Book Club: The Edge of Reason, 2, science for humans

Time to go back to Julian Baggini’s book, The Edge of Reason: A Rational Skeptic in an Irrational World, which I have began discussing last month. While the first chapter was about God and the rationality (or lack thereof) of arguments pro and against, the second one is about science and why it is far less rational that we are led to believe (especially by scientists).

The chapter opens with the observation that science is not an objective “view from nowhere” thing, but a sophisticated, yet fallible, human enterprise, fundamentally dependent on human judgment. As in the case of a poll Baggini cites from 1999: when 90 leading physicists were asked which interpretation of quantum mechanics they thought was best, 4 voted for Copenhagen, 30 for Many Worlds, and 50 said either none of the above or undecided. Clearly, which available model is preferable is a question of subjective judgment, not empirical fact (as the very word “interpretation” strongly suggests…).

The point Julian is making throughout the chapter is simple, yet controversial: “My aim is to show how accepting the role of judgement in science in no way undermines it, but it does require us to rethink how we assume reason works.”

One of his best examples comes again from quantum physics, when he notices that Schrödinger’s and Heisenberg’s competing theories were not only equally compatible with the empirical data, but in fact had been shown to be mathematically equivalent. They were, therefore, both empirical and mathematically underdetermined. Which one you preferred came down to subjective judgments having to do with beauty, simplicity, or even, possibly, whether you personally liked one physicist better than the other.

The general idea that theories are never, by themselves, uniquely determined by the evidence is known today as the Duhem-Quine theses, but Baggini points out that the concept goes back at least to John Stuart Mill, who wrote in A System of Logic that a hypothesis “is not to be received as probably true because it accounts for all the known phenomena, since this is a condition sometimes fulfilled tolerably well by two conflicting hypotheses.” Mill thought that this was commonly accepted by “thinkers of any degree of sobriety” (I love the turn of phrase!).

As Julian observes, scientists are often dismissive of the Duhem-Quine theses, charging that it is difficult enough to come up with one reasonable theory to explain the data, let alone dream up multiple alternatives. But “this misses the point. The value of the underdetermination thesis is not to make us seriously consider all alternatives to the most powerful and tested scientific explanations. Its value is that it makes it clear that even when the evidence appears overwhelmingly to support one theory rather than another, there is always a gap, however small, between what the evidence requires we conclude and what we actually conclude.”

Another factor that conjures in hiding the role of human judgment in science is the fiction that there is such thing as a quasi algorithm-like thing called “the scientific method.” While philosophers of the early part of the 20th century kept searching for it, the consensus nowadays is that it doesn’t exist. Yet scientists themselves help perpetuate the myth, both in references to the phantomatic method in introductory textbooks, and also by creating “the false impression of a regular, orderly method by writing up their findings in ways which gloss over the real messiness of discovery.”

Indeed, as Baggini stresses, if one looks at how science is actually done — rather than described in simplistic idealizations — it is a highly messy activity where “quirks and deviations” from the official picture of rigorous experiments and straightforward deduction are not exceptions, but rather the norm.

There are many documented cases in the history of science when a scientist persisted out of sheer obstinate conviction of being right, an attitude we associate instead with pseudoscience. My favorite example among those cited by Julian is Boyle, who was “persistent in holding to his theory when observation refused to confirm it. On 49 occasions he tested his hypothesis that smooth bodies that stuck together in air would come apart in a vacuum, without success, yet succeeded on the 50th attempt.”

What, then, distinguishes a brilliant scientist like Boyle from a crank? Good judgment of his intuitions, arising from experience as well as brilliance. Not much else, really.

Moreover, scientists — especially physicists — tend to rely on theory more than empirical evidence, even when the empirical evidence appears to contradict the theory. Eddington, the astronomer that confirmed Einstein’s theory of relativity in 1919, famously declared that “It is a good rule not to put overmuch confidence in the observational results that are put forward until they have been confirmed by theory,” a striking reversal of the usual idea of how science works.

The temperament and gut feelings of scientists play a major role early on during discovery and initial verification, notable cases include Einstein’s suspicion of quantum mechanics (“God doesn’t play dice”) and Heisenberg finding Schrödinger’s theory “repulsive.”

Here is one gem from the chapter: “Another of Einstein’s remarks is extremely revealing. He once said, ‘I find the idea quite intolerable that an electron exposed to radiation should choose of its own free will, not only its moment to jump off, but also its direction. In that case, I would rather be a cobbler, or even an employee in a gaming-house, than a physicist.'” This is an expression of a strongly emotionally held aesthetic judgment. Nothing to do with physics or mathematics as ordinarily understood, or the quest for truth, for that matter.

And of course beauty and aesthetics are not, in fact, guarantors of truth: “As George Ellis and Joe Silk point out, ‘Experiments have proved many beautiful and simple theories wrong, from the steady-state theory of cosmology to the SU(5) Grand Unified Theory of particle physics, which aimed to unify the electroweak force and the strong force.'”

Scientists — below the surface, mostly in private or informal exchanges — even disagree on major issues of epistemology and metaphysics. For instance, “Bohr … completely rejected scientific realism. ‘There is no quantum world. There is only an abstract quantum mechanical description,’ he said. ‘It is wrong to think that the task of physics is to find out how nature is. Physics concerns what we can say about nature.'” Needless to say, a number of his colleagues disagreed vehemently, thus unwittingly engaging in philosophical debates about the nature and scope of their discipline.

By the end of the chapter, Baggini concludes: “The success of science should not lead us to believe that it provides the model for all reasoning; rather that the domain of science is one which is especially conducive to the use of reason. … We need a more expansive notion of what it means to be rational, one which includes all the elements that are left out when we focus only on the strictly formal and empirical ones. At the heart of this notion we need to place judgement.” Indeed.

292 thoughts on “Book Club: The Edge of Reason, 2, science for humans

  1. Bunsen Burner


    ‘The last major figure to make that claim was not a philosopher but a physicist – Erwin Schrödinger.’

    Do you remember where you came across this fact?


  2. Haulianlal Guite


    ||Mach, influenced by his reading of Kant said that this made him realise that the concept of the noumenon was superfluous.||

    Yes, that’s how many philosophers tend to treat the noumenon, not realizing that, whether informative or not, it is a transcendental category required by our reasoning of the phenomena. The phenomena cannot just hang or dangle out here by itself; the datum that our senses register must come from somewhere. Or we have to conclude in perhaps a solipsistic manner that the datum comes from us within (which however is not an option either since Hume demolishes any non-transcendental conception of the self).

    ||It has always seemed to me that this is akin to saying that when I enter into what I call a forest I stumble around and never encounter anything but trees and never a forest. This may have been Hume’s point, I don’t remember.||

    Hume’s point about the bundle theory of perception, as I remember it, is that the sensations are not grounded anywhere insofar as our reasoning goes. But we have to believe in the existence of the self, at least as a bundle – not because of reason as such, but passion. Therefore reason is and will always be a slave of the passion.

    The only way to escape this skepticism is the transcendental route. I have not heard of any other argument that comes even close to demolishing Hume.

    ||Even if I were the Solipsist, there are thoughts that come unbidden to my mind which must originate from outside my consciousness and so there would still be an external world of sorts.||

    This is Kant’s transcendental argument, exactly!

    ||You don’t have to presume the existence of a world in order to do empirical science. Schrödinger said there was no physical universe, Einstein said there was and Bohr left it open. Didn’t stop them collaborating on physics.||

    You maybe right here, I’m inclined to agree on this.


    ||I much prefer the later Wittgenstein’s strategy.||

    Again, the way I see Wittgenstein’s private language argument is that it already presupposes the existence of a common language, and the argument works something like a reductio (that is, private language is impossible because all languages are public). This to me is another form of the transcendental argument, and bolsters Kant’s point instead.

    ||I think you’re pretty clearly wrong. You can’t have the relevant conversation — or entertain the relevant doubts — without those things standing fast.||

    Yes, we have to hold those as standing fast, but only for the context of the investigation. In another context we may doubt its existence entirely. Indeed, if we can meaningfully discuss the principles of a chess game and how the game gets played without imputing ontological properties to chesss, maybe the same game can be played of the world.


    ||In contrast, the scientist’s procedure is one of investigation, wanting to find out what actually happened, how the thing works, or in Maxwell’s expression, “What’s the particular go of that.” The motivation is the solution of problems, and usually solving one problem only leads to a train of other related problems, so the quest is endless.||

    I totally agree! Philosophy will never end because science will keep throwing up new philosophical conundrums to add fresh spins to our old perennial problems, and science will never end because there maybe an infinite number of ways to model reality, each with their own philosophical theory-laden assumptions and implications.

    Liked by 1 person

  3. Markk


    To meaningfully talk about anything existing or not existing, it needs to be put in a context. To claim that the world is made of ideas is one thing; to deny it exists entirely cannot be done as there is no context from which to do this.


  4. Haulianlal Guite


    That’s exactly Kant’s point, which I believe is valid. All talk of existence or non-existence already presupposes the existence of something, which lays down the ground for discourse.

    My response to Dan, however, is in the specific context of empirical science. Science doesn’t need to assume any existent entity for its models and equations to work. If it can model all occurences of dreams, there is no reason the same principle does not apply for what we ordinarily call reality. Science is metaphysically neutral in this sense.

    Though, for the transcendental (that is, trans-empirical) reason Kant already made a point about, which you repeat, we know something that he calls the noumenon, exists. Its only that the noumen is irrelevant to science as such. Maybe that’s what Mach meant when he say the noumenon is superfluous.


  5. Robin Herbert

    Hi Bunsen Burner

    “Do you remember where you came across this fact?”

    Yes, you can find it in chapter 4 of Schrödinger’s “Mind and Matter” as well as in the epilogue of “What is Life?”.

    As it happens I have both of these on my phone right now. The second is a little long to quote but the first goes :

    Sherrington says: “Man’s mind is a recent product of our planet’s side.”

    I agree, naturally. If the first word (man’s) were left out, I would not. It would seem queer, not to say ridiculous, to think that the contemplating, conscious mind that alone reflects the becoming of the world should have made its appearance only at some time in the course of this “becoming” should have appeared contingently, associated with the very special biological contraption which, in itself, quite obviously discharges the task of facilitating certain forms of life in maintaining themselves, thus favoring their preservation and propagation: forms of life that were latecomers and have been preceded by many others that maintained themselves without that particular contraption (a brain). Only a small fraction of them (if you count by species) have embarked on ‘getting themselves a brain.’ And before that happened, should it all have been a performance to empty stalls? Nay, may we call a world that nobody contemplates even that? When an archeologist reconstructs a city or a culture long bygone, he is interested in human life in the past, in actions, sensations, thoughts, feelings, in joy and sorrow of humans, displayed there and then. But a world, existing for many millions of years without any mind being aware of it, contemplating it, is it anything at all? Has it existed? For do not let us forget: to say, as we did, that the becoming of the world is reflected in a conscious mind is but a cliché, a phrase, a metaphor that has become familiar to us. The world is given but once. Nothing is reflected. The original and the mirror image are identical. The world extended in space and time is but our representation (Vorstellung). Experience does not give us the slightest clue of its being anything besides that—as Berkeley was well aware.

    Liked by 1 person

  6. Markk


    I personally think just assuming the world exists is a better strategy than trying to prove it via transcendental arguments, in the same way that we must assume we have free will, that the human mind is adequate and so on.

    Liked by 1 person

  7. Haulianlal Guite


    ||I personally think just assuming the world exists is a better strategy than trying to prove it via transcendental arguments, in the same way that we must assume we have free will, that the human mind is adequate and so on.||

    That will be to stop doing philosophy altogether, and to begin doing empirical science, or step outside the academia and just be a man of world affairs. Which one can, of course.

    For the constantly nagged philosopher however, who may not be satisfied even for a moment with this proposal to abandon the principle of sufficient reason altogether, and for whom such questions are like bad dreams and unsettling nightmares, yet reaching the ecstatic highs drugs swoon us with at the same time, such assumptions must continue investigated, with passion, vigour and intellectual courage. And demolish thereby the transcendental deduction itself (in case you hate the argument) and coming up with something else to justify the assumption, or to show why and how this said assumption fails.

    [p.s: I really don’t understand this hatred some people have of transcendental arguments. How does it matter what form an argument takes? When its so clear many transcendental arguments are valid. For example, the fact that you type means there is someone or something that can type. What’s to hate in that?]


  8. Haulianlal Guite

    Markk, this is why I really love the old philosophic adage that “one must learn to live before one can philosophize”, and that “only beings that are neither man nor gods must philosophize!” 😀

    Liked by 1 person

  9. Coel

    Hi marc,

    … what I think they might be saying is that talk of existence and such, even before humans appeared, is necessarily a mind-dependent exercise.

    Agreed, talk of such things would be mind-dependent, but the physical existence would not be.

    Hi Bunsen,

    Are there philosophers who dispute the existence of the physical world before humans arrived to create concepts? How do they deal with an evolutionary perspective where modern concept forming humans did not arrive all at once in some kind of single event?

    I’m going to take a wild stab that, if there are such philosophers, they would reply that the two different issues belong to completely separate “language games”, and that asking how they fit together as you’ve just done is a category error.


  10. Coel


    Now the reason this makes absolutely no sense is because the calculation about the limits of large structures …, is done using inflation!

    Sure, the calculation is done using {one particular variant of inflation} + {a whole lot of other stuff}. If the observed structures are bigger than predicted by that package, then it means one has to change something in that package. It doesn’t mean one has to throw out inflation in its entirety. As I’ve been saying all along, there is plenty of flexibility in the models, they are simply not that well tied down yet.

    There is a continual dialogue between the data and the models, and at the cutting edge of research there are always discrepancies between the two (that’s why it is “research”). One should never over-react to any one issue such as this.

    And, as I’ve also said, ditching the “inflation” element from the above package would likely make the discrepancy worse, since inflation helps in producing very large structures (since making small things very large is exactly what it does).

    Thus, overall, you’re taking a way too simplistic account of how science is actually done, and a way too simplistic account of what it takes to falsify a broadly-defined theory with lots of scope for variants.


  11. Robin Herbert

    Hi HG

    “this is why I really love the old philosophic adage that “one must learn to live before one can philosophize”, and that “only beings that are neither man nor gods must philosophize!”

    Only women then? 🙂


  12. Markk


    I certainly don’t hate transcendental arguments, I think I’m just approaching things from a different direction.

    We must have free will, there must be an external world, we must be capable of knowing it, and our logic must be adequate for this task. Those are non-negotiatiable starting points without which human knowledge collapses in a heap. Therefore any philosophy that undermines them must be rejected.


  13. Robin Herbert

    Hi Coel

    “I’m going to take a wild stab that, if there are such philosophers, they would reply that the two different issues belong to completely separate “language games”, and that asking how they fit together as you’ve just done is a category error.”

    No Coel I am almost certain there are no philosophers who try to blend Wittgenstein and Berkeley.

    Anyway, as I said doubting the existence of the world is rather old hat in philosophy these days.

    Doubting the existence of the world seems more something scientists do these days. Neil DeGrasse Tyson puts it at about 50/50.

    Liked by 2 people

  14. Haulianlal Guite


    ||Sure, the calculation is done using {one particular variant of inflation}||

    I don’t think that’s entirely accurate. Different inflationary models come up with different estimates of the theoretical limits, but that even the apex upper limit estimate is still too low that these 6 structures I mentioned, violated all estimates. Indeed, I used only the great GRB wall example because it is the most violent and dramatic refutation of the CP and therefore, all inflationary models.

    ||{a whole lot of other stuff||

    Of course this is true of all theories, no single theory has only a single stuff. But that’s not relevant. As long as the inflationary models are critical in coming up with those estimates, then it means something is wrong with the models as well, whatever else may have been wrong too.

    ||There is a continual dialogue between the data and the models, and at the cutting edge of research there are always discrepancies between the two (that’s why it is “research”). One should never over-react to any one issue such as this.||

    Now one can surely appreciate better why the peripatetics cautiond against overreacting when the bull called Galileo charged in with only the parsimony argument and the underdetermined principle of relativity to challenge the most entrenched unassailable theory of their time! Though the case of inflation is orders of magnitude of less significance.

    ||And, as I’ve also said, ditching the “inflation” element from the above package would likely make the discrepancy worse, since inflation helps in producing very large structures (since making small things very large is exactly what it does).||

    This is irrelevant. I’m not here to tell scientists what they should ditch and retain. My only point is that they can’t retain inflationary models and claim to revere, preach and accept falsification at the same time.


  15. Haulianlal Guite


    ||Only women then? 🙂||

    If by women we mean the irksome, inconvenient gadflies who keep hovering – unwanted and unsung – to bite the rosy asses of some smug scientists and hypocritical theologians at the same time, and shout out the emperor’s nakedness for what it is! 😀


  16. Coel


    … because it is the most violent and dramatic refutation of the CP and therefore, all inflationary models.

    It doesn’t matter how many times you assert that, it’ll still be wrong. You seem to think that if you assert enough times that a refutation of the cosmological principle would refute all inflationary models, then it’ll magically come true. It won’t, and it isn’t true, it is wrong.

    As I’ve explained several times, inflationary models are entirely ok with violations of the CP (as applied to the scale of the observable universe). Indeed, inflationary models are about phase transitions in the early universe, and phase transitions predict topological defects, which could well be the source of large-scale structure that cannot be explained by CDM-clumping. (E.g. )

    Thus, as I keep trying to tell you, inflation helps explain large-scale structures! Inflation helps explain violations of the CP! To your wrongness you are now adding incorrigibility.


  17. Haulianlal Guite


    ||We must have free will …||

    Not necessarily (though I believe in freewill, and I accept Kant’s argument about its necessity for morality, that ought implies can). Have you heard that story about the determinist stoic, Zeno, and his slave? This shows how the freewill/determinism debate may under certain interpetations, have no unique empirical significance (apart from assigning moral culpability).

    ||there must be an external world …||

    Again, for scientific purposes it may not be needed. The only reason I can think of why it should be needed for any purpose is the religious. Not sure if this will be to your liking though!

    ||we must be capable of knowing it||

    Again, no. We must be capable of modelling, yes, because we actually do that. But does our modelling entail its truth? That’s the realist/instrumentalist debate.

    ||any philosophy that undermines them must be rejected.||

    I’d say the reason any science must be rejected should be because it is not useful enough for some instrumental purpose, and any philosophy must be rejected should be only and solely because it is false in some important sense. So long as the possibility exists that we are just simulations, brain in vats, maya and so on, let’s continue investigating as part of the bigger philosophic endeavour to change the world, and our ways of seeing it (indeed, my greatest area of interest is political philosophy, and even my interest in philosophy of science and metaphysics such as the ones we discuss, is a consequence of my interest in political philosophy and ethics, where I believe Kant has been right in the essentials all along).


  18. Haulianlal Guite


    ||Thus, as I keep trying to tell you, inflation helps explain large-scale structures! Inflation helps explain violations of the CP! To your wrongness you are now adding incorrigibility.||

    This is just unbelievable. I have been saying so many times right from the beginning of this debate that I gave the example of the great GRB wall precisely because CP is violated, and because CP estimates are all derived out of inflationary models, these models are wrong too!

    How many times do I have to repeat, “yes I get it, I get it long before that inflation models are postulated to explain the rise of large-scale structures! Its just that the size of these 6 structures are far too large they violate the upper-bound estimates!”

    And for the last time I shall repeat this (after which you can go on believing whatever you want): when inflatin was first theorized, it was done so to explain the overall homogenity of the universe and the relative anistotrophy of the universe (why there are large-scale structures). And till the early 1990s, all was fine. All the discoveries faithfully stayed within the upper bound theoretical estimates that the models proposed for how large a structure can be.

    Then, one by one, observational cosmologists began discovering new structures that, erm, first lie on the boundaries of what is theoretically possible. Till came 2013, when this great GRB wall could no longer be explained away as a boundary case. Its a flat-out refutation of the highest upper bound estimate that inflationary models predict for how large a structure can be. Now all these estimates are based on various models, and to the extent that these models predict these estimates, because we have a structure which lies outside this, the models (all inflationary models) stand falsified.

    Now you can say inflation is far too precious a theory to be destroyed by data, and retain it for other reasons. No argument with that. What one can’t say is to claim adherence to falsification and data faithfulness in doing so.

    Now if you still refuse to accept the argument, nothing much I can add to that. Peace!


  19. brodix

    To take the physicists view, all there is is this energy either flashing about, or vibrating at the speed of light, slowing with interference. So all the solid stuff is relationships between energies, even kicking rocks. Beyond that, it is effect; Time, temperature, people, planets, etc.


  20. Markk

    I had a laugh at Robin’s comment above about mixing Wittgenstein and Berkeley.

    As for my own philosophy? A mix of Aquinas and Karl Marx,

    Five ways to prove that God exists and is the opium of the people.

    Liked by 1 person

  21. Coel


    And for the last time I shall repeat this …


    Now if you still refuse to accept the argument, …

    I don’t. Inflationary models do not depend on the cosmological principle being true on the scale of the observable universe.

    … nothing much I can add to that.

    Oh good!


  22. Massimo Post author


    “The only way to escape Berkeley’s idealism and Hume’s skepticism, imo, is to take the transcendental route.”

    “I much prefer the later Wittgenstein’s strategy.”

    Good choice. Me? I go with Hume: when you are concerned with such philosophical matters, take a walk, have dinner with friends, play some backgammon. You’ll find that the problems have simply gone away…

    Liked by 4 people

Comments are closed.