Sunday, March 3, 2013

The Problem of Induction is Not Just for Naturalists

Something that really gets under my skin is when a theist latches onto a general problem in philosophy and uses it as a means of 'discrediting' the naturalist or atheist worldview. Often times the eternally on-going debate over the foundations and justification of morality is the focus, while other times logic itself is the concern, as in the Transcendental Argument for God. However, there are also those theists who raise the problem of induction as a flaw in the naturalist worldview

Inductive reasoning is, to put it simply, reasoning from a specific case to a more general case. Most life depends on water to exist, so it seems that any new life form we discover will likely depend on it too. Induction differs from deduction in providing less robust conclusions. Deductive logic deals with possibility, while inductive logic deals with probability. We use induction to make inferences about the world around us all the time. It can be found in marketing and advertising, history, psychology, sports, and even science.

The problem of induction is commonly associated with David Hume, who questioned how induction might be justified in his 1748 work, An Enquiry Concerning Human Understanding. What is it that permits us to generalize from a particular, to connect our experience of life depending on water with the proposition or expectation that new life forms will depend on water as well? Hume noted that we make this connection not by deduction, but by induction, the very method we're employing

Centuries prior, Sextus Empiricus gave a brief and fitting summary of the problem:

When they propose to establish the universal from the particulars by means of induction, they will effect this by a review of either all or some of the particulars. But if they review some, the induction will be insecure, since some of the particulars omitted in the induction may contravene the universal; while if they are to review all, they will be toiling at the impossible, since the particulars are infinite and indefinite.

The problem of induction relates to what is usually called "the uniformity of nature". Science and logic seem to rely on the assumption that the natural world is in such a state of uniformity as to allow us to make reasonable inferences. If this assumption is untrue, then a great deal of what we think we know is wrong. But if this assumption is true, it would seem to call for an explanation, one that induction itself can't provide without begging the question.

Several responses have been made since Hume, with the most notable being Karl Popper. As a philosopher of science, Popper's approach primarily centers around scientific method, which he saw as bearing the responsibility of finding and correcting errors, rather than making inductive inferences. In fact, he argued that induction is a mere myth, and the way by which we actually create knowledge is through conjecture and criticism. Popper's solution has received plenty of criticism too, with the most substantial being that it doesn't really solve the problem of induction. His theory of falsification can go to show that, based on past experience, some scientific ideas are false, but it can't demonstrate when a scientific idea is true, a point he even conceded (Popper, 1979).

There is also the response of Hans Reichenbach, who offered a pragmatic justification of induction. His argument is strikingly familiar to Pascal's Wager in that he suggests that if we go with induction and it turns out to be true, we stand a better chance of success then if we were to go with any alternative reasoning, regardless of whether induction is true or not. Like Pascal's Wager, though, this pragmatic approach is only valuable in that it might point us to a useful strategy; it cannot tell us whether or not that strategy is founded on anything true (unless one accepts a pragmatist theory of truth, which comes with its own additional objections).

I am one of many people who concedes that the problem of induction has yet to be satisfied on epistemic grounds. Things like the uniformity of nature are really assumptions we make based on inferences for which we do not have solid justification. This is a real problem in philosophy.

Does the theist manage to avoid this problem by positing the existence of a god? In an article critiquing secular responses to the problem of induction, author James N. Anderson states at the tail end, "Of course, a Person for whom universal a priori knowledge of the very constitution of the universe is attainable (and perhaps even essential) would be an invaluable ally in such an epistemological predicament — especially so if that Person were inclined toward revelation of Himself and His universe." (Anderson, 2000)

Numerous assumptions are packed into this single sentence, but it is quite typical of the standard theistic belief regarding justification of induction, I've found. God created the universe and knows all about it, god has revealed himself and his works to human beings, and human beings like Anderson have been able to comprehend it as his god intended. These beliefs are fervently defended by apologists as being more than simple assumptions, but the ways in which they are defended are telling.

How do you know god created this universe? Look at the complexity, beauty, and order of nature, they might say. How do you know god has revealed himself to us in scripture? Look at the inerrancy of a text like the bible and the fulfilled prophecies in it. How do you know you've accurately received the revelation of god? I feel his presence and see confirmations of my faith in my life.

These are not ad hoc defenses of theistic beliefs. They make frequent appearances in the apologetic efforts of men like William Lane Craig, Gary Habermas, Josh McDowell, Alister McGrath, N.T. Wright, and many, many others. The beliefs that Anderson and his fellow theists see as giving justification to their use of induction are themselves justified by induction. There is no avoidance of the problem by tossing a deity into the mix and pretending that its existence would resolve everything. Note that even if the theist's god does exist, he is still left with the responsibility of justifying his own beliefs, a task which he undertakes by inductive reasoning.

In short, there is at present no escape from the problem of induction, not for theists or non-theists. When I engage in debates over this matter with believers, I call attention to this fact and invite them to see something like the problem of induction as a philosophical question to be tackled in mutual effort. We're both affected by this, so let's stop pointing fingers and look for a solution instead of trying to gain a foothold over the other. If induction has a secular solution, it will be available to theists as well, and should make no significant dent in their faith. But if the only thing you're interested in is a 'solution' that permits you to taunt the other side of the fence about how superior your worldview is, then it's no wonder your arguments revolve so much around the consistency of views than around the truth of those views. The irony is that consistency is a criterion that very often (perhaps always?) employs induction.

So what do we make of this problem? Is it a cause to stop drawing inferences altogether, to worry that life has no meaning, and to go sit in a dark corner and contemplate suicide? No, I dare say it's not. However, it is a cause to make us think more carefully about the inferences we draw, to consider that life may not have some of the meanings we ascribe to it, and to open ourselves up to new and different possibilities. Maybe induction is not perfectly justifiable because we tend to think of it in the wrong way, as if it ought to be more deductive or more absolute. Probability theory is not about making flawless predictions, but it can and does yield fairly reliable and practical results time after time, in discipline after discipline. Thus, I see our lack of a satisfactory answer to the problem of induction as no cause for drastic alarm. It is nonetheless a problem that merits interest, so that we don't forget our limitations and we continue to use caution in the conclusions we come to.


Sources:
Anderson, Secular Responses to the Problem of Induction, Proginosko.com (2000).
Popper, Objective Knowledge: An Evolutionary Approach (Clarendon Press, 1979), p. 7

6 comments:

  1. It seems like this blog post attacks only the weaker sort of pro-theism argument a theist could make vis-à-vis the problem of induction.

    Here's a deductive argument one could make:

    (1) If atheism is true, acceptance of induction is not warranted.
    (2) Induction is warranted (e.g. we’re warranted in thinking the sun will probably rise tomorrow).
    (3) Therefore, atheism is false.

    In practice, we believe induction is correct reasoning on the basis of intuition, i.e. it just seems true to us, intuitively. On my theistic worldview, God created both the universe and us with induction in mind, designing our cognitive faculties (via superintended evolution) in such a way that when they function properly we intuitively apprehend basic truths like elementary truths of logic and induction, thereby resulting in God-given intuitive knowledge. If God did exist, God doing this would be (at least) reasonably possible, and so knowledge of induction is reasonably possible on theism.

    This, obviously, is not an option for the atheist, and so the situation for the atheist and theist have vis-à-vis the problem of induction is not necessarily symmetrical. The theist has a solution to the problem of induction (assuming warrant entails justification), and if the atheist has no epistemic solution to the problem of induction, this scores points for theism.

    When the argument from warranted induction (as we could call it) is put in the deductive form above, we should be careful about what objections are and are not relevant. For example, a remark like "The beliefs that Anderson and his fellow theists see as giving justification to their use of induction are themselves justified by induction" may be true, but it does nothing to show that any premise of the deductive argument is false or unjustified, and if the premises are justifiably true, the conclusion follows whether one likes it or not.

    ReplyDelete
    Replies
    1. You're right that this post only goes after the weaker argument. However, I think the deductive formulation you've given is also open to substantial objections.

      What is most interesting is how quickly you move from ideas about the intuitions we have to cognitive faculties and then to God designing such faculties. There are numerous non-theistic accounts of, for example, moral intuitions, such as those proposed by Michael Huemer, Russ Shafer-Landau, David Enoch, Erik Wielenberg, and others. Theists like Wes Morriston have even defended their responses to the moral argument, and so I think it's a contestable assertion to claim that atheists can have no appeal to reliable intuitions, whether moral in nature or the kind that could support inductive reasoning.

      But I would also note that the lines you connect from intuitions to cognitive faculties to God's involvement are not only necessary in order for you to declare that theism prevails, they do not follow from deductive reasoning. This is why, despite your last paragraph, my point still stands. It in fact *does* challenge premise (2) when the case you're making for induction's being warranted is on the grounds of such presumptions, each inferred from the other. To call this a solution is premature, to put it mildly, since the existence of an allegedly God-given faculty like you suppose is far from self-evident.

      There is also something dubious about properly basic beliefs being considered a grounding for anything like induction. They do not seem to provide that kind of grounding as much as they are just an end of explanation. Precisely what it is that justifies the relationship at issue is often left either unclear or unstated. And again, there has been little worthwhile argument as to why properly basic beliefs should be reserved only for theists, particularly when religious differences cause big problems with respect to how one could possibly mediate between such claims.

      Delete
    2. Theists like Wes Morriston have even defended their responses to the moral argument, and so I think it's a contestable assertion to claim that atheists can have no appeal to reliable intuitions, whether moral in nature or the kind that could support inductive reasoning.

      Moral beliefs are not inductive reasoning, however, and the most likely explanation on atheism for our intuition for induction seems like it wouldn't deliver warrant. On atheism, our intuition is likely due to evolution (induction is how we learn and adapt to our environment, and evolution gave us this induction intuition because it worked so well in the past for our evolutionary ancestors in promoting adaptive behavior). But induction only needed to work in the past for us to have the intuition now, and so we have the intuition for induction working in the future only because induction working in the past resulted in that intuition. If the circular reasoning "Induction was working in the past, therefore it will likely work for the future" fails to warrant induction likely working for the future, then the fact that "Induction working in the past yielded the intuition that it will likely work in the future" isn't going to do much better in delivering warrant. Consequently, with atheistic evolution our intuition apparently fails to deliver warrant for induction working for the future.

      But I would also note that the lines you connect from intuitions to cognitive faculties to God's involvement are not only necessary in order for you to declare that theism prevails, they do not follow from deductive reasoning. This is why, despite your last paragraph, my point still stands. It in fact *does* challenge premise (2) when the case you're making for induction's being warranted is on the grounds of such presumptions, each inferred from the other.

      My grounds for induction being warranted isn't really inferred from anything; rather it seems to be properly basic, accepted on the basis of intuition, i.e. inductive reasoning just seems correct to us, intuitively. The point of my theistic account for how induction might be warranted on theism was only to show that induction being warranted is reasonably possible on theism. As far as theism prevailing tout court, this would seem to follow from the deductive argument (recall the conclusion of the deductive argument).

      If you want to deny premise (2), that is a relevant objection. Still, this maneuver seems to have a high intellectual cost. If you believe that you have an inadequate epistemic basis for induction you’re not rational in accepting it. More concretely, if you think you have an inadequate epistemic basis for accepting the sun will probably rise tomorrow you’re not rational to believe it. Which is more reasonable to abandon? "I am rational to believe the sun will probably rise tomorrow" or "atheism is true"? I don't think abandoning the former is a rationally acceptable option, and rejecting premise (2) would seem to undermine much of our scientific knowledge.

      Delete
    3. I think you may have misunderstood what I was suggesting with reference to our moral intuitions. I was not in any way claiming that moral beliefs are inductively reasoned. All I was doing was saying that if moral intuitions can be accommodated on an atheistic worldview, so too can our intuitions about induction. I'm sure you might dispute this, but my point was rather that there are actual arguments out there which do put a damper on your brisk leap from intuition to theism. There are those who take an evolutionary approach, as you describe, but this is also not all that is available to the atheist, particularly if we entertain the possibility of brute facts or supervenience relations.

      But induction only needed to work in the past for us to have the intuition now, and so we have the intuition for induction working in the future only because induction working in the past resulted in that intuition.

      Yes, this is, in essence, you simply restating the problem of induction. What's curious to me is that you talk as if theism resolves the problem, when there is really nothing to show that it does. Not only do you have to grapple with this aspect of the problem, but you have additional inferences you've drawn about God and God's motivations at play. It isn't that you simply appeal to properly basic beliefs and are done with it - you reject Cartesian demons, the corruptibility of human knowledge post-Fall, and other problems for varieties of theism, while assuming God would care to give us reliable intuitions about induction. In order for you to support premise (2) while also affirming premise (1), you need to do better than just appealing to intuitions. You need to show why those intuitions make more sense on theism than on atheism. And it could still be that they make sense on neither, especially if the problem actually hasn't been solved as of yet.

      I won't dispute that the problem of induction is a problem, which is what your last paragraph explains. But tied to your claims about intellectual cost are bigger questions about whether or not we need to have a full and rational account of induction to be rational in the beliefs we arrive at from it. Maybe so, maybe not. We are not perfectly rational beings, though, which leads me to be skeptical of the strong claim you're making here. There are, in fact, many things we help ourselves to that are groundless in the strictest sense, but this idea that we should therefore have to abandon them is a relic of Enlightenment thinking has proven increasingly problematic ever since. I am not one of those atheists who thinks reason is everything, or who objects viciously to properly basic beliefs, but I do think theistic claims often purport to be more powerful than they really are. I am not so convinced by the simple fact that there exists a particular or general theistic tradition that leads you to believe God would want to secure our knowledge of some phenomenon, to jumping to assume that this is the most rational course of belief we can undertake.

      Delete
    4. I think you may have misunderstood what I was suggesting with reference to our moral intuitions. I was not in any way claiming that moral beliefs are inductively reasoned. All I was doing was saying that if moral intuitions can be accommodated on an atheistic worldview, so too can our intuitions about induction.

      That was how I interpreted you, hence beginning my reply with "Moral beliefs are not inductive reasoning, however, and the most likely explanation on atheism for our intuition for induction seems like it wouldn't deliver warrant." I then explained why I believed this to be so.

      Yes, this is, in essence, you simply restating the problem of induction. What's curious to me is that you talk as if theism resolves the problem, when there is really nothing to show that it does.

      I think it "resolves" the problem in the sense that on theism it's reasonably possible for our intuition to deliver warrant for induction, i.e. via God-given intuitive knowledge. Did I prove my theistic just-so story was actually true? No; again, I was just pointing out it was reasonably possible on theism. My purpose for doing so was to provide an "escape hatch" for a person who wants to affirm premise (2) and is willing to deny atheism to do it.

      In order for you to support premise (2) while also affirming premise (1), you need to do better than just appealing to intuitions. You need to show why those intuitions make more sense on theism than on atheism.

      Sort of.

      Logic is like magic; if the premises are justifiably true, the conclusion follows, regardless of whether or not I meet any other demands you might make of me. I think rational intuition suffices to justify premise (2); if you don't think so, we'll agree to disagree, and if you're willing to pay the heavy intellectual cost in denying premise (2), e.g. abandoning "I am rational to believe the sun will probably rise tomorrow," that's your prerogative. Premise (1) has quite different support from mere intuition, and so it would have to be argued against on different grounds. I think both premises are justified, and as long as both premises are justifiably true, the conclusion follows regardless of whether I've failed to meet any other demands.

      That said, by way of rebuttal, you could argue that the justification for premise (1) also implies that premise (2) is false or that it fatally undercuts the justification for (2). Does it? My account of how on theism induction might be warranted for us wasn't proven, but it permits one to accept both premises of the argument since my justification for (1) doesn't apply to my theistic just-so story for how intuition might deliver warrant for us, and my just-so story is reasonably possible on theism. So I don't think my justification for (1) fatally undercuts my justification for (2). Given that, if both premises are justifiably true, the conclusion follows. All hail the glory of logic! ☺

      But tied to your claims about intellectual cost are bigger questions about whether or not we need to have a full and rational account of induction to be rational in the beliefs we arrive at from it.

      At the very least, it seems ill-advised to say premise (2) is false because then it means we're not warranted in thinking the sun will probably rise tomorrow. That intellectual price seems too costly to me.

      The best atheistic response I can think of is to bite the bullet and accept a pragmatic justification of induction (e.g. it'll work if any method can, so why not use it?) rather than an epistemic one. Still, denying premise (2) is pretty costly, and I'm not prepared to abandon so much scientific knowledge quite yet.

      Delete
    5. I think it "resolves" the problem in the sense that on theism it's reasonably possible for our intuition to deliver warrant for induction, i.e. via God-given intuitive knowledge. Did I prove my theistic just-so story was actually true? No; again, I was just pointing out it was reasonably possible on theism.

      And that's perfectly fine, but my intent has been to suggest that this is no less the case on naturalism than it is on theism. Part of my argument is that theism helps itself to "extra steps", if you will. I think if we have intuitions, or properly basic beliefs about induction, those could do the work of warranting our use of induction without reference to a god. A god, as I mentioned, complicates the picture. If we're going to suppose there must be an end of explanation somewhere, I see no reason why it can't end there just as well as it could end, for some folks, with God. You haven't proven your just-so story and neither have I, and with a deductive argument, all that's needed for a rebuttal is a claim of possibility.

      Logic is like magic; if the premises are justifiably true, the conclusion follows, regardless of whether or not I meet any other demands you might make of me. I think rational intuition suffices to justify premise (2)...

      Yes, I'm familiar with deductive arguments, and I think the remarks I just made above give some idea of where I stand on this. My hesitation has been more about what precisely you mean when you say induction is warranted. You even put "resolved" in scare quotes, which seems to suggest even you are a little hesitant to say the problem has been put to rest.

      That said, by way of rebuttal, you could argue that the justification for premise (1) also implies that premise (2) is false or that it fatally undercuts the justification for (2).

      I did not make this argument because I don't think (1) implies this. I'm happy to admit that not all I've said in our exchange has been directly addressed at refuting your deductive argument. It's because I know you go further than the argument in not only rejecting atheism, but in affirming a theistic account, that I've made other comments on that theistic account, too. I admire your dedication in sticking to the logic of the argument, though. ;-)

      Delete