What Can We Know About The World Without Looking At It?

One last thought on all this God/cosmology stuff before moving on.

The crucial moment of our panel discussion occurred when John Haught said that he couldn’t imagine a universe without God. (Without God, the universe couldn’t exist.) It would have been more crucial if I had followed up a bit more, but I didn’t because I suck (and because time was precious).

Believing that something must be true about the world because you can’t imagine otherwise is, five hundred years into the Age of Science, not a recommended strategy for acquiring reliable knowledge. It goes back to the classic conflict of rationalism vs. empiricism. “Rationalism” sounds good — who doesn’t want to be rational? But the idea behind it is that we can reach true conclusions about the world by reason alone. We don’t ever have to leave the comfort of our living room; we can just sit around, sharing some single-malt Scotch and fine cigars, thinking really hard about the universe, and thereby achieve some real understanding. Empiricism, on the other hand, says that we should try to imagine all possible ways the world could be, and then actually go out and look at it to decide which way it really is. Rationalism is traditionally associated with Descartes, Leibniz, and Spinoza, while empiricism is associated with Locke, Berkeley, and Hume — but of course these categories never quite fit perfectly well.

The lure of rationalism is powerful, and it shows up all over the place. Leibniz proclaimed various ways the world must work, such as the Principle of Sufficient Reason. Lee Smolin uses Leibnizian arguments against string theory. Many people, such as Oxford philosopher Richard Swinburne, feel strongly that the world cannot simply be; there must be a reason for its existence. Paul Davies believes that the laws of physics cannot simply be, and require an explanation. William Lane Craig believes that infinity cannot be realized in Nature. Einstein felt that God did not play dice with the universe. At a less lofty level, people see bad things happen and feel the urge to blame someone.

But the intellectual history of the past five centuries has spoken loud and clear: the dream of rationalism is a false one. The right way to attain knowledge about the universe is ultimately empirical: we formulate all the hypotheses we can, and test them against data. (Making decisions about which hypotheses best explain the data is of course a knotty problem, but that’s for another time.) Broad a priori principles are certainly useful; they can help guide us in the task of formulating and testing hypotheses. But that’s all they do — if we get lazy and start thinking that they grant us true knowledge of the world, we’ve gone off the rails.

A common manifestation of the rationalist temptation is an insistence that a certain state of affairs cannot merely exist; it must be explained, we must find a reason for it. The truth is that, if things are a certain way, there might be a reason for it, but there might not be. Both are hypotheses that should be examined. I personally have a strong feeling that the low entropy of the early universe is an unusual situation that probably has a deeper explanation — it’s a clue pointing towards something we don’t understand about the universe. But I’m careful to distinguish that I don’t know this to be true. It’s perfectly conceivable that the universe simply is that way, and there is no deeper explanation. Ultimately the decision will be made by constructing comprehensive theories and comparing them to data, not by scientists stamping their feet and insisting that a better explanation must be found.

An inquisitive five-year-old might bombard you with an endless series of “Why?” questions. Sometimes you encounter an older version of this five-year-old; someone who, when you say “I have finally formulated a successful unification of all the laws of physics!” will insist on asking “But why is it that way?” If you say “it just is,” they will say “that’s not good enough.” That’s the point at which you are allowed to turn the tables. Just start asking, “Well why isn’t it good enough? Why do I need a deeper level of explanation for how the world is?” Not that it will actually change their attitude, but it can be personally satisfying.

Favorite targets for people insisting on deeper explanations include the existence of the universe itself (as Haught was indicating) and the particular laws of physics we observe (as Davies argues). The proper scientific attitude is to say: well, there may be a deeper explanation, or there may not. Before we go out and actually look at, the universe could very well be many things. It could be a single point. It could be a line or a plane. It could be non-existent. The universe could be a fiber bundle over a Riemannian manifold, an n-dimensional cellular automaton, a trajectory in Hilbert space obeying Schrödinger’s equation, a holographic projection of a conformal field theory, the dream of a disturbed demon, a layered collection of natural and supernatural dimensions, someone’s elaborate computer simulation, or any of a million other things. It could be unique or multiple, meaningful or intrinsically purposeless. It could be brought into existence by something outside itself, or it could be sustained by a distinct being, or it could simply be. If you personally find some of these alternatives unsatisfying, that is a matter for you and your therapist to work out; reality doesn’t care. The way we will find out the truth is not to insist that it must be one way or another; it’s to understand the likely consequences of each possibility, and line them up with what we actually observe.

You can see why a rationalist line of reasoning would be attractive to the theistically inclined. If you have God intervening in the world, you can judge it by science and it’s not a very good theory. If on the other hand God is completely separate from the universe, what’s the point? But if God is a necessary being, certainly existing but not necessarily poking into the operation of the world, you can have your theological cake without it being stolen by scientific party-crashers, if I may mix a metaphor. The problem is, there are no necessary beings. There is only what exists, and we should be open to all the possibilities.

None of this is to say that there is no room for logic or reason in understanding how the world could possibly work. “2+2=4” is a true statement in any possible world, once we specify the definitions of “2” and “+” and “=” and “4.” But that doesn’t mean it’s a true statement about anything that actually happens in the world. The universe might very well have been something where there weren’t two collections of two things to add together, nor sufficient computing power to perform the arithmetical operation. Once we accept some hypotheses about the world (through comparing their predictions to reality), we are allowed to use reason to draw inferences from those hypotheses. (That’s kind of what I do for a living.) But step one in that process is to be open to which sets of hypotheses are actually relevant to the real world.

The temptation of rationalism can be a hard one to resist. We human beings are not blank slates; not only do we come equipped with informal heuristics for making sense of the world we see, but we have strong desires about how the world should operate. Intellectual honesty demands that we put those desires aside, and accept the world for what it actually is, whatever that may turn out to be.

  1. Sean, thanks for this insight. I often have to deal with various points of view in my astronomy outreach. The outreach I do crosses all social-economic groups and the religious spectrum. In about 1 in 3 outreach events, I encounter a group who wants to debate these issues. The insight here helps me straighten out my own thoughts with ways I might use to address their questions. It was enlightening. I appreciate your insights and your ability to put it into words that I can understand.


  2. Thanks for this, Sean. This topic came up a lot in the office after a recent article which mentioned religion- the debate spilled over across all our social networks. I found myself holding my tongue because I didn’t quite have the words to express exactly what this blog touches upon!

    Teacher: “This is how we observe it happening.”
    Student: “Why does it happen?”
    Teacher: “Whatever the ‘why’ is it doesn’t change what we observe about ‘how’.”
    – Mike @ Questional

  3. Sean, doesn’t it really just get down to belief. God cannot be proven or disproven, it is strickly a matter of belief. As Bertrand Russell said “If God does not exist man would have to invent him”. Man wants psychologically to believe in something to explain that which man does not understand. And people want to believe that life on Earth is not the end. It is a need. Not rational but needed.


  4. Jim– the word “rationalism” appears nowhere on the Less Wrong post you linked to. Don’t confuse “rationality” and “rationalism,” they are completely different things!

  5. God cannot be proven or disproven, it is strickly a matter of belief.

    Belief in what, though? An invisible person with magical superpowers that very conveniently never demonstrates magical superpowers or personhood?

    As Bertrand Russell said “If God does not exist man would have to invent him”

    Voltaire, you mean: Si Dieu n’existait pas, il faudrait l’inventer.

    Man wants psychologically to believe in something to explain that which man does not understand.

    The entire point of the post was to argue that the only thing that actually works at explaining anything is the scientific method, and coming to a conclusion — believing in something — without evidence, is useless.

    And people want to believe that life on Earth is not the end. It is a need. Not rational but needed.

    This (and Voltaire) notwithstanding — while religion may be common, it is not universal.

  6. There are obviously necessary beings. We are. We are our own evidence, and any theory has to include us.

  7. I would call this a straw argument if there weren’t so much validity to it. Most do think at these simple levels. For those capable of deeper thinking, no search for knowledge can be free from both empiricism or rational thought and the mistake is to believe it must be one or the other. Rationality must not be a substitute for empiricism, nor can empiricism be free from it. Why must always remain a question for empiricists and reason must never be the answer for rationalists. Knowledge has no end, nor reason a bound.

  8. There are obviously necessary beings. We are.

    You’re not understanding “necessary” in the sense used by philosophers. The argument is that all beings are contingent — that is, there are prior factors which lead to them existing. You do exist, but you might not, if those prior factors did not align to lead to you existing.

    The theololgical argument goes, the chain of contingent factors cannot continue infinitely (note: this is what they say; I don’t think anyone has shown it to be necessarily true), so there must be something that has no contingent priors; this something is necessary for everything that follows from it. All factors are ultimately contingent upon it, while it is not contingent upon anything else.

    OK, got all that? So, this necessary whatever-it-is — this is called “God”, by theololgians.

    Now, anyone with a brain would ask, “Wait. Why are we calling this necessary whatever-it-is ‘God'”? And the real answer is, Thomas Aquinas wanted to rationally defend belief in God, and in doing so, failed to notice that he was arguing irrationally.

    There might be some ultimately necessary whatever-it-is, upon which everything that exists is contingent upon — I don’t claim certainty on the matter. But there’s no reason to call it “God” unless you’re a theololgian who wants to play word games in defense of theistic belief.

  9. Sean,

    When you say “there are no necessary beings”, I’m curious how you know this–have you gone out and looked at the universe in order to determine this, or are you applying rationalistic thought processes to reach that conclusion?

    Now it seems to me, given your focus on empiricism, that you’re talking about *epistemic* necessity, in which case you’re absolutely right–even Haught should have been able to wrap his mind around the idea, “What if I didn’t have any knowledge relating to the existence of God? How would I explain the universe?” But I still have the feeling that Haught had in mind ontological necessity, and that is (as I said in the other comment thread) where you two were probably talking past each other.

    More apropos of this post, though, a conclusion of ontological necessity can be reached at least in part *by observing external circumstances* and seeing that there are things to observe, and observers to do the observing. And if we see that every individual thing that we observe has an (external) cause, then it is reasonably (though perhaps not absolutely) certain that the collection of all things we observe similarly has an (external) cause. To borrow your phrasing, not only is 2+2=4 true in any possible universe, but we are also actually observing *in our universe* two collections of two things that can be added together. So to claim that belief a la Haught is based on rationalism without any empirical checks placed upon it is not quite right.

    But then again, what’s particularly tricky in the case of ontological necessity is that “the observable universe” taken as a whole provides but one data point, which makes broader extrapolation rather ill-advised, as a general rule. Still, I’d be interested to know if you think the epistemology/ontology distinction impacts your argument at all.

  10. I just don’t think the notion of “ontological necessity” is in any way useful. I can imagine many different worlds, with completely non-overlapping sets of characteristics. I see no difference between someone claiming that a concept is “ontologically necessary” and someone admitting that they can’t imagine a world without that concept. The correct prescription is to work to improve one’s powers of imagination.

  11. Dear Sean,

    Just to clarify, I used the principle of sufficient reason (PSR) in the paper you cite and others to argue that any successful unification of physics must be background independent. This is not “against string theory” as many experts would agree that any precise, non-perturbative formulation of string theory must be background independent. The lack of such a formulation is a persistent weakness of string theory, but it may be one that can be overcome. I have certainly put a lot of effort into trying.

    In fact, most approaches to quantum gravity are as background independent as general relativity. It means that classical spacetime must emerge, and recently there is a lot of progress with showing that this happens, in spin foam models and causal dynamical triangulations. It is also the background independence of loop quantum cosmology that yields clean results on the elimination of cosmological singularities.

    Of course, in science experiment tops general principles. Philosophy can be at best an heuristic guide. But still, historically the PSR has been an excellent heuristic guide to the construction of theories that did pass experimental tests. The best example is general relativity. Another is statistical mechanics. It forced people to make sense of the atomic hypothesis, when pure empiricists were arguing atoms didn’t exist and matter was just what it was. This led to the 20th century explosion of knowledge.

    I think there is a good case to be made that the PRS remains a good guide now. For one thing, it tends to concentrate our work on theories that are closed, explanatorily, which means they are more likely to lead to falsifiable predictions. This was the point of my cosmological natural selection scenario. It also explicitly excludes the troubled, paradox prone, cosmological scenarios with an infinite number of copies of every thing, person and event in the universe, as these massively violate the principle of the identity of the indiscernible. The latter is a consequence of the PSR. Seeing how much intellectual gymnastics is going into wresting with, but so far not resolving, the measure problem in the multiverse of eternal inflation, one can appreciate what good advice the PRS is.

    There is much more to say about this, it is central to my current writing projects and I look forward to discussing with you sometime soon.

    Best wishes,


  12. Although I agree that overall empiricism trumps rationalism, I think it is a mistake to present the two as dichotomous opposites. Of course, at its worst, rationalism completely ignores empiricism: as has regrettably happens in far too often in my field (linguistics) where the Chomskian approach is to rationally deduce how human language ‘must’ work, and then declare actual data ‘epiphenomological’ and simply ignore or deny contradictory observational findings. (It could be added that in his politic writings, too, Chomsky exhibits a tendency to favour his theory-driven narratives over actual, inconvenient events. In this sense the conspiracy theorist and the religious believer overlap somewhat!).
    However, conversely, observation unmodified by reason can lead to science being reduced to a merely cataloguing activity. Psychology, long cowed by the embarrassment of Freud, has to certain extent reduced itself in this manner.
    I should add that, initially, Chomsky was extremely beneficial to linguistics, shaking it out of exactly this type of butterfly-collecting approach and presenting it with some meaty underlying questions to tackle. However, the answers he posited to these questions were constructed purely from rationalism, and fail to correspond well with actual observation leading a whole chunk of the discipline, who refuse to drop this paradigm, into a parallel world where the topic of investigation is not human speech as observed but idealized utterances drawn circuitously from their theories.
    A rationalist spark, therefore, the idea that ‘this may be the case because…’ is necessary to move beyond purely correlational results to hypotheses concerning causation, and the consequent novel testable predictions. The error I think is in not recognizing that when data and theory collide, it is the theory that must give way.
    If this all sounds a bit Popperian, well, that’s ‘cos it is. Unashamedly so! Whether he actually did this or not, Popper’s anecdote about starting a lecture course by instructing his students to take out their papers and pencils, and then commanding “Observe!” illustrates all this nicely: the students (apparently) immediately asked “observe what” and “how”, etc., giving Popper a nice little parable about how even the simple act of observation is actually theory-laden.

  13. Maybe I’m missing something but how is this different than when 500 years ago the village priest was telling his parishioners to stop asking why; that’s just the way God made it? Somebody better be stamping their feet and demanding a better explaination or be making apologies to lots of dead people because we’ll know no more of the core reason than they did. At least their reasons gave some sort of comfort to people.

  14. Without a healthy dose of rationalism all of our scientific observations become a collection of coincidences.

  15. So, Sean, do you think that the universe can come from literally nothing? If so, why? IF there is nothing, there are no laws of physics. So how do you make a universe?

  16. You need a being that transcends space and time in order to create space and time from a state that did not have space and time. That is, if there is no multiverse, and our space and time had a beginning, then you need a being that transcends space and time to create our universe because the laws of physics do not somehow exist outside space and time, so our universe couldn’t have been caused by a “quantum fluctuation”. Nor have you answered my question, “Why something rather than nothing if the energy content of nothing is still 0?” Hawking’s explanation is not an explanation at all.

  17. So, did the narrator make the argument that the universe came into being from nothing? Or was the point that one cannot talk about a cause because time itself came into being with the Big Bang, so there was no “before” in which a cause occurred?
    Unfortunately, the show didn’t make it clear that we really know nothing about the laws of physics at that tiny scale (quantum gravity). Hence, the statement that the universe could have come into being without violating any of the KNOWN laws of physics is incorrect since we don’t know what those laws are at that tiny scale.
    Sure, one can say that the total energy of the universe is 0, so a universe can pop into existence without needing any energy or violating energy conservation. But if the universe, for example, came from “nothing”, then why did the universe come into being at all? The total energy of “nothing” is 0 as well. Thus, why isn’t there still nothing? And one cannot make the argument that the universe could have popped into existence from nothing as a result of quantum mechanics, because there IS no quantum mechanics if there is nothing. And even if, somehow, the universe could have come into being from nothing, why THESE laws and constants, and not others?
    So I’m surprised that, to my knowledge, the show didn’t mention the multiverse idea which could resolve these questions by saying there may exist a “universe-generating” mechanism that can make a universe from a pre-existing universe. If a multiverse had always existed “for all eternity”, that could completely do away with the need for a “first cause”. One could just say that the multiverse had always existed with the set of laws that it does, laws which also include some sort of universe-generating mechanism. But that still doesn’t address the question, “Why those laws and not others?” Unfortunately, our knowledge of the laws of physics is still incomplete, so we cannot say for sure whether such a mechanism exists, nor whether other universes exist.
    All they referred to was the notion of virtual particles randomly coming into existence from the vacuum of space-time, but this is just an analogy. The vacuum of space-time is not nothing. It is part of the universe.
    So concluding that these ideas render a creator unnecessary is, in my opinion, misleading and incorrect.
    If the universe came from nothing, then on what grounds can we say that the universe arose from a random quantum fluctuation? If there’s nothing, literally nothing, then the idea “quantum fluctuation” makes no sense — it doesn’t even exist, it’s not possible. If you have nothing, you stay with nothing.
    Does there exist a being which transcends space and time? No? Why not? Can you use science to answer the question?

  18. Couldn’t have said it better myself. I know, because I tried to in a comment to the previous post of this subject, and didn’t. Well put.

    @ Phil: “You need a being that transcends space and time …” No, some of us really don’t. More to the point, you don’t get to decide this on behalf of the universe. That was sort of the point of this whole post. That’s even if your “being” had any explanatory power at all, which it doesn’t. (Hint: exactly how does your being accomplish what you claim it does. If it’s just “magic” with no explanation, I could as easily assume the universe itself is “magic”, couldn’t I?)

  19. Sean,

    I’m with those who think you’ve posited a false dichotomy.

    Of course, real progress isn’t made simply by never leaving the comfort of our living room; just sitting around, sharing some single-malt Scotch and fine cigars (not that this sounds so bad 😉 ), thinking really hard about the universe.

    Conversely, empiricism is inadequate because scientific theories explain the seen in terms of the unseen and the unseen, you have to admit, doesn’t come to us through the senses. These both are extremes and neither is correct.

    However, you go on to say that empiricism says that we should try to imagine all possible ways the world should be, and then actually go out and look at it to decide which way it really is. I agree with this but don’t think it’s empiricism, at least in the traditional sense.

    In fact, when you say that we should try to imagine all possible ways the world should be and then go out and see if it’s right, what you are saying is that we begin by conjecture, and then we criticize and test the conjecture in the world, trying always to arrive a more correct explanation. That’s the scientific method.

    Conjecture looks a lot like rationalism, but its more than that, because it’s always based on prior conjectures that have been tested and criticized in the world. And criticizing and testing the conjecture in the world looks a lot like empiricism, but it’s more than that, because its based on our then current best explanations.

    Neither label does the scientific method justice.

  20. JimV

    “Hint: exactly how does your being accomplish what you claim it does. If it’s just “magic” with no explanation, I could as easily assume the universe itself is “magic”, couldn’t I?”

    But can the universe create itself from nothing? No. The only thing we know about the universe is that it is governed by a set of laws and principles. If you don’t have a universe, you don’t have these laws or principles. So if you don’t have a universe, nor laws and principles, then how does one explain the origin of the universe from nothing? Laws of physics don’t somehow exist, just as Plato’s forms exist somewhere. The laws are what we come up with to describe the universe and predict its behavior. They don’t exist outside the universe, nor do they have any kind of power to create one.

    I don’t know how this being (God) accomplishes it. All I know is that, unless there is a multiverse that had always existed, the laws of physics can’t tell us how our universe was “born” if, indeed, it had a beginning.