Atheists: Occam's Razor is Not Your Friend

I’ve often wondered about atheists use of Occam’s Razor in theological discussion. And particularly as a way of battering anti-scientific theists. It seems to me to be an incredible irony.

Occam’s Razor is expressed in many different ways. It suggests that explanations with fewer numbers of causes, or with causes of lower complexity, are better. It might be said, as Newton did: “We are to admit no more causes of natural things than such as are both true and sufficient to explain their appearances.”, or as Occam himself did: “entities must not be multiplied beyond necessity”.

If you can say anything about “God did it” as an explanation; its that it is pretty darn simple. Certainly the explanation of why human beings are as they are is a darn sight simpler as “Because God made them that way”, or “Because God wanted them to be that way”, than the necessarily scientific explanation of the highly contingent history of biological life and its evolutionary branching.

I’ve heard atheists go on, at this point, to claim that Occam’s razor should apply to new ‘entities’ or new ‘agents’. So an explanation with fewer agents is always to be preferred. But this is crap. Nobody would take seriously an explanation, for a find of pottery, of how natural forces formed the clay in the right shapes, engineered that they were exposed to forest fires, and had the patterns applied to them by a series of insects running across their face. We explain stuff by positing actors, often many of them, when it is the best explanation.

So Occam’s razor might be useful in some technical senses in science (it is better to tie a hypothesis into existing knowns than to build a tall theoretical framework for a new observation). But in theology? No, I’m afraid it is rather a shot in the foot.

There are very good reasons not to accept theological explanations. In particular they are non-predictive. But if you honestly put your faith in Occam’s razor, you should be a theist, as far as I can see.

About these ads

55 Comments

Filed under Uncategorized

55 responses to “Atheists: Occam's Razor is Not Your Friend

  1. I agree that there are tons of reasons why a theistic god is really really unlikely for this universe, but I think you’re maybe short-changing Occam’s Razor a bit here. Actually, what Occam’s Razor boils down to (as do so many things) is Bayes’ Theorem – the more “entities” you add in, the more you diminish the probability chain and therefore your final odds ratio of your proposition being correct vs incorrect. The real application of OR (Occam’s Razor, not Odds Ratio!) is in assigning any specific event to the agency of a theistic god vs it having a PNE (Perfectly Natural Explanation). If an apple falls from an apple tree, the probability of a PNE is pretty high in comparison to the probability of an ATG (Act of a Theistic God), and that is even before you start unpacking (as Bayes insists you MUST do) the probability chain that bolts a theistic god onto the basic grounded proposition that we can take the universe itself for granted.

    So I would suggest that OR is still pretty serviceable here – if our theistic friends can come up with a specific definition of “God” that we can start testing, I think we can assign it a low probability. I would hate to see you fall into the vacuous Augustinian trap of pretending that God can be “simple” – these theologians pretend that concepts and words are the same thing, and by manipulating and eliding the one, they affect the other. Of course it doesn’t work that way :-)

  2. Ian

    “the probability chain that bolts a theistic god onto the basic grounded proposition that we can take the universe itself for granted.” Bullseye! And exactly my point. Why would you assume that for a theist God is an additional explanation on top of a PNE? *Of course* it is for you. But on its own terms the PNE is an additional and unneccesary term bolted onto the divine explanation.

    The two options aren’t PNE+God vs PNE. They are PNE vs God. And God, in the way many theists seem to understand it (and with some justification, theologically) is a far less multifarious explanation than all that fancy laws of physics. Maybe you can find liberal Christians who think that their God-explanations are completely additional to PNE (maybe Zero1 is such a person, possibly), but those aren’t the people I see in this kind of debate.

    Why did the apple fall from the tree: because of the entire contingent history of the biosphere caused apples to evolve to use gravity in that way to move reproductively, combined with the hydrodynamics of the entire weather system causing a particular movement of wind that took advantage of a classical condensation of a random quantum event that caused the bond with the branch to weaken at that precise point in time.

    Or God decided you needed a snack.

    Of course they’re wrong. But I think they’re wrong for very different reasons than OR.

  3. Boz

    I’m not at all sure about this, but here goes.

    The Razor is a principle that suggests we should tend towards simpler theories until we can trade some simplicity for increased explanatory power. (source:http://en.wikipedia.org/wiki/Occam%27s_razor)

    ‘god did it’ is very simple, I agree.

    however, ‘god did it’ has zero explanatory power. So, occam’s razor prefers any other theory, that has more than zero explanatory power.

    Having said this, I (atheist) feel my confirmation bias senses tingling. So I am very unsure.

  4. i argued this back to an atheist who was trying to use the razor and trump up binary logic and atheism as the “true” stance. funny thing is William of Ockham was a Franciscan monk. so the “God did it” would fly okay with him. this is why history is so important to know. if things were as they appeared, education would be unnecessary.

  5. Ian

    @Boz – I agree completely. Thinking about your comment, I think it hinges on what is meant by ‘explanatory power’. To a theist, explanatory power is teleology: why something happened is defined in terms of the purpose of its happening. So God is much much more explanatory powerful than nature: the apple fell because God wanted to give me a sign. That explains much more than the random, contingent physical explanation. To a materialist, of course, it explains nothing, because we’re looking for causal explanation. If you replace ‘predictive power’ for your ‘explanatory power’ then I think you have something that isn’t a linguistic bias: objectively science provides better predictions than theology. Similarly if the theologian used ‘teleological power’, then they win hands down. Science cannot compete with that.

    I might post on the two whys soon, it is an interesting topic, I think.

    @Zero – I love your last sentence! I am always slightly worried about reading too much into the ‘faith’ of historical figures. Newton was religious, for example. Because the cultural context was very different. Would William of Ockham be an atheist if he lived today? Maybe. Would Richard Dawkins have been a Christian if he lived then? Almost certainly. I take your point that he didn’t seem to find a contradiction in his method and his professed faith. But if he had, would we know about it? Unlikely. So yes, I’m agreeing with you, but I tend to be skeptical of that line of reasoning.

  6. oh yeah, didn’t mean to infer that he’d be a rigid catholic or whatever if we brought him here today. he was a really interesting guy, didn’t quite fit in his time, a great thinker, yet very Franciscan, a theology i line up a lot with, aside from the celibacy/poverty aspect. love the service parts and even their theology of atonement. those guys were way ahead of the curve on many things. no telling where Ockham would land today, but from what i know about him and his faith from historical readings he left behind and about him, his theories don’t lend themselves automatically to atheism nor to evangelical Christianity.

  7. elitistb

    The issue is that God, as an infinitely powerful deity, is by definition infinitely complex. Thus, anything less complex than “infinite” would be a preferable answer. “God did it” is the most complex answer that could ever be given, because it requires the infinitely complex God as an assumption. If a theist were to actually use Occam’s Razor in a logical fashion, they’d have to admit this.

    Take a car. If you think “God did it” is an extremely simple answer, then it would be the answer. Who made this car? “God did it”, because that explanation is simpler than the huge complex variety of industries that came together to produce that car. Your pottery example is heavily flawed, because we already KNOW about how pottery can be made, just like we know about cars. We’ve seen it done countless times before, and we have a working theory of how it is created. Even without closely looking at the pottery for tool and process marks, assuming the pottery was made by an intelligent being (most likely a human, though maybe we could be surprised on that) would be reasonable. Now, if we know all that about pottery, and one person says “A human made this”, and another said “god made this”, Occam’s razor would favor the human as being the less complex answer.

    Occam’s Razor, like much of logic, does indeed have no place in theology.

  8. Ian

    Hello elitist, and welcome to the blog.

    The issue is that God, as an infinitely powerful deity, is by definition infinitely complex.

    Well, straight off the bat you seem to be assuming

    a) the complexity of a being has a lower bound at its powerfulness (a very odd line of argument to introduce as if it were self-evident)

    b) that the intrinsic complexity of an actor is significant to OR (not something that philosophers of science tend to argue, in my experience: are less complex animals more likely to cause natural phenomena than more complex animals?)

    Your pottery example is heavily flawed, because we already KNOW about how pottery can be made, just like we know about cars.

    I think you’re mistaking me for someone who believes that God is an explanation. I’m not. You’re right, but Occam’s Razor what you’re arguing at that point.

    Now, if we know all that about pottery, and one person says “A human made this”, and another said “god made this”, Occam’s razor would favor the human as being the less complex answer.

    Yes, but that’s not the situation we’re talking about, is it? It is the choice between pottery made by people and pottery made by natural processes (i.e. I said nothing about God in that example).

    Occam’s Razor, like much of logic, does indeed have no place in theology.

    I think what you’re doing is starting from the point that OR is your friend, then finding ways to try to interpret it to make the point you want to make. OR doesn’t do what you want it to. And that’s okay.

    What I think you’re missing, is that OR is really quite a weak criteria. You can’t roll it out for *everything* that is true and say it holds. In fact for many many things that are true, OR does not hold – the more complex explanation is true. That’s fine. That’s sciences job to find out. I went through 8 years of doing science at university, and I never used OR to decide a case. I’m sure it informed my intuition, but among the hundreds of tools I used to argue my dissertation, OR was not one.

    We normally caveat OR with something like “all other things being equal…”. But they rarely *are* equal. Human beings are far more likely to be behind pottery than natural processes, because we know human beings make pottery, and we can estimate the absurd probability of a bowl being formed, fired and decorated by chance. OR would suggest that fewer actors is a better suggestion, but that gets totally trumped by the evidence.

    Occam’s Razor, like much of logic, does indeed have no place in theology.

    Where OR *is* useful in these discussions is to differentiate between situations where God is an additional hypothesis. So Doctors and Drugs healed Johnny of his infection, or Doctors, Drugs and Prayer healed him. There OR suggests the former. But as we’ve seen, OR is so weak, that it doesn’t tell us much, and certainly doesn’t settle the argument (I’ve seen it often used that way in debate “but that can’t be true, by Occam’s Razor, fnar, fnar.”). And in fact, there’s no reason to go near OR in this case, since we have excellent statistical evidence that prayer doesn’t heal people.

  9. elitistb

    I’ve not done quotes in replies like these, so I’m hoping that quote tags work. If they do not, my apologies.

    I think you’re mistaking me for someone who believes that God is an explanation. I’m not.

    If you quoted someone, and I said “your quote”, I’m not saying you were the one who made the quote, I’m saying you’re the one who brought it up. The same applies to the pottery example. I’m not saying you are arguing it, I’m saying it is “yours” in the sense that you made it an object of discussion.

    OR would suggest that fewer actors is a better suggestion, but that gets totally trumped by the evidence.

    and

    We explain stuff by positing actors, often many of them, when it is the best explanation.

    I should have read your article at least a third time, as the first two I missed the implications of what you were saying. Basically, you are arguing against a straw man of Occam’s Razor. I’m not saying you’ve done this intentionally, simply that what you are arguing about is NOT Occam’s Razor, but an incorrect variation of it.

    Occam’s Razor states that one should tend towards the simplest explanation, UNLESS the added complexity actually adds explanatory power AND still fits all the known evidence. Occam’s Razor NEVER preferences a simpler theory that has less evidential support than a more complex theory that has more evidential support.

    It is quite a bit stronger than you have stated. However as noted it is just a tool, and any use of it needs to be justified.

    Well, straight off the bat you seem to be assuming

    I’m actually assuming neither of your options. A is incorrect, because it isn’t the lower bound we’re concerned with, but the upper bound. Saying “God did it” is a nonexplanation, because he could have done it in an infinite variety of ways, including convincing a human to do something for him, creating molecules from thin air, or reversing gravitron flow through the 16th dimension. It provides no explanatory power because it includes all explanations. B is also wrong, but this is because you’ve miscast OR as giving preference to the simpler explanation in opposition to evidence.

    Let’s move away from your pottery, which as far as we know only humans make. Let’s take something that looks like a beaver dam, something that both humans and beavers could be an explanation for. If you see such a construction, what species do you assume made it? By Occam’s Razor, you’d preference it being a beaver construction, because we know at the last that beavers DO build dams of that fashion, and they do it regularly. Humans add complexity, because humans do not generally build dams in that fashion and we have no evidence that preferences humans with just the knowledge “it looks like a beaver dam”. With no other information, that’s where it ends, with the statement “Things that look like beaver dams are more often constructed by beavers than humans.” If there is other evidence, such as all the logs having obvious axe marks or very flat ends indicating a saw, then the beaver explanation is no longer as well supported, and you shift over to humans when you use Occam’s Razor.

  10. elitistb

    Ugh, that didn’t work very well at all.

  11. Ian

    Added the blockquotes – it is html-esque, so do <blockquote> … </blockquote> to wrap quotes.

    A is incorrect, because it isn’t the lower bound we’re concerned with, but the upper bound

    If power puts an upper bound on complexity, then God could be infinitely powerful but not infinitely complex. Which was the opposite of your assertion.

    B is also wrong, but this is because you’ve miscast OR as giving preference to the simpler explanation in opposition to evidence.

    Nope. OR gives a preference to a simpler explanation, in the absence of other differentiating evidence. In the presence of differentiating evidence (as in every example you’ve given) OR doesn’t get a casting vote. I know you understand that, the point I’m trying to make is that you should be willing therefore to say: OR can be wrong. The complex explanation can be right. That isn’t some deep down vindication of OR, or a different ‘stronger’ way to use OR. It is just that the evidence leads you to the correct decision: a complex cause is true.

    By Occam’s Razor, you’d preference it being a beaver construction, because we know at the last that beavers DO build dams of that fashion, and they do it regularly. Humans add complexity, because humans do not generally build dams in that fashion and we have no evidence that preferences humans with just the knowledge “it looks like a beaver dam”

    Seriously, that isn’t Occam’s Razor. Its true, but it isn’t OR. Its just balance of probabilities.

  12. elitistb

    Now it looks fine.

  13. elitistb

    If power puts an upper bound on complexity, then God could be infinitely powerful but not infinitely complex. Which was the opposite of your assertion.

    I’m not entirely certain I follow how something infinitely powerful could not be capable of infinite complexity. If power = upper bound of complexity, infinite power = infinite upper bound of complexity, it would seem to me. Saying “god did it” is an infinite answer, because he could have done it infinitely many ways (for potentially infinitely many reasons).

    The complex explanation can be right.

    Absolutely the complex explanation can be right, and OR be wrong. But you don’t know that until you have additional evidence which warrants the more complex explanation. I have never heard anyone using Occam’s Razor who says “This is the simpler of the two, thus it MUST be that way”. Anyone who does use OR that way is using it incorrectly, as it merely states that you should preference the simpler in the absence of differentiating evidence.

    “Everything should be kept as simple as possible, but no simpler.” Positing new agents, without justifying their existence by evidence, is not warranted.

  14. elitistb

    I see your point about my beaver example, so let’s change it up so that it is more in the spirit of adding additional agents.

    You see a beaver dam. Which of these sounds like a better explanation, and why?
    1. Beavers built it.
    2. Beavers were told to build it by humans, and thus beavers built it.

  15. Ian

    @Elitist – Ah, I see. So I think you have upper bound and lower bound mixed up. Upper bound means the maximum that something could possibly be (but it could be lower, maybe much lower). So specifying an upper bound tends to make you think something is smaller. So I think what you’re saying is that power is the *lower* bound of complexity (i.e. if you’re infinitely powerful you must be no less than infinitely complex). Which was what I accused you off initially :) And its fine to want to claim that. But it isn’t self-evident, and it doesn’t seem to me to be very useful if it was, because you’d still need to justify why the internal complexity of an agent had any bearing on OR.

    Saying “god did it” is an infinite answer, because he could have done it infinitely many ways (for potentially infinitely many reasons).

    But you’re assuming that the people who say “god did it” don’t know exactly why and how God did it! Of course they don’t, but they claim to.

    I think the problem is that you’re assuming that the person you’re debating with has fundamentally the same model of the world as you do. They don’t. If they did, then your conclusions would be natural. If you had their model of the world then you’d see just how wrong your naive materialism is :)

    Seriously, my point is that you have to start either on common ground (in which case it seems to me OR would favour divine explanations) or you end up both hammering each other over the head with stuff that has no bearing at all on the other person’s worldview.

    I have never heard anyone using Occam’s Razor who says “This is the simpler of the two, thus it MUST be that way”

    I have, many many times. This is the normal use of OR in a/theology debates, in my experience.

    Positing new agents, without justifying their existence by evidence, is not warranted.

    But God is only ‘new’ if you start from the position that he doesn’t exist. If you start from the theistic position, then God is a fine and very simple explanation for anything.

    What I think you’re actually saying, quite rightly, is that there is no good evidence for God. True. But that isn’t OR again. We never get as far as OR, because the two explanations aren’t equal. If they were, then OR would suggest the simpler one would win. And the simpler one is where all natural explanations are subvenient on God, rather than (as I believe, and I guess you do to) that God is subvenient on natural processes.

    1. Beavers built it.
    2. Beavers were told to build it by humans, and thus beavers built it.

    But given that we know beavers don’t take instructions from humans, why is this OR?

    OR would indicate the correct result in this case. If you judged this on OR you’d get the right answer. But can you see that that is largely arbitrary? Since the important piece of information here is nothing to do with simplicity? For example:

    1. Beavers built it because of the drop in water levels caused by the formation of an artificial damn upstream caused by the additional silt outflow of a tributary caused by the construction of a factory in the next state which was made economically viable due to a pork-barrel concession extracted when the senator agreed to back a struggling president’s unpopular legislation.

    2. Beavers built it because a passing hiker told them to.

    I hope you’ll agree that 2 is still less likely, for exactly the same reasons. Even though 1 now does include human agents, and a bunch of other stuff too.

    Your example also grazes on a very interesting aspect of OR, which Shane alluded to, but I deliberately didn’t pick up. The probability of a conjunction is never more than the probability of its terms [i.e. p(A and B) <= p(A), p(A and B) <= p(B)]. We can pick up that thread if you like. It is one way to mathematically state something like OR. It works for your example, but not for mine.

  16. elitistb

    Upper bound means the maximum that something could possibly be (but it could be lower, maybe much lower).

    I might be thinking of it the wrong way. If you only use a “who/what” as the explanation, and not include the “how”, then you can make no assumptions on that “how”. It seemed to me, at the time, that you could not limit yourself to the lower bound, but had to assume it could be any of the full set of possibilities, hence why the upper bound was the important factor.

    But you’re assuming that the people who say “god did it” don’t know exactly why and how God did it! Of course they don’t, but they claim to.

    Here seems to be an issue between how you and I are approaching this. I perfectly understand their side. I’m saying this is why I would not accept that. I agree with you that Occam’s Razor is of dubious use in theological arguments with anyone who would accept “god did it” as an acceptable explanation. Unless, of course, the use of it is not for the opposition.

    You stated that you have never used OR to decide a case. Have you had a case that had two explanations of equal power but unequal complexity? And even then, just as with the process of science, OR can be wrong, especially when there is unknown information. I think I mistook “weak” as “shouldn’t be used” as opposed to “should only be used in appropriate circumstances.”

    I have, many many times. This is the normal use of OR in a/theology debates, in my experience.

    This is a very inappropriate use for it, and that is a sad state of affairs.

    But God is only ‘new’ if you start from the position that he doesn’t exist.

    New is not the right word, I see that now from your response. “Unnecessary” is probably better, if you have another perfectly reasonable explanation. I know this won’t convince people making these arguments. Most of these discussions I have had are in no way related to convincing them, it is either for my own benefit or the benefit of people who have simply never thought about the situation, such as people who are repeating the argument as opposed to actually making it themselves.

    But given that we know beavers don’t take instructions from humans, why is this OR?

    You don’t think they can be trained? We don’t know, with just that information, whether or not the beavers take instructions from humans, in truth. But beavers are perfectly capable of building dams on their own. Postulating that humans told them to, without any evidence, is just fanciful conjecture. “Entities must not be multiplied beyond necessity.”

    It works for your example, but not for mine.

    I agree with this as well. As I stated above, we know how pottery is made, the two original explanations were not equal, and Occam’s Razor ends at the “are the two explanations of equal explanatory power”, step. This is why I said it was flawed, stating that the pottery was made by natural forces is an unequal explanation. Hence my statement further on that a choice between “God made this pottery” versus “Humans made this pottery”, OR would preference the latter.

    I think your original article confused me. You start off speaking about OR, then said:

    If you can say anything about “God did it” as an explanation; its that it is pretty darn simple.

    which really threw me for a loop, as this is a non-explanation. “God did it” is a statement of who, but not why or how. It explains nothing, and I’ve been trying to find out how this had any relation to OR. I still can’t really see why anyone would use OR in this situation, and I cannot even recall the last time I’ve heard it used in a PNE vs GOD situation, as usually at that point the simple “how” is raised.

  17. elitistb

    And I forgot the closed blockquote statement around the last quote I used.

  18. Ian

    (blockquote fixed)

    It explains nothing

    Not to you it doesn’t! But it explains everything to a certain kind of theist.

    Some theists are continually frustrated by the fact that science purports to explain stuff but it is maniacally obsessed with how questions and never says why. To that mindset all the physical, scientific verbiage in the world explains literally nothing.

    But if you take a scientific explanation on its own terms, against a theological explanation on its own terms. The theological one will almost always be simpler. That’s all my point is.

    Have you had a case that had two explanations of equal power but unequal complexity? And even then, just as with the process of science, OR can be wrong, especially when there is unknown information. I think I mistook “weak” as “shouldn’t be used” as opposed to “should only be used in appropriate circumstances.”

    No, exactly. I can’t think of any situation where I’ve genuinely had two possible explanations for the data, and the only way I can possibly tell which is better is determine which is more complex. Even the thought of doing that seems somewhat laughable. Certainly not something to hang a paper off!

    In science (in my experience) explanations have consequences. So to determine which of two explanations is true, you carefully predict the consequences of each, then go looking to see which consequences are present.

  19. Pingback: Why Are We Here? | Irreducible Complexity

  20. Neil

    What a fatuous argument. Why is “God did it.’ the simplest explanation? What about “it just appeared out of thin air”. Or “Charlie Chaplin conjured it all up one day over breakfast”? I’d love to hear how you feel that “God did it” is simpler than my two alternatives. Please don’t try to use logical argument to justify faith, you’re using the wrong tools.

  21. Ian

    I think you assumed, rather than read what I was trying to say, Neil.

    Who said anything about simple*est*? Why would you think I would think it is any simpler than “it appeared out of thin air?

    It is perfectly possible to use logical arguments to justify faith, regardless of whether you tell people to or not. If faith were not amenable to logical argument then, it also wouldn’t be refutable by logical argument, as far as I can see. To me it’s obvious faith lies in the domain of this that logic can address. There could be good and bad arguments for and against faith. (hypothetically, I suppose, I’ve not found many good arguments for faith). But Occam’s razor, as well as being a generally very week argument for anything, is an exceptionally weak argument *against* faith.

    Please reread the comment thread. And possibly the ‘about’ page. I appreciate you taking the time to comment, but it doesn’t help if you get exorcised by a position nobody here is taking.

  22. Neil

    You were trying to argue that Occam’s Razor is not a valid argument to use against faith and used the idea that “God did it” as the simplest possible explanation. I merely questioned whether it was anything of the sort.

    I don’t think faith is refutable by logical argument in the same way that the morality of my preference for filter coffee over espresso is not definable by the binomial theorem.

    We agree that there’s no point in applying Occam’s razor to faith at least. I just go further in asserting that there’s no point in applying any logical argument to it at all. No matter how much evidence we amass that explains the majesterial beauty of natural processes, there will always be a little room for God to lurk in the shadows of scientific unknowns. Why upset him by trying to winkle him out of his cubby hole with the ill-suited weapon of logic?

  23. Ian

    Okay, you’ve obviously read the thread now. You seemed to be arguing that God wasn’t a simple explanation for everything.

    On logic. Why concede logic? If someone says things that are logically contradictory, I’m happy to call them on it. Sure they are quite capable of saying something like “well its a different kind of logic”, or (the favorite) “that’s a paradox” or “its a mystery”. Fine. I’ll let them concede that their religion is illogical.

    Besides, OR isn’t a logical argument anyway. Its an epistemological heuristic. And heuristics get trumped by facts. And the facts are on our side, why I don’t have a problem conceding that the heuristic breaks down some times. Its the very definition of an heuristic.

  24. Neil

    What facts would they be?

  25. Ian

    We know what causes a very large class of phenomena. They are so well established as to be reasonably facts.

  26. Neil

    Sorry, you said, “facts are on our side” and I don’t understand what you meant.

  27. Ian

    I mean a heuristic (a rule of thumb) doesn’t trump evidence. And while the heuristic (I contend) favours a simple “magic did it” explanation, it is trumped by the facts of how it actually happened. Which are non-magical.

    And non-magical explanations, are what I consider to be my team.

  28. Neil

    No. Sorry, I’m still none the wiser. I’ll just leave it at that I think and not trouble you again. Nice talking to you.

  29. Ian

    Thanks for stopping by!

  30. Occam’s razor is not meant to be a positive theorem. Instead it is meant to rule out possible explanations if they are overly complex and a simpler one exists. It is entirely dependent on specific applications. Using occam’s razor on a broad philosophical question like ontology or epistemology is certainly a poor application. You are certainly correct in that much.

    Science isn’t really an atheists friend either, using that reasoning. After all, how can any scientist comment on something that necessarily exists outside of one’s frame of reference?

  31. Ian

    “After all, how can any scientist comment on something that necessarily exists outside of one’s frame of reference?”

    That really depends on how you define God. For the vast majority of believers, God is defined in ways that would very definitely overlap science’s frame of reference. And science has been pretty good so far at showing the contrary. If you mean a God that is so defined as to be totally independent of any empirical evidence, then fine. Not only do you have a God totally alien to almost all the world’s theists, but you have one that is by definition irrelevant to anything that could possible be observed.

    Seems to me science is an atheists best friend. Because, at the end of the day, all science is, is a way of finding out what is true independent of who’s looking, and what is a function of human perceptual biases and imagination.

  32. help3434

    “heuristics get trumped by facts” What do you mean trumped by facts? I though Occam’s Razor was a tool to explain the facts. Of course if you are arguing with a theist about say, how human life started, you would have make sure that the theist knew all the facts that you did before bringing up Occam’s Razor.

  33. Ian

    ““heuristics get trumped by facts” What do you mean trumped by facts? ”

    I just mean that a rule of thumb is an approximation that doesn’t always hold. There are times when the simplest explanation isn’t right.

    OR differentiates between two different explanations for the same set of facts. We often get new facts (we’ve had a lot of them since most religions came up with their explanations), and those facts win.

  34. Patrick

    Occam’s razor always favors a naturalistic explanation. Theism requires that we posit the existence of supernatural entities in addition to the natural ones. Naturalism says that the natural entities are all that exists. QED.

  35. Ian

    Say’s who? Why does a supernatural entity require a natural explanation? I think you’re regurgitating propaganda here, rather than thinking about it.

  36. Patrick

    Sorry, but you misunderstood the argument. I will try to clarify.

    Theism posits the existence of two kinds of entities:
    1. the natural, for example, humans
    2. the supernatural, for example, God

    Naturalism only posits the existence of one kind of entity:
    1. the natural

    Occam’s razor favors the explanation that requires the smallest ontological commitment. In other words, the fewer things that we have to posit in order for the explanation to work, the better. Therefore, Occam’s razor favors the naturalistic explanation because it requires that we posit only the existence of the natural whereas theism requires that we posit the existence of both the natural and the supernatural.

    I’m a philosopher and have studied philosophy of religion, logic, and metaphysics among other things.

  37. Patrick

    I’ll try to anticipate one of your potential replies. If you are going to say that God is natural, then God is governed by all the same natural laws that govern everything else in nature. But this is clearly incompatible with the notion of God, because he created the laws of nature. How could he be governed by laws which he himself creates? He might elect to obey such laws after creating them, but at the very least he was free of such laws before he created them, so there must have been some other set of laws which governed God during that time. So, logic dictates that God must be governed by a separate set of laws than the set of laws which governs nature. So you now have to have two different sets of laws, one for supernatural things like God and one for natural things like humans. And thus you get two fundamentally different kinds of being.

    Btw, by “law” I mean an inviolable principle of a thing, not something which one is morally obligated to obey, in case you were wondering.

  38. Patrick

    E.g., gravity is a “law” that governs natural objects.

  39. Patrick

    Actually, that’s not the greatest example, but I think you get my point re: laws

  40. Ian

    Patrick, Thanks for following up. I did misunderstand what you were saying. I thought you were simply reaffirming what I was arguing against without engaging the argument. Sorry for that.

    So, okay, you’re using Occam’s Razor at a different level to the way it is normally used. It is normally used with regard to explanations, rather than ‘isms’. But that’s fine, we can talk about explanatory frameworks.

    Given that I agree with you on the conclusion, I may not be the best person to argue this, but I note that a significant amount of theology posits that the natural world supervenes on the supernatural, in a way that is ontologically innocent. To that extent, they posit only one class of explanation also. Again: I don’t agree, it just isn’t obvious at all to me that Occam’s Razor has much to say here.

    Also, I think you should be careful not to let non-theists off the hook by saying “theism posits”, since there are a whole lot of non-theistic believers in the supernatural you could snag with your argument.

  41. Patrick

    “…a significant amount of theology posits that the natural world supervenes on the supernatural, in a way that is ontologically innocent.”

    Some theists might try this, but I don’t think it really works.

    If we want to say that the natural supervenes on the supernatural, then we get something like:

    living things
    chemical processes
    physical forces
    supernatural forces

    If we are naturalists, then we get something like:

    living things
    chemical processes
    physical forces

    This example is obviously simplified, but it doesn’t matter how detailed you wish to be. The theistic ontology will always include that extra “level” of reality which is absent in the naturalistic ontology. So I think Occam’s razor actually does work against theism.

  42. Ian

    Well, its obvious you don’t think it works. Neither do I. But they do. And now you’ve switched back to explanations from explanatory systems again. Like i say, its not my belief to expound, but you’re now back saying stuff I covered above, without engaging my arguments. So I’m not sure what you’re expecting from me.

  43. Patrick

    Maybe I’m not getting your argument.

    Surely when someone says that “God did it,” they are not denying the existence of chemical, biological, and physical forces. They mean that God manipulates these things in a particular way. This means that they are adding an additional entity beyond what is required for the explanation. So, Occam’s razor applies.

  44. Patrick

    Some theists do try to deny chemical, biological and physical forces (evolution being one example), but if they do this then they are greatly reducing the explanatory power of their theory, because God’s way are unknowable to us. If you want to deny the existence of natural forces and say that God does everything, that’s basically like saying “I don’t know how X happens, but God exists!”

  45. Patrick

    After thinking some more, I think the matter hinges on how one views teleology. If one rejects teleology then saying that God did it explains exactly nothing, in which case Occam’s razor would favor naturalism. A review of why teleological explanations have been mostly rejected in science might be helpful here.

    Something to keep in mind – Occam was a theist but he held that belief was a matter of faith and could not be justified through logic.

  46. Ian

    “I think the matter hinges on how one views teleology” I agree on all fronts. If one rejects teleology, then one is effectively rejecting the theistic explanation generally.

    “A review of why teleological explanations have been mostly rejected in science might be helpful here.” I also agree that teleology is specious. The point of this post was to say theistic explanations are de-facto simpler (unless, as you say, you a priori set the rules of the game to exclude theistic arguments that are simple). The subtext was that Occam’s Razor is a heuristic: there are plenty of good reasons to reject theistic explanations, some of which you’re now touching on (explanatory power, invalidity of the explanation, let’s also add lack of evidential basis and lack of internal consistency).

    “greatly reducing the explanatory power of their theory” In the technical scientific sense, yes. But I contend that, we are cultured (perhaps even predisposed) to think teleologically, so scientific explanations can feel very powerless unless you’ve been systematically retrained to think scientifically. Several times I’ve heard the objection that you have to spend literally years studying ever more layers of explanation, but science never fully explains anything. Hence why a theistic explanation is both far more simple and far more satisfactory to (I’d say) a majority of people.

  47. Patrick

    “The subtext was that Occam’s Razor is a heuristic: there are plenty of good reasons to reject theistic explanations, some of which you’re now touching on (explanatory power, invalidity of the explanation, let’s also add lack of evidential basis and lack of internal consistency).”

    So maybe it’s that Occam’s razor works to eliminate theism once we take these other factors into consideration. By itself, Occam’s razor isn’t enough. But when we consider lack of evidence, logical contradictions, and lack of explanatory power then Occam’s razor works. For example, a theory might be simpler than another but contain logical contradictions, so Occam’s razor would actually favor the more complicated theory since it says “do not multiply entities beyond necessity.” In this case, a more complicated explanation is necessary in order to avoid logical contradictions.

    I wonder if it is possible to use Occam’s razor on teleology itself. I think a purely teleological explanation clearly has less explanatory power since this leaves out material, efficient, and formal causes. An explanation that includes all four causes might seem to have more explanatory power, but it also has more logical problems. Teleology only makes sense when describing the actions of a rational actor. To suppose that nature exhibits purpose (aka teleology) is either a faulty analogy or begging the question. The logical problems that result from supposing that teleology applies to nature means that an explanation which includes all four causes (when applied to natural phenomena) is both more complicated and contains logical fallacies. Therefore Occam’s razor would favor an explanation that omits the teleological. Occam’s razor would eliminate the teleological in all cases except when attempting to describe the actions of a human.

    Additionally, it seems that teleology might presuppose the design argument. If this is so, then we could reject teleology on the basis that the design argument fails (faulty analogy, begging the question).

    This has been a fun exercise.

  48. Patrick

    To expand a bit on my last point:

    Teleology only makes sense when describing the actions of a rational actor. Teleology makes perfect sense when describing the actions of a human.

    In explaining a natural phenomena, to say “god did it” is to say that “this was the result of a rational actor.” To say that “this was the result of a rational actor” is the same as saying “this action was the result of design/purpose.”

    Therefore, to say that “god did it” when explaining a natural phenomena is to say “this natural phenomena has design/purpose.” In other words, saying that “god did it” presupposes that the event being described exhibits design/purpose. So the starting point for teleological explanations is that nature exhibits design/purpose.

    And now we are at the argument from design, which fails because it is both begging the question and a faulty analogy.

    It’s early for me, so I may have missed something, but I think I’ve covered my bases.

  49. Ian

    “For example, a theory might be simpler than another but contain logical contradictions, so Occam’s razor would actually favor the more complicated theory ”

    So here you seem to be using “Occam’s razor” to mean “sensible reasoning” or some such. The motivation for this post was that a lot of atheists seem to use it this way: Occam’s Razor is right and therefore supports their view. This is what I disagree with.

    Occam’s Razor is a heuristic, a rule of thumb, that states the simpler explanation is usually right. This is not always true. When it isn’t true, that is okay, Occam’s Razor doesn’t hold there. Even when it is true, we don’t rely on it. You can’t publish a scientific paper saying “therefore by OR, here’s the answer”. It is a general heuristic.

    So you can say, ah well only if we exclude all other possible ways of determining which explanation is right, then OR is true. In this particular case, such and such an explanation is logically inconsistent, so it doesn’t get to play, and so on. But that is only making OR valid by restricting the domain of its application massively.

    You can take this further. It can be elevated to a law within empiricism in a weaker form: that if two explanations are empirically indistinguishable, then the more complex one is unnecessary. But note that doesn’t mean the more complex one is wrong.

    In theological arguments, as we’ve seen, it is possible to structure one’s argument so that it either side appears more simple, or is more simple under certain assumptions. And such a restructuring can make OR right or wrong.

    So while I agree OR is useful, and a good heuristic, we must recognize it is a very loose heuristic, and doesn’t support building firm conclusions based on it. As a debating point with theists it is tendentious, as we’ve seen, because it is just a way of wrapping up one’s presuppositions (lack of teleology, for example), and hiding behind an unrelated heuristic.

  50. Patrick

    “So you can say, ah well only if we exclude all other possible ways of determining which explanation is right, then OR is true. In this particular case, such and such an explanation is logically inconsistent, so it doesn’t get to play, and so on. But that is only making OR valid by restricting the domain of its application massively. ”

    I don’t think this is overly restricting the use of OR, I think that this is the correct use of OR. To apply it more broadly is to misuse it. OR might make more sense if we refer to it by its other name, “The Law of Parsimony.” This helps to make it clear that the point is not that the simpler explanation is always preferred, it is that, all other things being equal, the simpler theory is preferred. “Do not multiply entities beyond necessity.” It is necessary for an explanation to explain all the facts, be logically consistent, have evidence, etc. People who ignore these factors and apply OR to eliminate a theory are not using OR correctly.

    A second point is that OR could be directed toward causality itself to eliminate teleological causes from explanations of natural phenomena (i.e., things which are not caused by human intent). Arguing that there is design/intent in nature relies on a faulty analogy between natural things and man-made objects. It is clear that man-made objects have teleological causes. But we have no reason to think that natural objects exhibit anything resembling the design found in human artifacts. Natural objects do not resemble artificial ones; the universe could just as easily be likened to a giant inert vegetable as to a giant mechanical device. There is nothing that compels us to see the universe/nature as exhibiting intent/purpose.

  51. Ian

    I suspect we’re in danger of violently agreeing! Yes, I agree with all this.

    “all other things being equal” – I think is the key, and is the ‘correct’ use of OR (or at least being most similar to its formally correct cousin, is most like to be right). But all other things are very rarely equal, and therefore OR is probably not as useful as it appears. And if you can’t show that all other things are equal, then OR is as likely to give you the wrong answer.

    As for your last paragraph, again, no arguments from me. I’ve no desire to defend the teleological perspective. I think a good summary of my view is that the supernatural position is wrong, not because of OR, but because of the ‘other things’ that definitely aren’t equal!

  52. Keith

    Isn’t accepting Occam’s Razor as something valid when explaining the real world, in itself, also adding another entity to the explanation?

  53. Ian

    It is an interesting point. I’m not sure it is quite there as a slam-dunk self-defeater, because OR isn’t an explanation exactly. But I think there’s probably an important point about the amount of intellectual machinery we use to think about problems, and whether the amount is itself a source of error. Probably is.

    But I think post-hoc rationalisation is an order of magnitude more powerful than any technique we can marshall against it, and the consequences of failing by using it are so severe, I generally think that the more heuristics for avoiding it, the better.

  54. John Cowne

    Firstly, I love the tone of this interchange. Congratulations to ‘both sides’ for your respectful yet strong defences of your positions. Secondly, as a Christian, I have often wondered about OR’s relative value when we argue at the level of our presuppositions. I find the most ‘frugal or economic’ explanation posited in the ‘Kalam Cosmological Argument’ for Deity vs non-Deity convincing. In doing so, my ‘bias’ peeks through the heuristic cracks as I form my model of ‘first cause’ or ‘agency’. I’m reminded of the caution to new medical students in their developing of diagnostic skills: “When you hear hoofbeats, don’t expect to see zebras”. Of course that would have to be revised (reversed) if you live in Kruger National Park (Africa). “God did it” is seen by atheists as naming a horse (the Natural Explanation) a zebra (God). Yet “the Most ‘Natural’ Explanation” does not satisfy my understanding of ‘irreducible complexity’, given any time-scale that any scientific discipline wishes to use.

  55. Ian

    Thanks John. I don’t really blog any more, but I wrote a post a while back about cosmological arguments: http://irrco.wordpress.com/2010/01/19/the-form-of-the-cosmological-argument/

    I’m not sure what you mean by ‘irreducible complexity’ in the last sentence. With that and ‘time-scale’ you seem to be seguing into an evolution argument.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s