According to non-doxasticism, religious faith is possible in the absence of belief. Instead of belief, non-doxastic faith is based on epistemically weaker attitudes such as acceptance (Alston, 1996), assumption (Howard-Snyder, 2017a) or hope (Pojman, 1986). It is generally acknowledged that a religious life can be based on non-doxastic faith rather than doxastic (i.e. belief-based) faith.

There is a strong tendency in the literature to stress the similarity between non- doxastic and doxastic faith. This tendency becomes even more pronounced when it comes to Christian faith. It has been argued that non-doxastic faith is so much like doxastic that the two can be covered by one single conceptual analysis (Howard-Snyder & McKaughan, 2022; Howard-Snyder, 2013; 2017a), and that non-doxasticism is compatible with accounts of faith found in the Bible (Howard-Snyder, 2017a) or in the writings of Mother Teresa (Howard-Snyder & McKaughan, 2022; McKaughan, 2017). William Alston has argued that a non-doxastic Christian can live a fully involved religious life, in a way which does not include make-believe or pretence:

The person who [non-doxastically] accepts the doctrines is not necessarily inferior to the believer in commitment to the Christian life, or in the seriousness, faithfulness, or intensity with which she pursues it. The accepter may pray just as faithfully, worship God just as regularly, strive as earnestly to follow the way of life enjoined on us by Christ […] The non-doxastic accepter can be as fully involved in the [Christian] forms of life, and not just on an ‘as if’ basis […] It is only the unthinking assumption that belief exhausts the possibilities for a positive attitude towards the articles of faith that gives rise to the judgement that the accepter is engaging in an elaborate make-believe or pretense. (Alston, 1996, pp. 17–18, italics in original)

Alston’s remark that non-doxastic faith is not a matter of make-believe is presumably meant to demarcate the position from fictionalism. Fictionalism treats religious language as fictive and religious life as a game of make-believe (Eshleman, 2005; Le Poidevin, 2019). Since non-doxasticism builds on a realist conception of religious language, it should not be denied that the intellectual side of non-doxasticism is much more in line with traditional, belief-based religion than it is with fictionalism. However, it is wrong to assume that also the religious life and inside perspective of the non-doxasticist is in line with doxastic religion.¹ I assert what Alston explicitly denies, namely that a committed non-doxastic religious life requires extensive use of imagination and pretence. Prominent features of a religious life require belief to be properly enjoyed, and the non-doxasticist must use her imagination to keep up with the believer.² I discuss two such features: supramundane religious experiences (with a special focus on experiencing holiness) and the love of God.

I also reject the idea that non-doxasticism must be kept separate from fictionalism. To live a satisfying and rationally consistent religious life, the non-doxasticist needs to act and talk as if certain propositions regarding herself are true (propositions like “I experience holiness” or “I love God”), even though she believes or even knows that these propositions are false. In short, she needs to adopt local fictionalism.

Before beginning, two preliminary remarks. First, to keep in line with mainstream accounts from the current literature, I focus on religious life in a theistic tradition centred on the individual’s relationship with God (understood according to perfect-being theism). While I personally think non-doxasticism has much to offer in relation to other religious views as well, this focus will make my point versus Alston and other philosophers who stress the “ordinariness” of non-doxastic faith as forceful as possible.

Secondly, it is common to distinguish between relational (or operational) faith and propositional faith (Alston, 1996, pp. 12–14; Schellenberg, 2005, pp. 125–126). I am interested in the kind of faith which functions as the cognitive basis for a religious life, so by “faith” I mean propositional faith if not explicitly stating otherwise.

In the following section “Religious Non-Doxasticism,” I introduce non-doxastic faith in more detail, and in “Fictionalism, Imagination and Pretence” I proceed to clarify the involved notions of fictionalism, imagination and pretence. It is followed by two sections where I argue for the importance of imagination in a non-doxastic religious life: “Supramundane Religious Experience” and “The Love of God.” In the last major section “The Need for Local Fictionalism” I present my argument for supplementing non-doxasticism with local fictionalism.

The core idea of non-doxasticism is that faith can be based on cognitive attitudes which are epistemically weaker than belief,³ such as acceptance (Alston, 1996), assumption (Howard-Snyder, 2017a), hope (Pojman, 1986), or voluntary assent (Schellenberg, 2005). The exact details vary between accounts. Some philosophers hold that the cognitive attitude needs to be very belief-like, in that it should involve verbal affirmation (Alston, 1996, p. 4) and/or inwards assent (Schellenberg, 2005, pp. 134– 137). This requirement has been denied, not only by critics of non-doxasticism (Scott, 2020, pp. 96–98) but also by some prominent defenders (Howard-Snyder, 2017b, pp. 151–157). Since the points I wish to make are general and not tied to any specific account of non-doxasticism, I set matters of affirmation and assent aside and only assume what I take all proponents to agree upon, namely that the cognitive attitude of faith must allow for rational action and rational thought.

I am not sure whether there is a natural language term for this commondenominator attitude (what do you call it when someone adopts p as ground for thought and action without being inclined to affirm p?). It does, however, come quite close to Howard-Snyder’s “belief-less assumption.” In lack of a better option, I am therefore going to use “assumption” and say that when someone uses p as ground for thought and action, she assumes that p.⁴

Non-doxastic faith is commonly analysed as a composite attitude, and the cognitive attitude is but one of several conditions in such a conceptual analysis. There are least two further conditions which almost all accounts include or entail. Since a complete analysis also includes the non-doxastic cognitive attitude itself, and sometimes further conditions as well, these two conditions must be understood as necessary but not sufficient. We can therefore say that necessarily, if S has non-doxastic faith that p… Conative Pro-Attitude (CA):

S has a conative pro-attitude towards p, such as desiring its truth.

CA requires faith to be directed at a positive possibility (I can under normal circumstances have faith that I will survive my cancer, but not that I will die in horrible agony). Non-doxastic faith is often taken to be voluntary, and CA provides the required motivation. The conative attitude reflects the subject’s preferences “all things considered.” On occasion we might have “dirty desires” which do not reflect our considered preferences. If I intentionally quit smoking, I might desire to have a cigarette even though I evaluate having one negatively. In such circumstances, I might have faith that I will succeed in never smoking another cigarette, but I might not have faith that I will have a cigarette.

Proponents of non-doxasticism apply different notions of epistemic possibility, so the understanding of EP varies greatly. Some understand epistemic possibility in the widest possible sense, as p being “neither known nor justifiedly believed to be false” (Schellenberg, 2009, p. 8). Others opt for a very strict epistemic requirement (Alston, 1996, p. 14; Howard-Snyder, 2017b, p. 154). They hold that non-doxastic faith that p requires the subject to take p to be at least as likely as any other option, so that you can only have faith in the most likely alternative. There are also those who take the middle ground and require that p is a real possibility with a “non-negligible chance of being true” (McKaughan, 2013, p. 13), or a Jamesian “live option” (Palmqvist, 2023b, p. 653). I, once again, only assume what all proponents seem to agree upon, namely that the subject is in an important sense agnostic regarding p, lacking both belief and disbelief.

This formulation of non-doxasticism builds on a notion of categorical or outright belief and disbelief, where having a belief that p is qualitatively different from regarding p as probable.⁶ Having outright belief means that “your mind is made up that p” (Lee, 2023, p. 1094) and that you “simply take p for granted, treating it as a starting point for further reasoning” (Wedgewood, 2012, p. 312). It involves to “feel it to be the case that p” (Alston, 1996, p. 5) and to include “that p” in one’s general picture of the world. For convenience, outright disbelief that p can be treated as outright belief that -p.

Religious fictionalism and non-doxasticism are both belief-less approaches to religion. The crucial difference is that the former does not require any sense of epistemic possibility. Fictionalism is fully coherent with both atheism and naturalism. While the non-doxasticist treats her religious worldview as an epistemic possibility, the fictionalist treats it as fiction, and where the non-doxasticist relates to the religious worldview by assuming its truth as ground for thought and action, the fictionalist relates by means of imagination and pretence, so that her religious life becomes an elaborate game of make-believe.⁷

Imagination and pretence are contested topics in contemporary philosophy, and defending a full-blown account lies outside the scope of this paper. I will only highlight some important distinctions and relevant features of the kind of imagination and pretence involved in a belief-less religious life. The distinction between imagistic and attitude imagining proposed by Langland-Hassan (2020) serves as a suitable point of departure. Imagistic imagining is about visualising, about conjuring up images for the mind’s eye to see. This contrasts with attitude imagining, which is about entertaining an idea:

[Attitude] imagining is a kind of thought process that allows us to step outside of what we really believe to consider mere possibilities […] The emphasis is on the capacity of imaginings to enable rich, elaborated thought about the possible, fictive, pretended, and fantastical, without any attached stipulating that the thoughts are image or picture-like. (Langland-Hassan, 2020, p. 5, italics added)

Imagistic imagining is central in all kinds of religion, since religion concerns “what no eye has seen,” a reality of immense significance beyond the limits of our senses. Imagistic imagining is even more important in belief-less religion since a subject without belief will have to image-imagine things the believer experiences (or so I argue in the next section).

While not particularly important for religious believers, attitude imagining is central in fictionalism, and I will demonstrate that it is also crucial for non-doxastic faith. Attitude imagining is about considering “mere possibilities,” by which I take it that Langland-Hassan means logical possibilities, which includes both fictive⁸ and epistemic possibilities.

Many philosophers hold that imagination is a “belief-like” state, which means that some of the characteristics of belief hold for imagination while others do not. Just like with belief, it is impossible to imagine a contradiction with a singular act of imagining (Shinbahabu, 2016, p. 112). Unlike belief, imagination is voluntary in the sense that it can be instigated at will. There is some controversy over whether imagination is limited in the same way as belief when it comes to inferences. Is the imaginer to infer and imagine what a believer in the same position would infer and believe? (Stock, 2017, p. 177). While I am unsure regarding this view in general, it clearly applies to the kind of imagination relevant for present purposes. Both the fictionalist and the non- doxasticist aim at imagining the world as it would be if a religious worldview were true. They, therefore, want to imagine all that the believer believes, including what the believer believes from inference.

While imagination refers to a mental state or process, pretence concerns the subject’s actions in the world. Archetypical examples involve games of make-belief, acting and deceiving. In order not to overcomplicate things, I use the basic definition offered by Elizabeth Picciuto and Peter Carruters (2017, p. 317): Pretence (def):

To pretend that p is to act as if p (without believing it) while imagining that p.

This definition should not be taken to imply that you need to imagine that p in the strong sense of immersing yourself in a fiction where p is true. While immersion most often makes pretence more satisfying (especially in the religious case), pretence only requires that you use your imagination to figure out what to do. If you pretend to be a lion when playing with your kids, you need to use your imagination to work out how a (kids-friendly) lion might act.

Regarding the qualification “without believing it,” it is important to see that pretence only requires a lack of belief that p, it does not require disbelieving that p, or the falsity of p. One can pretend what is true as long as one does not believe it to be true, and there is certainly nothing wrong with pretending that epistemic possibilities are actual. If unconvinced, consider the following example. A soldier is reported missing in action. The soldier’s family accepts that they cannot justifiably believe that he is either dead or alive. To keep their spirits up, the family starts to play a game they call “Daddy is coming home,” in which they pretend that the soldier is alive and currently on his way home. This is a perfectly sensible instance of pretence (and it would only cease to be sensible if the family learnt that the soldier is actually alive and on his way home).

Given the previous assumption that all thoughts about the merely possible are instances of imagination, it certainly seems that non-doxastic faith involves imagination and pretence already at the conceptual level. In non-doxastic faith, an epistemic possibility p is used as a ground for thought and action. When it comes to using p as ground for thought, it is natural to interpret this as an instance of attitudinal imagination. If so, we should also interpret acting on p as a matter of pretence, since the subject will be acting on what she imagines.

That non-doxastic faith seems to involve imagination and pretence is an important observation at odds with the picture painted by Alston. In the following sections I am going to argue that the non-doxasticist needs to make use of imagination and pretence in a way which goes beyond the attitude imagination and pretence involved in non- doxastic faith itself.

Arguably, there are at least two distinct kinds of religious experiences. The first kind are experiences of something completely otherworldly (in theistic religion, most often God) and the second are experiences of natural things as having religious meaning and/or religious properties (like an experience of a hike as a pilgrimage, a church as the house of God, or of nature as God’s handiwork).⁹ Let us call the first class supernatural and the second supramundane. Supernatural experiences are rare theophanies most religious people never have, while supramundane experiences are a common part of religious life. In this section, I argue that supramundane experiences are unavailable to the non-doxasticist, and that the best she can do is to imagine.

I take the following considerations to apply to all kinds of supramundane experiences. However, to keep the discussion manageable, I focus on a specific kind, namely the experience of holiness.¹⁰ In such an experience, an object is perceived of as being part of a supramundane order of reality while at the same time being part of ordinary reality. It is an enrichened experience, in which an additional, sacred dimension is glanced through the mundane. In his famous account, Mircea Eliade expresses this as follows:

By manifesting the sacred, any object becomes something else, yet it continues to remain itself, for it continues to participate in its surrounding cosmic milieu. A sacred stone remains a stone […] but for those to whom a stone reveals itself as sacred, its immediate reality is transmuted into a supernatural reality. (Eliade, 1957, p. 12)

Since our concern lies with non-doxastic faith, we should be careful to interpret holiness in a way which leaves open the question whether it exists mind- independently (otherwise, an experience of holiness would be a reason to believe, which would complicate the picture in more than one way). A straight-forward way of satisfying this requirement is to understand experiences of holiness in terms of theory-laden observation.¹¹

The view that observations are theory-laden, common in the philosophy of science, denies the traditional picture according to which we observe things “as they are” before adding a layer of theoretical interpretation. Instead, it suggests that we observe things in accordance with our pre-understanding, so that experience is permeated by theory and background beliefs.¹² On this picture, the faithful experience holiness (like in a sacred place or holy artefact) because they already believe that there are holy things, while the non-believer does not have such experiences due to a lack of proper background beliefs.

It seems reasonable to suggest that theory-laden observations require belief. Non- doxasticism requires epistemic possibilities which are too unlikely for belief to be possible. Our observations of the world do not conform to what we hold to be possible in this sense, which is fortunate, because otherwise we would frequently hallucinate. Unsure whether your wife is home when you enter your house? If epistemic possibilities gave rise to theory-laden observations, you would start experiencing her even if she is not home.

An objector might agree with this but claim that adopting a non-doxastic attitude like assumption will make a difference. However, it seems doubtful that assumption can have that kind of effect. If you are unsure whether a boat on the horizon is green or not, you do not experience the boat as green, and non-doxastically assuming that it is green does not affect your experience. Furthermore, non-doxastic attitudes are voluntary, so if they did give rise to theory-laden observation, it would be possible to change one’s visual experience by will. Let us say that I assume the boat on the horizon to be green, and then, for whatever reason, I assume it to be yellow. Clearly, my visual experience will not change as a consequence. In all, we have strong reason to reject any view suggesting that a non-doxastic commitment would enable us to experience the world as conforming to unbelieved epistemic possibilities.

Perhaps the objector presses her point and claims that even if assumption cannot change how we experience things in the short run, it might still do so in the long run. If I assume that p intensely and for a long time, does it not seem plausible that I might eventually start “seeing things” as conforming to p? Assuming that the imagination is the only cognitive faculty by which we deal with the non-actual, my response is that when a subject starts “seeing things” because of her non-doxastic assumption, she is in fact already using her imagination – she imagines seeing things (because if she is not imagining and actually sees p, that cannot be due to her assumption). Imagining seeing things and having a theory-laden experience are quite different things.

A clarification might be in order. I do not deny that epistemic possibilities can influence what we perceive. For example, if you hold that it is epistemically possible that there are ghosts, and that a particular house might be haunted, this might well affect your experience when in the house. You will pay more attention to the creaking of floorboards, changes in temperature, and other things you associate with ghosts. What I am denying is that assuming epistemic possibilities can give rise to theoryladen observations of ghosts.

A subject with doxastic faith believes that there are holy things, which is a requirement for theory-laden experiences of holiness. While making a voluntary commitment and assuming that p “there are holy things” will allow the non-doxasticist to use p as a ground for action and thought, it will not allow her to experience the world as containing holy things. When we fail to have the right kind of experience, our best (and presumably only) option to come anywhere near having that experience is through imagistic imagination. It therefore seems to me that the only way the non- doxasticist can get close to the inner experiences of the doxasticist is by imagining things to be holy.

Consider the following example. Pete and Tom are visiting the (allegedly) holy city of Jerusalem. Pete has doxastic faith and believes that the basic Christian picture of the world is correct. His experiences are theory-laden with Christian beliefs and with what he can infer from these beliefs. Only when the world does not conform to the Christian picture will Pete be forced to engage in conscious interpretation. Tom has non-doxastic faith. Being an agnostic, he regards the Christian worldview as an epistemic possibility which he assumes to be true (i.e. treats it as ground for thought and action). Tom will not experience the world as Pete does. What happens automatically for the believer will take conscious effort for the non-doxastic assumer. He must bring the Christian worldview to mind with an act of will (which might be harder or easier depending on his level of religious understanding and the vividness of his imagination), and actively interpret the world as conforming to it.

Arguably, Pete and Tom will experience Jerusalem differently. Pete will have a theory-laden experience of the city as holy. Tom will only experience a cramped and crowded Middle Eastern city, and he must make a conscious effort to start acting and thinking about it as holy. However, he will not experience it as holy. Tom must imagine what he does not experience, i.e. that Jerusalem is a holy city, permeated by divine reality. Of course, I am not arguing that Tom is rationally or morally obliged to imagine. What I suggest is that such imagining is the only feasible way for him to come near the inner side of the religious life of a believer.

Being in a loving relationship with God is central to theistic religion, and especially so in its contemporary Christian forms. The relationship itself can only be real if God exists, which is what the believer believes and the non-doxasticist assumes, but what about the non-supernatural part of the relationship, i.e. the loving attitude of the human subject? Can a non-doxasticist love God like a believer, or does she need to imagine?

A preliminary. There is much philosophical disagreement over the analysis of love, and the matter only gets worse if we also take theological views into consideration. Some say love is fundamentally a wish for union, others that it is a kind of exclusive attention, still others that it is a kind of robust caring, to mention some important views (Helm, 2021; Brümmer, 1993; Frankfurt, 2006). In order to sidestep most controversies, I will only presuppose that love involves a positive attitude directed at a certain object, which affects thought and behaviour.

However, there are two issues where I must take a stand. Some philosophers deny that love is an emotion. It seems obvious to me that love must have an emotional component. Love seems incompatible with never having positive emotions towards the object of love since without such emotions, love reduces to benevolence. I suggest we include a positive emotion (or at least a disposition to form such emotion on occasion) as a necessary condition for love. In my argument, I will also assume that emotions are (at least partly) responses to an object having certain qualities. While this account of love has been denied by philosophers like Frankfurt (2006), the idea that God is worthy of our love in virtue of possessing divine qualities like perfect goodness is firmly embedded in theological tradition, so it should not be overly controversial in a religious context.

It seems clear that you cannot love someone if they do not exist.¹³ When a Harry Potter fan loves Harry Potter, what she loves is a fiction and not a person. This is easy to see, because in Harry Potter’s case there is no real person to love. Even if the Potter fan is deluded like Don Quixote and believes that Potter is real, this would not change the fact that she is in love with a fiction.

It might seem equally clear that you cannot love someone if you believe that they do not exist.¹⁴ An atheist fictionalist who claims to love God surely seems to be in the same situation as the Harry Potter-fan, unable to love anything more than a fictional God or at most the idea of God. However, while her disbelief prevents her from consciously loving any actual God there might be, there is always the issue of mistaken identity to consider. You might love Clark Kent and be unaware that you actually love Superman. Likewise, if God exists and is the highest good, an agnostic or even an atheist might love the highest good and thereby love God without realising it. In this text, however, I am interested in a love of God which can occupy a central stage in a religious life. Clearly, you cannot give the love of God the central role in your religious life if you fail to realise that it is the love of God. The possibility of loving without believing due to mistaken identity can therefore be safely set aside.

A non-doxasticist importantly lacks disbelief as well as belief, and she assumes what she does not believe. The question we must ask is therefore twofold: can you love someone over whose existence you are agnostic, and if not, does adopting a non- doxastic attitude like assumption enable you to do so?

The linguistic data seems to suggest that we refrain from talking about love if we are agnostic concerning the existence of what we would love. Couples trying for children generally go through periods where they are uncertain whether their attempts have been successful. They do not say “We do not know if we are pregnant, but we love the baby” – no one claims to love their unborn children before the pregnancy is confirmed. Likewise, when people start dating, they will generally go through a phase where they feel that they need to get to know each other better before saying “I love you.” Arguably, that is because they are yet agnostic concerning whether the person they date has the qualities they look for in a romantic relationship.¹⁵ The linguistic data, therefore, seems to suggest that in order to love, you must believe that X exists and has the qualities which makes X a suitable object of your love.

That belief that X exists is a prerequisite for love makes perfect sense if we accept that love has an important emotional component, and that emotions are (at least partly) “evaluative-cum-motivational responses to an object” (Helm, 2021). If love is partly a response to an object with the right qualia, it is hard to see how this response could appear in someone who is uncertain regarding the existence of the object in question.

Now to the second question. If an uncommitted agnostic cannot love God, can an agnostic love God by adopting non-doxastic faith? Such faith, we must remember, means that she takes the theistic worldview to be true in thought and action. Part of what she takes to be true is that God exists and that, as a perfect being and ultimate foundation of our existence, God is worthy of love. But assuming that God is worthy of love does not bring about the same kind of emotional response as believing. A desirable epistemic possibility might invoke hope and even a longing for its truth. By assuming such a possibility to be true you might fall in love with its central ideas (in the theistic case the idea of God), and you would certainly be in a position to start loving God if you came to believe, but I cannot see how assuming the theistic worldview to be true would make loving God possible for the non-doxasticist.

Consider an analogy. Kevin has gone on two dates with Lea, and he hopes that she possesses the loveable qualities he looks for in a woman. Even though he is still uncertain, he adopts non-doxastic faith that she has these qualities. This is certainly a good stance, since by acting as if Lea has the required qualities, Kevin makes it more likely that he will discover these qualities if they are present. However, if Kevin starts to love Lea based on these non-doxastic assumptions, before getting to know her, we should say that his love is premature and that he is fooling himself. You cannot properly love someone based on how you hope they are, not even if you make this hoped for picture an article of faith. While Kevin’s picture of Lea might well turn out to be true, we should not say that he really loves her (rather than a picture in his head) before he is justified in believing her to be who he hopes she is.¹⁶

A religious life based on non-doxastic faith would hardly be satisfying if it were to exclude the love of God. The non-doxasticist therefore needs to imagine that she loves God, and since she cannot very well imagine loving God without imagining that God exists, she needs to imagine that as well. Since love importantly has a behavioural side, she also needs to pretend that she loves God. While perhaps not being rationally required in a strict sense, imagination and pretence in this case seems crucial to maximize the similarity between a doxastic and non-doxastic religious life.

I have argued that a subject of non-doxastic faith needs to employ her imagination in order to come as close as possible to a belief-based religious life. In this section, I am going one step further. I will argue that a rational, non-deluded agnostic seeking a religious life akin to that of a believer must cross the border between non-doxasticism and fictionalism. She needs to treat as true some propositions which she actively disbelieves. Since non-doxasticism requires live possibility, she needs to adopt a fictionalist stance regarding these propositions.

The propositions I am referring to concerns the fictionalist own inner states and experiences. Examples include the ones at the centre of discussion in previous sections, like “I love God” or “I have an experience of holiness.” These propositions do not express epistemic possibilities to the agnostic, and yet she needs to base thought and action upon them. Therefore, while she can be a non-doxasticist regarding the religious worldview at large, she needs to be a fictionalist concerning herself in relation to that worldview.

Consider once again the issue of loving God. By taking up non-doxastic faith, the subject resolves to assume the truth of theism as a basis for thought and action. Part of what she is to assume is that a perfect being worthy of her love exists. She will realise that if she believed it existed, she would love it. That means that if what she assumes were to be true, her actions and thoughts would be motivated by that love. However, it is not even a live possibility that she loves God. Although she would love God if she could believe what she now only takes to be a live possibility (i.e. that God exists), she does not currently love God. She knows that the proposition “I love God” is currently false. Therefore, she cannot non-doxastically assume that “I love God.”¹⁷ Her only option is to turn to fictionalism and pretend to love God even though she knows she does not. She thereby adopts fictionalism, not as a general stance, but locally to be able to integrate the proposition “I love God” into her otherwise non-doxastic religiosity.

Consider again the case of experiencing holiness. When Tom the non-doxasticist assumes that Jerusalem is a holy city, he must also assume that he himself experiences Jerusalem as being permeated by God’s reality. But he knows that he does not experience Jerusalem in that way, and therefore he assumes something he knows to be false. Acting (i.e. pretending) as if something one believes to be false is true is incompatible with non-doxasticism and only possible on fictionalism. Therefore, Tom needs to be a fictionalist regarding his own experiences.

The case generalises to all instances where assuming a religious picture to be true in thoughts and action would lead the subject to the realisation that she would in fact be a different person in some respect if she were to believe what she only assumes. For example, if she is a non-doxastic Christian, she might realise that believing to be a forgiven sinner would make her a humbler person. But she does not believe that she is a forgiven sinner, and she knows for certain that she is not humbler than she in fact is. Since it is not an epistemic possibility that she is humbler than she in fact is, she cannot non-doxastically assume “I am humbler than I in fact am.” When she pretends to be humbler than she in fact is, she has crossed the border to fictionalism and is now involved in a local fictionalist game of make-believe.¹⁸

Here is the argument put a bit more formally: P1:

If S disbelieves that p, the only way she can act and think as if p is true is through fictionalism.

P1 primarily follows from the fact that non-doxasticism requires epistemic possibility and is incompatible with disbelief (as stated by EP in the necessary conditions for non- doxasticism). This means that if S disbelieves p, non-doxasticism is unavailable and fictionalism is the only option left if she wants to live a religious life.

It might of course be questioned whether fictionalism is truly the only way a subject can act and think as if something false is true, since, for example, we might be reluctant to think of kids’ games as fictionalism. However, it should be clear from context that we are only considering religiously relevant options. Also, I am not denying that there are non-realist alternatives to the qualified sense of fictionalism (championed by Le Poidevin (2019) and others) which I relate to in this text. I therefore propose a broad reading of P1 where fictionalism is understood as any non-realist stance which allows a disbeliever to act and think as if religious propositions were true, in a religiously meaningful way.

P2 is simply what I have argued for in this section. It is perhaps the least obvious premise but given present considerations I find it hard to escape. It is formulated in terms of theism, presumably it generalises to most religious worldviews – what religious view would make no difference whatsoever to what you think about yourself if you believe it to be true?

P3 is a substantial assumption, requiring a careful defence. First, however, a qualification. The argument is only supposed to target non-doxastic faith in which the subject assumes p to be true as a ground for thought and action and thereby strives to emulate the religious life of a believer. I do not wish to deny that there might be other possible kinds of non-doxastic religion which are not based on such faith. Neither do I wish to deny that such religion might be somehow satisfying. For example, hope is also a non-doxastic attitude on which a kind of religiosity can be based. If you hope that p it is generally unwise to think or act as if p is true,¹⁹ and so large parts of a traditional religious life will be out of bounds. However, a hopeful agnostic can still love the idea of God, religious stories could be inspiring to her, and some traditional spirituality would be available – it is an often-repeated fact that agnostic prayer is both possible and rationally acceptable (Draper, 2009, pp. 210–211).

Now, the question is whether a non-doxastic faith where the subject does not pretend any ISP to be true could be psychologically satisfying? I would say that pretending some ISPs to be true is not only required for psychological feasibility, but for properly using theism as a ground for thought and action without pretending to have the relevant inner states. Because can you really be said to act and think as if it is true that there is a perfect being worthy of your love while refusing to think and act as if the existence of this being has affected you? Arguably, part of what it is to act as if God exists is to act as if motivated by the responsive love a person is supposed to feel when she believes to be loved by her Creator. Consider an analogy. Bob and Steve are two live-action role-players, acting out an encounter between a knight, played by Bob, and a dangerous troll, played by Steve. As a role-player, Bob is to assume in thought and action that Steve is a dangerous troll. However, much like a non-doxasticist who refuses to adopt local fictionalism, Bob resolves to act and think as if Steve is a troll without acting and thinking as if he himself is a person who has encountered a troll. I cannot see how this could be done. How could Bob act in a way which is appropriate without pretending to be a person who has encountered a troll? To act appropriately, Bob needs to figure out whether to be afraid or challenged or threatened by the troll, and what course of action that reaction would lead to. Only then can he act on the assumption that Steve is a troll.

Furthermore, a non-doxasticist who does not pretend to have the inner states of a believer will have two different and inconsistent grounds for thought and action: the theistic worldview she non-doxastically assumes to be true, and her beliefs about her own thoughts and inner states. There will be inconsistency in thought. For example, the subject would have to think of God as worthy of all our love, but still not think of herself as loving God etc. There will also be many occasions where the different grounds will suggest different and conflicting actions. For example, should the non- doxasticist love her rude and ignorant neighbour? The theistic worldview, on which she is committed to act, implies that she should love others since she is herself loved by God. But she does not believe that she is loved by God, and she hates her intolerable neighbour. Presumably, there will be many occasions like this, where the assumed religious view and her true inner states suggest different courses of action, leaving the non-doxasticist unsure what to do. By adopting local fictionalism concerning her ISP, the non-doxasticist will be able to circumvent this problem. I therefore conclude that we have good reasons to accept P3, and that the argument seems to hold.

I have argued for two major points. First, that non-doxastic faith requires a heavy use of the imagination to make possible a satisfying religious life and maximize similarity with doxastic faith. Second, that non-doxastic faith must be complemented with local fictionalism, since the non-doxasticist needs to treat as true some central propositions regarding herself which she disbelieves. I will end by pointing out some consequences of these conclusions, as well as some further issues they raise.

Acknowledging the need for imagination and local fictionalism should make us hesitant towards the prevalent tendency to stress the similarity between doxastic and non-doxastic faith. If non-doxastic faith differs in these important ways, we have good reasons to question the assumption that they express the same faith-attitude or that they can be captured by the same analysis. The current analysis therefore gives us reason to question accounts such as the one presented by Howard-Snyder (2013), where propositional faith can be based either on either belief or a non-doxastic attitude.

Instead, we should consider the view that non-doxastic faith and doxastic faith are not different species of the same basic attitude. On the assumption that faith ordinarily is belief-based, non-doxastic faith will appear to be a “technical attitude” or “philosophical invention,” an imagination-based ersatz alternative for those lacking belief. While this might suggest that non-doxasticism is closer to fictionalism than commonly acknowledged, it does not imply that the former reduces to the latter. Non- doxasticism is still a robustly realist approach to religion. Acknowledging non- doxastic faith as a “technical attitude” means that we can improve it via conceptual engineering, to suit our theoretical needs. As considerations in the last section suggest, we might want to re-work our understanding of non-doxastic faith to include fictionalist elements.

I have recently argued that religious fictionalism requires role-playing (Palmqvist, 2023a). The current considerations suggest a similar conclusion regarding non- doxastic faith. However, the role-playing perspective on fictionalism comes with some problematic conclusions. Most importantly, it implies that fictionalists experiencing spiritual growth are subject to Quixotical delusion when they fail to separate the spiritual life of the role-character from themselves. Would the same problem affect non-doxastic role-players, or should we view them differently because they mostly relate to epistemic possibilities?

The current results also give rise to interesting questions regarding what the case would be if the religious view of the non-doxasticist who employs local fictionalism is in fact true. In such case, she would pretend to be the person she would have been had she believed the truth about the world. Would such a person be aligned with reality? Would a non-doxastic Christian who pretends to be in a loving relationship with God be in an actual loving relationship with God if God exists?

In all, a hybrid position which combines non-doxasticism with local fictionalism seems promising, but it is also an approach which raises further questions. Much work remains to be done before it can be fleshed out in detail.