetry-based phenomenology or string theory. Given the crises (or, at least, huge question marks) in these two areas we currently face, there seems to be a serious problem in the community. Usually such times of uncertainty as to the direction of future research offer wide opportunities to young people, in the prime of their careers. To grab these opportunities a certain reorientation and re-education are apparently needed. Will this happen?

Maybe it’s already too late. In a more recent assessment, Dawid writes:8

Many physicists may wish back the golden old days of physics when fundamental theories could (more often than not) be tested empirically within a reasonable period of time and a clear-cut empirical verdict in due time rendered irrelevant all tedious theoretical considerations concerning a theory’s viability. Empirical science, however, must answer to the situation it actually faces and make the best of it. A sober look at the current situation in fundamental physics suggests that the old paradigm of theory assessment has lost much of its power and new strategies are already stepping in.

There’s more. Scientists have a duty of care to a public that has developed an unprecedented appetite for popular science. This is an appetite that was greatly enhanced by the success of Stephen Hawking’s A Brief History of Time and has been fed by some excellent science writing, not least from Greene himself.

I haven’t done the research, but I very much suspect that if you were to ask a randomly selected group of scientifically literate readers about the theories we use to describe and understand the universe, many of these readers will likely tell you something about superstrings, hidden dimensions and the multiverse.

In truth, today these theories describe nothing and add nothing to our understanding, because this is metaphysics, not science. These theories do not form part of the accepted body of tried-and-tested scientific theory used routinely to describe our physical world, the kind used at CERN in the hunt for the Higgs boson. As Nobel laureate Tini Veltman claimed, paraphrasing Wolfgang Pauli, these theories are ‘not even wrong.’9

Now readers of popular science might just want to be entertained with the latest ‘Oh wow!’ revelations from contemporary theoretical physics. But surely they also deserve to know the truth about the scientific status of these theories. I think Danish science historian Helge Kragh hit the nail squarely on its head when he observed, in a review of John Barrow and Frank Tipler’s The Anthropic Cosmological Principle:10

Under cover of the authority of science and hundreds of references Barrow and Tipler, in parts of their work, contribute to a questionable, though fashionable mysticism of the social and spiritual consequences of modern science. This kind of escapist physics, also cultivated by authors like Wheeler, Sagan and Dyson, appeals to the religious instinct of man in a scientific age. Whatever its merits it should not be accepted uncritically or because of the scientific brilliancy of its proponents.



I believe that contemporary theoretical physics has lost its way. It has retreated into its own small, self-referential world. In search of a final ‘theory of everything’, theorists have been obliged to speculate, to cross the line from physics to metaphysics. No doubt this was done initially with the best of intentions, the purpose being to get back across the line carrying some new insight about the way the universe works that would provide an empirical test. Instead, the theorists have become mired in a metaphysics from which they can’t escape.

We might ask if there’s any real harm done. I personally think there’s a real risk of lasting harm to the nature of the scientific enterprise. Admitting ‘evidence’ based on ‘theoretically confirmed theory’ is a very slippery slope, one that risks undermining the very basis of science. In the meantime, the status of this fairy-tale physics has been mis-sold to the wider public.

We’re in crisis, and we need a time-out.


Farewell to Reality: How Modern Physics Betrays the Search for Scientific Truth is published in the US by Pegasus Books, New York. I’d like to acknowledge a debt to Columbia University mathematical physicist Peter Woit, and especially his book Not Even Wrong: The Failure of String Theory and the Continuing Challenge to Unify the Laws of Physics, Vintage, London, 2007.


1 J.P.A. Ioannidis, ‘Why Most Published Research Findings are False’, PLoS Medicine, 2(8), e124, August 2005.

2 Albert Einstein, ‘On the Generalised Theory of Gravitation’, Scientific American, April 1950, p. 182.

3 Philip K. Dick, from the 1978 essay ‘How to Build a Universe that Doesn’t Fall Apart Two Days Later’, included in the anthology I Hope I Shall Arrive Soon, edited by Mark Hurst and Paul Williams, Grafton Books, London, 1988. This quote appears on p. 10.

4 Werner Heisenberg, Physics and Philosophy: The Revolution in Modern Science, Penguin, London, 1989 (first published 1958), p. 46.

5 Steven Weinberg, personal note to the author, 13 January 2013.

6 Richard Dawid, ‘Underdetermination and Theory Succession from the Perspective of String Theory’, Philosophy of Science, 73/3, 2007, pp. 298-332.

7 M. Shifman, ‘Frontiers Beyond the Standard Model: Reflections and Impressionistic Portrait at the Conference’, arXiv:1211.0004v2, 14 November 2012.

8 Richard Dawid, ‘Theory Assessment and Final Theory Claim in String Theory’, Foundations of Physics, 43/1, 2013, pp. 81-100.

9 Martinus Veltman, Facts and Mysteries in Elementary Particle Physics, World Scientific, London, 2003, p. 308.

10 Helge Kragh, Centaurus, 39, 1987, pp. 191-194. This quote is reproduced in Helge Kragh, Higher Speculations: Grand Theories and Failed Revolutions in Physics and Cosmology, Oxford University Press, 2011, p. 249.