The evidence crisis

220px-Calabi-Yau-alternateby Jim Baggott

Thanks to a kind invitation from the Simons and John Templeton Foundations and the World Science Festival, last Friday (30 May) 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 behavior 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, 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 one of the few universes 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 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.” Amen.

In the end

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 risk of lasting damage 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.

_____

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.

Jim Baggott completed his doctorate in physical chemistry at the University of Oxford and his postgraduate research at Stanford University. He is the author of The Quantum Story, The First War of Physics, and A Beginner’s Guide To Reality. Most recently he published Farewell to Reality: How Modern Physics Has Betrayed the Search for Scientific Truth.

[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.

196 thoughts on “The evidence crisis

  1. It must have been an interesting symposium MP. For what it’s worth, what is said here seems spot on to me. The problem is that a fundamental theory must be metaphysical. So either we stick to a non-reductive physics or we venture into metaphysics. ‘Twas ever thus. But I’m not sure it would be impossible to get back again. Doesn’t Ulrich Mohrhoff show us how to do it?

    Like

  2. Hi Jim,

    … the theorists have misappropriated and abused the word ‘discovery.’ So, they ‘discovered’ that elementary particles are strings or membranes. […] I want you to be clear that these are not discoveries, at least in the sense of scientific discoveries.

    Do string theorists actually use the word “discovered” in this way? Aren’t physicists well aware that before string theory can be regarded as confirmed it needs some good empirical evidence and predictive success?

    At the moment it does not have that, which is fine in a way, because what string theory currently is is mathematics, an exploration of mathematical structures that may or may not prove to have physical relevance. Often in history mathematical exploration has run ahead of physics and empirical evidence, so allowing some leeway there is not a problem. But isn’t everyone well aware that in the end it needs empirical verification?

    Like

  3. This was a very well written article which makes a very good case for the problems besetting theoretical physics. It’s hard to disagree with it too much, especially with the view that theoretical physicists have an obligation to be clear on how well-confirmed their ideas are.

    However, while I agree with and acknowledge the problem, there are a few reasons why I don’t see much of an alternative to the status quo (apart from encouraging theorists to be more forthcoming about their uncertainty).

    1) As stated, it’s the only game in town at the moment. It remains a possibility that these avenues of research will yield empirical results.
    2) If there are very compelling theoretical reasons to believe something, this is more than speculation and it is appropriate that it be represented as such, even as we acknowledge that we have no way of achieving empirical confirmation.
    3) We may be blurring the line between physics and metaphysics, but unless you want to charge metaphysics with pointlessness this is not necessarily a bad thing as long as we make it clear what we are doing.
    4) We must remember that just because an idea is untestable does not mean that it is false. For instance, if the anthropic principle is indeed the explanation for fine-tuning, as it may well be, then a search for a single-universe explanation of the constants is doomed to failure. That doesn’t mean we shouldn’t try, but it also means we should take seriously the idea of a multiverse.

    Like

  4. Hi Coel,

    I meant to pick up on the usage of the word ‘discovery’ also. I’m not sure Jim is really fair here, although I guess we can say there are discoveries in conceptual or mathematical space even if we don’t know that these apply to the real world.

    Like

  5. Few questions. Why should any independent reality, and there can be many, be any more accessible to everyday language purported to report purely subjective experiences – aka metaphysics – then any other human activity? There can be an independent reality of tooth fairies as well, let’s remember.

    Can any “metaphysical” statement ever be disproven or proven? Since the answer is no, by definition, why aren’t these statements just more magical thinking “Mind over matter.” – or really just everyday language over matter.

    Also by definition, any statement claiming to be testable by evidence must be disprovable.

    Finally, isn’t the “crisis” of unexplainable observations and measurements always how knowledge advances? Knowledge being the prediction of measurable future events.

    It seems a too clever marketing trick for philosophers to claim where there is uncertainty in language and empirical statements metaphysics and philosophy should be rushed in – like religious folks and the “god of the gaps” rhetorical trick.The author is proposing a “philosophy of the gaps” argument.

    Give the “gaps” a little while and some will be spanned.

    Like

  6. BMM, I thought hard about whether to approve this comment. On the one hand, I don’t want to impede the flow of discussion, especially from people who disagree with my positions (or, in this case, the guest author’s). But please be consider that my patience is not infinite. Your comment has actually precious little to do with the essay itself, and Jim is most certainly not advancing a “philosophy of the gaps” argument, whatever the heck that may be. Cheers.

    Like

  7. The big question for me on this issue is what is the best and most productive course of advancing science or physics in this case. I can see the force of the arguments on each end, with string theory being the only game in town, it makes sense to work towards it. However, it can also be argued that too many resources in terms of talented people and money has been spent in that direction with little to show for it. Maybe putting string theory on the back seat will give physicists a new way of looking at the problem.

    This is of course tricky because neither strategy is guaranteed to lead to success but considering we have been trying at String theory for decades now and have not made any progress, I would lean towards putting it in the back seat but not getting rid of it completely.

    Like

  8. I think it’s fair if that distinction is not made very clear when presenting these ideas to the public. When Briane Greene talks about discovery of strings in a popular TV documentary, it’s understandable that people will view it as the same as electrons or maybe something closer to Higgs Boson, which was much more likely to be tested empirically but people might get the impression that strings will be just as easily and directly proven as well.

    Like

  9. Except I don’t know how putting string theory in the back seat will give physicists a new way of looking at the problem. Unfortunately, we don’t automatically get new ideas by setting old ideas to the side, especially when we’re talking about a population of intelligent, dedicated and motivated physicists and not one individual stuck in a dead end. There are already many, many people trying to find viable, testable alternatives to string theory, and until they succeed it and other untestable ideas such as loop quantum gravity will remain the only game in town.

    Like

  10. I know of one way: it would free faculty lines, postdoc and graduate student positions, as well as research grants, to explore other options. During the debate Greene was extremely disingenuous when he painted an idyllic picture of the research community, in which individual scientists are completely free to choose what to work on, without any constraint whatsoever. Please.

    Like

  11. Interesting to compare this with the recent posts by Coel Hellier, and the comparison suggests to me two possible ways of rescuing current theoretical physics from the charge of physical meaninglessness. (1) It may be that the number of possible coherent theories that are compatible with known physics turns out to be very small, so that the laws and minimum set of parameters needed to specify our Universe are shown to be less than completely arbitrary; I think we would all agree that this would count as progress in our understanding of reality (this, I think, was at one time the hope with string theory); (2) It may turn out that some current theory will later be found to have testable consequences, just as the Big Bang theory turned out (as Hellier describes) to be testable in ways that Lemaitre could not have envisaged.

    I leave it to those more knowledgeable than me to evaluate these suggestions.

    Like

  12. We also don’t automatically get new ideas or progress by throwing all of our eggs into a basket that has not produced anything for decades now. I think you also are underestimating how much the current popular theory can influence young researchers into thinking this is the only viable way to go about the problem. That to me indicates that they are not pushing themselves as hard to think about alternatives because they are more convinced that string theory will lead the way. It’s possible but it’s certainly not as likely as say 30 years ago, when I would have been far more optimistic about string theory.

    Like

  13. Coel,
    Aren’t physicists well aware that before string theory can be regarded as confirmed it needs some good empirical evidence and predictive success?

    I am puzzled. Is this the same Coel who has spent the last few days arguing that we don’t need “some good empirical evidence and predictive success” for the speculative multiverse hypothesis?

    Have you reversed your position or do you apply less stringent conditions to your favoured hypotheses?

    You carry on to say
    But isn’t everyone well aware that in the end it needs empirical verification?
    Have you really changed your position to that extent?

    Like

  14. Re getting young academics to work on alternatives: in every field I know much about, the most arcane, non-intuitive, complex the theory, the more specialized and unreadable the journals that publish the stuff, the more jobs. Also, never underestimate the role of faculty, who will after all review and approve the degree, in “guiding” students toward their own perspective, even if privately the faculty may acknowledge little or no results other than jobs. No idea how to break this pattern.

    JG

    Like

  15. We need to be careful not to misrepresent string theorists such as Brian Greene. For example, here are some remarks from Brian Greene about 3 months ago (on reddit).

    “There is much new in string theory. […] But what we lack is any connection to experiment or observation. And that is the only way to know if these ideas are correct.”

    “Not too long ago I was at home listening to a radio program about science when the host described me as a “staunch believer in string theory.” I almost hit the roof. I don’t believe in string theory. I don’t believe in anything that doesn’t have experimental support, observational evidence. So, similar to your description I would say this:

    “I think that string theory represents the most compelling approach we currently have for melding quantum theory with Einstein’s general relativity. And that’s a big deal. The failure to unite these two theories has been a large gap in our understanding for decades. But the union of which I speak is on paper. We don’t yet know if the theory is correct. And indeed, there are other approaches (eg loop quantum gravity).

    “I don’t work on those other approaches because they seem less promising to me. But nature is what decides what’s right and what’s wrong. And as of today, the jury is still out….”.

    Like

  16. Hi Massimo,

    Check out Smolin’s Trouble with Physics, it includes a sociological analysis of funding and hiring in fundamental physics.

    You are right that there is an issue here, but it is also the case that fashions change. The people deciding on academic appointments and grant awards are actually pretty sensible in the end.

    Like

  17. Although I’m very sympathetic with the author’s main point, I don’t think he has a very adequate discussion of what constitutes “evidence.” All me to play Devil’s advocate. In what sense, exactly, does our current empirical data not serve as (perhaps inconclusive) “evidence” for these metaphysical theories? It’s not as if these superstring theorists are making stuff up COMPLETELY out of whole-cloth – they’re trying to find the most plausible explanation of our currently available empirical data.

    Now, I suspect Jim (if I may) might say something like this: “Ah, but our currently available empirical data is compatible with a number of different theories. Perhaps M-theory is a reasonable explanation for our currently available data, but there might be other possible explanations. Until we have more evidence, we have no reason to think that M-theory is a better explanation of the data than some alternative, empirically-equivalent theory.”

    But the obvious response here is to point out that theories are ALWAYS underdetermined by empirical data. Take, for example, what Jim writes about the Higgs boson: “There could in principle have been other possible explanations for the masses of the W and Z particles….In 2012, establishing a correspondence between the empirical data produced at CERN and theoretical predictions for the behavior 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.” True, the realization of a prediction gives us grounds to believe that the Higgs field really exists, but you can *still* say “there could in principle have been other possible explanations” for the new experimental results. There are *always* infinitely many incompatible theories that have the same empirical predictions. Even when Einstein’s empirical predictions for relativity were realized, you could say “there could in principle have been other possible explanations.”

    Of course, one might reply that some explanations are more reasonable or plausible than others, even if they have empirically equivalent predictions. But now you’re playing right into the hands of the string theorists. They would probably argue that, although there could “in principle” be alternative explanations for our current empirical data, they are giving the most plausible or reasonable explanation. So they would say string theory is supported by “evidence” in the same way that the existence of the Higgs field is supported by “evidence.” The main difference between the two is just the *order of discovery* – in one case, the theory came first and predicted new empirical evidence, whereas in the other case the evidence came first and the theory was produced afterwards.

    To be clear, I agree with Jim that a theory’s ability to predict *novel* empirical results (instead of simply being consistent with current empirical results) is extremely important. But it’s not at all obvious to me *why* it should be important (as far as evidence is concerned). Whether an empirical result is discovered before or after the proposal of a theory seems to hang too much on contingent historical and psychological factors.

    Sorry to be so pedantic: this is all a rehash of basic issues in philosophy of science. But that’s kind of my point. As much as I want to agree with the author’s position, it seems to me that he hasn’t even begun to address the oldest and most basic objections.

    Like

  18. Hi labnut,

    I am puzzled. Is this the same Coel who has spent the last few days arguing that we don’t need “some good empirical evidence and predictive success” for the speculative multiverse hypothesis?

    No, I did not argue that. Indeed I argued that we do need empirical evidence and predictive success for the multiverse hypothesis. What I have said is that that empirical evidence can be indirect and that if a theory is validated by some of its aspects being both falsifiable and empirically verified then it isn’t necessarily a problem if other aspects are not falsifiable. There is no difference at all between my stance in that article and my comments on this post.

    Exactly the same applies to string theory. The problem with string theory is that currently no aspects of it have been empirically verified or shown to have had predictive success, and thus is it much more speculative than the eternal-inflation-multiverse theory, which has done a very good job of predicting CMB fluctuations (whatever one thinks of BICEP2 and the dark energy prediction).

    [Can I also suggest that we try not to derail this comment thread. It seems to me from our past interactions that I say things multiple times and you read me as saying something different multiple times.]

    Like

  19. Labnut,

    He hasn’t changed his position at all. He is being perfectly consistent, but you have been misunderstanding him.

    Coel never once said that we have grounds for accepting eternal inflation as true.

    He made two claims which are consistent with what he has said here.

    1) We should accept eternal inflation as a scientific hypothesis, not metaphysics or unscientific speculation.

    2) We do in fact have some rather indirect, limited evidence for inflation.

    There is no conflict between these claims and the agreement here that we need to be clear that these ideas remain unconfirmed and that further evidence is required before they can be accepted as true.

    Like

  20. Coel, I’ve served on NSF panels (different field though), and while it is certainly the case that I’ve largely seen good faith efforts by smart people to fund the best science, it was also pretty obvious to me how fashions, sociological factors, and even individual psychologies clearly affected the outcome.

    Like

  21. DM,
    This was a very well written article which makes a very good case for the problems besetting theoretical physics. It’s hard to disagree with it too much, especially with the view that theoretical physicists have an obligation to be clear on how well-confirmed their ideas are.

    You put it well.
    To this I would add that speculation is the normal precursor to good science. It is the necessary exercise of the mind wherein we explore conceptual possibilities. There is nothing disreputable about informed speculation that is grounded on good science and we should not be afraid to admit to speculation.

    But, and this is the big but, we must always be careful to label our speculation as being just that, speculation. This whole debate is about a labelling problem. We must be careful to apply the correct label.

    Like

  22. Hi Massimo,
    I agree (seen it myself also), but the “fashions, sociological factors, and even individual psychologies” tend to change over time rather than be permanent fixtures, so in the end the problem is got round.

    The process is of course far from perfect, but then everyone involved is human.

    Like

  23. Of course. Does that mean we aren’t justified in level critiques to improve it? (I know, that’s not what you are saying, but you are coming close, in my reading.)

    Like

  24. Isn’t the real crises in science that we continue to collect bunches of data from observations (e.g., associated with hypothetical things like so-called dark matter, dark energy, …) and for which there is no single, coherent theory, no theory at all, or insufficient, half-baked theories to account for them?

    Like

  25. I agree that the critiques do a valuable service (similar to, for example, the role of MOND in comparing to dark matter models).

    What I would disagree with is the suggestion in the article that “There is simply no way back to empirical reality from here”. If string theory continues to be unproductive then fashions will change and physics will move to other ideas.

    Like

  26. I think the author meant “no way back to reality” from this way of doing science. Remember that a number of string theorists are actually advocating radically rethinking how we do science and what we consider “evidence.”

    Like

  27. Hi C Lqrvy,

    a theory’s ability to predict *novel* empirical results (instead of simply being consistent with current empirical results) is extremely important. But it’s not at all obvious to me *why* it should be important …

    As you say, given any set of facts, one can invent an infinite set of models that are compatible with those facts. Just on probability grounds, however, it is very unlikely that any random one of those models then explains novel facts that it wasn’t designed to explain, since there are vastly more ways of being wrong than right.

    Thus if such a model does explain a fact that it wasn’t designed to explain, then that is a strong indication that the model resembles the truth, since only the truth will explain any and all novel facts.

    Colloquially summed up as “Oh, what a tangled web we weave: When first we practise to deceive!”.

    Like

  28. Mr. Baggott, thank you and Massimo for this contribution.to Sciencia Salon. I’ve read a couple of your books and appreciate the clarity of your style. It is reflected again in this well-written and well-organized piece. I want to focus on a few sentences in the section entitled “In the end”:

    “I believe that contemporary theoretical physics has lost its way. It has retreated into its own small, self-referential world. . . . . We’re in crisis, and we need a time-out.”

    In your opinion, what practical measures might the community of theoretical physicists take to refind its way since it seems that many do not agree with your conclusion? Or is Massimo’s statement “free faculty lines, postdoc and graduate student positions, as well as research grants, to explore other options” a necessary first step?

    Like

  29. I fully agree with Jim Baggott as he says that many of contemporary theoretical physics has lost its way. But no matter how they have retreated into its own small, self-referential world, the fact is that they have a deep impact on millions of people that are interested in physics. These people can’t understand the difference between physics and metaphysics, so they are helpless. The tendency of some physicist to raise their thesis as religious issues seems a psychological trait which is deep rooted in the subconscious mind. I don’t know if the weak version of the anthropic cosmological principle is a question of faith and allows a non-religious study, but seems obvious that the emotional aspects of some scientific thesis and ideas sprout from a sort of religious or cuasi-religious landscape. At the end scientists are ordinary people with different beliefs and disbeliefs and to deal with is not easy, although it would be desirable to set up an epistemological scheme to clarify the question. As the classics said, virtue is in the middle ground

    Like

  30. Coel, you said
    I argued that we do need empirical evidence and predictive success for the multiverse hypothesis.
    I am very happy we agree on that point. There is of course the small matter of absence of empirical evidence and lack of predictive success.

    I note that you said ‘need’ and not ‘have’. I presume this means you agree we do not yet have empirical evidence and predictive success for the multiverse hypothesis? I also note you call it a ‘hypothesis’ which, I suppose, reflects this understanding?

    In that case I need to remind you of your exact words, which seem to contradict this understanding.
    First you said:
    I thus argue that claims about a multiverse are entirely scientific

    Later you went on to explain what being scientific meant:
    we are being scientific if the information content of a model is motivated by and justified by empirical evidence. We are being unscientific if our model has information content that is not justified by empirical evidence.

    In other words, you are arguing that claims about the multiverse are scientific, because it is justified by empirical evidence and does not have information content that is not justified by empirical evidence. When you qualify the claim as being ‘entirely scientific’ you mean it is true in every respect, bar none.

    That is a very strong claim indeed. Do you still stand by it?
    Is it really true that the multiverse hypothesis is
    1) justified by empirical evidence? (justified is a strong claim)
    2) does not have information content that is not justified by empirical evidence? (another strong claim)

    These are strong claims. Do you still stand by them?
    Please do not use the derailing argument. This is directly related to Jim Baggot’s post because it nicely illustrates the issues that he talks about.

    Finally I want to reiterate an earlier point. What this debate is really about is choosing accurate and honest labels that truly reflect the state of the art. We must not inflate our claims or misrepresent them by using inaccurate labels.

    Like

  31. Hi labnut,

    I presume this means you agree we do not yet have empirical evidence and predictive success for the multiverse hypothesis?

    No. As I explained at length in my article and the comments to it, we have *some* empirical evidence and predictive success for the multiverse hypothesis, though well short of sufficient to consider it secure.

    I also note you call it a ‘hypothesis’ which, I suppose, reflects this understanding?

    You suppose wrongly; physicists generally don’t read much into different words like “hypothesis”, “theory”, etc.

    When you qualify the claim as being ‘entirely scientific’ you mean it is true in every respect, bar none.

    You really do come up with the most astonishing interpretations of what I write! By “entirely scientific” I mean within the province of science at the cutting edge and the sort of thing that science should be investigating. At the cutting edge, many of the models and ideas of science turn out to be wrong.

    [Can I suggest that if you want to analyse what *I* have said, rather than what Jim Baggott wrote, then out of respect to him it might be more appropriate to do so on the comments to my article?]

    Like

  32. DM,
    see my reply to Coel above.
    I want to make a procedural point, in this instance. Both you and Coel frequently make the accusation ‘you don’t understand’.
    1) That is a form of put-down. It is out of place here and many forums forbid that kind of accusation.
    2) That is often the retort of those who are stuck for answers.
    3) Make your point clearly and explicitly. Quote the other person’s words, produce the facts and show where they are wrong.

    Like

  33. labnut does have a point, though really I should have been the one to make it. Please let’s refrain from that sort of writing and focus on the very enjoyable, vigorous intellectual debate this site is attempting to offer (and succeeding, to some extent!).

    Like

  34. Labnut, Massimo,

    I really don’t feel that these comments are fair.

    Both you and Coel frequently make the accusation ‘you don’t understand’.
    1) That is a form of put-down. It is out of place here and many forums forbid that kind of accusation.

    Firstly, I’m not saying that there is something about the world that Labnut doesn’t understand. I’m saying specifically that he was misunderstanding Coel, which is demonstrably true as we can see in the ensuing exchange. Labnut was under the impression that Coel was arguing that we should accept inflationary theory as true, but this was not in fact the case.

    2) That is often the retort of those who are stuck for answers.

    I was very clear to explain exactly the nature of the misunderstanding, and I fell my claim was vindicated.

    3) Make your point clearly and explicitly. Quote the other person’s words, produce the facts and show where they are wrong.

    I was very specific about my point. I claimed that Coel never once said that inflationary theory should be accepted as true. On Coel’s side, there is no citation I can provide to support this point because I can’t prove a negative. On your side, I was replying to a short message which was all about your interpretation of Coel’s position so I didn’t see the point of quoting you.

    I stand by what I said. I accept that perhaps I could have found a more diplomatic way to say it.

    Like

  35. Coel,

    I agree that we shouldn’t misrepresent Brian but your quote themselves further my point. Clearly people view what he is saying differently from what his actual views are, which is at least partially his responsibility, especially as a public speaker.

    He doesn’t make himself clear enough when he presents his ideas and potential of string theory that this is pure speculation at this point. Brian is a really good public speaker and writer so he really sells the ideas and even though he occasionally adds in “We need to test this”, he argues very forcefully that this is likely true.

    Also, I’m pretty sure he refers to string theory as a discovery in his NOVA special, I’m at work right now but I’ll run through it later today to get a direct quote. In the meantime, here is a quote from a Columbia University page about his work on string theory.

    “String theory also says that the fabric of space can rip, an important discovery made by Greene and a colleague a few years ago, and that there are more than three space dimensions — perhaps as many as 10”

    Like

  36. No, the article is certainly not “spot on”. It presents an extremely oversimplified view of a complicated topic and completely mischaracterizes the theories that physicists are studying.

    Contrary to what Baggott has written, string theory is not really analogous to the standard model of particle physics, and one shouldn’t expect it to be testable in the same sense that the standard model is testable. The standard model is a particular phenomenological model that postulates specific properties for the particles observed in nature, whereas string theory is more of a general framework for constructing such models.

    It would be more appropriate to compare string theory to quantum field theory, the general formalism that physicists use to understand elementary particles. Just as the standard model is a particular example of a quantum field theory, one can construct particular models of fundamental physics based on string theory.

    Models of physics based on string theory have been used to explain aspects of the real world with varying degrees of success. Many string-inspired models of particle physics have already been ruled out at the LHC, while certain stringy models of cosmology may help to explain the results of the BICEP2 experiment, assuming those results hold up. Outside of particle physics and cosmology, string theory been used successfully to explain many aspects of strongly coupled quantum field theories, and this has implications for our understanding of nuclear and condensed matter physics.

    As soon as one specifies a particular model, string theory provides plenty of testable predictions, and it has contributed a great deal to our understanding of fundamental physics. But it is meaningless to talk, as Baggott does, about general predictions of string theory. One could just as easily say that quantum field theory is scientifically meaningless since there are no testable predictions common to all quantum field theories.

    Like

  37. Jim Baggott has described the problem with luminous clarity. Massimo has written also about the demarcation problem. It seems to me there are two demarcation problems. One is the boundary between good science and pseudo science. The other is the boundary between good science and speculative science.

    Pseudo science is a carbuncle that grows on good science, infecting society. It is always harmful. Speculative science, by contrast, is a necessary and healthy substrate to good science. But it is a dangerous thing. Unrestrained it becomes an invasive cancer undermining good science. Until now there has been a simple discipline that has kept speculation in check, preventing it from invading good science. That discipline was the insistence on anchoring conclusions in reality, by using empirical verification or falsification. This was the immune system that kept speculative science in check.

    For a variety of reasons that discipline is being loosened. One of the reasons is that science is coming up against the really hard problems. The hard problems are frustrating, they don’t admit to easy, quick or cheap solutions. In a world starved of grants and impatient for more spectacular breakthroughs it becomes tempting to do an end-run around the problem. This is what Sean Carroll is doing when he proposes that the falsifiability criterion should be relaxed.

    Then there is a deeper problem, an ethical problem. The metaphysical anchor of science was a bone deep belief in the value of truth. We were not just pursuing the ‘reality of things-in-themselves’, we were also pursuing an ideal of ‘truth’. This ethical ideal shaped the way we conducted science. It dictated care, honesty, completeness, thoroughness, fairness and transparency. This, as much as empiricism, gave us confidence in science.

    Sadly, the ethical landscape has changed in the last 60 years. Today the dominant ethical ethos is a kind of de facto moral consequentialism. Now the act is no longer wrong, only the consequences are wrong. This subtle change is unmooring science, disconnecting it from the hard principles of empiricism and truth, leaving it adrift in a sea of moral relativity. One of the results of this is the proclamation of wishful thinking as good science.

    Like

  38. An excellent piece, Jim. You quote Steven Weinberg as saying that we have to go with string theory “because it is the only game in town.” The metaphysical assumption here is that there IS a game in town: that is, the supposition is that there is a unified theory of all four fundamental forces. But there may not be any such theory.
    indeed Einstein/s general theory of relativity shows gravity is *not* like any other force – more than that, it’s not even a force, it is an artefact arising from spacetime curvature. It is therefore perfectly possible there is no unified theory of all the forces, because gravity is quite different from the rest. In that case string theory/M theory goes out the window, as do all aspects of the multiverse hypothesis that use the idea of string theory to justify physics being different in different universe domains in a multiverse. Hence the anthropic justification for all this untestable stuff can be called into question as well: it’s missing a link to experimentally established physics. Rather it’s all based in untestable metaphysical assumptions and self-referential theory, as Jim points out.
    Note I’m not referring here to inflation per se, which has proved a very productive and remarkable theory in terms of observational predictions. It’s firstly the *chaotic* version of inflation that is not proven either directly or by being based in known and tested physics, and secondly the further assumption that physics is different in different inflationary bubbles (if they do occur).

    Like

  39. That last comment can be construed as a discovery about what string theory says, rather than a discovery about how nature is.

    Like

  40. Reblogged this on SelfAwarePatterns and commented:
    I reviewed Jim Baggott’s book a while back, saying that I thought he had made many crucial points that I hoped the theoretical physics community would heed. Science, by getting away from evidence, risks entering a kind of Neoplatonic phase. I still say speculation is fine, as long as it’s clearly labeled, but when the speculation starts to be called “discoveries”, there is a problem. In this post at Scientia, Baggott covers many of the same points.

    Like

  41. the supposition is that there is a unified theory of all four fundamental forces. But there may not be any such theory.

    Given that physical theories are models of nature, surely the only way in which there cannot be a unified model of all aspects of nature is if nature itself is not internally consistent, which seems a rather weird concept.

    Like

  42. Coel, I don’t think so. Scientific theories could be empirically adequate models, but not necessarily approximating truth (the anti-realist position). Even if they do approximate truth, there is no guarantee at all that limited epistemic agents like us could in fact manage to produce one unified theory, as opposed to a bunch of more or less disconnected ones. Which, of course, is plainly the case at the moment, unless you can provide us with a quantum theory of economics, or evolutionary biology, or whatever.

    Like

  43. Hi Massimo,

    Scientific theories could be empirically adequate models, but not necessarily approximating truth (the anti-realist position).

    It which suggestion I put on my scientismists hat and declare that the idea of “truth” and “reality” that is not approximated by complete empirical adequacy is rather bizarre, and that anything that is causally disconnected from our empirical world is not “real” (or is at best some meta-reality).

    Even if they do approximate truth, there is no guarantee at all that limited epistemic agents like us could in fact manage to produce one unified theory, as opposed to a bunch of more or less disconnected ones.

    I agree with you there, though humans being unable to find a unified theory is different from no such theory being possible.

    Like

Comments are closed.