Ever since the Enlightenment, philosophers and theologians have argued over a challenge to religious belief known as the “problem of evil.” Very briefly, how can God be both omnipotent and benevolent, yet allow things like the Holocaust or the 1994 genocide in Rwanda? If this apparent inconsistency cannot be resolved, God must either not exist or be radically different than conceived by most believers. As evident from the entry on this subject by Michael Tooley in the Stanford Encyclopedia of Philosophy, there is an enormous literature on controversy. I believe it is fair to say that Professor Tooley is convinced that the problem of evil constitutes a decisive reason to deny the existence of a benevolent God.
I confess that although I have spent a fair amount of time thinking about this subject, I have not studied the relevant scholarship. Therefore, what I propose below may be subject to compelling objections or counter arguments of which I am unaware. Nevertheless, as a theist, I offer it as my personal answer to this challenge. I will focus on section 7.2 of Tooley’s essay, which analyzes a proposed rebuttal grounded in free will, and section 7.5, discussing possible solutions based on religious beliefs.
The free will defense is that, as summarized by Tooley, “because of the great value of libertarian free will, it is better that God create a world in which agents possess libertarian [that is, the commonsense understanding of] free will, even though they may misuse it, and do what is wrong, than that God create a world where agents lack libertarian free will.” To cut to the chase, I am underwhelmed by Prof. Tooley’s reasons for rejecting this obvious response to the atheist.
As will become clear in short order, his unelaborated use of the term “better” is highly problematic. Better by what standard? Moreover, aren’t there states-of-affairs that are better in some ways and worse in others? There is intense, unresolved controversy regarding what are the true or correct set of moral values. Many philosophers, especially those working in the Kantian tradition, hold that persons have dignity as a result of their personal autonomy, and that respect for this attribute is a paramount value. At least for them, a world with free will is “better” than almost any possible world without it.[1]
Putting this issue aside for just a moment, Tooley starts with the claim that “no satisfactory account of the concept of libertarian free will is yet available.” I beg to differ. My friend Danny Frederick provided just such an account in his “Free Will and Probability,” Canadian Journal of Philosophy 43 (1): 60-77 (2013), available (gated) here, and (ungated) here.[2] The Canadian Journal is widely read by philosophers, yet no one has even attempted a refutation of Danny’s argument, so at least for now I feel entitled to dismiss this point.
Tooley then observes that even if persons have free will, God cannot be benevolent because He stands idly by while some of us employ it to commit great evil: “Indeed, very few people think that one should not intervene to prevent someone from committing rape or murder. On the contrary, almost everyone would hold that a failure to prevent heinously evil actions when one can do so would be seriously wrong.” Well, how exactly should God do this, without implicating other important values?
Is Tooley suggesting that every time a person is about to commit a horrible crime, a big hook will descend from the sky and haul the miscreant away? If this sounds preposterous, then what does he propose in its place? The problem is that it would become clear over time that we have been relieved of the responsibility for addressing the scourge of criminality. Of course, since God is by hypothesis almighty, he could surely find a way to deceive us in this matter (or any other). We could have been created with less natural intelligence, or our consciousness might be repeatedly modified, like a lab rat, so that we never discover this “ever helping hand.”
Thus, God could make it so that there would be no need for humanity to identify the causes of egregious criminality and ameliorate them as best we can. Instead, God will do this for us, like a mother making sure that her young children eat the right food, dress warmly enough when they go outside, avoid malevolent strangers and so on. But this loss of responsibility necessarily comes with a loss of dignity; we cannot be trusted to eventually eradicate great evil, and so are no longer masters of our own fate. I suppose some may regard this paternalistic world as “better” than the present one, but I do not, and Tooley offers us no reason why we should.
Tooley also argues that “the proposition that libertarian free will is valuable does not entail that it is a good thing for people to have the power to inflict great harm upon others.” Thus, he suggests that God might have created a world where persons had “a conscience that led them, when they had decided to cause great injury to others, and were about to do so, to feel that what they were about to do was too terrible a thing, so that they would not carry through on the action.”
However, this proposal is subject to the same objection voiced above; that is, how could such harm be prevented without vitiating other values? Surely if we are not to be robotic, some sin and harm must be permitted by even a benevolent God. But, says Tooley, there is a line that a loving God should not allow persons to cross. Would it be crass to ask Tooley to draw this line for us? Would it not be completely arbitrary?
Moreover, this proposal obliterates the moral difference between a petty criminal whose own character prevents him from doing worse than (say) robbery and the would-be serial killer whose manufactured conscience now limits him to minor crime. Even if God pulls off this deception without our noticing, it is clearly paternalistic. We have been relieved of the responsibility for preventing the rise of monsters like Hitler and Stalin, as they have been legislated out of existence. And, this scheme robs us of the pride and sense of accomplishment we should feel if we somehow manage on our own to make it possible for “the lion to lie down with the lamb.”
Finally, Tooley notes that such things as natural disasters, horrible diseases, and the like are not the product of human choice, so “an appeal to free will provides no answer to an argument from evil that focuses upon such evils.” Well, of course not. A killer tornado is not “evil” in the way that a Ted Bundy is evil, so it would be rather silly to insist that it be permitted to retain its free will. However, the same considerations already cited are relevant here. Is there no value in permitting humanity to conquer disease and minimize the consequences of natural disasters through our own efforts?
There is an even deeper problem with the problem of evil argument, as Tooley presents it, which is that it fundamentally begs the question against the theist. It does so by evaluating the possibility of a “benevolent” God using only the tools of science and philosophy.[3] Thus, Tooley acknowledges that a religious explanation might exonerate God’s benevolence, but only “if one thinks that one’s religious beliefs are ones that it is reasonable to accept.”
Yet, it is an essential element of religious faith that God’s wisdom and understanding are infinitely superior to our own. As stated in the book of Isaiah, “For My thoughts are not your thoughts, neither are your ways My ways, says the Lord. As the heavens are higher than the earth, so are My ways higher than your ways and My thoughts [higher] than your thoughts.” Accordingly, if God is omnipotent and, by implication, omniscient, divine “benevolence” may be radically different than comprehended by human reason. Accordingly, the argument formulated by Tooley has nothing to say to believers.
Of course, this would not matter if the belief in God were itself intrinsically irrational. You can’t reason with crazy people, and if theists fall into this category, no argument can prevail. But without rehearsing all the details of this controversy, I am among those who hold that the belief in God is no more or less crazy than disbelief. Anyone who thinks otherwise should carefully consider the “fine-tuning” argument, outlined here.
In short, the so-called problem of evil is not a persuasive argument against a benevolent God, both because of the response to this challenge based on free will and because it fundamentally begs the question against the theist.
EDIT TO ADD: For the traditional Jewish perspective on this age-old question, look here. (thanks Earl Vernon).
_______________
[1] See, e.g. Nozick’s defense of libertarian rights.
[2] He also blogged about it on this site, here.
[3] In a 2001 interview, Saul Kripke, one of the world’s preeminent philosophers, was asked about his Judaism (the exact question was not disclosed in this article). His reply: “I don’t have the prejudices many have today, I don’t believe in a naturalist world view. I don’t base my thinking on prejudices or a world view and do not believe in materialism.” Or, to quote Hamlet, “There are more things in heaven and earth, Horatio, than are dreamt of in your philosophy.”
Hi Mark,
Thanks for your kind words on my paper.
I have not read Tooley’s paper but, based on your account, what he says seems not only question-begging but even a little silly. He accepts that a world with beings with free will is better than a world without them. From that it follows that a world in which people do evil is better than a world with no people and thus that God will ‘stand idly by’ while people do evil. Tooley cannot raise that as an objection once he has accepted that it is better that there are beings with free will. Of course, given that there are created beings with free will who do evil, we can expect that other created beings with free will will try to stop them. But God will not try to stop them because, as Tooley has conceded, it is better for there to be beings with free will than not.
Tooley’s claim that “the proposition that libertarian free will is valuable does not entail that it is a good thing for people to have the power to inflict great harm upon others” is plainly false. Free will MEANS that people have the power to inflict great harm. It does not entail that they WILL inflict great harm; but it makes it very likely that some of them will (a bloody racing certainty).
The injustices that follow from free will do, though, demand some action from a good God, one would think; but theists usually invoke an after life in which all debts are paid to solve that problem.
Best,
Danny
Hi Danny:
I am pleased to be able to cite your paper, as it deserves a very wide circulation. And I doubt that Tooley or anyone else has a refutation.
I believe Tooley is assuming for the purpose of argument that free will has great value, but not so much that it overrides every other consideration. Thus, a benevolent God would limit it, and its effects, in the ways he proposes. I agree that this is no longer “free will” in the way we usually conceive it, and I still think this move fails, but I’m not sure it begs the question. Let me know if you think I am misreading him.
Much as I like Danny’s response (and his article) it does not really address the issue of social context. Free will – using his definition – must consist not only of those things which we undertake as individuals but also those things which we permit others to undertake. I would like to see his analysis of free will as passive permission rather than simply action – do the same processes apply or do they require knowledge of the consequences of a failure to act to be considered as acts of will?
Hi Laurie,
It’s nice to see you on here. We should go for a beer – except that I do not do that any more (if you can believe that).
I’m not sure that I understand your question. So I will say a few things and hope that they are relevant.
Free will means the ability to act or not to act; i.e., one is not determined one way or the other. If I see someone is about to do something, then, since I have free will, I can either try to stop him or not bother. So, freely desisting from restraining others is a possible exercise of free will.
Now let’s complicate it a bit. Jones is about to mug an old lady. I see him approach her. I could stop him or I could let him continue. Suppose I let him continue. That is an exercise of my free will: I freely refrained from stopping Jones. But the reason I did nothing to stop him was that I had no idea that he was going to mug that old lady. For all I knew, she might have been his grandmother and he might have been about to greet her. So, while it is true that I refrained from stopping Jones mugging that old lady, it is not true that I freely refrained from stopping Jones mugging that old lady. My one action of staying put was at the same time a free action of staying put (because I intended to stay put) but a non-free action of letting Jones mug the old lady (because I had no intention to let Jones do that).
How can one and the same action be a free action of one type and an unfree action of another type? Well, every action we perform is an action of an infinitude of types. For example, my rasing my hand to scratch my face, may also be my (unintentional) disturbance of the flight of some nearby insect, my (unintentional) making a bid at the auction I am attending, my (unintentionl) irritation of the man next to me who hates to see people scratching their faces, my (unintenional) moving of my hand when the clock was striking 4.00, and so on. But it was an intentional action of scratching my face. So one and the same action is intentional as one type but unintentional as other types. Now we do freely only what we do intentionally. I freely raised my hand to scratch my face, but I did not freely irritate the man next to me, even though that irritation was a consequence of a free action of mine (and thus even though irritating him was an action of mine, though a non-free one).
But there would be no action at all without an undetermined act of will. If I had not willed my hand to move, and if my willing had not been undetermined, then, even if my hand had (somehow) moved and scratched my face, I would not have performed an action of moving my hand.
Does that respond to (or perhaps answer) your question, or have I been barking up the wrong tree?
I’d have to read Tooley to get a better understanding of what he is saying; but I’m afraid I do not have time to do that just now. But the bits of what he says that you recounted irritated me, as you probably surmised. (Contemporary philosophers generally irritate me because I do not accept the assumptions that they seem to think obvious.)
I think that any limitation on the damage that beings with free will can do would have to be built into the laws of nature, rather than requiring special divine intervention. If God is omniscient He can foresee all the possibilities and if He is omnipotent He can set up the best constraints within ‘the system of the world’ to limit the damage. But even an omnipotent being cannot do what is impossible, so the best possible constraints GIVEN FREE WILL may leave open the possibility of the Holocaust and other horrors. Besides, why should God intervene in the here and now if there is an afterlife in which all can be made well again? Perhaps the point of this life is to learn some hard lessons.
But these are fairly random thoughts. I am not a theologian, or even a theist.
Random or not, that seems pretty much on the money to me.
I am neither a theist or theologian, nor have I read Tooley’s article, so all in all, I feel I’m well qualified to talk on this subject (feeble humour alert!). Firstly, if Tooley thinks that the problem of evil defeats the existence of God, that would indeed be foolish. The issue is more generally formulated in terms of the thought that it counts against the existence of God – in other words, it is nonetheless defeasible. A couple of thoughts. I think it is foolish to imagine that the way God would act more benignly would be to intervene in some way just at the point that the evil was about to be committed. What he might have done, say, is arranged for us to be created (via evolution, say) just so that we were a bit more like St Francis of Assisi and a little less like Joseph Stalin. In other words that human nature were a bit more benevolent than it is. Secondly, I think that the arguments relating to free will do provide a riposte to the Tooley-type arguments. Not a great riposte, but at least a potential one. But, I think that the issue of natural disasters, which is somewhat downplayed in this article, is of more concern to the theist. A few years ago, an Indian ocean Tsunami was responsible for about a million deaths, plus a great deal of suffering. Could not the geology have been somewhat different and the disaster averted? Again, this point can be defeated, but it does merit proper consideration. A final point: you say that belief in God is no less irrational than disbelief (hope I paraphrased this correctly). I don’t think this is entirely correct. You believe in something because you have a reason to believe. If you do not have sufficient reason, its not 50/50. In that situation, the correct default is disbelief. This does not suggest you have equal and opposite reasons to disbelieve – only that the reasons to believe are not sufficient for belief.
Hi Simon:
Thanks for the interesting comments. I am afraid that you and I will not put to rest this age-old conundrum, so I’ll venture only a couple of thoughts regarding the last point you raise. I’m not sure I understand how the “reason to believe” idea gets applied in this context. By way of example, suppose two scientists disagree over whether a hypothesized sub-atomic particle exists or not. As of yet, there is no direct empirical evidence either way, but both are making inferences based on competing theoretical systems. I can’t see that either one has a (better) “reason to believe” such that his position becomes the default until the question is settled by observation. Unless one starts with a certain assumption, i.e. materialism (see the Kripke quote), I think the debate over God has the same form. I don’t think the skeptic is entitled to assume that materialism is true, nor is the theist entitled to the contrary assumption. But, perhaps I have missed your point, in which case feel free to pursue this further.
Like Simon, I think it is the natural disasters that are more troubling to the theist, because the free will defence does not apply to them. There is, of course, the Leibnizian defence: God can do only what is possible, and the best possible world is one that nevertheless contains disasters. It is an ad hoc defence because there is no way for us to know whether a world without natural disasters would be metaphysically possible.
On the question of reasons for belief, Popper argued that if it is rational to believe only if we have a sufficient reason to believe, then it is never rational to believe anything. The problem of induction shows that to be the case with respect to empirical propositions; and the ancient sceptical infinite-regress arguments show it to be so with respect to all propositions. He sometimes quoted E.M. Forster: ‘I do not believe in belief.’
At other times he was more cautious. He acknowledged that there was a background to rationality. Before we can become critical, indeed, before we can be persons, we need to have a whole set of beliefs accepted without reasons. Some of them are inborn, being part of our biology, others are acquired culturally by assimilation from the people around us. But a rational person is one who is prepared to subject any of those inherited beliefs to criticism and to replace the believed propositions with better ones that are held tentatively (until they are in turn refuted). Notice that to show that one proposition is better than another is not to prove or justify it or to make it rational to believe it – all of that is impossible. So we do not move from irrational or arational beliefs to rational ones; we move from irrational or arational beliefs to hypotheses, and, if we are lucky, to hypotheses that can be rated as better than their alternatives, given the current state of debate.
Some evolutionary theorists have argued that belief in God is inborn in us and that it is an adaptation that helps us to survive and thrive. They also point out that this says nothing about whether such belief is true. Insofar as we are rational, we will subject that belief to criticism and, if we continue to hold it because we think it better than the alternatives, we will not believe it but, rather, hold it as the leading hypothesis currently available.
So, on this account of rationality, belief in God (indeed, belief in anything) is not rational. But it would be rational to accept, tentatively, the hypothesis of the existence of God if it made better sense of the world than alternative hypotheses. I am agnostic, not only in the sense that I neither believe nor disbelieve in God, but also in the sense that I hold neither the hypothesis that God exists nor its negation. I cannot make overall sense of the world either on the supposition that He does exist or on the supposition that He does not.
Hi Danny,
With respect to the indictment of God regarding natural disasters, I believe there is a problem not addressed by Tooley. That is, the same natural laws that benefit us in most cases are responsible for earthquakes, etc. So engineers rely on principles relating to pressure, friction, and so on in designing car brakes, but these same forces when at work in tectonic plates cause massive earthquakes. If these laws applied in car brakes and mild quakes, but not in deadly ones, scientists would discover this, and wonder why. There would be no scientific answer, and so eventually we would see the hand of God, and there goes religious faith, something God (if he exists) might reasonably wish to preserve.
Although I don’t ride the Popperian train to its final destination, I completely agree that agnosticism is the correct philosophical position re God.
Hi Mark,
Engineers use law-statements lifted from scientific theories in solving their practical engineering problems. Scientists then discover that, contrary to earlier suppositions, those same laws cannot be used to explain earthquakes. They conclude that something is wrong in contemporary science. What is it? It could be any number of things. Perhaps our descriptions of tectonic plates are false, perhaps our estimates of the relevant masses and forces are mistaken, perhaps there are some forces acting in the situation that we do not yet know about (possibly new kinds of forces, possibly well-know kinds of forces generated by currently undiscovered objects, etc.); or perhaps the law-statements that we have taken to express natural laws are false. Different scientists would suspect different things and go about trying to amend different parts of contemporary theory to resolve the problem. If one of them is a Newton or an Einstein, he may propose a new theory containing law-statements that contradict the law-statements of contemporary science. The new theory would be tested and would either be refuted or survive. Suppose it survives. We then get a scientific revolution with its new approach to explaining old phenomena, which lasts until the next set of problems materialise. But what we won’t get is scientists saying that the failure of the currently accepted law-statements to predict what is going on in earthquakes shows that God is meddling. The failure of contemporary scientific theory does not show that God interferes, it just shows that contemporary scientific theory to some extent sucks and needs to be improved. That, at least, would be the scientific presumption.
Hi Danny,
Permit me to respond with a thought experiment. Imagine a new religion comes on the scene. Its proponents claim that they have unique access to and a special relationship with God. Now an unexpected thing happens. Whenever natural disasters strike, followers of this religion are always spared. Typhoons, earthquakes, floods, etc. wipe out entire communities, but the few adherents of this new faith invariably emerge unharmed. This phenomenon persists for hundreds of years.
The world’s greatest scientists propose countless theories to account for this anomaly, but all are readily falsified. It seems clear to me at least that at some point, people (including these scientists) would conclude that the best explanation for what they are observing is that the members of this faith really do have a special relationship with the almighty. This religion then gets mighty popular.
Obviously I have constructed the best case for my position, but I don’t see why this scenario differs in principle from the one I proposed regarding tectonic plates. Given enough time (a million years, for example), scientists will rule out all plausible explanations other than that God is indeed meddling, or at least this is certainly a possible outcome. And, God might have good reason for not wanting us to believe that He is manipulating things in our favor.
Hi Mark,
The reason why religious theses are not accepted as part of science is that they do not conform to scientific standards. There are two parts to that point. The first part is that, generally, religious theses are not falsifiable: they do not make or entail statements that could be refuted by an observation that we, with our actual powers of observation, could conceivably make. An example of an unfalsifiable religious claim is ‘God exists.’ The second part is that, while some religious theses are falsifiable, they are also falsified, and their religious adherents to not respond by revising the thesis in a way that both saves it from falsification AND predicts something new which turns out to be the case. An example of such a claim is that the end of the world will take place in a specific year. The claim has often been made and refuted. Its adherents revise the claim to pick a later year. That saves it from the refutation; and it even predicts something new; but the new prediction also gets falsified in turn.
Your thought experiment suggests a religious thesis that is falsifiable and that survives severe tests. As you formulate it, it is not scientific: ‘these people have a special relationship to God’ is unfalsifiable (its content is too weak to entail any falsifiable prediction). But the thesis: ‘these people have a relationship to God that ensures that they will survive any natural disaster’ is falsifiable. If it remains unfalsified through all the natural disasters you describe, it would be a successful scientific theory. Of course, like any scientific theory, it would be susceptible to being replaced by a better scientific theory that explains why it is false but successful. Kepler’s theory that the planets have elliptical orbits was the most successful astronomical theory ever proposed, until Newton came along with a better one, which entailed that the planets do not have elliptical orbits but, due to the numbers and sizes of planets and the sun in our solar system, their orbits turn out to be approximately elliptical (especially if we do not have good telescopes).
This situation is very different from your earlier thought experiment about tectonic plates. The move from ‘the law-statements that explain all these phenomena do not explain earthquakes’ to ‘God must be meddling’ is totally ad hoc. It saves the currently accepted law-statements from refutation but without making any new empirical prediction. That is pseudo-scientific procedure. If we ever got to a stage where scientists made that move, we would have come to the end of science.
So, religious hypotheses COULD be scientific. That is a point I make in my article ‘A Puzzle About Natural Laws and the Existence of God’ that was published in the International Journal for Philosophy of Religion’ in 2013, available (ungated) here:
https://www.academia.edu/426573/A_Puzzle_about_Natural_Laws_and_the_Existence_of_God
But theologians have not troubled to try to make their theories scientific and, even where one of their theories turns out to be falsifiable, they make ad hoc moves in response to falsification. Of course there will be some scientists who will not accept a religious hypothesis even if it is scientific. But that is nothing unusual. Many scientists refused to accept Newton’s theory long after it had been seen to be remarkably successful because they thought that the posited force of gravity was ‘occult.’ Einstein would not accept quantum theory because he thought that God does not play dice. Being a good scientist does not mean accepting the (currently) best explanation; in fact, the better scientists are those who do not accept the (currently) best explanation but try instead to replace it with a better one (though they should also acknowledge that the currently best explanation is CURRENTLY the best).
All of the above contradicts the most firmly held assumptions of contemporary epistemology (which is an intellectual slum).
Thanks for the clarification, Danny. Perhaps because I’m too thick, I don’t get the dissimilarity in the two cases. It seems to me that both examples make reference to God’s intervention in the natural world (which cannot be directly observed), and both hypothesis gain traction only because all other putative explanations are eventually falsified. So, in the natural disasters case, scientists would not initially leap to the conclusion that God was saving his followers. They would first exclude random chance, the possibility that the followers were somehow causing these events, etc. Only at the end, out of necessity, would they come to take seriously the possibility of divine intervention.
So too with tectonic plates. Scientists would first try all other theories that they can formulate, and only after falsifying these would they consider the possibility that God is altering these physical laws to prevent tragedy. Obviously, it would take longer for them to get to this place in the second example, but it appears to be the same in principle.
Incidentally, much the same process seems to be at work in giving the “fine tuning” argument credibility. From the link I provided at the end of my piece:
Fred Hoyle, the astronomer who coined the term “big bang,” said that his atheism was “greatly shaken” at these developments. He later wrote that “a common-sense interpretation of the facts suggests that a super-intellect has monkeyed with the physics, as well as with chemistry and biology . . . . The numbers one calculates from the facts seem to me so overwhelming as to put this conclusion almost beyond question.”
Theoretical physicist Paul Davies has said that “the appearance of design is overwhelming” and Oxford professor Dr. John Lennox has said “the more we get to know about our universe, the more the hypothesis that there is a Creator . . . gains in credibility as the best explanation of why we are here.”
Edit to add: Perhaps we are getting hung-up on what constitutes a “scientific” hypothesis, versus a rational inference. The point I am making about having reasons to believe in the possible intervention by God only depends on the latter being an appropriate tool, not the former.
I think you are right, Mark, that at least part of the difficulty here consists in different understandings of what science is. I will try to make my understanding clearer.
The difference between the two God hypotheses turns on the distinguishing features of science (as explained by K. Popper): falsifiability of statements and non-ad-hoc responses to falsifications.
It is ad hoc to try to save currently accepted law-statements from falsification from data about earthquakes by introducing the hypothesis that God is meddling. The reason it is ad hoc is that the hypothesis that God is meddling in the earthquakes entails no new falsifiable predictions of its own. It is not independently testable. The only reason for accepting it would be that it saves currently accepted law-statements from falsification. That is pseudo-scientific. If we can appeal to ad hoc hypotheses whenever a theory gets falsified, then no theory would ever be refuted and the biggest bunch of crap would rate as just as scientific as Newton’s theory (which, despite its astounding success over two centuries, was eventually refuted and replaced by relativity theory).
In contrast, the theory that a specific group of people will be unharmed in any natural disasters in which they get caught up is falsifiable. If we see any one of them get killed in such a disaster, the theory is falsified. Since the predictions of the theory are so surprising (we would be astounded if they were not falsified), the fact that the predictions turn out not to be falsified means that the theory rates as scientific. Of course, if, contrary to your supposition, the theory gets falsified, but its adherents rescue it from falsification by introducing ad hoc hypotheses, then we are back in the realm of pseudo-science.
The fact that God’s intervention in the world cannot be directly observed is not the important point. Newton’s theory referred to forces, which cannot be directly observed: we directly observe motions, which may be the effects of forces, but not the forces themselves. To be falsifiable a theory must be inconsistent with a conceivable observation statement, so it needs only to entail a statement the negation of which describes something that could be directly observed. (I am here using your terminology of ‘direct observation,’ which on another occasion I would object to, because all observations are theory-laden – but let’s leave that on one side here).
So, ‘God is meddling here’ is unfalsifiable because ‘It is not the case that God is meddling here’ does not describe something we could conceivably observe. But ‘God will save these people (including Pete and Joan) from being harmed in any natural disaster’ is falsifiable because, for instance, ‘that landslide just engulfed Pete and Joan’ is a conceivable observation statement.
There would never be any necessity for scientists to invoke an explanation alluding to God. It is a POSSIBILITY that a falsifiable hypothesis about God could do the explanatory work; and if such a hypothesis is proposed and survives testing it would rank as a scientific theory (even though atheistic scientists would have nothing to do with it, just as in the late-seventeenth and early-eighteenth centuries lots of scientists would have nothing to do with Newton’s theory because of its invocation of an ‘occult’ force of gravity). But there is always the possibility of coming up with a different theory. At no point do we reach a situation when it can truly be said: the only possibility left is that God is meddling. There is an infinite number of rival explanations for any set of data (that is demonstrable from the axioms of classical logic – as Nelson Goodman showed with his ‘grue’ hypothesis).
I cannot comment on the ‘fine tuning’ argument, Mark, as I have never studied it.
I agree that not all explanations are scientific explanations. Early cosmological explanations were metaphysical rather than scientific; but they were proposed AND criticised by pre-Socratic philosophers. All science develops from philosophy: it takes a lot of work to get from a first sketchy stab at a solution of a problem to something that is falsifiable. We can rate non-scientific theories as better or worse according to how much they explain, how simple they are, whether they are consistent and whether they explain more than they were proposed to solve.
I hope that makes things at least a little clearer.
Hi Danny:
That does help, thanks. A couple of follow-up questions. With respect to the natural disasters case, are you saying that if the theory if not falsified (i.e. believers are never killed), this indicates that the reason they are saved is because of their special relationship to God? If not, then with respect to providing a reason for believing in God, the two cases are the same it seems.
In the earthquake case, how would you even formulate a scientific theory alluding to God? Is it always unscientific to posit God as an explanation even after falsifying ALL non-God explanations?
How would Popper evaluate the following case. You meet a guy in a bar. After a number of drinks he tells you that he can manipulate the laws of physics as he likes. You tell him that he’s cracked. He offers “proof.” He tells you to go home and select any coin you choose, and to flip it in a manner that you believe will ensure a random distribution of heads/tails. He tells you that the first hundred flips will all be heads, and that after this, the conventional physical laws will be back in force.
On your way home, you go far out of your way to stop at a store where you are sure you will receive in change a coin that your acquaintance could not possibly have manipulated, yet the first 100 flips are indeed “heads,” after which it is roughly 50/50. You consider various explanations for this phenomenon, other than that the guy in the bar is telling the truth, but after extensive investigation you are able to exclude them all to a reasonable certainty (using standard scientific methods). Would Popper conclude that the leading scientific explanation of this event is that your “friend” can indeed manipulate the laws of physics?
I have posted my reply below, Mark, as it is a bit long, and the page-width here is rather narrow.
I generally agree with Danny on the subject of belief, but here is one rather obvious point: Belief (or otherwise) seems only to play a part when there is lack of certainty. We don’t say we believe in Pythagoras’ theorem – we know that, for right angled triangles, the square on the hypotenuse is equal to the sum of the squares on the other two sides. We know this because the theorem is a result of deductive logic. But belief does pertain in issues like belief in God and also science, since the latter is based on inductive methodology.
Now to the specific point about belief in God that you asked me to comment on, Mark. I don’t think the two cases you have cited are analogous. In one case we are looking at a belief of the type X is A. If you tell me you believe X is A, my first question is what reason (not evidence, necessarily) do you have to believe that X is A. You provide me some reasons. but I do not think they are sufficient to induce belief. I therefore take the position that I do not believe that X is A. This position is not the mirror image of your belief – that would be the thought that X is ¬A and that is not something I claim. That is why it is my opinion that non-belief is a neutral default position. We may think of Bertrand Russell’s thought experiment about a Royal Worcester tea set orbiting in the asteroid belt. (Was it Royal Worcester? – I may have the wrong brand!). Do I believe it – not unless you provide sufficient evidence, I don’t. (Of course, you may object that the belief in God is more realistic than in the tea set, and that Russell therefore biases the argument – intuitively, I agree, though how one would assess probabilities is beyond me.) So, I adopt what I believe is a neutral position – that I do not believe in the tea set. Note: I do not adopt the belief that there is no tea set.
Now consider the two scientists. The situation is different. One adopts the belief that Y is B, the other that Y is C. Let’s assume these are mutually exclusive. Again, we must weigh the evidence. A dialogue ensues. One may convince the other or they may maintain their respective positions. A third party may decide that they don’t believe either, because neither has provided sufficient support. This replicates the neutral position described above, whilst the other two’s beliefs do not. In this regard, I think my position re belief in God is similar to Danny’s, though just what atheism is and just what agnosticism is, is, I think, subject to some debate. Ayn Rand seemed to think my position to be more akin to atheism (though she was living in her own little morally certain world, so who can tell!)
Hi Simon:
We may be in agreement, since you say, “I think my position re belief in God is similar to Danny’s, though just what atheism is and just what agnosticism is, is, I think, subject to some debate.” If what you mean is that agnosticism is a better supported view than either theism or atheism, then I concur. I certainly don’t claim to have “evidence” for God in the form of empirical data. And I think the tea set example is distinguishable, because we have powerful independent reasons for doubting its existence (how would it get there, etc?). On the other hand, since God, by hypothesis (as the author of all scientific laws), only makes himself available for empirical confirmation to the extent He wishes, I don’t see how science can exclude the possibility of God.
Yes, I understand your point about the tea set comparison. And I certainly wouldn’t suggest that belief in God is incompatible with science. But, at present, I am not persuaded to believe in the existence of God.
Fair enough.
Hi Simon,
I disagree with much of what you say, but I will try to be brief.
There is no certainty about anything. You give the example of Pythagoras’ Theorem. It is provable from the axioms of Euclidean geometry, but according to general relativity theory, space is not Euclidean but Riemannian, so Pythagoras’ Theorem turns out to be false. You may infer from that that the axioms of Euclidean geometry are false.
You also refer to deductive logic. But there are alternative systems of deductive logic and laws which hold in one do not hold in another. For instance, in classical logic we have ‘p or not p’ and ‘if p and q, then p.’ But the former is denied in intuitionist logic and the latter is denied in connexive logics. Which logic is correct? That is a matter of dispute: it is uncertain.
You say that science is based on inductive methodology. But there is no inductive methodology. Science proceeds by conjecture and refutation.
But I agree with you that disbelief that p is not equivalent to belief that not p; or, as I would prefer to put it, that non-acceptance of p does not equate to acceptance of not-p.
Yes, you are correct, but this is why there is endless debate in philosophy about increasingly minor details and why people switch off. I think my point about the difference between deductive logic and inductive logic illustrates an important aspect of the issue in hand and I stand by it if only in a coarse grained way that recognizes that some objections can be made. I have lost interest in debates of how many angels can fit on the head of a pin. I prefer (this is just a personal view) to try and grasp important issues, where we can make some genuinely interesting progress, whilst acknowledging that we do not also have to say the final words on any subject. I don’t agree with your criticism of the term “inductive methodology”. The method is observations to patterns to theory. If you prefer to call it an approach, that’s cool too.
Hi Mark,
The natural disasters case. Here is the theory: ‘These people have a special relationship with God that ensures that they will never be harmed in natural disasters.’ It is falsifiable. For example, ‘two of those people just got killed in that landslide’ is inconsistent with it and describes a situation we could conceivably observe. Given what we know of the world, we all expect the theory to be falsified, over and over again: it seems to be a no-hoper. Yet, we are supposing, attempts to discover actual observable situations the descriptions of which are inconsistent with the theory continually fail: it remains unfalsified despite serious attempts to falsify it. That makes it a successful scientific theory.
You ask: does that indicate that the reason the people are saved is because of their special relationship to God? No, it does not. Every successful scientific theory is (at most) the best that is currently available. One of the jobs of a scientist is to try to replace such theories with better ones. Newton’s theory was startlingly successful for about two hundred years. Does that indicate that the reason the planets (and other bodies) move as they do is because of the force of gravity? No. Newton’s theory has been replaced by general relativity, according to which there is no force of gravity. All of our scientific theories are hypotheses and their success (if they are successful) is temporary – at least, so long as scientists keep doing their job.
You ask: so, with respect to providing a reason for believing in God, the natural disasters case and the earthquake case are the same? Yes. The success of a scientific theory never provides an epistemic reason for belief in the theory. Hume pointed that out in his discussion of induction. Of course, we may have non-epistemic reasons for belief. If someone wants to avoid thinking too hard, he has a reason for believing whatever he is told by people in positions of authority, for example.
The earthquake case. Scientists are presented with the problem that a successful mechanical theory fails to explain earthquakes. So they have to make a change to currently accepted theory. Not necessarily the successful mechanical theory. It may be that other theories used in conjunction with the mechanical theory to produce the falsified predictions are at fault; for example, theories about tectonic plates, estimates of the mass of the relevant bodies, and such like. Someone makes a proposal: the assumption so far taken for granted that God does not interfere with the day-to-day workings of the world is false; and it should be replaced by the hypothesis that God interferes with earthquakes.
The scientists ignore that proposal. And they are right to do so. Why? Not because it mentions God. But because it yields no new falsifiable predictions; and because saving a current theory by introducing a hypothesis that yields no new falsifiable predictions is ad hoc (it explains nothing new). The hypothesis yields no new falsifiable predictions because the following is not an observation statement: ‘this earthquake had no interference from God.’ That is not the kind of thing we can observe. The hypothesis is ad hoc because it offers an explanation for why current theory fails to explain earthquakes, but the offered explanation does not explain anything else. The hypothesis is introduced solely to save current theory. That frustrates the growth of knowledge and is thus unscientific.
You ask: how would you even formulate a scientific theory alluding to God? The natural disasters case gives a possible (not actual) example of how that would be done. It is possible rather than actual because it is not enough for a theory to be scientific that it be falsifiable: it must also survive serious attempts to falsify it. We assumed in the natural disasters case that the theory did survive serious attempts to falsify it; but that is just a possibility. In the earthquakes case, we could try the following. ‘God always interferes in earthquakes to ensure that no more than half-a-dozen humans are killed.’ That is falsifiable; it is also falsified. We could try: ‘God always interferes in earthquakes to ensure that no more than a billion humans are killed.’ That is in principle falsifiable and it is not yet falsified; but I do not think anything could count as a serious attempt to falsify it, because human populations are not concentrated enough to produce a situation that would falsify it. If human populations continue to grow, however, there may come a time when it seems there are real risks of it being falsified; so if it then survives falsification, it should count as scientific. Except that there is a problem with it, namely, that it seems arbitrary or senseless. Why would God permit up to a billion humans to be killed in an earthquake but no more? But perhaps interesting answers could be proposed to that question.
There is, though, a problem with the two falsifiable God hypotheses that we have considered; and I think that problem is one that you recognise, though I do not think you have stated it explicitly. In each case, there is a simpler hypothesis that would do the same explanatory work. In the natural disasters case, we could replace ‘These people have a special relationship with God that ensures that they will never be harmed in natural disasters’ with: ‘These people will never be harmed in natural disasters.’ The empirical consequences of the two theories are the same with regard to those people and natural disasters; so, unless the God component in the first hypothesis does some additional explanatory work somewhere else (which it could do), the second hypothesis would be preferred. The same goes for the two hypotheses: ‘God always interferes in earthquakes to ensure that no more than a billion humans are killed’ and ‘In earthquakes no more than a billion humans are killed.’
But it is worth noticing an apparent parallel. Physical theories explain observable phenomena in terms of the properties of physical objects. But it would be ontologically more simple to drop talk of physical objects and thus replace current physical theories with corresponding theories that talk about observable phenomena only (‘sense data’ or such like). That, indeed, was the stance adopted around century ago by the old neo-Kantians, positivists and early logical positivists. One of the problems was that it did not actually make physical theories simpler: it made them impossibly complex. So that programme of simplification was not workable. But eliminating references to God from successful scientific theories that mention Him (such as Kepler’s theory and even Newton’s theory in its full exposition) is a possible simplification.
You ask: is it always unscientific to posit God as an explanation even after falsifying all non-God explanations? No. As I said above, there seems to be no reason in principle why explanations referring to God should not be falsifiable, survive serious attempts to falsify them and not be reducible to simpler theories which are their explanatory equals. But it is always unscientific to try to rescue current theories from falsification by means of a supposition which, when conjoined with current theory, yields no new falsifiable predictions. Further, even if all currently available non-God explanations have been falsified, there is an unlimited number of new non-God explanations to be discovered, with a bit of imagination, ingenuity and perseverance.
The case about the man in the bar who claims that he can manipulate the laws of physics can be a bit complicated. For instance, it is not impossible that an unbiased coin turn up heads for each of the first hundred tosses; it is just not very likely. It seems even less likely that it should happen after a man has predicted it would happen; but that is still possible. The consequence is that it seems impossible that probability statements should be falsifiable. Popper made a pragmatic proposal to deal with that, effectively regarding the observation of an event that a probabilistic theory claims to be extremely unlikely as refuting the theory. I have never concerned myself with the details of that proposal, or any of the discussions of it, so I cannot say much more about it. But let us suppose the following, for the sake of argument:
(i) it is a physical law that an unbiased coin will, in the long run, tend to turn up heads half of the time;
(ii) an unbiased coin turned up heads for the first one hundred tosses;
(iii) if (ii) occurs we should regard (i) as falsified.
The man in the bar claims that (i) can be rescued from this falsification by adding to (i) the following clause: ‘except when that man in the bar announces otherwise.’ If we add that clause to (i) we need also to add a corresponding clause to (iii), so that (ii) does not falsify the modified (i) if the man has made his announcment.
Does that modification of current theory make any new falsifiable predictions? Yes it does. It predicts that whenever the man announces a falsification of the unmodified (i), there will be a falsification of the unmodified (i) [but not, of course, of the modified (i)]. Suppose that we test that new prediction severely, by having the man make his announcement in all sorts of different cases that we would expect him to be unable to control in any ordinary way. If the predictions are not falsified, then the modification to theory concerning that man’s extraordinary powers would be a successful scientific hypothesis.
You ask: would Popper conclude that the leading scientific explanation of these infringements of previously accepted law-statements is that the man in the bar can indeed manipulate the laws of physics? Well, I don’t know; but I think he should do, on his own theory. But assuming that he would do so, he would also say that, as in all cases of successful scientific explanations, we should try to replace this successful scientific explanation with a better one.
Thanks, Danny, that’s all I have for now.
Just one further comment, Simon. You say:
“I don’t agree with your criticism of the term “inductive methodology”. The method is observations to patterns to theory.”
But that is one of the things I was denying when I said that there is no inductive methodology. Science does not begin with observations. It begins with a problem. Scientists then propose theories to solve the problem. The theories are criticised, amended and, at least sometimes, developed into a form in which they are falsifiable. At that point they can, in principle, be tested. So scientists then either look for refuting instances, or more usually, design an experiment that they expect to yield a refuting observation. Designing the experiment sometimes takes as much imagination, ingenuity and perseverance as coming up with a new scientific or proto-scientific theory. So, observations are important; but they come in at a late stage, as a way of testing a theory.
Further, the patterns are usually provided by the theory and are not observable. Newton’s law of gravity, for instance, says that any two bodies exert a gravitational attractive force on each other, and it includes a formula for calculating the size of the force. But there is no pattern in the phenomena corresponding to that law. Very few bodies can be observed to attract each other. Sure, magnets do; but magnets also repel each other. The order that science discovers in nature does not correspond to observable patterns, so it cannot be derived from them. It is, rather, hypothesised, a theoretical construction. Once we have a hypothesis, we can try to discover ways of testing it. so the order is not ‘observation-pattern-theory’ but ‘problem-theory-test-observation.’
Hi Danny:
Okay, so I lied. This may not be directly related to your comment here, but it has nevertheless provoked a question: what is the connection, if any, between Popper’s scientific epistemology and the existence or not of necessary truths. As I understand it, Popper denies the existence of such truths but, if they exist at all, they are by their very nature not empirical claims. If necessary truths could be falsified by observation or experiment, they would not have this status. So, even if one disagrees with Popper on this issue, as I do, could you nevertheless follow his approach with respect to science, which is usually thought to pertain only to empirical matters?
Perhaps an example would help. I believe that the statement, “In standard math [the kind we use to calculate the change we get at a store] the next largest whole number following 1 is 2.” If that claim is shown to be false, I suspect that it will not occur as a result of an empirical observation or test. So, at least until my example is called into question, how would holding my view interfere with the scientific method?
Hi Mark,
Sorry for the delay, but I have not been online much today.
Popper argued that the laws of nature that are discovered by empirical science are necessary truths. They are not logically necessary: they are a posteriori necessities. Saul Kripke made a big point about the existence of a posteriori necessities (from around 1970 onwards), particularly his ‘Naming and Necessity’ (which is available for download at various places online). But his angle was for the most part different from Popper’s. Statements of natural laws are often expressed as universal generalisations: every A is B. But such a generalisation might just happen to be true because of accidental circumstances. In contrast, a law of nature is necessarily true. So the proper representation of a statement of a law of nature is: necessarily, every A is B.
If a statement of the form ‘every A is B’ is testable, then so is the corresponding statement of the form ‘necessarily, every A id B,’ because the latter entails the former. You might say that the former is simpler and so should be preferred; but Popper says no, because a law of nature is, by its nature, a necessary truth (though one that can be known only empirically).
Popper argues for the necessity of laws of nature in Appendix *x of his ‘Logic of Scientific Discovery’ which can be downloaded from here:
http://strangebeautiful.com/other-texts/popper-logic-scientific-discovery.pdf
That book is not very readable unless you have a strong background in symbolic logic and are familiar with the state of philosophy of science in the early 1930s (it was written in 1934, though translated into English only in 1959, along with a bunch of new appendices). But Appendix *x (one of the new ones) is very readable and also, I think, very illuminating. I strongly recommend it. I suspect you will find it an eye-opener.
When you say that Popper must deny that necessary truths are falsifiable, you are mixing up epistemology and ontology. The ontological question: are laws of nature necessary? Yes. The epistemological question: can we know laws of nature a priori? No. So if we have a falsifiable statement of a law of nature, it may turn out to be false? Yes. So the statement was not necessary? That’s right. So how could it express a law of nature? It never did. We only thought it did until we falsified it.
Popper thought that truths of logic and arithmetic (but not geometry) were both necessary and knowable a priori (though I think he wavered a bit on the a priori question). I reject that because I don’t think there can be any a priori knowledge.
Clearly, Popper did not think that the view that the truths of logic and arithmetic are knowable a priori interferes with scientific method. I suppose I do think that, in that I think that scientific method is applicable to logic and arithmetic too. But not directly. More in the way that Euclidean geometry was falsified by general relativity. Some scientists have claimed, for example, that the logical law of excluded middle is refuted by quantum theory; but others (including Popper) reject that claim.
Hi Danny:
Thanks a lot. This is very helpful, and I will try to read appendix *x. It seems that on at least one important question, I agree with Popper and disagree with you. Well, I always said that guy was a genius! 🙂
I thought Schopenhauer took care of the problem of Evil by simply saying human good is not on the agenda of the “Will.” And later he had some kind of second thoughts in which he added that the Will has some kind of higher agenda in which the Good is the goal.
Hi Avraham:
You have me at a real disadvantage here. I think I read parts of Will and Representation back in college, which was during the Garfield administration. As I distantly recall, Schopenhauer was firmly in the continental tradition, and was thus approaching this problem with a completely different methodology than employed by analytic philosophers. So, I’m not sure how to respond. Perhaps you are having a little fun with me?
Mainly Schopenhauer does not think human good is the goal of the Will. But in a later letter he indicated a higher aspect of the Will that does intend the Good. But in any case he was before the Analytic tradition. That started with an idea of Frege and went downhill from there.
Schopenhauer was basically with Kant. He just did some modifications. But in any case I think your approach looks good. I just have not had time to study it.
Thanks. I visited your blog, and liked what I read. I suspect we have a lot in common in our personal philosophies. Please stay in touch if you can.
Your approach is fruitful. The preservation of freewill is necessary for the moral plane to exist. And to the Ramchal Moshe Chaim Lutzato that is necessary for the very purpose of the creation of the universe. Though I admit I did not see the exact connection when I was reading his works.
I totally agree. The assumption regarding free will is not only necessary for the attribution of moral responsibility but, following Nozick, is the primary reason that human beings enjoy a special moral status.