Why Scientific Realism Wins

In my last post I quoted Stephen Hawking’s defense of Positivism. He even goes so far as to suggest that there is no all encompassing view of reality but instead only “a family of interconnected theories, each describing its own version of reality…” (p. 70)

But accepting Positivism as the true nature of reality has consequences.

A famous real-world example of different pictures of reality is the contrast between Ptolemy’s Earth-centered model of the cosmos and Copernicus’s sun-centered model. Although it is not uncommon for people to say that Copernicus proved Ptolemy wrong, that is not true. …the real advantage of the Copernican system is that the equations of motion are much simpler in the frame of reference in which the sun is at rest. (p. 71)

Boy, are you ready to accept this? That the earth is no more revolving around the sun then the sun is revolving around the earth and that the only real reason we believe the earth revolves around the sun is because the math is simpler that way?I confess, there does seem to be a sense in which this is true. Einstein’s General Relativity does not allow for a favored coordinate system. Therefore, it is not strictly true that the sun does not revolve around the earth according to General Relativity because it would always be possible to formulate an earth centric theory that gives the exact same predictions.

This is where the famous – and mostly misunderstood – Occam’s Razor comes in. Occam’s Razor does not claim that the simpler theory is the right one. If I had a dime for every time someone tries to claim that…[1]

Occam’s Razor is at it’s best when it makes a much more modest claim. This claim is that given two theories that make all the same predictions, you should choose to use the one that is mathematically simpler to calculate because it’s the “best” by tautological definition. [2]

Of course, as Hawking points out, “simplicity is a matter of taste.”

Therefore, what is wrong with creating a Positivist view of reality, especially if it might turns out to be correct? (Though presumably we’ll never know that for sure.)

Roger Penrose vs. Stephen Hawking: Schrödinger’s Cat

But it seems that Positivism – if fully embraced – has some negative consequences. As I mentioned in this previous post, Positivism seems to be most useful when you just want to make predictions and aren’t so worried about what those models really mean about reality. The model is reality, if you will. Nothing more.

In an extended debate/lectures between Stephen Hawking and Roger Penrose Hawking said the following:

These lectures have shown very clearly the difference between Roger and me. He’s a Platonist and I’m a positivist. He’s worried that Schrödinger’s cat is in a quantum state, where it is half alive and half dead. He feels that can’t correspond to reality. But that doesn’t bother me. I don’t demand that a theory correspond to reality because I don’t know what it is. Reality is not a quality you can test with litmus paper. All I’m concerned with is that the theory should predict the results of measurements. (Find online here.)

Schrödinger’s cat is a classic physics thought experiment where you put a cat in a box and arrange for a vile of poison to break inside the box depending on whether or not a quantum particle is detected. Since quantum particles are in “superpositions” (i.e. are in all locations at once in some sense) until “observed” then quantum theory predicts that the cat will actually be both dead and alive simultaneously until “observed” (i.e. we open the box) and then it’s “wave function” collapses and it will now be either dead or alive.

Contrary to popular belief, Schrödinger proposed this thought experiment to demonstrate that there is something wrong with quantum theory. But today physics take seriously the idea that until observed the cat is in a superposition of being both dead and alive simultaneously.

So in a sense, Hawking is correct. In fact Quantum theory makes the right prediction about Schrödinger’s cat: it predicts we’ll either see it as dead or alive, but not both.

On the other hand, quantum theory seems to suggests rather disturbing possibility about reality with no real answers to what they mean. Does this, for example, mean that that the cat is actually both dead and alive and there are two conscious versions of ourselves, one looking at a dead cat and one looking at a living cat? This is the infamous ‘many worlds’ interpretation of quantum physics. Or does it mean that consciousness creates reality? If so, why? (And if so, then are photographs conscious, since they count as observations too?)

Penrose’s point is that the theory is incomplete either way. There is some sort of significant explanation gap that needs to be filled. Hawking isn’t concerned about the explanation gap, so he feels no need to fill it.

Hawking’s Inconsistency

It seems to me that Hawking’s Positivist view point is inconsistent at best. Consider the fact that he’s done considerable research into black holes. He has pointed out that according to our best theories black holes should give of radiation? But how is this possible if the black hole, by definition, can’t allow anything to escape from it?

Hawking looked more carefully at what quantum theory tells us about the vacuum of space. It predicts that space is actually full of positive and negative particles constantly appearing and then canceling each other out. This should, according to Hawking’s theories, happen right near the event horizon (i.e. border) of a black hole. The positive particles would thus jump away from the black hole and the negative ones would fall in. The end result would be that the black hole would give off radiation and in fact would shrink or evaporate over time because of all the negative particles falling into it.

This theory disturbed many physicists because it had been believed that quantum information could not be destroyed. But if thing falls into a black hole and then the black hole evaporates (over a seriously large period of time) that would imply that quantum information could be lost.

Now the first thing we should recognize is that we’ve never made any observations of black hole radiation before. It is hard to believe a true full throated Positivist would even bother to make up theories about what black hole radiation is like. What is the point when, by Positivist definition, the model is only “connect[ing] the elements of the model to observations”? What observations are we connecting it with regarding Black Hole radiation?

I would submit that this proves that Hawking isn’t really a full blooded Positivist after all. He is forced to resort to scientific realism to make scientific progress.

And, Of Course, The Model Does Matter

Secondly, there is an alternative to Hawking’s theory that navigates the issues better, at least in some cases. Under the alternative theory the radiation of a black hole is actually caused by p-branes in the black hole forming waves that peak above the black hole’s event horizon. These peaks then become particles and escape. In Hawking’s book, the Universe in a nutshell, he points out that:

The mathematical model of black holes as made of p-branes gives results similar to the virtual-particle pair picture described earlier. Thus from a positivist viewpoint, it is an equally good model, at least for certain classes of black hole. For these classes, the p-brane model predicts exactly the same rate of emission that the virtual-particle pair model predicts. However, there is one important difference: in the p-brane model, information about what falls into the black hole will be stored in the wave function for the waves of the p-branes. (p. 127)

In plain English what Hawking just said is that there are two equally good ways of understanding black hole radiation, but one preserves quantum information and one does not. Since the preservation of quantum information is still (at least according to our current best theories) a hard fast law of physics, this suggests that the p-brane model is superior to the virtual-particle pair model because it fits better with the rest of the laws of physics. (i.e. doesn’t violate them.)

Again, it seems to me that Positivism has come up short. Unless we take a Scientific Realists view, we can’t actually assess the obvious: that these two theories are not equal.

Scientific Realism is Superior – Even If Wrong

Which brings me to my key point. Even from a Popperian / Scientific Realist view of the reality, we can’t ever know for certain if there really is a single all encompassing view of reality or not. For nothing within Popperian epistemology allows us to “prove” a theory to be true.

But Popperian David Deutsch does make a compelling argument, that for me, destroys Hawking’s arguments lock-stock-and-barrel. As Deutsch points out, a Positivist view of reality has no explanatory power. Deutsch, commenting on the deep relationship between Scientific Realism, comprehensibility, and algorithmic compression, argues:

If, for instance, we want to understand why the world seems comprehensible, the explanation might be that the world is comprehensible. such an explanation can, and in fact does, fit in with other explanations in other fields. But the theory that the world is half-comprehensible explains nothing and could not possibly fit in with explanations in other fields unless they explained it. It simply restates the problem and introduces an unexplained constant, one-half. (The Fabric of Reality, p. 351)

Try to wrap your head around that statement for a moment. Then do this: try to refute it. The logic of it is impeccable. Thus we are forced to eject Positivism regardless of whether or not it is true:

To understand our best theories, we must take them seriously as explanations of reality, and not regard them as mere summaries of existing observations. (The Fabric of Reality, p. 350)

It is not that we know the Kuhn / Hawking / Positivist view of reality to be false. That can never be proven one way or another. We eject it because, even if they are right, it explains nothing. Even Hawking is forced into a Scientific Realists view when he wishes to make scientific progress. This is because he still has to come up with theories – conjectures – and then take them seriously and honestly believe that they represent reality. This will lead to criticism, either from ourselves or (more likely) from others. This is the only way we can make scientific progress.

In other words, if Kuhn is right then Kuhn’s theory predicts that we must start with the assumption that Kuhn is wrong.

For me, this is sufficient refutation of Kuhn’s final conclusions of a Positivist worldview. I fully accept that the rest of Kuhn’s theory is basically correct. I do not see most of his theory as at odds with Popper’s. Only in Kuhn’s final conclusion, that Positivism might be a reality, do I object. But even then I primarily object on the grounds that, if Positivism is true, this it is what we might call a ‘useless truth.’ Because the untruth is superior to the truth, we should merely go with the untruth even if it’s wrong.

In other words, I believe in Scientific Realism because it just makes sense to. I have faith in the power of this belief regardless of whether or not it is true. (Though obviously I believe it is in fact the truth.)

Therefore, Scientific Realism is the undisputed winner, at least in the Positivist vs. Scientific Realism debate.


[1] Real life example: Joseph Smith is a fraud because according to Occam’s Razor, the simplest theory is the correct one and it’s easier to believe he made it all up then real ancient American angels came to him.


Unfortunately even famous scientists get this wrong. Carl Sagan popularized Occam’s Razor in his book (and movie) Contact. Unfortunately he didn’t actually understand it, so he gave it the interpretation above: it’s easier to believe we humans made God up than that God actually exists, therefore there is no God as per Occam’s Razor. Granted, Sagan goes on to then refute this view of Occam’s Razor but without ever pointing out that actually it’s not Occam’s Razor in the first place.

[2] There is more to Occam’s Razor that what I am saying here because any extraneous parts to a theory should then have to be explained. Yet that will not be possible before the two theories being compared start to diverge in their predictions. Therefore the simplest theory is also the ‘best’ in the sense that it has less to explain. But this is a post for another time and isn’t relevant to the current subject.

12 thoughts on “Why Scientific Realism Wins

  1. “Boy, are you ready to accept this? That the earth is no more revolving around the sun then the sun is revolving around the earth and that the only real reason we believe the earth revolves around the sun is because the math is simpler that way?I”

    If I had a dime for every time someone tried to interpret positivism that way… 😉

    Of course positivists think there is a difference between those two models! One model is very elegant and gives us highly accurate predictive results while the other does not.

    But the realist wants to ask further, what is the explanation for one model working better than the other? The answer, they claim, is reality-as-it-really-is.

    The positivist does not necessarily reject this answer as much as it does the question. Since there is no getting outside of all models to observe what causes some models to fit better than others, any answer must be nothing but a metaphysical posit without any empirical backing of any kind whatsoever. The idea of reality-as-it-really-is is a cog which does no work in the explanatory mechanism in that it leads to no predictions of any kind. Accordingly, the positivist sees no reason to “multiply entities” beyond what is necessary and simply drops the whole concept altogether.

    I am, however, inclined to agree with Bruce in suggesting that realism is better, even if it is false.

    Dewey and Rorty construe reality-as-it-really-is as the modernists inability to fully let go of God. They would allow that theism might (might!) be better than atheism even if the former is false and the latter true. This is exactly the situation for scientific realists and their religion of science, according to them. Reality-as-it-really-is does absolutely no work, even if our beliefs in it do motivate us, so to speak.

    It is my position that the craft of science necessarily involves rules outside of those having to do with predictive accuracy and explanation. Realism might not do any explanatory work, but the craft of science involves so much more than the disinterested accumulations of explanations isolated from all other forms of life. Realism might not play a role in scientific explanation, but it sure plays a role in scientists.

  2. The suggestion that realism is essentially a modernized version of theism raises an important question for those who are also theists in the traditional sense:

    If reality-as-it-really-is is a source for normativity which provides no empirical data whatsoever, then it essentially functions as a god which is not the god of the bible. In other words, realism brings the theist dangerously close to polytheism.

  3. Jeff G,

    A few points.

    First, I don’t disagree with you on your retort to my response to Hawkings. But I challenge you to back read your retort into what Hawkings actually said. I don’t see it there. Perhaps Hawkings makes a weak Positivist? Or maybe don’t really understand it entirely?

    Further, you are buying into the idea that science is about observation and prediction — both Positivism creeds. I buy neither. You already read my posts on both of those, so I’ve already explained why I think those are incorrect views of science.

    Again, perhaps there is some legitimate sense in which the term “Realist” can apply to both “Posivists” (or some of them) and also Popperians. But the two forms of “realism” are mutually exclusive and both can’t be equally right.

    And, in fact, both are NOT equally right. Based on an explanation to explanation comparision, we can drop Positivism in favor of the Popperian view of Realism for the reasons outlined above.

    But, of course, I don’t personally really drop it entirely because I still find it useful at times. Sometimes just saying “do the calculations and shut up about what it means” has appeal to me.

    I think you might want to take a harder look at Poppers form of Realism. It really does serve as a refutation of Positivism even in its strongest forms. And it removes the whole idea that science is specifically about observation or prediction and replaces it with the idea that science is about finding explanations.

    Thus “reality as it really is” is cooked into it (it’s not a true explanation if it’s only purpose is to make predictions) and that is why the positivists were off track. (Because they missed the point by focusing only on predictions.)

    Popper epistemology assumes there is a “real reality” for the sake of finding explanations. But it makes no assumption we could ever actually find it out. And if we did, we’d never know it for sure. But the existence of a “real reality” must be assumed for our explanations to be taken seriously.

    You make the argument that this can’t be rigth based on the idea that science is about prediction. But since Popper rejected that idea (and I believe he was entirely right to do so) this is now a questionable argument and can’t be used to refute Popperian epistemology without being circular.

    I am not sure what to say about Realism as polytheism. I still don’t really see the connection.

    Yes, I think there is Something-Like-God about “realism as it really is” at times. But for all I know, that might be because that Something-Like-God is in fact God. So I don’t see why it would have to be a move to polytheism at all. In fact, I think people that believe in God must of necessity believe in some form of “Realism as it really is.”

  4. “One model is very elegant and gives us highly accurate predictive results while the other does not.”

    Just noticed this. Actually, you have to assume that you have two models and they both give exactly equivalent results. If you aren’t assuming that, there is the easy way to choose: Go with the one that matches reality better.

    But if the two are identical in results, then the idea of picking it “because it’s more elegant” will ultimately have some level of subjectiveness to it. And I think this is Hawkings point.

    The problem is that Hawkings ignores this regularly. And has to. So as much as he argues that it’s a workable approach, it doesn’t seem to be. At some level, two identical models have a bad tendency to turn out to have non-identical results. As with the case of how black hole energy works — yes both has identical mathematical results. But they did not have identical repurcusions with other theories. So we can tentatively choose between them in a theory to theory comparision (and without ANY observations) by selecting the one that doesn’t demolish our other theories. (i.e. to reject Hawking’s own theory in favor of the better one.)

    This is Popper’s point. We can do theory to theory comparisons and based on a wide variety of selection criteria, pick the best. It is not soley a matter of observation or prediction — neither of which applies to this example.

  5. Yeah, I can’t help but think that his version of positivism is different from the forms I’ve encountered (logical positivism, comtian positivism, positivistic economics, etc.) which simply focus on empirical observation being the sole source of information about the world.

    I call myself a neo-pragmatist of sorts, so I don’t limit science to explanation or prediction. That said, I still don’t see anything that could ever, in principle, set realism apart from non-realism. All observation and behavior can be built into the non-realist position without further invoking reality-as-it-really-is.

    As a way of appreciating the epiphenomenal character of realism we can cash it out in terms of Bishop Berkeley’s idealist version of empiricism. He held that there was no such thing a matter, only ideas which continued to be exist and maintain their systematic relations to one another thanks to God. All we have to do is replace “God” with reality-as-it-really-is and see how empty and metaphysical the realist thesis is.

    Furthermore, since Mormons do not equate God with reality or the foundation of all being, etc. a Mormon realist is thus left with two gods: one, the god of Mormonism and the other the god of scientific realism (aka a materialist version of Bishop Berkeley). This polytheism, I suggest, is the very embodiment of the science/religion debate. The scientific realist, who tends toward scientism, is morally compelled to believe whatever reality-as-it-really-is tells him and the Mormon is compelled to believe whatever God tells him through the prophets. The Mormon realist now must decide which god he will give priority to.

  6. “…simply focus on empirical observation being the sole source of information about the world”

    Okay, let’s look at this.

    First, this is Hawking’s point. So he at least go that much right.

    Second, you keep saying that a non-realist position can be equivalent. But that is not so.

    Yes, anything a realist gets out of “reality as it really is” can eventually be incorporated into a non-realist position. If that is all you are saying, I agree.

    But the positivism and non-realist has not the slightest motiviation to take our explanations (science or religious) seriously enough to follow it through to it’s logical conclusions. THAT IS HOW IT”S INFERIOR! Not in terms of whether or not it can or can’t be incorporated into a non-realist or even positivist position. The inferiority of positivism and non-realism is that they are inert and useless compared to realism. And they only ever produce results by first assuming realism temporarily.

    Try to address that. For that seems to me to be their primary weakness.

    Let’s compare this to the Berkely idea of God being all that is required. Granted, this must be true for the classic (Catholic Protestant) view of God. Which leaves an uncomfortable question: Why did God even bother with the laws of physics at all? They have no use now.

    From a Poppperian standpoint, we can dispensed with Berkely’s position because it’s as inert as Positivism or non-realism. It tells us nothing. It’s not an explanation, it’s an explanation spoiler. Nothing more. Physics just exists as a bunch of laws because God thought it would be a good joke to make us think we lived in a lawful and comprehensible reality.

    Now it might indeed be true! Maybe God just on a whim happened to decide “I’m going to have laws of physics for no particular reason.” But if you want to use those arbitrary laws (which God might change tomorrow for all we know) you *have to assume they are real* at least long enough to follow them through to logical predictions.

    So it literally doesn’t matter if Berkely is right or not. We are not dispensing with Berkely based on some measure of the truth of falseness of his statements. We’re dispensing with him based on the fact that his statement is utterly useless even if true.

    This is how epistemology actually works in real life. We can talk about things like Positivism or how a realist position can be incorporated into a non-realist position till we turn blue. And it doesn’ tmatter one whit because they are still useless as explanations. (Well, non-realism is useless. I’ll admit Positivism has some uses if you are just trying to apply science and not expand it.)

    That’s it. That’s all there is to dispensing with these other views. They count for nothing even if true, so we can stop talking about them and move on with the assumption that they are wrong. (Note how this works with the God example. Physics aren’t the real underlying reality, but we are still right to assume they are, at least for now.)

    Does this mean there is a reality as there really is out there? Don’t know, don’t care. I only know that to take an explanation seriously and follow it to it’s logical conclusions is the sole and only way to find the sorts of logical problems that require making up new and better theories. Therefore “reality as it really is” is cooked in, but not really required. It’s “assumed” but not “required.” I have no interest in arguing whether or not “reality as it really is” exists because if it doesn’t, I still have to believe that it does.

    If it turns out at any point that physics is a whim of God, then science counts for nothing. And no explanations do. And no comprehension does. All is incomprehensible. Even what we thought we comprehended was actually incomprehensible because it relied on something incomprehensible. And God will prove the utter alien and incomprehensible nature of reality by changing the laws of physics tomorrow. So who cares about knowledge at all?

    I reject that God by the way.

    And that I do reject that God means I both believe in reality as it really is (or at least I assume it for all intents and purposes) while also believeing in the Mormon God. And, in fact, belief in the Mormon God requires me to on faith believe in “reality as it really is” precisely because the Mormon God is lawful and not whimsy like the classic God. So this is why I reject your view of a choice between science and the Mormon God.

  7. Jeff,

    I write too much in comments. Let me summarize the main points as I see them:

    1. This post is a positivism example of how we can and do choose between explanations (theories) based on criteria other than observation and prediction. So I do not see how this isn’t an effective theory to theory refutation of Positivism in favor of “Realism” (i.e. Popper’s realism.)

    2. Berkely’s God, even if true, can’t be explored or explained or comprehended. So there is no point in talking about this God. And even if true, we have to still believe Physics laws are true and explore and explain them as if they are true. So we have also just refuted Berkely’s God using the same means we refuted positivism.

    Refutation isn’t really a refutation that we’ve somehow disproven them. It’s a refutation in this case based on the fact that Realism is superior to them even if true.

    Last arguement: Berkely’s God isn’t a case against Realism. It’s a case against Classical Theism. The Mormon belief in a lawful God suggests that Berkely’s argument is problematic. Even if you are a Mormon that believes God did create the laws of physics, you’d have to still believe God has a reason (explanation) for why because Mormons believe in a lawful God. Therefore physics, while contingent in some sense, is still also based on some underlying set of reasons that are laws. So there are Eternal Laws that are real and so realism is correct.

    So my feeling is that I should reject your argument that realism is at odds with Mormonism and instead suggest that Mormonism demands realism.

  8. “Go with the one that matches reality better.”

    This is where I follow Rorty in getting off the realist bandwagon. What, exactly, is being matched up here? You make it sound like science is simply a game of Guess Who? played against nature. For starters, we can never know what reality-as-it-really-is looks like independent of when we are looking at it. Secondly, our descriptions of the world aren’t pictures or some other things which “matches up” with that unobservable reality-as-it-really-is.

    Let me now back up a bit a describe a version of non-realism which doesn’t sound as transparently absurd and unmotivating:

    Let us view reality-as-it-really-is as a kind of dough and our concepts are the shapes with which we cut up this dough. Aristotle cut up the dough in terms of teleological ends which all objects pursue. Galileo cut it up in terms of matter in non-teleological motion. To be sure, Galileo’s “cutters” have served us much better than Aristotle’s did, but this is no reason to think that the latter’s were “false” in any way. Aristotle’s science was never falsified at all, and we have no reason to believe that, given even time and effort, it couldn’t have reach analogous versions of the theories which have been reached with Galileo’s science.But then, we don’t really use Galileo’s science either, thanks to Einstein and quantum mechanics.

    My point is that we have used various sets of tools to approach, navigate and control the world around us. Some sets work better in some contexts than others do. But this does not mean that any set is somehow “truer” or “a better match” than any others.

    Of course your response is to point out that I am acknowledging the existence of a dough. But non-realists do not deny that there is a non-human something which “pushes back” so to speak. Rather, they deny that there is a unique and universal way in which this dough ought to be carved up or that the dough comes pre-carved up on its own. Indeed, this very analogy is itself a mere tool for approaching the scientific approach to the dough – and there is no reason why other, better tools cannot do this same job even better. Realism, however, is not that tool.

    I’m not saying that atoms, etc. don’t “really” exist. I am saying that it is a contingent fact about us humans that we have chosen to carve up reality in terms of atoms, etc. rather than some other set of conceptual tools.

    This is the solution to the mind/body problem, an area where I think most Mormons will be uncomfortable embracing realism; mental concepts and material concepts are simply different ways of engaging the social world around us. Neither one is more “fundamental”, “emergent” or “true” than the other. We are not “really” matter or mind; instead, these are the concepts with which we happen to engage the world around us and there’s nothing deeper to it.

    Scientists, then, are not the guardians of truth – the priests by which we approach the god of reality-as-it-really-is in the scientific religion. Rather, they are simply the R&D department for the rest of us, dedicating their time and efforts to the noble task of improving – or even inventing – new conceptual tools by which we can better engage, navigate and control the world around us. They have no more use for the myth of reality-as-it-really-is than engineers do. Some conceptual tools simply work better for our purposes in the same way that some cars simply drive faster.

  9. The way I see it, the pragmatist is perfectly comfortable comparing one theory to another to see which is better for our purposes. But the realist keeps wanting something more.

    It’s like the person who isn’t satisfied with knowing how many pesos, euros or rubles the dollar is worth… they want to know how much a dollar is worth in *real* money.

Comments are closed.