‘Unfortunately for the good sense of mankind, the fact of their fallibility is far from carrying the weight in their practical judgment, which is always allowed to it in theory; for while every one well knows himself to be fallible, few think it necessary to take any precautions.’Footnote 1 Or so says Mill in On Liberty. According to Mill, when – all too rarely – a man does allow reflection on his fallibility to raise in his mind the question of his justification for one or more of his beliefs, he misguidedly supposes that the only precaution he needs to take is to appeal to the beliefs of his neighbours. So, Mill continues,
… in proportion to a man's want of confidence in his own solitary judgment, does he usually repose, with implicit trust, on the infallibility of ‘the world’ in general. And the world, to each individual, means the part of it with which he comes in contact; his party, his sect, his church …. Nor is his faith in this collective authority at all shaken by his being aware that other ages, countries, sects, churches … have thought, and even now think, the exact reverse. He devolves upon his own world the responsibility of being in the right against the dissentient worlds of other people; and it never troubles him that mere accident has decided which of these numerous worlds is the object of his reliance, and that the same causes which make him a Churchman in London, would have made him a Buddhist or a Confucian in Pekin.
The conclusion Mill himself was keen to draw from such considerations was political: such fallibility in man's understanding of the good justifies a liberal pluralism. For reasons that are both epistemic (the free interchange of conflicting ideas will advance us on the road to the right ideas) and pragmatic (we generally do more harm than good when we interfere with those whose actions affect no-one's good but their own), the state should allow any experiments in living that do not result in one individual's liberty infringing that of another. But as well as their political implications, which may or may not follow in so direct a way as Mill suggests, Mill's considerations also seem to point to a conclusion in religious epistemology. Despite Mill's not having drawn it out, there is an argument implicit in his comments that is germane to both externalist and internalist understandings of the epistemic justification of religious beliefs, even though some of these understandings would not wish to use the term ‘epistemic justification’ to refer to whatever it is that they suggest must be added to true belief for it to count as knowledge. In this paper, we shall articulate this argument; examine how it challenges those religious believers who would wish to claim their religious beliefs as knowledge; and consider what they may do to meet this challenge.Footnote 2
Mill's own talk of fallibility might make one think that reserve about claiming knowledge must rest on an extreme infallibililism with regard to the nature of knowledge, the rational and well-informed ‘Churchman’ being reluctant to claim knowledge in a certain area given that there remains the possibility of error in his judgements in that area. However, in fact, as we shall see in unpacking it, no such exigent view of the requirements of knowledge is being supposed in Mill's argument. Rather, the argument is that when one observes the actual geographical and historical groupings of people holding particular religious beliefs, one realizes that the processes that give rise to religious beliefs and the contents of the beliefs to which they give rise have nothing straightforward to do with the facts that would make one of those sets of beliefs true. These religious truth-making facts do not cause in any straightforward way people to have true beliefs about them. (We shall explore this notion of straightforwardness at greater length in due course.)
That being so, if one does happen to be in an area and time where the religious beliefs which have been formed in one by these processes are in fact true, the fact that they are true is not one that one can justifiably suppose has caused one to believe that they are via a process that is reliable in instilling true beliefs. Thus, if one is an externalist, one should not count them as knowledge. If one does happen to be in an area and time where the beliefs that one comes to concerning which are good and which are bad reasons for the holding of religious beliefs are true beliefs about which are good and which are bad reasons, that too appears to be a matter of luck. Thus, if one is an internalist, one should be chary of claiming one's beliefs about the level of justification enjoyed by one's religious beliefs as knowledge. Knowing that these things are a matter of luck, the Churchman – even one who has been on a course in the philosophy of religion and thus thinks of himself as having reasons for his religious beliefs – has no ‘higher-order’ reason to think that what strikes him as a good lower-order reason for continuing to hold his religious beliefs is more likely to be a good lower-order reason than what would strike a Buddhist or a Confucian as a good lower-order reason for their continuing to hold their religious beliefs. Thus, an externalist should abandon claims to religious knowledge and an internalist should abandon at least his claims to know that he knows his religious beliefs to be knowledge. This, I suggest, is Mill's argument in a nutshell.Footnote 3 Now we shall look at it in more detail.
Let us start by granting for the sake of argument that the Churchman in London is right in his religious beliefs: there really is a God of the sort he believes in and this God has done all that this Churchman believes him to have done. Mill's argument appears to be that, even granting this, the actual processes by which the Churchman has come to his beliefs are such as to be discomforting for any claim on his part to know these things. The reason is that the same processes operating on people no more dilatory than he in Peking have led them to ex hypothesi false beliefs. It is, one might say then, only luck that the Churchman is right, an accident of his personal and perhaps social history. Drawing out the luck inherent in the Churchman's happening upon the truth is, we may take it, the point of Mill's reflections to the effect that had the Churchman been sent as an infant to Peking and brought up by devout Buddhists or Confucians there, he would by now have held what are – ex hypothesi – the wrong religious beliefs just as devoutly as he now holds what are – ex hypothesi – the right ones. Had Islam conquered the West in the Middle Ages, then, without leaving London, this Churchman – now presumably a ‘Mosqueman’ – would, had he still existed, have come to hold different and ex hypothesi wrong religious beliefs, finding the ‘world’ with which he was then in contact supporting him in his adherence to Islam against any ‘infidel’ churches that still survived elsewhere on the globe.
Let us suppose for the moment that Mill is right in claiming that this sort of contingency is pervasive in the causes that lead people to their religious beliefs; we shall come back to reflect on this premise later, but for the most part we shall aim to see what the religious believer who wishes to claim his or her religious beliefs as knowledge can do against Mill's argument if he or she accepts whatever empirical facts Mill is gesturing towards with his claims about the alternative histories that people could have led. If contingency playing this sort of role is sufficient to render any beliefs one comes to as a result of it – however true they might fortuitously happen to be – unjustified, then it follows that the religious beliefs of the Churchman in London are no more justified than those of a Buddhist or Confucian in Peking; indeed, none of them are justified at all. So, the question must now be, ‘Why might one think that this sort of contingency undermines epistemic justification?’. Needless to say, the answer to this question varies depending on the conception of epistemic justification with which one operates.
Sadly, from the point of view of providing a linear treatment of Mill's argument, there are many and diverse accounts of epistemic justification, some of which would – as already observed en passant – eschew the terminology of epistemic justification altogether. However, for present purposes we may use the mutually exclusive and exhaustive division between externalist and internalist accounts and we may reasonably suppose that, if we stay at a level of sufficient generality, we may make Mill's argument cogent to all religious believers, regardless of which variant of an externalist or internalist account of epistemic justification they endorse. Let us first look at Mill's argument as it is relevant to an externalist understanding of epistemic justification.
On externalism
According to Mill, it is not straightforwardly due to the metaphysical, other-worldly, facts that make the Churchman right (as we are supposing he is) that the Churchman believes as he does. Rather, much more mundane, this-worldly, facts about how the Churchman has been brought up are those which are causally responsible for him having the beliefs that he has. For some, causal, theories of epistemic justification, this alone will be enough to crush any claims by the Churchman that his religious beliefs are knowledge. But even for a more relaxed reliabilism, it seems as if it puts at least some pressure on any such claims. The Churchman no doubt believes that God created the universe. If he is right, then had God not created the universe, it is not true that the Churchman would still have believed this, for there would then have been no universe and thus no Churchman to believe anything at all. So it is not true that someone accepting Mill's claims concerning the manner of the proximate causation of religious beliefs need concede that had these religious beliefs not been true, then the people believing them would have believed them in any case; at least he or she need not concede this for all religious beliefs. But what is true is that the causal connections between the facts in virtue of which a religious belief is held (if Mill is right about the proximate causation of such beliefs) and the facts which make those religious beliefs true are not ‘straightforward’, as I put it. If Mill is right, then the Churchman should agree that it is much more mundane, this-worldly, facts about how he has been brought up that are the most proximate causes of the religious beliefs that he has.
He may of course maintain that the facts that make those religious beliefs true do play some part in the causal history of his beliefs; some of them concern the setting up of the universe and the setting up of general belief-acquiring processes of the sort that Mill has described and, we are taking it for the sake of argument, described accurately. This move on the part of the Churchman does not, however, seem to do anything to remove the worry concerning the epistemic justification of those religious beliefs on any plausible externalist understanding, for – assuming as we are that the Churchman is right – God has set up a universe whereby the processes which spring ultimately from his act of creation lead through divergent cultures to people having divergent beliefs, and in many (indeed most?) cases false beliefs, about him. Of course some end up not believing in him at all. It is true that had God not created the universe, then nobody would have believed anything about him, but it is also true that given the way the Churchman must admit God has set up the universe, lots of people (perhaps most of them) do believe false things about him and some don't even believe he exists at all. Thus, whilst the Churchman may maintain that had at least some of his religious beliefs been false, he would not have believed them, he cannot maintain that had they been true in slightly different circumstances (e.g. his being sent to Peking as an infant), then he would still have believed them. Thus, for someone convinced that epistemic justification requires some sort of ‘truth-tracking’ the Churchman's religious beliefs will not count as knowledge if Mill is right about the contingency of the processes by which they have originated. Or so it seems.
That Mill's argument threatens the externalist who wishes to claim his or her religious beliefs as knowledge would be challenged by some. Plantinga, for example, asserts that even if Mill is right about the mechanisms by which believers come to believe, a believer's religious beliefs
… could be produced … by a reliable belief-producing process. Calvin's Sensus Divinitatis, for example, could be working … in such a way as reliably to produce the belief that [the God of Classical Theism exists]; Calvin's Internal Testimony of the Holy Spirit could do the same for [the belief that the Christian variant of Theism is true].Footnote 4
Plantinga's bifurcation of the process of belief-acquisition into that generated by the sensus divinitatis and that caused by the internal testimony of the Holy Spirit complicates matters somewhat and for our purposes we give ourselves maximum room for manoeuvre if we allow ourselves to leave Plantinga and Plantinga's Calvin to one side and consider a Churchman, who we may nevertheless call ‘Calvin’, whose externalist epistemology we construct solely with the aim of giving him the best resources with which to resist Mill's conclusion that his religious beliefs are not knowledge. We shall see that Calvin as we shall construct him can consistently claim his religious beliefs to be knowledge, although some relatively heavy-duty metaphysics ultimately need to be introduced in order for him to do so.
Calvin asserts that he has received through the operation of the sensus divinitatis/the instigation of the Holy Spirit the correct religious beliefs, including the correct religious beliefs about how these beliefs have been instilled in him. Two options are open to Calvin in the way he presents the operations of the sensus divinitatis/the Holy Spirit. It could be presented as indifferent over whether it works in accordance with or contrary to the cultural promptings of society, performing what may be in effect Humean-miraculous acts of belief formation in those on whom it descends. However, depicting its operations in this way would be liable to render ‘Calvinism’, as we may call it, amenable to refutation by empirical observation if one does not in fact find Calvinist Churchmen popping up in Confucian temples and the like. Fortunately then, this is not the only option for Calvin to take; nothing in Christianity suggests that each belief-token or even many belief-tokens of Christian beliefs are caused by natural-law-violating miracles. As well as working in mysterious ways His wonders to perform, it may well be that God works in mundane ways our beliefs in Him to inform.
As we are trying to see what the believer who claims his or her religious beliefs to be knowledge may do to meet Mill's arguments without challenging his empirical premises, so let us suppose that Calvin goes down this second road, asserting that the sensus divinitatis/Holy Spirit ‘utilizes’ natural belief-forming mechanisms. If the Holy Spirit's operations are so understood, nothing Calvin says need be in contradiction to the facts of the geographical and historical distribution of the ex hypothesi correct religious beliefs, which are, we may take it, as Mill describes. Perhaps everyone in every culture has some sense of divinity (a sensus divinitatis one might call it), but God has decided in his wisdom to allow the contingencies of history to determine which cultures allow this to flower into a belief in the correct religion (Calvin's own) and which corrupt it so that it leads to belief in a false religion (Buddhism and Confucianism, for example). Perhaps some work may be done by bifurcating this process; within monotheistic cultures, sometimes the Holy Spirit instigates belief in full-blown Christianity and sometimes it fails to do so, leaving what we might call the residual sensus divinitatis to do its best to instil belief in monotheism.
One may think that the distinction between the operations of the sensus divinitatis and those of the Holy Spirit may be usefully introduced here, to reduce the worry that God would wish all humans to know of His plan of salvation, if one thinks such a worry would be reduced were one to picture God as leaving some sensus divinitatis intact in everyone, intact albeit marred after a historical Fall; post-Fall, one needs the Holy Spirit to come to correct religious beliefs fully. In any case, whilst the distinction between the sensus divinitatis and Holy Spirit is treated somewhat laxly in this paper, there is scope for tightening it up and deploying it in this fashion or others, if other parts of one's philosophical theology would benefit from it.Footnote 5Ex hypothesi, in Calvin himself the sensus divinitatis/Holy Spirit has performed optimally through the natural processes of belief acquisition to which Mill has gestured in his comments, leading Calvin fully to correct theological views. So far, so good.
Calvin now faces the issue of what to say of the seeming possibility that he himself might have not had his sensus divinitatis cultivated in the right way/the Holy Spirit descending on him effectively, had he been sent as an infant to Peking, for example. Let us first consider whether it is permissible for Calvin to allow this as a possibility and merely claim himself epistemically fortunate that it was not an actuality. Can he say, ‘Indeed. But so what? For many beliefs that have in fact been formed by reliable enough processes to elevate them to the status of knowledge, there are alternative histories that would have led to those processes not operating. Of Peking, I say simply, “There, but for the grace of God, go I”’? There are two levels of luck that need to be distinguished here. The first is the luck present (or not) in a belief-acquiring process pertaining to whether or not that process yields true belief, a less reliable process being more lucky if it's given rise to true beliefs on this occasion than a more reliable one would need to have been. The second is the luck present (or not) in the process by which someone comes to be subject to a process of belief acquisition that gives rise to certain outputs. May Calvin maintain that he is lucky in the second sense but not in the first? I shall argue that this is a legitimate way for Calvin to respond.
Calvin could be an indeterminist concerning the laws of nature, indeterminism allowing an opportunity for the sensus divinitatis/Holy Spirit to cause what would otherwise be uncaused quantum happenings in the brain and through them lead to religious beliefs without its performing Humean law-violating miracles. On this model, the actions of the sensus divinitatis/Holy Spirit would be similar to the most fundamental physical changes which souls are posited as performing on the standard libertarian account. In the case of the sensus divinitatis/Holy Spirit's actions, these quantum happenings would lead to correct religious beliefs, just as on the standard libertarian account, microscopic, indeed sub-microscopic, happenings directly caused by souls lead in due course to macroscopic free actions. Going down this road, Calvin could assert that there is a genuine metaphysical difference between the processes by which he has come to the correct religious beliefs – the quantum event (or events) in his brain that led in due course to these beliefs was (were) in fact caused by the sensus divinitatis/Holy Spirit – and the processes by which the Buddhist or Confucian has been led to incorrect beliefs – in their brains, superficially similar quantum events really were uncaused. This difference may serve as the hook off which Calvin can hang a difference in reliability between the two processes; for this Calvin, the truth-making facts are in very close causal connection with the beliefs of his that they make true and reliably produce those true beliefs in him; for Buddhists, the same does not hold.Footnote 6 Calvin may then agree with Mill that it is indeed just luck that he's the subject of a reliable – i.e. non-lucky – process, but assert that this higher-order luck doesn't make the lower-order process any more lucky – less reliable – or thus the true beliefs he comes to by it any less knowledge. Of course, Calvin need not put his cultural positioning down to luck; he may say that it is Providence; it is not a matter of mere good luck that he did not go to Peking as an infant, but a matter of grace. A model which saw some souls pre-elected to epistemic grace would mean that an apparently chancy distribution would not in reality be such. However, with this move Calvin would approach the ‘alternative’Footnote 7 solution to Mill's dilemma which we shall discuss next.
If we suppose that instead of indeterminism, Calvin takes a determinist stance, the danger is that he will, it seems, need to concede that the same processes lead the Churchman to his beliefs as lead the Buddhist and Confucian to theirs. There is however scope for Calvin to resist Mill's sceptical conclusion by focusing on how this process could not have led him – the individual who is in fact Calvin – astray, for Calvin may maintain that he is one of the Elect and the Elect as such are never led astray by this process, a process which does indeed lead to error in others, the non-Elect. Let us suppose then that this determinist Calvin asserts that the seeming possibility of him being sent as an infant to Peking is not a real possibility at all. Of course it's not; every seeming alternative history is not a real alternative if determinism is true. As God has pre-destined that the souls of the Elect will be born and continue on in situations where they will find themselves naturally caused to come to the correct religious beliefs and as Calvin is, ex hypothesi, one of the Elect, so it was in fact impossible for him to have been sent as an infant to Peking. There was no way that he, Calvin, being one of the Elect, was going to find himself with any false religious beliefs as a result of the natural belief forming processes that Mill gestures towards.
As a matter of empirical fact, we may observe that there are people who claim to be amongst the Elect yet differ from Calvin in some of the religious beliefs that they hold, indeed some of their religious beliefs are in contradiction to those of Calvin. Let us call such people heretic Calvinists. What should Calvin say about them?
Two moves are open to Calvin at this stage. Calvin could be inclusive, allowing that heretic Calvinists are amongst the Elect after all, even though insofar as their religious beliefs contradict his they are, he asserts (and we hypothesize correctly), in error. However, if Calvin goes down this road, then Mill's argument is germane at least to the religious beliefs over which the Elect can differ. The core Calvinist beliefs remain immune to Mill's worries, but the non-core ones (viz. those over which God has allowed the Elect to differ and hence the heretic Calvinists to be in error) are in the same position as are all beliefs for the non-Calvinist Churchman we have previously been considering. Alternatively, Calvin could be exclusive. He may assert that any differences in religious beliefs (from those held by Calvin) whatsoever are such as to indicate that one is outside the Elect and anybody whose beliefs are exactly co-extensive with those of Calvin is someone who thereby indicates that they are amongst the Elect. Thus can Calvin maintain that the sensus divinitatis/Holy Spirit's belief instigation is completely reliable (for the Elect) whenever it operates (as it only does for the Elect).
Calvin could be ecumenical enough to acknowledge that the sensus divinitatis has brought home the truth of God's existence to many who are nevertheless overall in error (that process is reliable whenever it operates and it operates in these heretic Calvinists too, assuming heretic Calvinists do at least manage to agree with Calvin on the central tenets of theism) whilst maintaining that the full-blown sensus divinitatis/Holy Spirit process which gives people all and only Calvin's beliefs (which is also reliable whenever it operates) only operates to this full extent in the Elect. Whichever way he goes on this issue, Calvin will admit that some people think they are amongst the Elect and share a lot of theology with Calvin, yet from the fact that they differ, these heretic Calvinists may be known (by the Elect, not of course by themselves) not to be amongst the Elect. ‘From the inside’, as it were, heretic Calvinists have no less reason for supposing that they're right in their claims to be the Elect than Calvin has for supposing that it is he who is right in his claims, but these heretic Calvinists are in fact wrong and Calvin is in fact right.
One might worry whether Calvin is not, after all, committed by this latest series of manoeuvres to the view that the belief that one is amongst the Elect is not caused in people by a reliable enough mechanism for knowledge; after all, whilst some of the people who come to the belief that they're amongst the Elect are right, some are wrong. Such a worry is misguided. If an indeterminist, then Calvin may say that the process by which the Elect come to the belief that they are amongst the Elect is a different one from that by which the non-Elect come to the belief that they are amongst the Elect. If a determinist, he may assert that for the Elects' souls the natural processes that lead to them thinking they are amongst the Elect are completely reliable. It is only for the souls of the non-Elect that these processes are unreliable. This last is not some empty stipulation about how the terms ‘Elect’ and ‘non-Elect’ are to be used. Which souls are amongst the Elect and which are not is a substantive issue; it is not that one's soul gets to be amongst the Elect simply in virtue of its being one in which the sensus divinitatis/Holy Spirit works optimally. Rather, it gets to be amongst the Elect by a prior act of divine fiat; it subsequently – perhaps of metaphysical necessity, but not of linguistic stipulation – gets to lead a life in which the sensus divinitatis/Holy Spirit works optimally through the natural processes to which Mill gestures. The heretic Calvinist's soul was not destined for (complete) truth, so it was not assigned a life history as a result of which it would only come to true beliefs. Thanks to the grace of God, matters are different for Calvin.
Another worry one might have with these manoeuvres is that they immediately raise the spectre of a potential defeater to what we may take it is another of Calvin's religious beliefs, the belief that God is perfectly good and all-powerful. Does God's pre-destining the souls of the non-Elect to error, for example, not stand in some prima facie tension with His perfect goodness? But here, as with the problem of evil in general, Calvin may avail himself of a number of strategies; the one his historical namesake took is but one. An attractive alternative would be to minimize the eschatological significance of being in the Elect; the Elect in the sense we are talking about here are first and foremost the epistemically Elect (those destined to religious knowledge this side of the grave); it is compatible with this sort of ante-mortem division between the epistemically Elect and non-Elect to maintain a universalism which sees both kinds of people ultimately saved.
Calvin can, I conclude, escape Mill's challenge, partially if an inclusivist, completely if an exclusivist. And he can escape it in two ways, depending on whether he is an indeterminist or a determinist. Navigating either of these escape routes forces him to carry some relatively heavy metaphysical commitments with him, but these are commitments that some who would wish to call themselves Calvinists have been willing to carry. Plantinga, the most prominent contemporary exponent of something akin to Calvinism, would, one presumes, take the indeterminist tack.Footnote 8 Gellman considers doing something similar. Gellman is working with the assumption that Judaism is the correct religion, so if we switch to that assumption and adjust Calvinism accordingly, we may say of a ‘Jewish Calvinist’ that if
… he believes, for example, that the essence of a person is his soul, and that souls are uniform across the Jewish–Gentile divide, then he will have reason to believe that a person who is born a Jew, could have been born a Gentile instead [and thus would have got significantly false beliefs]. Suppose, however, he believes … that … Jewish souls are essentially different from Gentiles' souls …. Then he may rationally believe that he, a Jew, could not have been born a Gentile.Footnote 9
Thus, he may rationally believe that he could not have been led to false beliefs by his Jewish sensus divinitatis.
In any case, we may conclude that if externalism is correct and if Mill is right about the processes by which people come to their religious beliefs, then if and only if one of ‘Jewish Calvinism’, ‘Christian Calvinism’, ‘Islamic Calvinism’, et cetera is correct, do the adherents of whichever one is correct have epistemic justification for their religious beliefs sufficient for them to count as knowledge. The ‘and only if’ has been defended only indirectly, but what we have said is sufficient for us now to be in a position to see that without some way of either portraying the process by which one's religious beliefs have been formed in one as metaphysically different from that operative in those with whom one disagrees (in the manner of the indeterminist Calvin) or portraying the people in whom an otherwise similar process operates as metaphysically different (in the manner of the determinist Calvin), i.e. without cancelling out lower-level epistemic luck at a higher level, something which requires relatively heavy metaphysical commitments of a Calvinist sort, Mill's argument must undermine religious knowledge on any externalist account of knowledge.
Once we appreciate that Calvin should recognize that an indefinitely large range of theists with whom he disagrees – as we have seen there can be Jewish Calvinists, Christian Calvinists, Muslim Calvinists, and many variants within each of these broad positions (the non-Elect in each case, for example) – and non-theists can make precisely the same kind of case for the epistemic justification of their conflicting beliefs as he can for the epistemic justification of his own, it is tempting to say that he ought to feel less entitled to maintain his claim to religious knowledge. However, this temptation is precisely the one which, qua externalist, one has already hardened oneself against. The worry here is in essence the same one that is sometimes raised about externalist solutions to the problem of the external world, viz. ‘But the person who is a brain in a vat would have been led to by parallel reflections to conclude that they knew that they weren't’. This then is a worry which one might have with externalism per se, not one which one might have with it particularly as applied to the case of religious knowledge. Nevertheless, it is a worry which many people will have and which naturally leads one to consider how things stand for the internalist. Initially, at least, it seems as if the internalist will have an easier time of it.
On internalism
If one did not think that people were at least sometimes capable of being moved in their beliefs by their appreciation of good reasons, then one would abandon philosophy as such. If then one accepts that beliefs being affected by reasons is a possibility, one may say that even after one has accepted that one has been caused in some non-rational manner to acquire one's religious beliefs, one could consider whether there is in fact some argument or evidence that should lead one to prefer to continue on with them or that should lead one to prefer to abandon them; find, let us say, that there is more reason to prefer to continue with them than to abandon them; and thus thereafter have religious beliefs that one has managed to elevate beyond their non-rational origins and thus, if epistemic justification requires some internal awareness of good reasons, in the epistemic justification of which one could now rationally believe. If, as presumably an internalist would suggest, epistemic justification is merely a matter of one's having a sufficient awareness of what are sufficiently good reasons for the belief in question, then – pending an investigation into whether or not there are any sufficiently good reasons for particular religious beliefs – one cannot know in advance that, even if Mill is right, the Churchman, at least if he's been on a suitable course in the philosophy of religion, might not yet have epistemic justification for his beliefs and thus they be properly counted by him as items of knowledge. What would Mill say to this? In answering this question on Mill's behalf, we go beyond the evidence that Mill himself has left us, but we can extrapolate: he would have drawn a comparison between the fields of science and some areas of philosophy on the one hand and religion on the other, a comparison unfavourable to religion.
One can perhaps truly say that had Ptolemy been transported forward in time having penned his last words on his own theory and then been exposed to the sort of evidence and reasoning that one is exposed to on a course in astronomy in the present day, he would have conceded that Copernicus, rather than he, was right. However, it is harder to believe that the Churchman can say anything similar about the Buddhist or Confucian or the Buddhist or Confucian say anything similar about the Churchman. Enrolling a confirmed Churchman in a course in Buddhist metaphysics now may be expected to lead to a greater understanding of the intricacies of this system on his part, but not a conversion to it. A practising Confucian may come to be educated in London and graduate with a first class degree in philosophy and theology, but not be motivated by his learning to attend church thereafter. Furthermore, it seems likely that Mill would insist that had the Churchman been sent as an infant to Peking and then, later in life, gone on a course in the philosophy of religion run by his local temple there, he would now be exactly as convinced that he had been scrupulously rational in vindicating his Buddhist or Confucian beliefs as he is now, in London, convinced he has been in vindicating his Christian beliefs having gone on a course in the philosophy of religion run by his local seminary. If this is right, the Churchman must concede that even if he does have objectively good reasons for his religious beliefs in a way that the Buddhist or Confucian does not, and we may grant for the sake of argument that he does, the fact that he has been able to classify the objectively good reasons to believe as he does, as objectively good reasons, is a matter of luck; had he been sent to Peking as an infant, then what he now takes to be good reasons he'd take to be bad and what he now takes to be bad, he'd take to be good. In other words, our assessment of what is a good and what is a bad reason in the arena of the philosophy of religion (in contrast, one might suggest, to astronomy) seems to be determined less by the goodness or badness of the reason and more by facts of birth and upbringing.
Given that amongst philosophers of religion there is – Mill might maintain as a matter of empirical fact – as much controversy about what are and what are not good arguments for and against the existence of God as there is amongst religious and irreligious people about whether or not there is a God, it would, by another iteration of Mill's argument, be unreasonable to suppose that there are good reasons on one side or another that one has oneself – against the odds – been capable of detecting as such. Of course this has painted the contrast between the philosophy of religion and astronomy too sharply, but Mill might argue that blurring the contrast does nothing to elevate the philosophy of religion above the cultural swamp; it merely sinks astronomy. In fact, many scientists show themselves just as unwilling as philosophers of religion to give up the cherished models with which they have been brought up. Large issues in the sociology of knowledge open like pits before us. Let us skirt around them by asking whether, even if the bleakest view available here were right, it would matter. It is not obvious that it would. (In a moment we shall come back to examine the implausibility of maintaining that the bleakest view is right, in any case.)
The Churchman, if he is an internalist about knowledge, may concede that it is just a matter of luck that he has been able to classify as objectively good reasons for his religious belief the reasons that he has. Nevertheless, as long as he is right in so classifying them – these really are objectively good reasons for his religious beliefs – then that is sufficient for these religious beliefs to count as knowledge. The facts made mention of in the previous paragraphs, if facts they be, may suggest that he has no reason to suppose that he is right in so classifying them, for it may be that these facts show that, in the field of the philosophy of religion, whether or not one believes that a good reason is a good reason has nothing to do with its status as such but everything to do with accidents of one's history. But even if it is true that this Churchman cannot have good reason to classify his belief that these lower-order reasons are good as in itself an instance of knowledge, this does not undermine this Churchman's more basic religious knowledge. Even if it is true, it does not mean that he does not know, just that he does not know that he knows. The only principle that could take one from conceding that he does not know that he knows to his not knowing would be the principle that for genuine knowledge of p, one must know that one knows that p, but this principle, generating as it does an infinite regress, denies the possibility of knowledge for anyone but God (who we may take it could have an infinite nesting of beliefs), so we should surely eschew it.
All in all, it looks then as if the Churchman, if he is an externalist, had better concede to Mill that the fact that the religious beliefs that he holds are true beliefs rather than false is a matter of luck in a way that disbars him from claiming them to be items of knowledge unless he is a Calvinist. If he is an internalist, he may be able to claim his religious beliefs to be items of knowledge, if he has done enough theology, but his classification of theology, in contrast to the Buddhist or Confucian's metaphysics, as a source of knowledge, he should admit, is one the accuracy of which he cannot know. He should admit this if (and only if) he concedes that if he'd been educated in a Buddhist temple, he'd now be as certain that Christian theology was a compilation of bad reasons for false or meaningless beliefs as he is now convinced that Buddhist metaphysics is a compilation of bad reasons for false or meaningless beliefs. In seeking to see how the Churchman could yet respond to Mill's argument even if the ‘bleakest view’, as we called it, available here were correct, we have not questioned this concession, and we should note, at least in passing, that in fact there is good reason to question it.
The standards of rationality to which natural theologians appeal are not plausibly solely internal to theistic commitment; neither are the premises from which they proceed. That being so, theistic natural theologians would not at all agree that when they judge a natural theological argument to be sound, this is caused purely by theistic ‘acculturation’. In essence then, the internalist may challenge the ‘bleakest view’, but even were he or she to accept it, accepting it would not undermine his or her claims to religious knowledge, just any higher-order claims to knowledge of the knowledge status of his or her religious knowledge.
If Mill's argument as elaborated thus far is cogent, those who are irreligious, i.e. believe that no religion is true, can draw no comfort from the fact. The same reasoning applies, mutatis mutandis, to them. If Mill is right, then had the irreligious been sent by their parents to be brought up in a seminary in London or a Buddhist or Confucian temple in Peking, they too would by now have been Churchmen, Buddhists, or Confucians. If they are externalists, the irreligious cannot take Calvin's approach to solving this difficulty, so are actually in a worse positionFootnote 10. If they are internalists, then matters are similar for them as for the internalist religious believer. If we abandon for a moment our assumption that the Churchman is right and suppose instead that no religious beliefs are true, the fact that then makes right someone who believes this (a person who we may call irreligious) again does not seem to be a part of the explanation of why it is that he or she holds the right view. If there are objectively better reasons to believe that the irreligious view is correct than to believe any alternative, we have – on the bleakest view about how our assessment of reasons is culturally determined – no reason to suppose that we are in a privileged position to discover the fact; the actual causes of people believing this truth, if truth it be, seem to have nothing to do with its truth. Thus, if the sort of contingency Mill asserts really does undermine justification, then the justification for the irreligious view is more or equally undermined by the observation that the causes of people holding it lie in entirely mundane and contingent facts about how they've been brought up.
In its more modest version, Mill's argument can thus be seen to undercut the relevance of the philosophy of religion to religious believers (and indeed the irreligious): if the evidence suggests that people are caused to have (or lack) whatever religious beliefs they have (or lack) in ways that have nothing at all to do with the objective reasons for or against the religious beliefs to which they've been drawn (or from which they've been repelled), then the philosophy of religion is irrelevant to religious (and irreligious) believers. In its less modest version, Mill's argument undercuts the very possibility of the philosophy of religion: if the evidence suggests that people are caused to have whatever beliefs they do about the quality of the reasons for and against religious beliefs in ways that have nothing at all to do with the objective – good or bad – status of these reasons, then the philosophy of religion cannot even be pursued. The philosophy of the philosophy of religion may yet be secure. But a further iteration of Mill's argument may undercut that too. Hang on, one immediately thinks, if it does, won't that be for it to undercut itself? In fact, hasn't Mill's argument – at least if combined with what we have been calling the ‘bleakest view’ – already undercut itself? We shall close by focusing on this worry.
In conclusion
Isn't it the case, one might pointedly ask, that one's perception of whether or not Mill's argument provides one with good reason to abandon a knowledge claim in a particular area is just as (if not more) culturally determined as (or than) one's perceptions of the cogency of a particular argument in the philosophy of religion? If so, then this would undermine any claim one might initially have been inclined to make to know how one should be more chary of claiming to know having considered Mill's argument. However, the proponent of Mill's argument could counter this in two ways. First, he or she may take it on the chin. It is, after all, a permissible move for sceptics to demur from claiming that they know their scepticism to be well grounded.Footnote 11 Secondly, he or she may challenge the empirical claim. In fact, he or she might venture, the evidence suggests that people's assessment of the cogency of Mill's argument is not culturally determined; the bleakest view is not true of Mill's own argument. A proponent of Mill's argument may maintain that mathematics, natural sciences such as astronomy, and areas of philosophy sufficiently large to provide a stable base from which, with Mill's argument, one may question knowledge claims made in other areas of philosophy – including the philosophy of religion – are ones our epistemic access to which is independent of the sort of personal and social contingencies that afflict these other areas, areas in which we can consequently have less or no knowledge. Given that we are trying in the main to see what can be done to meet Mill's challenge without calling his empirical claims into doubt, let us therefore move on to consider a second way in which one might worry that Mill's argument has undercut itself.
Any consideration that applies to all views in an area equally cannot give one reason to accept or reject any in particular. So, one must be able to ‘divide through’ by any scepticism that Mill's argument would enjoin on us, in the same manner as can one divide through by the global scepticism that might be encouraged by a reading of the arguments in Descartes's Meditations. Having ‘divided through’, then everything must remain, relative to everything else, just as it was. This approach has, it seems to me, much going for it. But there are difficulties.
First, whilst very few can in fact be made to doubt that there is an external world that their senses inform them about more or less accurately by arguments such as those of Descartes, people can be made to doubt that religious reality is such as they supposed it to be by considerations such as those of Mill. I suggest that the historical Mill himself rather underrated the destabilizing effects that a Churchman's ruminating on the probability that, had he been sent as an infant to Peking, he would now be a confirmed Buddhist or Confucian, can be expected to have on him. Whilst it is perhaps true that the Churchman is unlikely from reflections on an argument such as Mill's alone to abandon his religious beliefs, Mill's argument could and indeed does sometimes play a part in leading people to abandon the religious beliefs with which they have grown up. (I believe it played such a part in the developing religious beliefs of John Hick, for example.) It can certainly play a part in leading to agnosticism in the technical sense whereby one is an agnostic just if one regards one's religious beliefs (which one may persist in with undiminished fervour) as not having whatever it is that must be added to truth to make them knowledge.
If the Churchman studies philosophy in general, he will either be drawn to an externalist account of knowledge, in which case (if Mill is right) he will not be able consistently to classify his religious beliefs as knowledge unless he embraces Calvinism, or an internalist account, in which case (if Mill is right) he may, if he studies suitable theology too, classify his religious beliefs as knowledge, but cannot (if Mill is right) know that he has correctly so classified them. Disbelieving in the existence of the external world is not perhaps something that is psychologically possible for many of us, but reclassifying our belief in it as not a proper item of knowledge is. Similarly, even if Mill is right that reflection on the peculiar contingencies with which people come to religious beliefs does not actually do much if anything to dissuade people from holding whatever religious beliefs they have grown up with and it is quite consistent with accepting the full force of Mill's argument to maintain that it should do nothing to dissuade us from holding them (applying equally as it does to the irreligious view, it leaves everything where it was), such reflection can lead to one's reclassifying one's religious beliefs or irreligiosity as not knowledge.
One point which emerges from considerations such as this is that it is possible for the religious believer to accept Mill's argument in its entirety and his or her religious beliefs persist completely unaltered thereafter even if at some level that which was previously unthinkingly classified as knowledge can no longer be so classified. For some, ‘I believe that p, but I believe that I have no more reason to believe that p than to believe that not p’ seems paradoxical in a sense that deprives one of the chance of ever truly saying it of oneself. However, I would contend that it is quite psychologically possible that this be true of one and one know it to be. Presumably those beliefs which one regards as properly basic beliefs have this feature. Indeed, even beliefs which are basic and, one may reasonably suppose, improperly so, can have this feature. (I might find myself at a horse race with a firm, but I recognize irrational, conviction that a particular horse will win.) It is then quite psychologically possible, indeed one may think probable, that the Churchman will stick with his ‘Churchy’ beliefs and the Confucian with his or her Confucian beliefs even after he or she has accepted Mill's argument. As observed, given that Mill's argument applies equally to the irreligious, it does not follow from accepting it that the Churchman (or indeed Buddhist or Confucian) should abandon his or her religious beliefs. If the beliefs of the Churchman are such as to predict some eschatological verification, then at some stage – e.g. the Last Judgement – he may find himself saying, ‘I believed that this would happen’, but, having accepted Mill's argument, unless he is a Calvinist, he should not say that he knew that this would happen or, if an internalist, perhaps knew that he knew that it would happen.
And if the religious believer relinquishes claims to knowledge, this does not disbar him or her from quite properly acting on the religious beliefs that he or she has this side of the grave, in the manner that relinquishing my claim to know (or even have any reason to believe) that a particular horse would win the race disbarred me (on various other assumptions) from betting on his doing so. In the case of religious beliefs, very plausibly one has to act on one or another (or on the irreligious hypothesis). Life is, if you will, a horse race on which one has to bet on something; what one might think of as failing to bet on any religion's being true is one's betting on the irreligious hypothesis being true and thus betting after all.Footnote 12 A parallel could be drawn with moral beliefs. The moral beliefs to which one is drawn very plausibly have the same contingent aetiology as one's religious beliefs. But one has to do something (even sitting still and ‘doing nothing’ is in this sense doing something) and when deciding what to do, one has start from wherever it is that one is.Footnote 13
This move in response to Mill's argument would represent then something of a ‘Humean’ sceptical solution to Mill's sceptical doubts; it would accept Mill's conclusion, but negate what one might have supposed to be its practical effects. No longer are certain items in the Churchman's noetic structure to be classified by him as knowledge, but – that act or reclassification having been accomplished – these items can – indeed must – still do all the work that they did when they were classified as knowledge. However, if our Churchman is a philosopher, then he can never think that abandoning a knowledge claim is an entirely happy solution, especially if he cannot even claim that he knows now that he does not know. This may make Calvinism appear more attractive to him than even its proponents can think he has internally-accessible good reason to find it. Or it may make natural theology as attractive to him as its proponents were already asserting he had internally-accessible good reason to find it.Footnote 14
