Evidence, Absence and Burdens

Murali

Murali did his undergraduate degree in molecular biology with a minor in biophysics from the National University of Singapore (NUS). He then changed direction and did his Masters in Philosophy also at NUS. Now, he is currently pursuing a PhD in Philosophy at the University of Warwick.

Related Post Roulette

78 Responses

  1. Fnord says:

    It is not the case that we have no evidence regarding the existence of a teapot orbiting the earth.

    Earth orbit is close enough that we have, to some degree, the ability to observe things there. That’s why the original Russell’s teapot thought experiment was that it orbited the sun, somewhere between the orbits of Earth and Mars. A vastly larger space, and much of it further away and/or concealed from direct observation. Given the advance of astronomy since Russell’s time, we could postulate a further distance; the possibility of a teapot orbiting the sun somewhere in the Kuiper belt. For which we have nothing like enough observation evidence to discern whether a teapot exists or not; if were a teapot in the Kuiper belt, we almost certainly would not have found it. And yet it would be absurd to suggest we should be agnostic, by your definition of estimating a 34-66% probability, of there being a teapot in the Kuiper belt.

    So, we come to this point:

    We could check with people in NASA for example. Did someone leave a tea pot in the air lock?

    When you are trying to determine whether there’s a horse in your room, is your first step to ask the local ranchers if any horses are missing?

    You’re getting at something you left out of your initial description of epistemic rationality: the issue of prior probability. The prior probability of a teapot in orbit around the earth is low. Now, if NASA had lost a teapot in earth orbit, that would be strong evidence for the existence a teapot, and so we could believe in it despite the low prior probability (likewise, if we had a picture of a teapot in orbit, etc). On the issue of whether a certain distant star is orbited by a planet, in the absence of observational evidence one should express agnosticism; on the issue of whether that star is orbited by a teapot, one should disbelieve it.Report

    • Murali in reply to Fnord says:

      The only reason why prior probability that a teapot orbiting the sun at the Kuiper belt is low is because we have a lot of evidence (presumably) that people did not leave any teapots in the satellites they launched there. So, again it is not mere absence of evidence, but evidence of absence of teapots in space. In fact, it is part of the concept of a teapot that it is something artificial and hence highly likely to be made by beings who may sometimes drink tea. The only tea-drinkers we know of are us. If any space-faring tea-drinkers other than us had come near enough to us, there is some serious probability we would have heard of them. Thus it is unlikely that we are being observed by tea drinking galactic overlords. Thus at least within the vicinity of our solar system and its immediate surrounds, any teapot would have had to come from us. Given that launches into space are relatively rare, it is unlikely that there is a teapot floating around in space.

      Now, if spaceflights were more frequent and Han Solo had to dump his collection of china to avoid detection by the galactic empire, we would expect to see some teapots in that vicinity.Report

      • Fnord in reply to Murali says:

        In fact, it is part of the concept of a teapot that it is something artificial and hence highly likely to be made by beings who may sometimes drink tea.

        Let’s say, for the sake of argument, that intent isn’t necessary. If an object that was indistinguishable from a teapot was purely by natural processes and the vagaries of chance, we’ll call that a teapot for that argument. Should we be agnostic about the existence of such in object in the Kuiper belt?

        Or, OK, let’s try a more down to earth example. You’re a detective, called to a murder scene. You haven’t gathered any evidence yet. Your partner says, “we have no evidence, so I’m agnostic about the possibility that this murder was committed by Jim H. Jones, of 123 South Pleasant St, has committed this murder. Obviously, this will change as we gather evidence, but until we check his alibi or find evidence pointing at someone else, I’d put 50% odds on Jones having done it.” Is your partner’s reasoning valid?Report

    • Stillwater in reply to Fnord says:

      The only reason why prior probability that a teapot orbiting the sun at the Kuiper belt is low is because we have a lot of evidence (presumably) that people did not leave any teapots in the satellites they launched there.

      No, that’s not it at all. It’s that there’s no reason to it *without* evidence of someone having left a teapot there. Given what we know right now, we would assign an infinitesimally small probability to that claim being true (whatever the probability of a logical possibility is, really) unless we had the type of evidence you’re referring to.Report

  2. Chris says:

    The can’t prove a negative thing is particularly difficult to get rid of. There are people who trace it back to the Scholastics, and it’s been common at least since the 18th century, but it’s very easy to show that it’s not true. If we’re just being formal, any positive statement can be reworded as a negative, and vice versa, and even if we want to stick with everyday speech, the highest levels of proof — showing that a statement entails a contradiction — are still possible with negatives. Even at lower, less formal levels of proof, if we sufficiently establish the scope of a negative claim, it’s possible to prove it within that scope to the same level that it is possible to prove positive claims with the same scope.

    The elephant in your room is a perfect example. The scope is well defined: you’re room. An observation of the objects of sufficient size to be elephants (even baby ones) is sufficient to prove the negative. If not, then such observation would be no more valid as proof that there is an elephant in your room.

    The teapot example is more difficult because the scope of the claim is broad enough to make the same sort of observation practically impossible, and we’re forced to fall back on our existing knowledge of teapots, space, etc. to make reasonable inferences. These allow us to conclude that the possibility of a teapot out there is pretty damn small, which, for pretty much any practical purpose, amounts to proof that there’s no teapot out there. It’s not perfect, but it’s pretty damn good.

    God is another issue, because like the teapot, it requires using existing knowledge, but unlike the teapot, that knowledge is the subject of much debate. Plus, it really does have an unlimited scope.Report

    • James Hanley in reply to Chris says:

      It’s not just god, but the supernatural in general. We can’t really prove that the horse in Murali’s bedroom doesn’t just turn invisible every time he looks for it.Report

      • Murali in reply to James Hanley says:

        Given the state of my room it wouldn’t be a surprise if an invisible horse regularly trampled on everything.Report

      • In that case, there’s evidence of presence, not absence of evidence 🙂Report

      • Chris in reply to James Hanley says:

        We can feel for the invisible horse, right? If feeling and seeing aren’t enough, then one of two things is the case: either we couldn’t prove with seeing and feeling that a horse is in the house, and therefore the negative proposition is not unique, and we’re radical skeptics, or the horse is invisible and intangible, and we can pretty easily show that it contradicts some true proposition about horses, and the negative is proven true.

        The problem with supernatural claims is that we’re talking about things for which empirical observation is insufficient because it can’t cover the entire scope of the claim, and observation alone isn’t going to get us to a contradiction the way continuing to strip away the physical properties of a horse did.Report

      • Chris in reply to James Hanley says:

        There are, of course, other tools at our disposal with respect to supernatural claims.Report

      • James Hanley in reply to James Hanley says:

        My preferred tool is mockery.Report

    • Stillwater in reply to Chris says:

      The elephant in your room is a perfect example. The scope is well defined: you’re room. An observation of the objects of sufficient size to be elephants (even baby ones) is sufficient to prove the negative. If not, then such observation would be no more valid as proof that there is an elephant in your room.

      Of course, this argument relies on a certain confidence that our observations constitute proof for certain claims. Could you be deceived in your beliefs? Hallicunating? Etceteraing? All this kind of stuff strikes me as an attack on empiricism, which it obviously is, really, but it doesn’t refute – or even challenge – empiricism in any way I can see. It merely introduces a lack of certainty (hence the use of the word “proof” as the bar which must be cleared) entailed by empirically determined beliefs. But empiricists are entirely happy with a lack of certainty, in particular about so-called “contingent” claims.Report

      • Chris in reply to Stillwater says:

        @stillwater sorry, just saw this. You’re right, that would be an attack on empiricism, and of course, if our senses aren’t any good for finding elephants, they’re probably not any good at saying for sure that one’s here either. So positive and negative claims become impossible to demonstrate empirically.Report

  3. Road Scholar says:

    As a general rule you are correct, yet it seems to me that it depends entirely on the nature of the claim. Consider the Simulation hypothesis (Matrix, Brain-in-a-Vat, etc.). In principle it’s possible that such a simulation could be imperfect. There may be programming errors or processor glitches and so if the hypothesis were true and such errors existed or could be made manifest under the appropriate conditions, the possibility exists that positive evidence could be gathered that the hypothesis were true. But if the hypothesis were true and no such errors existed then the simulation would truly be indistinguishable from its negation.

    This would truly seem to be a case where it’s impossible to prove a negative. But it seems like a degenerate case since there is no particular reason to believe it’s true and it’s really just a “sophomores getting stoned in their dorm room” sort of conjecture. Is it really the case that one must somehow positively refute, either by evidence or theory, every such random conjecture anyone can dream up to justify disbelief? Is Elvis alive? Is the world ruled by shape-shifting lizard people? Were the moon landings hoaxed? Are our lives and destinies governed by the position of the planets at the moment of our birth? Is President Obama a secret muslim agent intent on destroying America? Is the Earth hollow? I could go on (I used to listen to Art Bell a lot. 🙂 ) but I’m sure you get my point. There’s no dearth of half-baked, random-assed ideas floating around out there and the notion that I’m required to positively refute each and every one before reasonably assuming disbelief is… just exhausting to consider.Report

    • Murali in reply to Road Scholar says:

      Is it really the case that one must somehow positively refute, either by evidence or theory, every such random conjecture anyone can dream up to justify disbelief?

      You have to through evidence and theory show that the alternatives (taken together) are a lot less probable. My previous post covered this.Report

    • Kim in reply to Road Scholar says:

      Road,
      ” But if the hypothesis were true and no such errors existed then the simulation would truly be indistinguishable from its negation.”
      not strictly speaking true.
      the negation is a far larger range of solution space than the positive (aka it has to be simulatable — and in our present universe’s case, it is suspiciously simulatable.).Report

  4. DavidTC says:

    ‘We can’t prove a negative’ is really just a way for confused people to say ‘Occam’s Razor’.

    We’re going to operate off the assumption that there’s no horse in the bedroom because a horse being in a bedroom requires all sorts of weirdly complicated and specific things to have happened.(1) If you want to claim there’s a horse in the bedroom, I’m going to need to see it, otherwise I will assume it’s not there. Or, at the very least, *you* need to see it, and I need to trust you.

    Because, under Occam’s Razor, if you propose a horse in your bedroom, the simplest explanation for your claim is that *you’re factually wrong*.

    1) Barring some sort of pre-existing situation like you sleep on a ranch, and your bedroom is inside the barn, and horses can get in if your bedroom door is left open. Then, if you get home, and the bedroom door is open, and a horse isn’t visible in the rest of the barn, you can confidentially assert ‘Oh, there’s probably a horse in my bedroom again’. But I’d probably already know some of that, or you’d explain it to me.

    If we ought not to disbelieve in Zeus merely because we lack evidence*, disbelieving in Jehovah or Krishna because we lack evidence one way or another merely compounds the error.

    This rather indicates that your premise is wrong. Any conclusion that results in believing *all* Gods is rather goofy.

    We ‘ought’ to disbelieve in Gods because Occam’s Razor says they don’t need to exist to *explain all the evidence we have*, not because we lack evidence. There’s no such thing as ‘lacking evidence’ in any real sense, all that phrase means is ‘The stuff we know does not point to that specific conclusion’.

    Evidence is just what we call something when we think it proves a theory of ours. Saying that ‘no evidence’ disproves, or doesn’t disprove, a theory is not a correct way to say things. (E.g., God supposedly created a universe, and there *is* a universe, so that is, in principle, evidence of the theory of God to some extent. It’s just not very *good* evidence, because other things could have created the universe.)

    What disproves theory X is that theory Y is ‘stronger’. There are two ways for Y to be stronger:
    a) The things we know appear to point more to Y. This is what is called ‘more evidence’. If I see a horse in your bedroom, either there is a horse in your bedroom, or I am having very vivid hallucinations of the sort that I’ve never had before. I now believe, based on the evidence provided by my eyes, that there is a horse in the bedroom.
    b) Theories X and Y fit what we know somewhat equally, but X is much more complicated for no logical reason. Now, ‘complicated’ is somewhat vague, and often things we think are very complicated turn out not to be, and sometimes things actually *are* complicated. (Life, for example.) This is Occam’s Razor.

    P.S. And if we do believe in a specific God, we ought not to believe in the others, because multiple competing pantheons not only don’t need to exist, they’re are pretty explicitly incompatible according to their own rules. It’s like believing in the phlogiston *and* oxidation. Explaining how Zeus *and* Jehovah both exist requires all sorts of multiplying entities unnecessarily.Report

    • Guy in reply to DavidTC says:

      You can get a pretty good idea of simplicity by asking, “If I had to write a simulator for X, how much code would I need to write?”Report

      • DavidTC in reply to Guy says:

        No. That assumes nothing exists.

        Complexity would be how much *additional* code you’d have to add to an *existing* simulation of the universe.

        I..e, making a new hairbrush should be trivial, not an insanely complicated thing where you need to invent the ideas of mass and gravity and whatnot.

        And it also assumes laws of the simulation and the laws of the universe are the same. But they aren’t. The laws of the universe are quantum, and we don’t really have any quantum computers or have any good idea how complex it would be to program things in them. Things that are very easy to explain using quantum mechanics can became incredibly difficult to simulate.

        Not that simulating Newtonic physics is much easier.Report

      • Veronica Dire in reply to Guy says:

        @davidtc — Well, not exactly. We do have to assume a baseline, which is raw computation. I mean, if we are simulating this we are simulating it on something, some computational faculty. So the question becomes, what is our model of computation: raw Turning computation? lambda reductions? combinators? blah, blah, blah…

        Thing is, for all of those the answer you get is pretty much the same. I mean, sure it will be easier to multiply in a lambda space, since you have local bindings, and Church numerals are sexy. Likewise, if one proposes a 2-d Turing array things get easier compared with a 1-d Turing array. But on the other hand, these differences are subroutines. The real work happens a few levels up.

        I would be happy to discuss what work it would take in your average LISP. Not that I think the universe runs on a LISP, but computation is computation and it provides the desired insight.

        And talking about hairbrushes kinda misses the point. We are imagining what it would require to make all this from (almost) nothing, not what it would require to add something.

        (Add predictable section on hypercomputation here.)Report

      • Guy in reply to Guy says:

        @davidtc We can simulate quantum events pretty easily on a small scale (three seconds of googling). As a matter of fact, I know some people who do this kind of work (I’m more interested in astro, so I do other things). On a large scale, quantum fluctuations mostly factor out, so you usually don’t need to bother with them for your simulation. Newtonian physics are also not really that hard to simulate, in raw coding terms. In both cases, you can run into trouble with resources used during the simulation, but that’s the question at issue. And in any event, if we assume the same base universe for every explanation of a given phenomenon, we can subtract out the complexity of that base universe from the simulation. There’s no real difference between “additional code” and “absolute code” in terms of simplicity comparison.Report

      • DavidTC in reply to Guy says:

        @veronica-dire
        I mean, if we are simulating this we are simulating it on something, some computational faculty.

        I was talking about ‘less complex’ *in general*.

        You cannot simply define ‘less complex’, in Occam’s theorem, as ‘simpler to simulate’, because that path leads to utter stupidity. I.e., I’m never seen an atom, so the ‘simplest’ simulation of my life is one where *everyone is just lying about atoms*. (Whereas, in actuality, that would require all sorts of literally insane behavior of large parts of populations, and thus is well into ‘conspiracy theory’.)

        Simulations, of the sort that people are talking about when they tend to talk about simulations, of the sort that you and Guy seemed to be thinking of, are *completely* wrong to judge how ‘simple’ something is. In those simple simulations, walls are not atoms, they are mathematical planes that other simulated objects cannot pass through. In those simple simulations, a cat can fly around as easy a bumblebeee…you just give it the attribute ‘able to fly’. Those simulations tell you nothing about how ‘complex’ something is.

        If there was some sort of hypothetical simulation of literally every particle and thing that existed *except* the thing we were trying to explain, yes, how ‘easy’ something was to add to simulation would work as a rough measure of Occam’s razor, but a) how does such a simulation fit in the universe, and b) how exactly would we manage to know everything *but* what we were trying to figure out in the first place, and c) how on earth does this help us figure things out anyway?

        For Occam’s Razor, just imagine how easy it would be for such a thing to exist in the *actual universe*. That’s it. Take as much of the universe as you know minus that thing, and try to figure out what sort of and how many additional rules would allow that thing. If it’s none, you’re good. If it’s one new rule, you might have a scientific theory, so test it. Any more than that, and you’re probably wrong, at least according to Occam’s Razor.

        All throwing a ‘simulation’ in there does is confuse the issue.

        @guy
        We can simulate quantum events pretty easily on a small scale

        Yes, after working really really hard, they’ve come up with a possible way to simulate the interaction between four quantum particles, in specific limited circumstances. Four! That’s almost five!

        Why, if we keep this rate up, we’ll be able to simulate a single amino acid some day.

        Newtonian physics are also not really that hard to simulate, in raw coding terms.

        Are you serious? Have you ever heard of the three body problem? We can’t even simulate the gravity of the *solar system*.Report

      • Kim in reply to Guy says:

        DavidTC,
        the key word is simulate. If you give me five tries, I can do the Travelling Salesman problem too.Report

    • Murali in reply to DavidTC says:

      @davidtc ,

      Guy has the right idea. Ultimately it is about what code can you write with the fewest independent variables.

      Saying that God exists is like setting a particular variable G to 1. While saying that he doesn’t exist is like setting the variable G to 0. Code in which you don’t even use G is everything else equal, simpler. The value of G would then be indeterminate (or taking the principle of insufficient reason, 0.5)

      That is to say, the most defensible account of simplicity is about having fewer independent premises, not about having fewer objects in the world. There are simple, mathematical reasons as to why all else equal, a theory with fewer independent premises is more likely to be true than one with more independent premises.

      There does not seem to be a good reason to think that a world in which there are fewer objects is more likely than one in which there are more objects. That would involve all sorts of weird metaphysical assumptions.Report

      • Veronica Dire in reply to Murali says:

        That is why “God” is the big cheat. After all, what is the computational complexity of “God”? How much of the problem are you trying to hide away behind “G == 1”?

        It’s the ultimate “If he had this oracle then…”

        (On oracles: http://en.wikipedia.org/wiki/Oracle_machine)Report

      • Chris in reply to Murali says:

        Only a big cheat if you don’t have a sufficiently rich conception of God. Most of the powerful arguments for the existence of God attempt to show a sufficiently rich conception and that, given certain conceptions of things like necessity, contingency, and causality, that conception of God is in fact the simplest explanation (it is logically entailed by those other concepts).Report

      • Murali in reply to Murali says:

        @veronica-dire

        Which is why it is better to say neither G==1 nor G==0. If I can say everything I want to say without having to assign a 1 or a 0 then the simplest account of the world suspends judgment about whether God really does or does not exist.Report

      • Veronica Dire in reply to Murali says:

        @chris — I find such approaches to be pretty vacuous. It’s like, someone grabs the word “necessity,” but instead of understanding the broad possibilities found in (for example) Kripke semantics, they nail the concept down into a narrow thing and then say, “That could only be God.” Likewise for causality. Instead of a formal model of causality, such as Pearl’s, it is becomes vague metaphysical cheat card.

        But see, once you begin to apply a model such as Pearl’s, you suddenly find that model does two things: 1) it explains causality as we experience it in the natural world, and 2) it does not apply in any broad metaphysical sense. But now you have this: the existence of causality in the natural world need not apply in some extra-cosmic sense. So a question such as “What caused the universe?” is shown to have all kinds of hidden (and perhaps unjustified) assumptions.

        As a contrast, you have approaches such as mine, which says things such as, “Clearly computation exists in the universe, but that does not imply that the universe is fundamentally computational, so the notion (such as in this thread) that we can model it as such is an interesting exercise, but proves nothing.”

        I have yet to see a concept of “God” which is anything but vague words gathered together with hand waving.Report

      • Veronica Dire in reply to Murali says:

        @murali — I say “What does ‘G’ even represent in the system? Precisely what does it do, and why do we need it?”

        Myself, I take these three things as (kinda) axiomatic:

        1. I exists, along with others like me (other people)

        2. We live in a world together, we can communicate

        (Animals maybe fit into this also, somewhere. It’s complicated.)

        3. The universe appears to follow natural laws

        4. Math seems to work.

        Thanks basically it. On top of that I assume the natural sciences have formed real knowledge. From that I conclude two things: 1) evolution explains the origins of life, and 2) neuroscience (such as it is) explains the existence of consciousness, including mine.

        This includes little room for extra-neural consciousness. (Where is God’s brain?) Nor does it leave much room from brain independent values, emotions, purposes, etc. Which in turns leaves little room for ideas such as “transcendent love” or “universal morality.”

        Is there room in this for “G == 1”? Well maybe. But what does it do in the system? Make math work? Why can’t math just work on its own? Create the natural laws? What can’t natural laws just be?

        I don’t know why math works. Nor do I know why there are natural laws. But then, there might not be “reasons” for these things in the way we understand reasons. Which is fine. I guess. What does “God” add?Report

      • Chris in reply to Murali says:

        Yeah, I don’t see that at all. The logics that go into such arguments are pretty damn sophisticated, and can be pretty straightforwardly updated for modern advancements in logic and semantics.Report

      • Kim in reply to Murali says:

        v,
        Yes, but whose math?
        Our math is excessively convoluted for explaining even simple quantum mechanics. it could use a dramatic redesign.Report

    • Jim Heffman in reply to DavidTC says:

      “We ‘ought’ to disbelieve in Gods because Occam’s Razor says they don’t need to exist to *explain all the evidence we have*, not because we lack evidence.”

      This is a funny thing to say because the Occam in questions formulated the original statement as an argument in favor of God’s existence. He was arguing a position that modern thinkers would describe as Intelligent Design, i.e. “proposing that a system of such order and complexity as we see on Earth is entirely the result of random processes requires an impossible degree of coincidence, ergo it was intentionally created”.Report

      • DavidTC in reply to Jim Heffman says:

        This is a funny thing to say because the Occam in questions formulated the original statement as an argument in favor of God’s existence.

        There was a point in history where belief in Gods *was the ‘correct’ scientific theory*. (Or would have been, if anyone had any concept of scientific theories.) There simply was too much unknown, where saying ‘Some really big guy is moving the sun’ was an entirely reasonable explanation.

        And, of course, Occam wasn’t the first guy to think of Occam’s Razor, or the last.

        proposing that a system of such order and complexity as we see on Earth is entirely the result of random processes requires an impossible degree of coincidence, ergo it was intentionally created

        Occam had absolutely no way to calculate the probability of life existing, *or* to calculate how many different possible places there were for life to exist at, so that statement is basically the definition of ‘premature’. 😉Report

      • Chris in reply to Jim Heffman says:

        This is a funny thing to say because the Occam in questions formulated the original statement as an argument in favor of God’s existence.

        Nope.Report

      • Jim Heffman in reply to Jim Heffman says:

        “There was a point in history where belief in Gods *was the ‘correct’ scientific theory*.”

        I’ll remember this post next time someone starts lecturing me on how Global Warming is the correct scientific theory. I’ll be glad to quote it in support of my reply.

        And you aren’t really convincing me that Occam’s Razor did *not* come from a defense of religious belief.Report

      • Kim in reply to Jim Heffman says:

        Jim,
        Oh, dear, be my guest. Of course, we’ll be sure to give you a new nickname!
        Runga-Kutta.
        see, you actually have to have evidence to back up your theories.
        El Nino’s fizzling this year, you got a prediction better than Strictly Sober?Report

      • DavidTC in reply to Jim Heffman says:

        @jim-heffman
        I’ll remember this post next time someone starts lecturing me on how Global Warming is the correct scientific theory. I’ll be glad to quote it in support of my reply.

        Sounds like a plan: ‘My current beliefs are as smart as 14th century science! Global Warming is caused by WITCHES!’

        You know what’s made of wood? Carbon dioxide. Makes you think, doesn’t it?

        And you aren’t really convincing me that Occam’s Razor did *not* come from a defense of religious belief.

        I pointed out why Occam’s Razor lead to an *incorrect* solution in that specific instance, I didn’t say anything about the *origin* of it.

        I have no idea why Occam proposed it, nor do I actually care. (Chris seems to dispute your claim, I have no idea if he’s right.)Report

      • morat20 in reply to Jim Heffman says:

        I’ll remember this post next time someone starts lecturing me on how Global Warming is the correct scientific theory. I’ll be glad to quote it in support of my reply
        Replacing a theory with another requires (1) explaining all aspects that the previous theory explained as well or better than the previous theory and (2) explaining something the previous theory couldn’t.

        Einstein replaced Newton for objects in motion — Einstein and Newton agree on rocks falling on Earth, but Newton couldn’t explain Mercury and Einstein could.

        Now “god did it” as a scientific theory for why the sun rose is replaced by any theory that can (1) also explain why the sun rose and (2) can offer further understanding. Like, for instance, the “The Earth rotates around the sun, and revolves around it’s axis, and the moon rotates around the earth” which not only explains said sun rising, but also explains eclipses and the phases of the moon.

        When it comes to Global Warming, to replace it you will need to….explain a number of trends (rising temperatures, falling ice levels, etc) as well as explain what, exactly, is happening to energy as it enters an atmosphere with a changing mixture of gasses.

        Plus you’ll need to fit your explanation against historical data.

        There’s a reason climate change deniers often sound like the cranks who yell about radiometric dating — an awful lot of their case relies on people working with historical data. (It’s hard to claim the earth is 4000 years old when you have 65 million year old bones, and it’s hard to claim global warming isn’t happening when climate records show a steady and unprecedented rise. Especially when such rise is rather easily explained by basic thermodynamics — the heat capacity of various gasses and changing atmospheric proportions. There’s a reason the science has moved onto “where’s it going’ — what’s happening and why is solidly established.)Report

      • Chris in reply to Jim Heffman says:

        Occam’s Razor comes from his general philosophy. It does show up in one of its more explicit forms in Quaestiones, which is theological work, a commentary on a commentary on Aristotle’s Physics, in a discussion of the number of principles we should posit (an infinite or finite number), as part of a larger discussion of the nature and existence of God, though it’s worth noting that nothing in this discussion looks remotely similar to contemporary Intelligent Design theories, and it’s generally part of his larger project of separating philosophy and theology via arguing against the arguments for the existence of God used by other Scholastics (in this case, Scotus), most prominently various versions of the ontological (in this case, the cosmological argument). W. of O. was a fideist.Report

  5. Murali says:

    Is there room in this for “G == 1”? Well maybe. But what does it do in the system? Make math work? Why can’t math just work on its own? Create the natural laws? What can’t natural laws just be?

    if G==1 is not doing any work, neither is G==0. That’s my point. If someone were then to ask about the value of G, then all we can say is “don’t know”Report

    • Murali in reply to Murali says:

      sorry, this was in reply to @veronica-dire aboveReport

      • James K in reply to Murali says:

        @chris

        I think your last paragraph highlights the core of our disagreement – I simply don’t accept as valid philosophical approaches to solving empirical questions – and any question as to whether something exists or not is an empirical question. I don’t care about clever arguments about some anomaly in our intuitions about reality – if quantum physics proves nothing else it proves that the universe is utterly indifferent to how we think it should work.

        Ultimately I’m only persuaded by evidence – the example I gave was intended to be a small deviation from what physics already says since I was indicating the minimum I would expect to change my mind. Bigger change sot our existing models require more evidence.

        In any case, I think we’ve exhausted this conversation – we’re just arguing from too different starting premises.Report

      • Chris in reply to Murali says:

        Precisely my point, with one addition: your starting premise is unjustified and unjustifiable. It is, to put out bluntly, a leap of faith.Report

      • Stillwater in reply to Murali says:

        Chris, Is this

        and any question as to whether something exists or not is an empirical question.

        the premise you attribute to a leap of faith?

        I mean, I can understand getting all nuanced about what the word “exists” means, but if this is the premise you’re disputing I’m really confused as to how you would argue it. Without begging the exact same question you accused James of begging anyway.

        My solution? THis shit aint resolved philosophically but pragmatically.Report

    • James K in reply to Murali says:

      @murali

      1) Why is there a God term in your model in the first place? One only introduces a term in a model if that term does something. If a concept has no explanatory power the proper response is to zero it out and drop it from your model entirely. The alternative is to spend all your processing power on trivial hypotheses. Classical statistics deal with this problem with a burden of proof. Bayesian statistics uses prior probabilities, but both are just a form of “extraordinary claims require extraordinary evidence”. No epistemology is functional without a way of filtering out implausible hypotheses.

      2) You’re glossing over a lot of detail by just saying “God”. If you want to include gods in your model, you need to separately for every single kind of god that has ever been conceived of, plus some space for gods that haven’t been conceived of. Once you do that, you’ll see that the probability of any one of those gods actually existing is negligible.Report

      • Murali in reply to James K says:

        @james-k

        What makes a hypothesis implausible?

        What makes a claim extraordinary?

        Even if the probability of any one of these Gods existing is negligible, the disjunct of those events need not be. If we are trying to justify any one particular religion, you may very well be right about what the relevant hypothesis is, but that is distinct from the question of whether to believe that all of the God hypotheses are false.

        Consider the following situation. You are a detective on a cruise ship. One day, while out in the ocean, someone is found murdered in their bed. The murderer has to be one of the 1000 people on the ship. However, from the mere fact that for any individual person, it is nearly certain that that person is innocent, it does not follow that it is nearly certain that all of them are innocent.Report

      • Murali in reply to James K says:

        @james-k

        There is, perhaps, a pragmatic sense in which we only need to consider a term if it does something in our model of the universe. But, even when we are not employing it at all, we may still ask what the value of the term is even when it does not seem to be doing any work in our model. Setting it to 1 does not seem like a worse or better option than setting it to 0 (at least in the absence of evidence)Report

      • James K in reply to James K says:

        @murali

        What makes a hypothesis implausible?

        What makes a claim extraordinary?

        That depends on the epistemology you’re using but it should be driven by the information you’ve already got.

        Consider the following situation. You are a detective on a cruise ship. One day, while out in the ocean, someone is found murdered in their bed.

        The problem with this analogy is that it presupposes there has been a murder, it is begging the question.

        Setting it to 1 does not seem like a worse or better option than setting it to 0 (at least in the absence of evidence)

        There are firm grounds in Information Theory for removing under-performing variables from your models, parsimonious models are more likely to be correct ceteris paribus. If gods exist, there has to be a bunch of complicated physical laws that explain how gods work. If gods are physically impossible, there’s no reason to even ask whether any gods exist. That is the fundamental inelegance of theistic beliefs – they have to add a lot of new structure to the universe to act as a foundation for their favoured god(s).Report

      • Murali in reply to James K says:

        @james-k

        That depends on the epistemology you’re using but it should be driven by the information you’ve already got.

        Precisely. In the complete absence of information, everything looks equiprobableReport

      • James K in reply to James K says:

        @murali

        But there’s never a complete absence of information since you have some information already.Report

      • Chris in reply to James K says:

        That is the fundamental inelegance of theistic beliefs – they have to add a lot of new structure to the universe to act as a foundation for their favoured god(s).

        No, you don’t! Or rather, no anyone who doesn’t already accept your radical physicalism doesn’t.Report

      • James K in reply to James K says:

        @chris

        The physical laws that are part of my “radical physicalism” are things everybody needs because they predict the behaviour of observable phenomena. The extra stuff that non-physicalists talk about has no predictive power making it a superfluous addition to one’s model of reality.Report

      • Chris in reply to James K says:

        James, the second part is your radical physicalism. It is not simply an acceptance of physical laws, but an assertion, which appears to be taken as axiomatic, that nothing else (except, I assume, additions or revisions to currently known physical laws) has any epistemic utility, and therefore any epistemic or metaphysical validity.

        That’s cool, it just leaves you with a whole bunch of logical holes to fill. I imagine you aren’t particularly worried about those holes, of course, but it’s been my experience that the confidence and forcefulness with which people who think like you will proclaim the undeniable truth of their vulgar positivism is indirectly proportional to the amount of time they’ve actually spent engaging those holes. I find this more than a little bit annoying: you are so convinced that you are correct that there’s no need to think about problems people might raise. It looks a lot like something you’re likely to spend a fair amount of time decrying.Report

      • James K in reply to James K says:

        @chris

        What problems would you be referring to? It’s a little hard to reply without specifics.Report

      • Stillwater in reply to James K says:

        Well, for starters, one question would be: Are those physical laws exhaustive? Surely there’s a possibility that they aren’t, either as determined by future empirical studies or as a matter of explanatory power.

        But I take your point James, at least as I understand it. Empirical evidence and empirically based theories are much more well justified than any a priori (including religiously) based theory. If those two views were put in a ring together it wouldn’t be contest.Report

      • Road Scholar in reply to James K says:

        The problem with the God hypothesis in the final analysis isn’t so much about it being right or wrong or logical vs. illogical. It’s that it’s a huge cheat, basically a big cop out.

        Given that it can explain anything by just invoking a miracle, it actually explains exactly nothing.Report

      • James K in reply to James K says:

        @stillwater

        I’m almost certain that our existing physical laws aren’t exhaustive, but you can’t get knowledge out of ignorance – the mere fact we don’t know everything doesn’t make souls or gods or any of that stuff real. What we don’t know we don’t know but that’s no reason to believe in a single thing that isn’t justified by what we do know. If someone can show me any evidence of gods or souls or something supernatural then I’m prepared to change my mind, but not before.Report

      • Chris in reply to James K says:

        James, there is a huge literature on physicalism and materialism. Since you’re asserted it without argument and as Gospel Truth, as you’ve done many times here, I’m sure you’re aware of it. But since you’ve made one somewhat specific claim, that any posited non-physical entity, that is any entity that does not strictly obey the laws of physics, cannot possibly have predictive validity, I’ll ask you two questions related to problems for both prongs of your assertion: what is casualty, and what is the physical status of the first-person content (that is, the phenomenal content, the content only directly available to the conscious thinker) of conscious thought.Report

      • Chris in reply to James K says:

        It’s that it’s a huge cheat, basically a big cop out.

        You’re the second person to call it a “big cheat” in this thread, which makes me wonder if that comes from somewhere (or if you got it from Veronica’s comments above). Veronica further dismisses causality outside of an analysis of causal inference (like Pearl’s) as a “vague metaphysical cheat card.” I find these sorts of hand-waving dismissals interesting, because they always seem to come at the end of a claim, not an argument. That is, no one ever says, for a, b, and c reasons, x is a big cheat. They always say, in essence, “I believe a, b, and c,” therefore x (related to a, b, and c in an unspecified or also hand-waved at manner) is a big cheat.

        Of course, it’s possible to talk about what causality, even outside of a treatment of causal inference, without delving into what I imagine Veronica thinks of as “vague metpahysics,” as evidenced by all the work done by people like Salmon and Dowe (see here for a summary), all in the context of philosophy of science. And without a concept of causality: that is, what causing something else entails, not necessarily the statistical properties of data that allow us to infer causality, but what it is we’re inferring in the first place (unless you think it’s just those statistical properties, in which case, good luck with that). And the reason that is important is because the nature of causality entails a bunch of things about how different facts of the world relate to each other, specifically, the contingency of their relationships, which leads to inevitable questions about whether there are any non-contingent facts, and if so, what these would look like. It also leads to questions (and potential answers) to questions about precisely what predictability tells us about an explanation (e.g., if we can predict a state in the world entirely based on one model, does the model fully encompass what that state is?). It leads us to questions (and potential answers) about a whole host of problems in logic and philosophy of science, epistemology and perhaps even philosophy of mind.

        All of this is what you are essentially waving away when you refer to such things as a “big cheat,” because it’s precisely such things that go into virtually all of the sophisticated theological arguments, as well as the basic metaphysical questions (why is there something rather than nothing? why is the universe ordered? why is it ordered the way it is ordered? and so on).

        Look, I’m an atheist, but I find hand-waving dismissal of theology and metaphysics from a certain type of science-obsessed atheist no more compelling than I find hand-waving dismissal of materialism from theists. The fact that science does a very good job at measuring things and modeling measurements with, at least at some levels, high degrees of predictive validity (models of measurements that predict new measurements), does not mean that if you stick only to the sorts of entities on which science can speak, that is, what can be measured, you have a complete model of the world. Hell, you don’t even get a model of science itself that way. But you’ve dismissed the possibility of anything else, again, with a wave of the hand.Report

      • James K in reply to James K says:

        @chris

        To be clear, I am not declaring a priori that supernatural phenomena are impossible – I agree that would be unreasonable.

        What I am saying is that:
        1) Any supernatural phenomenon behaves very differently to anything we have observed to exist in the universe.
        2) That means that a model of the universe with supernatural phenomena in it has significantly more complex physical laws than one without.
        3) Given 1 and 2 there is a very low probability that supernatural phenomena are physically possible.
        4) Given the number of supposed supernatural phenomena that have been satisfactorally explained physically, I am also highly confident in the low probability I assign to the possibility of supernatural phenomena.
        5) I therefore do not generally take claims of supernatural phenomena seriously, and would not pre prepared to do so without a very large amount of evidence – a clear example of a phenomenon that cannot be explained with existing physics and for which no explanation seems possible that doesn’t include some kind of supernatural component.

        Meeting that threshold of evidence isn’t strictly speaking impossible, but it would be very difficult. Are you aware of any theologians (sophisticated or otherwise) who make a serious attempt at supplying evidence of this strength?Report

      • Chris in reply to James K says:

        @james-k

        Any supernatural phenomenon behaves very differently to anything we have observed to exist in the universe.

        Asserted without evidence and, I suspect, built on precisely the sort of circular reasoning I’m talking about.

        That means that a model of the universe with supernatural phenomena in it has significantly more complex physical laws than one without.
        Do you have a measure of how much more complex it is? You say “significantly,” but that word seems vague here. How much more complex? And again, since this is dependent on (1), and (1) requires a bunch of assumptions, it’s not clear to me that your system isn’t the more complex one (I recommend reading a book on medieval philosophy, seriously).
        Given 1 and 2 there is a very low probability that supernatural phenomena are physically possible.
        Even if (1) and (2) are true, this does not follow. Complexity is not the only factor in models that determine probability.
        Given the number of supposed supernatural phenomena that have been satisfactorally explained physically, I am also highly confident in the low probability I assign to the possibility of supernatural phenomena.
        Like what?
        5) I therefore do not generally take claims of supernatural phenomena seriously, and would not pre prepared to do so without a very large amount of evidence – a clear example of a phenomenon that cannot be explained with existing physics and for which no explanation seems possible that doesn’t include some kind of supernatural component.
        You’ve essentially done precisely what I accused you of: you started with an unquestioned physicalism/materialism, moved from it to some controversial positions even given physicalism, and then concluded that you were correct all along from the positions that you derived from the initial position that was in question in the first place.Report

      • James K in reply to James K says:

        @chris

        1) Now I’m honestly confused my what you mean by “supernatural”. If souls could be explained by particles in the Standard Model of physics then I wouldn’t call them supernatural. The same applies to gods. If there were gods, but they were just really advanced aliens I wouldn’t call them supernatural.

        2) I have to admit I haven’t tried to figure out the Komologrov complexity of souls, but at minimum you’re talking about adding a new fundamental particle / filed to the universe and that’s a pretty big deal.

        3) You’re right, complexity isn’t everything – predictive power matter a lot. But that’s where supernatural arguments fall down: what successful predictions have supernatural explanations made that were superior to natural ones? Without that even a tiny increase in complexity wouldn’t be worth it.

        4) Are you seriously asking me for examples of supposed supernatural events that have been found to have physical explanations? I could just say “everything”, but let me be a little more specific:
        A) Meteorological Events: including rains, flooding, strong winds and lightning.
        B) Astronomical Events: including the (relative to the Earth) motion of the Moon and Sun, eclipses and the appearance of stars and planets.
        C) Psychological Events: including seizures and schizophrenia.
        D) Political Events: including the birth and death of kings.

        I could go on, but I would die of old age before I finished.

        I’m becoming increasingly unsure what your point even is. If your point is simply that there’s a non-zero probability of some supernatural phenomenon existing then that’s true but utterly trivial. There’s a non-zero probability that Banjo the Clown is the one true god, but that fact alone doesn’t make it a hypothesis worth taking seriously. If you actually think there’s some specific supernatural phenomenon I should be taking seriously, then give me a reason to do so. Otherwise all I’m hearing from you is a bunch of what ifs and maybes and denunciations of my closed-mindedness. None of which is relevant to the question of whether any kind of supernatural phenomena exist.Report

      • Chris in reply to James K says:

        My point is illustrated in your (1). Your version of definition of phenomena is anything that can be explained by fundamental physics, and your version of the supernatural is yet another element in a physical model (a new fundamental particle). Your version of complexity is anything added to a physical model within its purview.

        You have ruled out gods a priori, and then claim to do so empirically. Like I said, define causality, and the deal with the problems it raises related to contingency, necessity, possibility, infinity, etc. If you can do that, I can respect your skepticism as a well thought out and logically sound metaphysics. Right now, it just looks like what works best for you, psychologically, and that’s cool, but it makes your confidence in the correctness of it suspect.Report

      • Stillwater in reply to James K says:

        James,

        I’m perplexed as well. I think you’ve done a good job of describing (or at least providing an account of) empiricism, one which I agree with, so I can’t precisely locate Chris’s objection. This whole conversation rests on the epistemological side of things, yes? (Can we agree on that?) And if so, then our current best explanatory theories are accorded great weight when evaluating alternatives, ones which include (from our pov) non-physical properties. Like supernatural properties apparently are.

        Chris, I’m with you about qualia and the like, but I’m not sure that’s a relevant consideration when talking about external phenomena, or observational (which is external almost by definition) evidence. Could you elaborate a bit on what you’re getting at?Report

      • Stillwater in reply to James K says:

        The problem with the God hypothesis in the final analysis isn’t so much about it being right or wrong or logical vs. illogical. It’s that it’s a huge cheat, basically a big cop out.

        I don’t look at it as a cop out, but I do think hypotheses like “there is a god” or “I’m a brain in a vat” or “I’m just part of a simulation game” suffer from the same sort of criticism, which is that they cannot be or refudiated by any empirical evidence whatsover. That’s why the logical form of the argument always kicks up a notch, going ever more meta, to stay one step ahead of either observational evidence or the explanatory power of currently held empirical theories.

        So the whole game strikes me as an attempt to introduce a type of logical possibility which exists outside the scope of empiricism to be used (either directly or indirectly) as an attack on empiricism. I think rationalism has a tight grip on some people, folks who like to believe that a priori rationication can reveal facts about the world. Murali’s focus on the simulation game strikes me exactly the same way: it’s logically possible that we’re part of the matrix and that what we view as the external world doesn’t exist. Deal with *that* empricists!

        Well, the answer – if there is one – is that on either horn of the dilemma (that I’m in or not in the simulation game) there is no reliable evidence to demonstrate a conclusion one way or the other. For example, if I’m not in the simulation game, then the probability that I am is infinitesimally small (whatever the probability of a logical possibility wildly inconsistent with the scope of my experience and the explanatory power of my accepted theories might be). And if I am in the game, there is absolutely no evidence (by definition) which could determine that I am or am not, but worse, there’s no reason to trust (otherwise) necessary logical rules upon which the simulation hypothesis even makes any sense.Report

      • Stillwater in reply to James K says:

        This little argument just struck me:

        Re: the simulation hypothesis: as the probability I assign to the belief I’m in a simulation game goes up, the probability of that belief being justified goes down, since if I’m actually in a simulation game, I cannot rely (ie., assign a probability of zero) to the reliability of any of the inferences I make from within the simulation.Report

      • Chris in reply to James K says:

        Empiricism is not necessarily a rejection of a priori reasoning and it does not necessarily entail scientism and materialism. James isn’t just laying out empiricism, which can be theistic. He’s basically saying any posited entity has to fit within the existing structure (epistemological and, I assume, ontological) of contemporary scientific physics. And his defense of this seems to be to assume physics is sufficient, as metaphysics, and then say anything else, because it is merely added to it, will add complexity and is therefore extremely unlikely. That’s not empiricism, that’s just circular reasoning to arrive where he wanted to be to begin with.Report

      • Stillwater in reply to James K says:

        Chris,

        Maybe we’re reading him differently, or I’m not reading his words with sufficient care, but I don’t think he’s making those types of claims, ie., about necessity or sufficiency regarding alternative theories or entities or properties or what not. I think he’s saying (could be wrong!) that given (an acceptance of) our current theories and their resulting explanatory power (which isn’t complete, he admits), a claim introducing additional entities or properties is justified only if there is empirical evidence warranting positing those property’s existence and that their introduction is justified by increasing the overall explanatory power of our theories.

        I take it he’s arguing from a very pragmatic conception of burden of proof here, one he believes (and I agree with him, if this is what he’s saying) empiricism and empirical theories have met, but which supernatural theories lack. I mean, surely an empiricist will admit that not all observable phenomena currently receive an explanatory account consistent with empiricism, but noting that absence of knowledge doesn’t justify (and certainly doesn’t entail) introducing the supernatural to close the gap.Report

      • Stillwater in reply to James K says:

        Aaaaand, after writing that, I realize I’m no closer to understanding your objection to James’ views on this stuff. So allow me to offer one way of distinguishing the views and locating what the source of disagreement is.

        It seems to me that James wants to view these types of issues from not only an empirical but also a pragmatic point of view. You, on the other hand, want to view these types of things from a philosophical pov and are criticizing him for leaving out important evidence which is, or can be, philosophically gleaned. And his response to your challenge is to accuse him of circular reasoning. Yet, I’m not seeing the circle, myself. His view rests on empiricism (of whateer form) as well as pragmatics, and it’s the pragmatic part of his argument that makes him the Immovable Object on this issue: unless there are pragmatically justified reasons for including supernatural properties into an explanatory account of The Way the World IS, he won’t budge. Implicit in his view, it seems to me, is a rejection of the philosophical grounds (and perhaps a priori grounds) upon which your argument is based.

        But I don’t see that as a circle, any more than accusing you of justifying your application of a priori reasoning by invoking a priori reasoning constitutes a circle.Report

      • Stillwater in reply to James K says:

        Oooops.

        And his response to your challenge is to accuse him of circular reasoning.

        Should be “and your response to his response to your challenge…”Report

      • Chris in reply to James K says:

        If your justification of atheism is “my world view works just fine for what I want out of life, without needing God added onto it,” I’m cool with that, whether it’s scientism or whatever. Hell, that’s why I’m a atheist.Report

  6. Jim Heffman says:

    ” It doesn’t follow from the existence of an asymmetry that negative claims deserve some kind of handicap or head start over positive claims.”

    That’s because there is no such thing as negative claim; there are only positive claims made as a direct response to previous claims.Report

  7. Jim Heffman says:

    Also, a statement about the validity of a claim’s proof is not itself a claim that requires proof.

    “Global warming is real and it’ll kill us all unless Americans stop driving cars and eating meat!”
    “The evidence doesn’t really support that conclusion.”
    “Yeah? PROVE THAT IT DOESN’T.”Report

  8. Mike Schilling says:

    the possibility of a teapot orbiting the sun somewhere in the Kuiper belt.

    The Kuiper belt took place on August 29, 1977 and did not involve a teapot. A gopher ball, perhaps.Report

  9. Jaybird says:

    If someone can show me any evidence of gods or souls or something supernatural then I’m prepared to change my mind, but not before.

    I’ll just throw this in here because it continues to irritate me:
    We can have a television show where there are characters called “soul collectors” who are able to tell right before people are able to die, who can collect their souls after they die, and then go on to communicate with these souls…

    And it’s pointed out that there’s no evidence that souls exist, maybe they’re just getting a snapshot of the mind right before death, and so on.

    What *WOULD* constitute evidence of souls in B5 if not the writers coming out and making those characters and having the souls uprise against a bad soul collector?

    Given that I don’t know the answer to that, I’m certain that there is nothing that could count as evidence for the existence of the soul in a world as mundane as our actual one.Report

  10. Matty says:

    I have to admit to struggling a little with the philosophy of this so I tried to convert it into my head into semi-practical advice on what to do and couldn’t.

    Let’s use the invisible horse analogy, you say there is an invisible horse in your room, if there is I need to step round it to get across the room. If we set the probability of there being a horse to 0.5 should I step to the side or not?.

    Going back to the God claim all the religions I’ve heard of require more than intellectual assent to the existence of a deity. They want particular actions and/or attitudes in response to that, these take time and effort away from other things. Assuming the probability of Zeus to be 0.5 is fine as an intellectual exercise but I can’t perform 0.5 of the daily sacrifice and this line of thinking gives me no guidance on whether I should sacrifice or not.Report

    • Murali in reply to Matty says:

      Even if Zeus exists, it doesn’t follow that you have to perform the daily sacrifice.

      Consider the following:

      p(zeus exists) = 0.5
      p(sacrifice|zeus exists) = 0.5
      p(sacrifice|zeus doesn’t exist) = 0
      p(sacrifice) = 0.25, which is closer to the disbelief range of things than the agnostic side. Getting into the specifics of religious duty requires assumptions over and above the existence of the relevant deity in question. These assumptions similarly lack evidence for and against them. The multiplicative effect of these assumptions ends up reducing the credence for specific religious requirements a much smaller value.Report