Academia.eduAcademia.edu

Can Belief in God Be Confirmed

1988

Digital Commons @ George Fox University Faculty Publications - College of Christian Studies College of Christian Studies 1988 Can Belief in God Be Confirmed? Mark S. McLeod-Harrison George Fox University, [email protected] Follow this and additional works at: http://digitalcommons.georgefox.edu/ccs Part of the Christianity Commons, and the Epistemology Commons Recommended Citation Previously published in Religious Studies, 1988, volume 24, pp. 311-323 http://www.jstor.org/discover/10.2307/ 20019289?sid=21106171698683&uid=24801&uid=3&uid=3739856&uid=3739256&uid=24803&uid=62&uid=2&uid=67 This Article is brought to you for free and open access by the College of Christian Studies at Digital Commons @ George Fox University. It has been accepted for inclusion in Faculty Publications - College of Christian Studies by an authorized administrator of Digital Commons @ George Fox University. For more information, please contact [email protected]. Rel. Stud. 24, pp. 311-323 MARKS. McLEOD CAN BELIEF IN GOD BE CONFIRMED? A basic thrust behind Alvin Plantinga's position that belief in God is properly basic is an analogy between certain non-religious (and presumably justified) beliefs such as ' I see a tree' and theistic beliefs such as 'God made this flower'.1 Each kind of belief is justified for a believer, argues Plantinga, when she finds herself in a certain set of conditions. Richard Grigg challenges this claim by arguing that while the non-religious beliefs are confirmed, beliefs about God are not. I wish to explore this challenge, clarify it and suggest that on one understanding it is irrelevant and on another it is false. Grigg writes: Plantinga points out that a belief such as the one that I had breakfast this morning is properly basic in certain circumstances, i.e. as long as I have no reason for supposing that my memory is defective. But note that we can trust beliefs such as (1) ['I see a tree'], (2) ['I had breakfast this morning'], and (3) ['That person is angry'] not only because we are unaware of defects in our experiential equipment but also because we constantly have outside sources for confirmation of such beliefs. Indeed, is it not only through such outside sources that we can become aware of a defect in our equipment? For example, when I return home this evening, I will see some dirty dishes sitting in my sink, one less egg in my refrigerator than was there yesterday, etc. This is not to say that (2) is believed because of evidence. Rather, it is a basic belief grounded immediately in my memory. But one of the reasons that I can take such memory beliefs as properly basic is that my memory is almost always subsequently confirmed by empirical evidence. But this cannot be said for a belief about God, e.g. the belief that God created the world (Grigg: 125, 126). ( 1 ) , ( 2 ) , and (3) the ' paradigm ' beliefs. Grigg's argument, briefly stated, is that the paradigm beliefs are legit­ imately properly basic because of some type of confirmation they have, while belief in God is not similarly confirmed. Since the analogy to which Plantinga appeals claims that the paradigm beliefs and beliefs about God I will call 1 R�ference s are as follows. (Alston) William P. Alston, 'Christian Experience and Christian Belier. (Plantm ga) Alvin Plantinga, 'Reason and Belief in God'. Both in Faith and Rationaliry: Reason and Bel�J in God, edited by Alvin Plantinga and Nicholas Wolterstorff (Notre Dame: University Press). Gngg) Richard Grigg, 'Theism and Proper Basicality: A Response to Plantinga', � ; P�ilosophy of Religion, t a11 mga's essay. International Journal xiv (1983), 123-7. For the development of the analogy discussed here, see 13-2 3 12 have the same epistemological status, the lack of confirmation for beliefs about God shows that the analogy fails. The nature of the confirmation for which Grigg calls, however, is not clear. Some clarificatory terminology and distinctions are in order. At one point Grigg seems to call for the confirmation of beliefs as when he suggests that ' I had breakfast this morning' is confirmed by there being dirty dishes in the sink. At another point he seems to suggest that it is the reliability of the belief-forming practice which stands in need of confirmation - 'my memory is almost always subsequently confirmed by empirical evidence ' (Grigg: 126). It will be convenient to call the confirmation of a practice's reliability the 'validation ' of a practice while reserving the term 'confir­ mation' for the confirmation of the truth of a belief. A second clarificatory point is that surely not every properly basic belief is confirmed. Hence, confirmation of a given belief is not necessary for its proper basicality. Two cases come to mind. First, some beliefs, even when we try to confirm them, fail to be confirmed and yet it does not follow that such a belief is not properly basic. The memory belief that I took a walk by myself in the woods yesterday may not be confirmed because no one else saw me, my hiking boots show no evidence of the walk, I didn't bring back any evidence of the walk from the woods, etc. So, even though many memory beliefs are confirmed, not all are, but such beliefs do not necessarily fail to be properly basic. The second case deals with the simple fact that many beliefs are not confirmed because we have neither the time, the interest, nor any special reason to d� so. Generally speaking I do not concern myself with the confirmation of my memory beliefs, unless there is some special reason to do so. For example, typically I do not worry about whether I ate breakfast this morning unless I am being asked by the physician just before she does surgery. If my memory seems vague on the topic I might then try to confirm or disconfirin my memory belief. But the failure of a given belief to be confirmed surely .does not entail that the belief is not properly basic. Two morals should be drawn here. First, the concern ought to be not that each and every properly basic belief is confirmed but that when, in general, one attempts to confirm a certain belief, it is confirmed. There will, of course, be some exceptions to the rule. This brings attention to the second moral, viz. that attention should be paid to the source of the belief to be confirmed - the practice - rather than the belief alone. An important relationship seems to hold between the confirmation of beliefs and the validation of the practice which generates them. For example, it may be that since many beliefs generated by a given practice are confirmed, the practice is validated. It might then be enough for a defence of Plantinga's theory against the confirmation challenge to show that if the practice from which a belief comes CAN BELIEF IN GOD BE CONFIRMED? is validated as reliable, then any belief generated by the practice, all other things being equal, can be legitimately taken to be properly basic. Here we find a potential explanation for the fact that we generally trust our beliefs even though not every belief can nor should be confirmed. Armed with these distinctions and this terminology we can understand Grigg to be calling attention to the following dilemma: either (A) a given belief-forming practice is validated because beliefs generated by the practice are confirmed relative to other beliefs delivered by the practice or (B) no belief generated by the practice can be confirmed until (i) the practice generating it is independently validated or (ii) the belief itself is confirmed independently of the new practice. This dilemma is the focus of the remaining sections, beginning with the second branch (B). II Stated :nore rigorously, branch (B) of the dilemma claims that for any belief p gener'ated by practice A, confirmation is due either to a belief (or set of beliefs) q generated by practice B (where Bis not the same practice as A) or to the validation of A which in turn is due to some belief (or set of beliefs) q which is generated by some practice B (where again B is not the same practice as A). In other words, the confirmation of a belief generated by a given practice is independent, in the final analysis, of the practice itself and, by extension, independent of beliefs generated by that practice. Now if Grigg is suggesting that the paradigm beliefs are confirmed by either of these two independent means he simply seems to be mistaken. For example, in the case of confirming memory beliefs I will nearly always, if not always, utilize other beliefs generated by memory itself. In order to confirm my memory belief that I had eggs for breakfast this morning I will have to remember how many eggs were in the refrigerator yesterday. The con­ fir mation of the belief is not independent of the practice which generates the belief in the first place. More importantly, however, if this independent confirmation were available for the paradigm practices there would be little motivation for Plantinga to move beyond classical foundationalism. If we had independent confirmation for beliefs which were self-evident or incorrigible or evident to the senses then our basic beliefs could serve as the kind of firm foundation desired by the classical foundationalist. There would be, in other words, an independent check of the truth of basic beliefs. But this check is not available, as the abundant examples from the history of scepticism show. Epistemologists have failed to meet the sceptic's charge and provide the independent link between our beliefs and the world. As a means of becoming more clear on the issue under consideration we MARKS. M0LEOD can explore briefly a suggestion due to William Alston. Some epistemic practices are what he calls 'basic practices'. A basic practice, he writes, is one that constitutes our basic access to its subject matter. [For example,] we can learn about our physical environment only by perceiving it, by receiving reports of the perceptions of others, and by carrying out inferences from what we learn in these first two ways. We cannot know anything a priori about these matters, nor do we have any other sort of experiential access to the physical world (Alston: r r 7). The practices which generate the paradigm and theistic beliefs may be just such practices. Since we have no other access to the objects about which we form the beliefs beside the practices which generate the beliefs, we cannot appeal to outside, independent information as a source of confirmation. So, just as the only access we have to the physical world is through perception, perhaps at the end of the day the only access we have to God is through the practice which generates theistic beliefs. The confirmation of both types of beliefs will rely on the practices which form them and in turn the validation of the practices themselves will rely on the confirmation of the beliefs generated by them. Grigg might argue that we could have access to those religous beliefs by using a discursive belief-forming practice (natural theology) and that this is what should validate the practice of forming theistic beliefs. But given the lack of success with natural theology (com­ parable to the lack of success in epistemology with confirming perception's deliverances by reasoning from beliefs about sensations (or from sensations themselves) to beliefs about the external, physical world), it hardly seems likely that we should turn there for validation of the practice or, by exten­ sion, the confirmation of beliefs. Whatever is demanded of beliefs about God, one should not demand more of them than that demanded of the paradigm beliefs. Confirmation of the independent variety of branch (B) is not available for either the paradigm or theistic beliefs. Since an independent check is not available, one should anticipate a kind of circularity in the confirmation of properly basic beliefs. Grigg's challenge collapses into branch (A) of the dilemma. III One further note of clarification is needed before we consider branch (A). Exactly which beliefs are to be confirmed? Grigg suggests that the belief 'God created the world' lacks the kind of confirmation the paradigm beliefs have. Now while one must admit that this appears to be true primafacie, the truth of the claim will not, in the final analysis, play the important role Gri gg attributes to it. The kind of belief to which he calls attention is what we can call a 'general belief' and this is the wrong type of belief on which to concentrate. Since the intuitive force of his conclusion rests on the nature of CAN BELIEF IN GOD BE CONFIRMED? the belief he singles out for consideration, his argument will lose much of its force once this focus is realigned. Rather than concentrating on Grigg's kind of example, Plantinga calls attention to beliefs such as 'God created the flower' or 'God is angry with my sin ' (see Plantinga : 80). What kind of beliefs are these? If the analogy is followed closely, a picture of properly basic beliefs develops in which the beliefs are not general claims such as 'all trees are green' and 'God created everything' but specific beliefs generated by specific experiences, e.g. ' I see a tree (here and now)' or 'God created this tree' .1 It is much less clear upon considering these examples that they lack confirmation. The intuitive force of Grigg's argument begins to wane once his somewhat misleading example is recognized as such. IV What then is the central difficulty standing in the way of confirmation for specific beliefs about God? The following example will help us here. Suppose I pray requesting of God a new Jaguar. I form the belief that God has provided me with a Jaguar. This belief stands in need of confirmation. The next morning I find a new Jaguar in my driveway, along with a letter containing the pink slip for the Jaguar registered in my name. A comparison of the Jaguar and breakfast cases will focus the discussion on the right issue. Grigg argues that the belief that I ate breakfast this morning is confirmed by the fact that there is one less egg in the refrigerator than was there yesterday. In this case I am able to predict that such and such a state of affairs will be the case given the truth of the belief to be confirmed. 'There will be one less egg in the refrigerator' will be true - ceteris paribus - if I ate breakfast this morning. But the prediction is possible only if it is also true that there are certain regularities in nature such that whenever one eats something it disappears from view. Grigg seems to have in mind confirmation which can, generally speaking, be done whenever one attempts to do so because there are certain true beliefs which link the belief to be confirmed to some predictable state of affairs. Likewise in the Jaguar example. If there is to be parity between the two cases there should be some type of prediction at hand which is possible because of the truth of certain linking beliefs I hold. Presumably the linking belief in the Jaguar case would be something like 'whenever I ask God for something he gives it to me'. The state of affairs which I could thus predict would be that in fact I do have a Jaguar registered in my name. This state of affairs confirms my belief that God has provided me with a Jaguar. 1 Plantinga does give the following example of a properly basic belief: 'This vast and intricate universe was created by God' (Plantinga: 80). It seems not to be the best example, however, for it breaks the ��logy between the paradigm beliefs and other beliefs about God. Although one could conceivably heve 'I see the world', it is not a belief which describes a literal case of perception. Two observations are in order. First, this type of confirmation relies upon discursive reasoning which, in the theistic case, amounts to natural theology. Let me explain. How does one know that the linking belief needed for confirmation is true? In the breakfast case the belief 'whenever I eat something it disappears from view' appears to be just a generalization from past empfrical experience in which every time I have eaten something it has disappeared. In fact, the confirmed and confirming beliefs involved in the breakfast case are them­ selves, when taken together, one more instance of confirmation for the linking belief or, alternatively, one more bit of inductive data which can be used as evidence for concluding that the linking belief is true. The linking belief is inferred from properly basic beliefs which describe empirical events or states of affairs such as 'eating breakfast' and 'seeing one less egg in the refrigerator than was there yesterday'. But the Jaguar case is significantly different. There is discursive reasoning in the background but while in the breakfast case the discursive reasoning begins and ends with claims about empirical objects or events, in the Jaguar case the reasoning begins with the presence of a physical object and ends with a claim about God. There is in the background of the confirmation an appeal to an independent source of information, viz. discursive natural theology. But, as already argued, this is ruled out. But suppose we reject the use of natural theology as part of the confir­ mation methodology (as we ought, given the argument in section rr) and move toward the suggestion given in branch (A) of the dilemma. Here we come to the second observation, viz. that even with the admission of cir­ cularity, confirmation for the belief that God has provided this Jaguar for me is not forthcoming. I lack the appropriate linking belief, for God does not give me whatever I ask for. This begins to raise the suspicion that the confirmation Grigg demands of theistic belief will not be forthcoming. There are then two central difficulties that any argument for the confir­ mation of theistic beliefs must overcome. First, one must avoid sneaking in the use of natural theology as a means of confirmation through the back door or provide an account of why this can legitimately be done. Second, one must explain the lack of predictive power of the theistic beliefs, arguing that it is not a feature necessary for confirmation of theistic beliefs. The possibility of using confirmation which relies on natural theology seems an unlikely one. The use of discursive reasoning which relies on empirical facts to justify the linking belief about God is nothing more than an attempt at having independent confirmation for the belief in question. Since the motivation for arguing that belief in God can be properly basic is that such justification is not available, it would be illegitimate (and quite likely unsuccessful) to use it here. Its use cannot be justified, thus it cannot be legitimately slipped in through the back door. It must be avoided. CAN BELIEF IN GOD BE CONFIRMED? Can the prediction problem find a solution through concentrating on confirmation of the circular variety mentioned in branch (A) of the dilemma? I have suggested that even if we grant a circular approach to confirmation, theistic beliefs remain unconfirmed. This claim stands in need of further analysis. Suppose we consider the belief ( I ) 'God created this flower' as a properly basic belief. If we follow the suggested route of circular confirmation this belief would have to be confirmed in (something like) the following way. Suppose one takes all properly basic beliefs of the form 'God created x' where x is some object in the world. Each would confirm (2) 'God created the world.' In turn, (2) could be used to confirm any belief of the form 'God created ' x and hence ( I ) would be confirmed. There is no appeal to natural theology here, for there is no move from some merely empirical data to a claim about God. This proposed line of confirmation fails, however. To see why, we first need a fuller account of the proposition held when one believes 'God created ' x . Such an account would go (something like) this: (3) 'The flower has the kind of (divine) beauty, intricate design, etc., which can only come from God's creation.' In other words, there is some feature of the flower which links the flower to God's creative touch. (There can be no inference from a merely natural feature to God's touch, for that would be natural theology again.) Now while (3) entails ( I ) (the former being a fuller account of the latter), (3) is not entailed by (2), for (3) has greater content that ( 1 ) but (2) does not have greater content than (3) nor does (2) claim the same thing as (3) or ( I ) . There must be some linking belief between (2) and ( I ) (captured in part by (3)) which indicates which features are linked to God's creative touch. More importantly, however, the linking belief must indicate when the confirming experience of sensing God's touch will occur or, more precisely, when we will recognize the feature which allows us to form the specific belief that God created the flower. The linking belief must refer to regularity. We can generalize from this and the Jaguar examples to the conclusion that there are no linking beliefs which enable us to predict when and in what way we will experience the handiwork of God in the revelatory way to which Plantinga points. One might even venture to claim that an empirical survey would reveal that such linking beliefs are not among the beliefs even of the most faithful theistic believers. So long as there is no linking belief which allows us to predict when we will have the appropriate con­ firming experience there will be no confirmation of the type for which Grigg calls. This then is the central difficulty in providing confirmation for theistic beliefs. c MARKS. M LEOD v My conclusion thus far is that the fundamental force behind the confirmation_ challenge is that predictive confirmation of theistic beliefs is not available. This problem is considered by Alston when he discusses the following disanalogies between what he calls 'perceptual practice' (PP) - 'our familiar way of objectifying sense experience' - and ' Christian practice' (CP) - 'the practice of objectifying certain ranges of experience in terms of Christian theology ' (Alston: 109). He writes: I believe that many people are inclined to take CP to be discredited by certain ways in which it differs from PP, by the lack of certain salient features of PP. These' include the following: ( r) With PP there are standard ways of checking the accuracy of any particular perceptual belief. If, by looking at a cup, I form the belief that there is coffee in it, I can check this belief for accuracy by smelling or tasting the contents; I can get other observers to look at it, smell it, or taste it; I can run chemical tests on it and get other people to do so. (2) By engaging in PP we can discover regularities in the behavior of objects putatively observed, and on this basis we can, to a certain extent, effectively predict the course of events (Alston: r 2 r). Alston further claims that I is simply a special case of 2. He writes: For our standard checking procedures in perceptual practice presuppose that we know a good deal about the ways in which things can be expected to behave in the physical world. Consider the appeal to other observers. Suppose I think I see a fir tree across from my house. What would count as intersubjective corroboration? Surely not any report of seeing a fir tree. If someone reports seeing a fir tree in Nepal, that will not tend to show that there is a fir tree across from my house. Nor will the failure of someone in Nepal, or across town, to see a fir tree have any tendency to disconfirm my report. Nor would it disconfirm my report if a blind man or one wholly preoccupied with other matters stands just where I was standing and fails to see a fir tree. The point is, of course, that only observers that satisfy certain con­ ditions as to location, condition, state of the environment (enough light), and so on can qualify as either confirming or disconfirming JTIY report. And how do we know what conditions to specify? We do it in the light of presumed regularities in the interaction of physical objects and sentient subjects. Persons in certain circum­ stances, and only in those circumstances, will count as possible confirmers or dis­ confirmers of my claim, because, given what we know about the way things go in the psychophysical world, it is only persons in such circumstances that could be expected to see a fir tree if there is one there .... Since r holds of a practice only if 2 holds, one can concentrate on the latter (Alston: r 22). One point of Alston's disanalogy, i.e. that CP lacks features I and 2 whi �e PP does not, is that the perceptual world is regular and on the basis of this regularity we can confirm or disconfirm our beliefs. For example, we can confirm the belief ' I ate breakfast this morning' because we hold the (accurate) beliefs that things do not pop randomly in and out of existence, CAN BELIEF IN GOD BE CONFIRMED? 3I 9 that things will be more or less as I left them unless someone has disturbed them, and that whenever I eat something it disappears from view. In short, the physical universe does not radically change without some reason or cause. Grigg's challenge follows a similar pattern. We do not find the regularities in religious experience and confirmation which we find in perceptual or memory experience. Theistic beliefs are not confirmed, and the practice by which theistic beliefs are formed is not validated, simply because the attempts at validation and confirmation depend on the regularity of the objects which the beliefs are purportedly about. They lack the possibility of what we can call procedurally predictive confirmation. God does not provide regular, consistent behaviour in which we can find patterns by which we can predict how or when he will reveal himself or truths about his relationship to the world. Alston admits that r and 2 really are disanalogies between PP and CP but claims they are irrelevant, for while the features discussed in r and 2 tend to show that PP is reliable, they are not necessary features of reliability. In fact, if features r and 2 were true of CP they would tend to show CP unreliable, given other features of CP. Alston writes: The reality CP claims to put us in touch with is conceived to be vastly different from the physical environment. Why should not the sorts of procedures required to put us in effective cognitive touch with this reality be equally different? Why suppose that the distinctive features of PP set an appropriate standard for the cognitive approach to God (Alston: l 28)? He continues by describing a possible state of affairs in which CP is quite trustworthy even though lacking features r and 2. He writes: Suppose, then, that (A) God is too different from created beings, too 'wholly other,' for us to be able to grasp any regularities in His behavior. Suppose further that (B) for the same reasons we can only attain the faintest, sketchiest, and most insecure grasp of what God is like. Finally, suppose that (C) God has decreed that a human being will be aware of His presence in any clear and unmistakable fashion only when special and difficult conditions are satisfied. If all this is the case, then it is the reverse of surprising the CP should lack [1 and 2], even if it does involve a genuine experience of God. It would lack l-2 because of (A).... Now it is compatible with (A)-(C) that (D) religious experience should, in general, constitute a genuine awareness of the divine; that (E) although any particular articulation of such an experience might be mist. aken to a greater or lesser extent, indeed even though all such articulations might miss the mark to some extent, still such judgments will, for the most part, contain some measure of truth; and that (F) God's designs contain provision for the co�r�ction and refinement, for increasing accuracy of the beliefs derived from rehg1ous experience. If something like (A)-(F) is the case, then CP is trustworthy even though it lacks features [ 1 and 2] (Alston: 129). I? other words, the epistemic practice of forming beliefs about God is different in an important way from the epistemic practice in which we form 320 beliefs about perceptual objects. The object of the former lacks the regularity of the object of the latter. While the dishes in the sink will remain in the sink and hence we can perceive them - ceteris paribus - when we try to, God does not necessarily act in this static fashion. But there may be an explanation for this lack of regularity, viz. God's revelation of himself is not confined by the regularities of the natural order. The lack of regularity in our experience of God, then, is no reason to reject the reliability of the practice. In fact, Alston claims, if the features listed above are true of God and the epistemic contexts surrounding belief in God, then the features not only show that r and 2 are not necessary for the trustworthiness of the practice of forming theistic beliefs. The featurns also show that if we did discover great regularity in God's dealings with us we would have a reason to distrust the deliverances of the practice. Thus, insofar as Grigg shares the point of view expressed by Alston's disanalogies Alston's strategy to free CP of the charges dealing with lack of confirmation will also be successful against Grigg's difficulty with theistic beliefs and the practice by which they are formed. The Alstonian reply is, simply stated, that the charge is irrelevant. So it should be with Grigg's charge. We can go ahead and admit that there is the disanalogy Grigg suggests between the paradigm practices and the theistic practice, viz. that the paradigm practices have objects which are regular in their relation to the knower while the theistic practice does not. But we can also claim that this is as it should be and the disanalogy is irrelevant. Of course nothing in Alston's claims entails that other kinds of confir­ mation are Impossible and here I wish tentatively to move beyond his suggestions. While Alston and Grigg are correct in their assessments of the possibility of procedurally predictive confirmation for theistic beliefs, there seems to be a type of non-predictive confirmation which may be available to some theistic believers for a certain kind of theistic belief, viz. general theistic beliefs. Suppose I hold the belief that God created the world. If I also form (by happenstance or God's grace which is not due to the regularity of God's activity) the belief that God created the flower then in some sense 'God created the world' would be confirmed. How can this be, given the argument above against procedurally predictive confirmation? Is it still the case that ( 2) 'God created the world', while seeming to entail ( r ) 'God created the flower' does not, for ( r ) is linked to (2) by (3) 'the (divine) beauty of the flower can only be explained by God's creation' and (2) does not entail (3)? This argument, I suggest, does not hold against non-predictive confirmation, for neither (3) nor any claim making reference to regular features of the object neither need be the link between (2) and ( r ) . This is the point of Alston's possible state of affairs in which God does not act predictably and yet is trustworthy. (3) is necessary for predictive confirmation, for it provides CAN BELIEF IN GOD BE CONFIRMED? 32I for the necessary regularity - the prediction that I will form the belief that God created the flower. It seems not to be necessary for non-predictive confirmation, however, for no such linking regularity is necessary. If I do form (I), it need not be linked to regular features of the flower. (I) need not be more fully describ(ld as (3) in order to allow for non-predictive confirmation. All that would be needed for such a non-predictive con­ firmation is an understanding of (2) which runs something like (4) 'God created everything.' If 'God created everything' is true, then it surely entails 'God created the flower'. But we need make no predictions of when or how we will form the specific belief which would confirm (2) or (4) if we do form it. A non-theistic example can help. It is widely believed that there are no unicorns. Nevertheless, suppose it is also widely believed - it shows up in all the stories about unicorns - that all unicorns eat eucalyptus leaves. Now there is no reason to expect confirmation of the belief that all unicorns eat eucalyptus leaves, for there are no unicorns, or so it is believed. Suppose, however, one discovers one day a unicorn in one's backyard eating euca­ lyptus leaves. Surely this would confirm, even if weakly, the belief that all unicorns eat eucalyptus leaves. One would never have predicted such con­ firmation but upon receiving it, it would have at least some prima facie plausibility as confirmation. If we add to the story that unicorns, like God, are somewhat circumspect in their self-revelations to humankind, only showing their faces when they wish, the case doesn't change. Some con­ firmation would be provided. Non-predictive confirmation of general beliefs seems at least plausible. Of course, even if all this is true I may never form the specific belief that God created the flower nor any other specific theistic belief. Nevertheless, if I do there seems to be no reason why I shouldn't take it to confirm the general belief. For all of this, however, the prima Jacie confirmation which might be provided may be outweighed by other considerations. There may be some confirmation that God created the world but there may also be disconfirmation, such as is provided by the presence of evil. This seems to the point behind Basil Mitchell's parable about the Stranger and the partisan. Be that as it may, however, another question needs asking. Can specific theistic beliefs be confirmed in a similar manner? The answer here is much less clear. I have suggested that beliefs of the form 'God created x' may non­ �redictively confirm 'God created the world' so long as no linking regularity is required. But for the specific belief 'God created the tree' to be confirmed, it must be confirmed by the general belief 'God created everything'. There must either be some link up between the confirmed belief and some other specific theistic belief via the general belief or the general belief itself must ronfirm. For example, either some belief of the form 'God created x ' ronfirms 'God created everything' which in turn confirms 'God created the . 322 c MARKS. M LEOD flower' or one is left simply with the general belief 'God created everything' confirming the specific belief 'God created the flower'. Now while we may be able to get the specific belief confirming the more general belief, to get the general belief confirming the specific is questionable. Just as it seems some­ what odd to suggest that 'all swans are white' confirms 'this swan is white', it seems somewhat odd to say that 'God created everything' confirms 'God created this tree'. In confirmations of general by specific beliefs there is a possible (deductive) inference from the general to the specific. This is not true with confirmations of specific by general beliefs. The specific belief 'God created the flower' clearly does not entail 'God created everything'. To return to Grigg's original charge, then, we might suggest that he did concentrate on the appropriate belief, the general one. But now there seems to be a somewhat plausible understanding of how such a belief might be non­ predictively confirmed and it is the specific beliefs which lack this somewhat strange and very weak kind of confirmation. Could we ever validate theistic practice? In other words, if I had many experiences in which theistic beliefs were formed could I move from the non­ predictive confirmation of individual beliefs to the validation of the practice? The answer to this question brings us to some final comments on the relationship between validation and the two types of confirmation. It was earlier suggested that perhaps when one confirms a sufficient number of beliefs delivered from a given practice, the practice is then validated. It seems clear enough that what connects the confirming instances and the vali­ dation of the practice is an inductive argument which relies on the regularity of the objects about which the beliefs are formed. Induction, resting as it does on our trust in the regularity of nature, provides grounds for the move from confirming instances to validation only if the objects about which the beliefs are formed are regular. If this is correct, then insofar as non-procedural, non­ predictive confirmation does not rest on regularity, the inductive inference will not be possible. We could never have validation of the theistic practice. This aligns well with Alston's claim that if we discovered that God was dealing with us in a predictable fashion, we would have evidence that the theistic practice is unreliable. To expect validation of the practice is to expect too much. We will have to be satisfied with non-predictive, non­ procedural confirmation of individual beliefs or perhaps, for most of us, no confirmation whatsoever. Thus, the requirement that theistic beliefs be confirmed (or that the practice leading to such beliefs be validated) in the sense that confirmation (or validation) requires predictive regularity is overly strong. The demand assumes that regularity of the object of belief is a necessary feature for confirmation of belief. In turn, this assumes that rational beliefs can only be formed about objects for which regular, predictive confirmation is possible. But what is necessary for rationality is not regularity but trustworthiness or CAN BELIEF IN GOD BE CONFIRMED? reliability. Alston provides an account of how a practice can be reliable without having a regular object behind it. I have provided an extension of his account, suggesting how confirmation of some beliefs about non-regular objects may occur. Of course, whether a given belief about a non-regular object is confirmed is an empirical question to be answered by whether or not one has the appropriate experiences. Grigg's challenge is either irrelevant or false. It is irrelevant if it rests on the claim that the object about which rational beliefs are formed must be regular, for regularity is not a necessary condition for confirmation nor for rationality. It is false if Grigg claims that no confirmation of any kind is even possibly available for beliefs about non-regular objects. The non-regularity merely entails a non-regular, non�predictive type of confirmation, if any at all. Hence, so far as this challenge is concerned, beliefs about God can be rational. VI The results of the essay can be stated in more positive terms. The confir­ mation of a rational belief is internal to a set of beliefs, the practice which forms that set, and, for beliefs dealing with regular objects, a set of back­ ground linking beliefs. Insofar as a person follows a certain kind of belief­ forming practice and forms a mutually cohering set of beliefs by that practice, the deliverances of the practice may be individually confirmed (when pos­ sible) and (when the linking beliefs are in place) the practices themselves validated. But there is a certain kind of circularity inherent in any kind of confirmation both of the paradigm beliefs and, if theistic beliefs are con­ firmed non-predictively, theistic beliefs. Theistic beliefs, of course, are not beliefs about regular objects and not everyone has a disposition to form theistic beliefs. Nevertheless, for those who do have that disposition, its deliverances can be confirmed when and if one has the right kind of experiences - perhaps if God is particularly gracious to you. Even if one is not given that special grace, however, one's theistic beliefs do have epistemic parity with beliefs delivered from dispositions we all share, at least so far as predictive confirmation is concerned, for predictive con­ firmation is irrelevant to theistic belief. The confirmation challenge fails. Westmont College Santa Barbara California, U.S.A.