Schellenberg's propositional faith

Schellenberg uses the term ‘propositional faith’ for the kind of non-doxastic attitude he claims is available when belief is not.Footnote ² He defines such faith in the following terms:

Although this non-doxastic attitude has much in common with the everyday notion of faith, Schellenberg's definition should not be viewed as a conceptual analysis. Rather, the definition must be understood as ‘technical’, specifying the non-doxastic attitude upon which Schellenberg's work is focused and which he labels ‘faith’.Footnote ⁴

According to (1), faith and belief are incompatible. This marks Schellenbergian faith as a purely non-doxastic attitude and by definition makes it available only when belief is not. However, as (1) stands a clarification seems to be in order. Schellenberg is very clear that faith cannot be rationally held if disbelief is justified. For faith in p to be rational, p has to represent an epistemic possibility for the subject. Since our interest here is in rational faith, this should be included in (1):

1. (1)* S lacks evidence causally sufficient for S to believe that either p or ~p.Footnote ⁵

(2) simply points out that one cannot have faith unless one considers the state of affairs in question an overall good thing. You cannot have faith that a meteorite will hit the Earth and wipe out humanity or that your loved ones will die of cancer.

Condition (3) is perhaps the most central, since it describes what it amounts to for a subject to have faith. It is to represent the world to oneself as including the object of one's faith. Schellenberg gives many examples of this kind of representing or ‘imagining’, such as when a runner imagines himself winning a race despite his disbelief in that victory, or when a depressed woman in therapy tries to imagine the world as a brighter place than she actually believes it to be (Schellenberg (Reference Schellenberg2005), 129–131).

The voluntary assent condition (4) describes the behavioural component of tenaciously and persistently representing the world to oneself as including a particular state of affairs. Since it is the aspect of Schellenbergian faith that will be in focus in the discussion to follow, I will present it in some detail. As a first approximation, the following explication by Schellenberg is illuminating:

Schellenberg further describes voluntary assent as ‘something like a disposition, though one intentionally initiated and sustained’ (ibid., 135). Faith is not supposed to be something momentary, but rather a prolonged effort requiring concentration and mental effort. However, as with other cultivated dispositions, it might evolve into a kind of habit (ibid., 135–136).

Schellenberg's voluntary assent is incompatible with ‘active scepticism’ about the object of faith, as opposed to the ‘passive scepticism’ of neither believing nor disbelieving required by (1). The question whether the object of faith or some competing claim is true is closed; one cannot investigate matters further while having faith.Footnote ⁶ In order to have faith ‘one must set aside all questions about the possibility of truth in competing claims’ (Schellenberg (Reference Schellenberg2009), 5, 82–83; quotation p. 82). This recurrent theme in Schellenberg's treatment of faith gets perhaps its clearest expression in his explanation of the kind of scepticism (the active sort) the sceptic has to give up to have faith:

According to Schellenberg, voluntary assent of religious faith is, in some important aspects, stronger than everyday voluntary assent. Everyday faith is context specific, an attitude picked up in specific, temporally limited situations in order to cope with some specific situation – often some crisis or other. Religious faith, on the other hand, is much less context sensitive and held more indefinitely (ibid., 135–138). Schellenberg uses the term ‘commitment’ to spell out the difference:

The conditions of Schellenbergian faith are now in place, but something should also be said about faith's object before continuing. In Schellenberg's view, it is not traditional religion but rather ultimism that is the proper object of non-doxastic faith. U ltimism is a kind of generic ‘smallest common denominator’ religious view devoid of details, which only states that there is a transcendent reality that is metaphysically, axiologically, and soteriologically ultimate.Footnote ⁷

To keep the discussion as general as possible, I propose we separate questions regarding the rationality of having a non-doxastic attitude like propositional faith from the question of what the proper object for such an attitude is. What I say about propositional faith and other non-doxastic attitudes should be taken to apply equally to faith in the ultimate and faith in a specific religious tradition (or in any other object worthy of worship one might reasonably think of) if not explicitly stated otherwise.

The objection: faith and the end of investigation

According to Schellenberg, the main concern when evaluating whether faith is rationally permissible is the following negative condition:

Schellenberg further concludes that the only aim that presents any challenge in this regard is the aim of being true to reason:

It seems clear that what Schellenberg has in mind could fittingly be described as pursuing epistemic goals and considerations; therefore, we might think of the challenge to faith in epistemic terms. With these considerations in mind, we can remodel the negative condition into an epistemic restriction (ER):

• ER: Faith is rationally permissible only if no epistemic aim (independent of the aim apparently calling for faith) that should all things considered be pursued can best be pursued by not having faith.Footnote ⁹

That Schellenberg embraces ER seems clear since he goes to considerable lengths to show how faith and reason are compatible. It is a basic assumption of his position that epistemic considerations are to have priority, and that non-doxastic considerations and attitudes are only warranted if the epistemic ones are inconclusive.Footnote ¹⁰ On all this, I agree with Schellenberg. Now to the objection.

In the light of ER the voluntary assent condition (4) of Schellenbergian religious faith seems problematic. As a first approximation of the problem, Daniel Howard-Snyder's observation is illuminating:

Unfortunately, Howard-Snyder makes the objection in somewhat exaggerated terms, since what Schellenbergian assent requires is not a final judgement on p, but an indefinite postponement of further investigations into the truth of p. However, the latter seems almost equally problematic with regard to ER. If Schellenberg's voluntary assent requires an indefinite postponement of further investigation, then having faith will amount to treating as true that which one does not believe to be true with no intention to clarify the matter. Given that it is a generally accepted epistemic norm that for any p, we should aim to know whether or not p; there is an obvious apparent inconsistency between ER and the voluntary assent condition even on a more faithful reading of the latter.

The problem is specific to religious faith and does not threaten everyday ‘secular’ faith, since the former demands a long-term commitment ordinarily not present in the latter.Footnote ¹¹ In everyday cases, doubt is put on hold for a delimited situation, often to achieve some specific goal. To use one of Schellenberg's favourite examples, there is surely nothing wrong with the runner who puts his doubts about his own physical form on hold while running a race. However, the race is a delimited situation and when the runner has crossed the finish line, he will once again actively attend to the evidence of his physical form. No longer will he tell himself that he is in a good enough condition to win even though he does not believe this. In the religious case, on the other hand, the situation of epistemic uncertainty in which faith might seem appropriate never ends. Unlike the runner, the subject with religious faith never reaches the finish line; she never resumes her active investigation of the matters she is agnostic about (or at least she never plans to do so).

Religious faith becomes even more problematic when we consider its scope. Borrowing a characterization from Robert Audi, we can speak of religious faith as ‘an overall stance that governs important aspects of human life’ (Audi (Reference Audi2011), 58). Voluntarily assenting to a religious picture of the world involves closing off investigation on matters concerning one's overall stance on important aspects of human life without planning ever to take up those investigations again. Yet investigating matters concerning one's overall stance on important aspects of human life is surely an epistemic goal worth pursuing. Therefore, it seems ER prohibits one from having Schellenbergian religious faith. The argument is simple:

• Premise 1: Faith is rationally permissible only if no epistemic aim (independent of the aim apparently calling for faith) that should all things considered be pursued can best be pursued by not having faith (ER).

• Premise 2: The epistemic aim of investigating matters concerning one's overall stance on important aspects of human life should all things considered be pursued and can best be pursued by not having faith.

• Conclusion: Faith is not rationally permissible.

How would Schellenberg respond to this objection? Obviously, he would challenge premise 2, since he embraces premise 1 and rejects the conclusion. Schellenberg has presented at least two lines of reasoning relevant to objections like this one. The first amounts to claiming that some ‘readiness’ for new evidence is still possible for a subject with faith, and the second that a new kind of religious investigation is available for the faithful.

The readiness for new evidence response

According to the first line of response, the subject of faith is still susceptible to new evidence at some level, even though active investigation has come to an end. Schellenberg is clear on what is required of this ‘readiness’ for new evidence:

Schellenbergian faith demands that evidence is not allowed to disturb ongoing assent and imagining, but ‘readiness’ is required to satisfy reason's demand that investigation not end prematurely. How can ‘readiness’ both allow appreciation of evidence in a way that is epistemically responsible, yet also prevent evidence from impacting on faith?

Schellenberg has not made any rigorous attempt to demonstrate that such a concept of ‘readiness’ is possible, but two different claims from his writing on faith and reason might be relevant. The first is that one still keeps an eye open for evidence even when having faith, while the second is that a subject having faith will still notice alternatives that forcibly present themselves.

Schellenberg repeatedly affirms that even though one is not actively continuing the investigation when having faith, one can still ‘keep an eye open’ for new and more successful arguments (ibid., 83). However, the notion of ‘keeping an eye open’ seems hard to reconcile with many of Schellenberg's claims about voluntary assent. Remember that Schellenberg holds that active scepticism must end, and only a certain form of passive scepticism may remain. Also recall how all questions about the possibility of truth in competing claims are to be left behind when one starts having faith. The notion of ‘keeping an eye open’ simply seems to imply too much inquisitiveness and investigative activity to be consistent with voluntary assent. In the expression ‘keeping an eye open’ a sort of voluntary half-heartedness is implied: a subject keeping an eye open intentionally refrains from the kind of total immersion that voluntary assent is supposed to bring about.

It might be objected that I have forgotten to take into account the dispositional nature of Schellenbergian faith. Having a disposition to imagine the world in a certain way does not imply that one is always doing so. Perhaps the subject can have another, complementary disposition of ‘keeping an eye open’ for new evidence, active at times when the faith-disposition is dormant?

When evaluating this objection, recall the way Schellenberg spells out the difference between everyday faith and religious faith (repeated here for ease of reference):

Schellenberg thinks that religious faith is more of an indefinitely held attitude than everyday faith. This is only natural when it comes to religious faith, since religion has a tendency to cover everything, or at least ‘one's overall stance on important aspects of human life’. Dispositions are commonly thought of as being activated by certain stimuli. In the religious case, intuition suggests that pretty much anything can act as such a stimulus. If so, and taking into account that the subject will typically seek to manifest their faith-disposition as much as possible in order to live their ‘long-term commitment to view the world a certain way’, one might reasonably wonder if there is any room left for any complementary disposition.

Furthermore, one might question the reasonableness in cultivating two contrary dispositions. Conflict is bound to happen, since presumably these dispositions will often activate at the same time. For example, if some new minor evidence is unearthed that is disturbing from the point of view of faith, should the subject have faith and ignore the new evidence, or should she ‘keep an eye open’ and take the new evidence into account? To avoid the epistemic challenge I have raised, priority would have to be given to the disposition to ‘keep an eye open’, but that would make it hard to maintain a ‘long-term commitment to view the world a certain way’. The conflict between voluntary assent and ‘keeping an eye open’ may be weakened somewhat on a dispositional reading, but it is far from dissolved.

The second notion of ‘readiness’ suggested by Schellenberg is the idea that even though one has faith, one will still notice alternatives that ‘forcibly present themselves’ (Schellenberg (Reference Schellenberg2009), 86). Faith does not make one blind to strong evidence against one's faith. But that an alternative might forcefully present itself is no good substitute for continuing one's investigation. When investigating, a whole range of much more subtle evidence becomes accessible than the strong kind that might force itself on a subject.

When it comes to religion, it is even unclear what such forceful evidence would consist of. If an ignored religious alternative is true, a direct divine revelation could presumably do the trick. Surely, even if you have Christian faith, a divine revelation from Vishnu might ‘forcibly present itself’ and give you a strong reason to doubt your faith. But if a non-religious alternative like naturalism is true, it is hard to see what spectacular naturalistic revelation could ‘forcibly present itself’ to the faithful.

I conclude that the two notions of ‘readiness’ suggested by Schellenberg are not capable of reconciling the concept of propositional faith with the demand from reason that investigation not come to a premature end. The notion of ‘keeping an eye open’, while probably acceptable from the standpoint of reason, is too strong to be compatible with voluntary assent, even on a dispositional reading. The notion of giving attention to evidence that ‘forcibly presents itself’ on the other hand seems consistent with voluntary assent, but it is too weak to meet the epistemic requirements.

To handle the present objection, Schellenberg would need to develop a concept of ‘readiness’ that is consistent with voluntary assent and which still allows some real investigation to continue. Since investigation presupposes active scepticism and voluntary assent requires putting one's active scepticism to rest, it is hard to see how that could be done.

An investigation within the bounds of faith

According to Schellenberg, taking up faith is not the end of investigation but a transfer from one investigation to another. To adopt propositional faith in the way imagined by Schellenberg, one is supposed to have already made a thorough investigation of religious claims and ended up an agnostic. In contrast, after adopting a faith, a new investigation will commence within the boundaries set by faith (ibid., 82–83).

This line of response presupposes that one's object of faith is ultimism rather than traditional religion. Schellenberg argues at length that ultimistic faith is highly suitable as a ground for further investigation into religious matters and that it will make us open to future discoveries in the field of religion. He envisages ultimism being used as a kind of research-paradigm in religious investigation, and the lack of details in the view is supposed to be temporary; they are blanks to be filled in by future generations of religious investigators (ibid., 37–40).

Schellenberg claims that ultimistic faith is indeed demanded by reason, as the optimal way of realizing and integrating many of the most important aims of human life. He identifies no fewer than ten such important aims, including several of epistemic importance.Footnote ¹² Since it is beyond the scope of this article to address them all, I will concentrate on what I take to be the two most important lines of thought presented in that discussion: the idea of ‘epistemically faith-contingent’ truths and the claim that faith comes with investigative advantages.

‘Epistemically faith-contingent truths’ are truths that require faith to be known. Schellenberg suggests that propositions about the ultimate might be faith-contingent. If so, faith in fact puts the subject in a better epistemic position with regard to the ultimate, enabling her to know truths unattainable in an uncommitted, general investigation. The idea originates from William James's discussion of the two epistemic duties of believing truths and avoiding falsehoods. Like James, Schellenberg uses this idea to advocate a truth-seeking stance. If this reasoning is sound, faith might even be a way out of epistemic uncertainty (ibid., 220–226).Footnote ¹³

Intriguing as the idea is, the highly speculative nature of Schellenberg's case for ultimistic propositions being faith-contingent makes me doubt its significance for rational considerations of faith. The claim is that, if ultimism is true, then truths about it might be faith-contingent. Nothing stronger than mere epistemic possibility seems presupposed. Since the existence of the ultimate itself is also thought an epistemic possibility, it is natural to think of the possibility of faith-contingent truths about it as a ‘second order’ epistemic possibility. To appreciate the level of speculation involved, consider an analogy. It certainly seems an epistemic possibility worth considering that there might be life on other planets. It is also an epistemic possibility that such life might hold the key to discovering the cure for cancer – who knows what secrets such lifeforms might harbour? Taking the first epistemic possibility into rational consideration seems reasonable enough; after all, astrobiologists tend to be successful in acquiring funding. However, no pharmaceutical company would seriously consider funding a space programme on the basis of this second-order possibility. Speculations about second-order epistemic possibilities are simply too removed from actuality to have much impact on rational considerations. Therefore, the second-order possibility of faith-contingent truths should not be given much weight by an agnostic evaluating whether or not to adopt Schellenbergian faith.

The possibility of finding faith-contingent truths is far from the only epistemic advantage Schellenberg claims for ultimistic faith. When discussing the ‘Understanding Aim’, Schellenberg presents at least four investigative benefits promoted by ultimistic faith. It provides motivation by providing a view of the cosmos that is both compelling and in need of further investigation. It brings balance to investigation, since the last centuries have been dominated by investigation taking naturalism as an unquestionable given. It brings unification in the sense that a religious framework can provide a unified view of the world. Finally, it brings moral support, since a religious commitment makes one more resilient (ibid., 124–129).

It should be noted that these considerations do not require that what one takes on faith turns out to be true. Even if it turns out to be false, faith could still serve these investigation-strengthening functions (granted, of course, that one does not learn of its falsity). This is a significant advantage, since faith requires only epistemic possibility and standardly is available even in cases where the probability is very low or unknown. By showing that faith comes with advantages that do not require truth, Schellenberg has indeed made a strong case for ultimistic faith.

However, it remains unclear if these advantages are enough to warrant the ‘leap of faith’ from a general investigation to one within the bounds of ultimistic faith. There are at least two major ways in which this ‘leap’ seems problematic. The first is the fact that a faith-bound investigation precludes naturalism, since voluntary assent requires one to ‘set aside all questions about the possibility of truth in competing claims’, and surely naturalism is a competing claim to ultimism. Even though this could be interpreted in terms of bringing balance, it seems a heavy price to pay to conduct an investigation that, from the start, precludes perhaps the most widely shared view on the subject – at least in the scientific and philosophical communities.

The second way in which the ‘leap’ seems problematic is the low probability one rationally assigns to ultimism prior to making the leap. This could be contrasted with the view of Lara Buchak – who, like Schellenberg, holds that faith requires closing one's investigation. According to Buchak, a faith commitment is generally made only after one has made a thorough investigation and, through that investigation, determined a reasonably high posterior probability for the proposition one has faith in (Buchak (Reference Buchak2017), 120–123). Schellenberg requires nothing of the sort. On his view, one's object of faith only has to be epistemically possible in the wide sense of being neither justifiedly believed nor disbelieved. Nowhere does he claim anything more for ultimism. Could reason really demand that we adopt a faith that is only possible in such a wide sense, even if it means ending our investigation regarding its truth?

Presumably, Schellenberg would meet my objections with the Sceptical Dominance Argument (Schellenberg (Reference Schellenberg2009), 173–182), which is his version of Pascal's wager, and by pointing to all the other aims I have not taken up for discussion, to show me how much I have to gain by adopting ultimistic faith. As already stated, it is beyond the scope of this article to address all the aims and arguments Schellenberg advances on this issue. For someone like Schellenberg or James – who emphasize finding, and living by, important truths – they may ultimately be persuasive. However, those of us who are at least as committed to avoiding falsehoods as we are to finding truths will presumably not feel equally enthusiastic about giving up active scepticism indefinitely to bet on something that is only a mere epistemic possibility. Therefore, one might wonder if there is any way to modify Schellenbergian faith to fit our more restricted understanding of the demands of reason.

Occasional faith and hope

In recent literature, the idea that faith can be non-doxastic seems to be steadily ascendant. Several philosophers have advanced accounts of faith that are explicitly compatible with doubt and active scepticism, thereby picking out attitudes available in situations of epistemic uncertainty.Footnote ¹⁴ However, their individual merits aside, none of these authors has followed Schellenberg in giving an imagination-based account of faith. Since I find the idea that faith consists in picturing the actuality of an epistemic possibility to oneself both fascinating and compelling, I will present a modified version of Schellenbergian faith that avoids the objections I have been raising against the original proposal.

The problem with Schellenberg's faith is its incompatibility with active investigation in combination with being a long-term commitment going on indefinitely. Since, following Schellenberg, I cannot see how voluntary assent to an imagined reality could be compatible with active scepticism, it seems the only way to redeem an imagination-based account of faith is to insert some kind of temporal limit, making faith occasional. This might seem strange, but since there is nothing wrong with having faith in limited situations in the secular case, it could simply be seen as a remodelling of propositional religious faith in the image of everyday faith.

In order to introduce a temporal limit, the problematic voluntary assent condition (4) of Schellenberg's definition needs to be reworked. We cannot remove it altogether, because without voluntary assent the imaginative representing, as required by condition (3), would not amount to much more than daydreaming. If simply imagining some possible state of affairs as actual were sufficient for faith, reading a novel would constitute an act of faith, so voluntary assent – the commitment to think and act as if the imagined possibility is true – is crucial.

Since (4) cannot be removed, we need to weaken it. The mentioning of commitment unique to religious faith has to go, and some requirement that faith might only be held in limited situations must be added. Replacing ‘committedly’ with ‘for the occasion’, we get the following new version of (4).

1. (4) _Occ: S voluntarily and for the occasion adopts a policy of assent towards that representation, or, more broadly, towards p.

The new (4)_Occ captures the kind of voluntary assent included in everyday acts of faith, but precludes problematic faith that goes on indefinitely. It is faith taking a legitimate break from investigation, not faith that abjures further investigation altogether.

Should we follow Schellenberg's original proposal for everyday faith, and think of occasional faith in terms of a disposition? Occasional faith is certainly context-dependent in the same way, since it is supposed to constitute a faith-reply one can make in certain limited situations. However, a disposition is naturally thought of as a tendency to respond in a certain way to certain circumstances, and we have already seen that such a faith-disposition might all too easily come into conflict with our investigative dispositions. We should, therefore, avoid thinking of occasional faith in dispositional terms. Instead, each situation where faith could be a suitable response should be viewed as a separate faith-opportunity, and the choice whether to have faith or not is to be made separately for each such occasion. In picking up Schellenbergian faith, one voluntarily commits to cultivating a disposition, while for occasional faith the voluntary decision is made separately on each occasion.

Even if occasional faith escapes the problems associated with ordinary Schellenbergian faith, one might ask if it is sufficiently substantive in a religious context, for what would that kind of religious faith really amount to? What limited situations are we talking about here? The most straightforward reading would probably include having occasional faith when performing religious rituals, visiting a place of worship, praying, and so forth. It is a faith adopted for the moment, when acting religiously, and perhaps a bit more long-term and sustained in times of personal crisis.

A kind of faith one only ‘wears in church’, as it were, might seem objectionable. But I do not think we should be too quick in rejecting it. I think faith-attitudes of this kind are pretty common, especially in communities with a high degree of secularization but where many important ceremonies in life are still held in a religious setting. For example, think of a religious funeral. When confronted with death in combination with some religious story of an afterlife, I think many unbelievers might feel some pull towards faith; an experience of ‘the will to imagine’ to use Schellenberg's term. And what is wrong, at least for the duration of the funeral and perhaps also when thinking of the dead afterwards, with giving in to that desire and ‘tenaciously and persistently represent[ing] the world to [oneself] as including that state of affairs’; that is, as including an afterlife for our dead beloved? In such a case, one obviously does not believe in the afterlife and would not commit oneself indefinitely to imagining its existence; rather one simply sets doubts aside from time to time and imagines the epistemic possibility of such an afterlife as actualized.

I think the main reason why this kind of occasional faith will strike many as odd is the old identification of faith with belief. Believing in this way would be highly objectionable. Surely there would be something deeply wrong with a person who believes in God when in church and when saying his prayers, but not otherwise. But that is not the issue here. Neither am I suggesting that a person who actively disbelieves is warranted in indulging in moments of faith. We are still only considering how one might respond in a situation of epistemic uncertainty, and so it is only the subject who neither believes nor disbelieves that is permitted to indulge in moments of occasional faith.

Another objection has to do with specifying the occasions for occasional faith. Unlike Schellenbergian religious faith, occasional faith is not supposed to be held indefinitely or even most of the time, but only in delimited situations. However, since religious faith tends to cover almost every area of human life, how can we delimit faith to singular occasions? Will not all situations be situations in which a faith response might seem legitimate? This is roughly the same objection I raised previously to explain why thinking of Schellenbergian faith as a disposition does not solve its epistemic problems. Occasional faith is able to meet this objection precisely because it is not dispositional. Since the subject is free to decide on each occasion if she is to have faith or not, it is up to the subject to manage the occasional faith in a responsible way that prevents any major interference with her epistemic endeavours. Exactly how much occasional faith a subject can have before the ‘legitimate break’ from investigation becomes problematic will presumably vary a great deal between individuals.

The obvious disadvantage of occasional faith is that it cannot be sustained for very long, leaving the occasions of faith as isolated events in the subject's life with no real connection between them. What we need is some further attitude that keeps these isolated events together and puts them into a context. The obvious candidate here is hope.

Hope is standardly taken to include among its necessary conditions some sort of desire and viewing the state of affairs hoped for as an epistemic possibility and nothing more. Hope precludes certainty. Presumably, a complete analysis would contain some further conditions. However, these two basic necessary conditions constitute a kind of minimum view that most philosophers engaging in the debate surrounding hope seem to share and no further analysis seems needed for our present purposes.Footnote ¹⁵

If hope is to be the complementary attitude to occasional faith, the ‘glue’ that keeps the isolated moments of faith together, we want to be sure that hope does not come with any epistemic disadvantage. Since we are looking for an attitude one can hold indefinitely, it is especially important that it is able to avoid the problems associated with Schellenbergian faith. And quite clearly, it does. When hoping that p, there is nothing preventing one from continuing to investigate the truth of p. Not only does hope seem compatible with having doubts, it even seems compatible with some degree of disbelief, as when we hope that we are wrong.

Are there any other epistemic problems associated with hope? Since hope only requires epistemic possibility and desire, and since the notion of epistemic possibility unproblematically informs us that the attitude of hope is non-doxastic and therefore available in a situation of epistemic uncertainty, only desire is left to consider. Desires can be epistemically problematic if they are allowed to bias our reasoning. For example, one might be tempted to assess evidence for things one desires too positively. But that is only a risk; desire does not inevitably lead to biased reasoning in the way Schellenberg's voluntary assent always comes with an end to active questioning. Risk alone does not make an attitude or an action unwarranted. For example, there is nothing wrong with believing even though there is always a risk of believing falsehoods, and there is nothing wrong with gathering new evidence even though there is always a risk of finding misleading evidence. Therefore, the risk of biased reasoning associated with desire does not seem enough to make hope an unwarranted attitude.Footnote ¹⁶

Another concern about the desire-component of hope is that it makes hope involuntary in an important sense. Even if the subject has some voluntary control over which desires are preferred and developed into hopes, the desires themselves remain entirely involuntary. Without a desire, one cannot have hope. Hoping for p is only open to those who desire p.

Fortunately, there are no restrictions to the effect that a response to a situation of epistemic uncertainty needs to be open to everyone in order to be rationally permissible. The fact that hope is partly involuntary does not mean that hope cannot be a legitimate response; it only means that it cannot be a required response. Hope might not be for everyone, but for those with the right desire, hope seems a perfectly legitimate and permissible response in a situation of epistemic uncertainty.Footnote ¹⁷

Since occasional faith can only be rationally held in delimited situations, and since most subjects feeling the pull towards occasional faith will presumably also feel some desire towards the actualization of that which is imagined, the combination of occasional faith with a prolonged hope will probably seem a reasonable and attractive position to many pro-religious agnostics. And as far as I can see, such a combination of attitudes is a perfectly rational response in a situation of epistemic uncertainty.

One final objection is worth considering. If, as I have claimed, hope is epistemically unproblematic and always available, why bother to add any kind of faith to the picture? Why not keep things simple by only taking on attitudes of hope? The answer to this question is that hope does have a fairly major disadvantage: it is a very weak ground for action. Acting on a mere hope would, in many circumstances, be utterly foolish. If you live in a besieged city in the middle of a war, you will almost certainly have a profound hope that there will be peace. But that does not mean that you dance in the streets and live as if the war has ended as soon as there is the slightest epistemic possibility of that being the case. Or, to paraphrase an example put forth by Daniel J. McKaughan, having a hope of winning the lottery does not constitute a good reason for spending as if the million-dollar jackpot was already yours (McKaughan (Reference Mckaughan2013), 113).

Even when acting on hope is not outright irrational or foolish, it could easily be prevented by some overriding reasons. Therefore, hope might not always be enough for performing religious actions or partaking in religious life. Occasional faith allows for action in a more robust way and makes possible the move from having a non-doxastic religious pro-attitude to participating in religious life. In other words, the attitudes of occasional faith and hope are complementary, which constitutes a good reason for including their pairing among our non-doxastic pro-attitudes.