PATCHWORK PUZZLES AND THE NATURE OF FICTION

Kathleen Stock has recently argued that Gregory Currie’s account of fiction is beset by two patchwork puzzles. According to the first, Currie’s account entails that works of fiction end up being implausible heterogenous complexes of utterances that furnish a fictional world and utterances that aim at representing the actual world. According to the second, competent engagement with a fiction can implausibly result in switching from one mental attitude to another – namely, belief and make-belief. In this paper, I argue for two main claims. First, that a few alterations to Currie’s account make it immune to Stock’s puzzles. And, second, that such a modified account presents clear advantages over the alternative one offered by Stock.


I. INTRODUCTION
Kathleen Stock recently argued that Gregory Currie's influential account of the nature of fiction was plagued by so-called 'patchwork puzzles' . 1 She then uses the existence of such puzzles to motivate the introduction of her own account of the nature of fiction. In this paper, I take as my starting point Stock's formulation of these puzzles and offer to recast Currie's account in a way that avoids their emergence. I then discuss Stock's own alternative and argue that my recasting of Currie's account fares better than her proposal.
The plan is the following. In the first section (II), I introduce Currie's account of fiction as well the patchwork puzzles which Stock claims plague it. In the second section (III), I amend Currie's initial account and dissolve Stock's patchwork puzzles. In the third, and last, section (IV), I engage critically with Stock and argue that my version of Currie's account should be preferred over her own account of the nature of fiction.

II. PATCHWORK PUZZLES
Two notions fundamental to the philosophy of fiction are a fictional work, such as Dostoyevsky's Brothers Karamazov, and the constituents (for example, sentences, propositions, or utterances) of such a work. Plausibly, a relation of conceptual fundamentality of some sort obtains between these two notions. This Patchwork Puzzles and the Nature of Fiction For their very helpful comments on earlier drafts of this paper, I thank Tobias Klauk, Niels Klenner, Julia Langkau, Kathleen Stock, Cain Todd, Alberto Voltolini, Manuel García-Carpintero, the audience of the EXRE 2018 autumn school in Champex, and an anonymous referee. Work on this paper was supported by the Swiss National Science Foundation, Grant No. 100012L_166619 / 1. Currie drives home this point by means of a series of thought experiments, the most telling being the following. An author, Smith, intends to produce a work of fiction. As a result, she produces what she takes to be genuine and faithful fictive utterances, that is, utterances whose contents she prescribes her audience to make-believe. But, unbeknownst to Smith, the content of her fictive utterances are true propositions about her life which she has psychologically repressed.
In other words, she mandates her audience to make-believe contents that truly report events of her life. This fact, Currie says, prevents Smith's work from being a piece of fiction because fiction does not amount merely to that which is the product of a fictive utterance but to that which is the product of a fictive utterance and for which make-believe can be correctly prescribed.
The content of Smith's work fails to meet the second condition. Insofar as the content of Smith's fictive utterances truly report events of Smith's life, makebelieve falls short of being the correct attitude one should take towards them.
What exactly should be added to a fictive utterance for its content to be a genuine candidate for fiction? For Currie, truth does not constitute the real matter and, hence, the reason for which Smith's work fails to be a fiction is not merely that its content is true. Rather, the problem is the kind of relation that Smith's story entertains with truth -namely, it is not only true but also nonaccidentally so.
The notion of non-accidental truth supplies the success conditions for fictive utterances. According to Currie, a hypothetical author could write a piece of fiction and, out of pure luck, its content could turn out to be entirely true. Such a case would still count as fiction, Currie claims, since the facts of which the fiction is true play no role in the author having produced a work of fiction with exactly that content. It is very different, says Currie, if the truth of content that pretends to be fictional is non-accidental. Indeed, in such a case, the world itself reliably supplies the work with its content and, as such, the content of the work cannot be correctly prescribed for make-believe.
From the notion of a fictive utterance and this second clause of non-accidental truth, Currie is able to define a second notion, fictional statement. 4 A fictional statement is a fictive utterance whose content is at most accidentally true.
Currie regards such contents as the genuine units of fiction and, incidentally, it is the fact that Smith's work lacks such fictional statements that explains its failure to be fictional.
We now have a pair of notions, fictive utterance and fictional statement.
Together, however, they are not yet sufficient to shed light on the distinction between fictional and non-fictional works. A legitimate demand on Currie's theory is that it should manage to go beyond that micro-level to deliver us an account of fictional and non-fictional works. According to Currie, there exists a path to the macro-level since a fictional work can be defined as that which contains a sufficient number of fictional statements. By contrast, a non-fictional work is one that contains none, or not enough. As he puts it: We can say that a work as a whole is fiction if it contains statements that satisfy the conditions of fictionality I have presented, conditions we can sum up briefly by saying that a statement is fiction if and only if it is the product of an act of fiction-making […] and is no more than accidentally true. 5 Many have agreed that an account along these lines represents a big step torwards the aim of providing a workable theory of fiction. However, it must face a problem which can be expressed as the incompatibility of two theses.
The first I shall call the Homogeneity Thesis (HOT): Homogeneity Thesis: works of fiction are homogeneous in the sense that the constituents of the fictional content of a work of fiction are all of a same kind.
This formulation needs unpacking in order to avoid misunderstandings. First, note that HOT is about fictional content, that is, what is true in a fiction. In other words, it merely claims that if something is a constituent of the fictional content C of a work W, then that thing must share a same nature, so to speak, with the other constituents of C -and, presumably, one that goes beyond the somewhat trivial, unenlightening sharing of the property of being a constituent of C.
Second, HOT being restricted to fictional content makes it compatible with two important facts about works of fiction. The first fact is that, as many have claimed and as we shall see below, some fictional truths can also be simultaneously genuinely asserted. This fact doesn't threaten HOT since its proponent can maintain that, on the whole, the constituents of the fictional content C of work W are on a par qua constituents of C while agreeing that some of these constituents can also be genuinely asserted by W. What HOT is committed to is the claim that the fictional content C of a work W cannot turn out to be the conjunction of two disjoint sets of utterances: those that are true in W and those that are genuinely asserted by W.
The second fact is that a work of fiction as a whole can be composed of more things than fictional truths, for instance authorial intrusions such as the following 5 Ibid. passage in Charlotte Brontë's Jane Eyre: 'A new chapter in a novel is something like a new scene in a play; and when I draw up the curtain this time, reader, you must fancy you see a room […]. ' Again, this neither threatens HOT insofar as it is restricted to the fictional content of a work of fiction, not its entire content.
The second thesis I shall call the Heterogeneity Thesis (HET): Heterogeneity Thesis: Fictions may be heterogeneous entities in the sense that their fictional content can be composed of two different kinds of utterances, fictive and non-fictive ones.
As should be obvious, HOT and HET are incompatible. Indeed, contrary to what HOT is committed to, HET claims that the fictional content of a fiction can be a patchwork of two different kinds of utterances, fictive and non-fictive ones and, hence, it is possibly non-homogeneous. 6 An account such as Currie's is wedded to HET. Indeed, as a result of introducing the notion of fictive utterance in the way he does, Currie opens up the possibility that most works of fiction are patchworks of fictive and non-fictive utterances.
As he puts it: 'A work of fiction is a patchwork of truth and falsity, reliability and unreliability, fiction-making and assertion. ' 7 Some scholars have argued, however, that this consequence is problematic and threatens the plausibility not only of Currie's account but of any bottom-up theory of fiction wedded to HET. Stock, for instance, writes: Yet it does not follow from the adoption of this [bottom-up] strategy that, other things being equal, we should happily accept a theory which entails that fictions standardly contain significant numbers of non-fictive utterances. I will assume that, where two otherwise equally plausible theories of fictive utterances have different consequences for the numbers of non-fictive utterances typically allowed as present in a fictional work, the one which posits fewer it better than the one which posits more of them. 8 Stock provides two reasons for her concern. First, she insists that we need to give a homogeneous treatment of the activity of writing fiction and that HET forces us to conceive it as a split between two distinct tasks: writing fictive utterances and writing non-fictive ones. As she puts it, not trying to diminish the amount of fictive utterances a work contains 'is to accept an uncomfortable split in an author's otherwise apparently homogeneous activity: now she produces one kind Patchwork Puzzles and the Nature of Fiction 6 I am grateful to an anonymous referee for pressing me to clarify HOT and HET and for supplying the Brontë example. 7 Currie, Nature of Fiction, 49. 8 Stock, 'Fictive Utterance and Imagining' , 145. of utterance, now she produces the other, yet both serve to furnish a fictional world, somehow. ' 9 In other words, Currie's account forces us to see the production of utterances which aims at depicting the real world and the production of fictive utterances as two activities distinct in kind which somehow achieve the same goal, the furnishing of a fictional world. This, however, creates an 'uncomfortable split in an author's otherwise apparently homogeneous activity' . The idea amounts to the following. When an author is crafting a fictional world, she does not take her activity to be composed of two distinct sub-kinds of activity: first, crafting fictive utterances and, second, crafting non-fictive ones. On the contrary, her activity amounts to the seamless crafting of a work of fiction by multiplying fictive utterances. Yet, though the question of how fictive utterances can furnish fictional worlds is rather straightforward, Currie leaves us in the dark as how nonfictive utterances can perform the same task. How can the production of both kinds of utterance collaborate to furnish a fictional world? Following Stock, I shall refer to this as the first patchwork puzzle of fiction.
Her second reason arises out of what we may call an attitudinal ascent from utterances to the attitudes they prescribe. Indeed, as we have seen, a central pillar of Currie's bottom-up strategy is that fictive and non-fictive utterances prescribe different attitudes -one of make-believe and one of belief, respectively. Hence, if fictional works are indeed patchworks of fictive and non-fictive utterances, then correct engagement with them may result in a patchwork of attitudes: some bits of the work we end up make-believing while other bits we end up believing.
Stock remarks, however, that this constitutes an implausible description of our engagement with works of fiction. Though we might end up believing propositions as a result of engaging with works of fiction, a plausible account of our engaging with them should account for the fact that we seem to be taking a uniform attitude towards any proposition true in the world of a work of fiction, whether really true or not. As she puts it: At best [Currie's account] leaves a problematic gap in our understanding of the practice of reading fiction, since, one assumes, a difference in kind of utterance demands a difference in response; yet if this is so, not only are we unclear what those prescribed responses are, exactly, but we are unsure how they are supposed to interact. It would be to a theory's advantage if this slightly schizophrenic picture of fictional works could be avoided. 10 Following Stock again, I shall refer to this as the second patchwork puzzle of fiction. 9 Ibid., 146. As we shall see below, Stock not only identifies these puzzles, but she also sets to dissolve them. But, in the next section, I shall provide an alternative way to get rid of these puzzles, closer to the spirit of Currie's original account. At a later stage, I shall explain why my recasting of Currie's account should be preferred over Stock's new proposal.

III. DISSOLVING THE PUZZLES
The two puzzles discussed in the previous section demonstrate that there is a tension between the different bits of Currie's account. In this section, I dissolve the puzzles while preserving the spirit, if not the letter, of Currie's account.
I proceed as follows. First, I identify a set of reasons that I claim explain why Currie framed his theory in this way. Second, I locate precisely the source of the puzzles in Currie's account and propose a way to get rid of them. Third, I evaluate my account with respect to the reasons identified at step one.
Three main reasons may be given to explain the shape of Currie's theory. At the most general level is the claim, argued by elimination, that a pragmatic account constitutes the best way to formulate a theory of fiction. Of course, this only explains the most general features of the account and not Currie's reason for his endorsement of HET, which, as we have seen, is at the source of the existence of the patchwork puzzles. We need two further elements to provide a my explanation.
First, he endorses the allegedly bedrock claim that a fiction must contain at least some utterances whose contents are made up. We deem it bedrock because it is hard to imagine how one could argue in favour of the claim that this is genuinely a part of the concept of fiction. The best one can do is to point to a certain practice of classifying works along these lines and reveal one's intention to carve out a notion of fiction which takes this fact into account. As we have seen, Currie's own version of this intuition is that a fiction must contain at least some fictional statements, that is, fictive utterances whose contents are at most accidentally true.
Second, even though the primary function of works of fiction is to be vehicles for fictional truths, they also sometimes communicate truths about the real world.
To use an often-quoted example, the opening of Tolstoy's Anna Karenina -'All happy families are alike; each unhappy family is unhappy in its own way'admittedly makes a claim whose range is supposed to cover not only the fictional families depicted in that work, but families in the real world as well. Currie's own way to deal with this fact is to treat such an utterance as a non-fictive one.
In sum, then, Currie's account can reasonably be seen to offer a theory of fiction that meets three desiderata head on: it is a pragmatic account that provides Patchwork Puzzles and the Nature of Fiction a satisfactory explanation both for the presence of made-up contents and for the claim that works of fiction can also be vehicles for genuine truths. But the price to pay for such an accommodation is quite high since it cannot avoid endorsing HET. I will propose a way Currie's account could meet these desiderata for a fairer price. Before turning to this, I first locate more precisely the source of the tension in Currie's account. In order to do so, I strengthen a bit the case made by Stock by reformulating her worries in the form of an argument. Its conclusion is no doubt stronger than Stock's initial worries but putting things in this form will turn out to be helpful as a means to structure the subsequent discussion.
The strengthened argument from the first puzzle against Currie's account goes as follows: (1) The activity of the author of a work of fiction is to compose a work of fiction; (2) to compose a work of fiction is to furnish a fictional world; (3) a fictional world can be furnished both by fictive and non-fictive utterances; (4) a fictive utterance aims at furnishing a fictional world; (5) a non-fictive utterance aims at depicting the real world; (6) an utterance that aims at depicting the real world cannot simultaneously furnish a fictional world; (7) hence, whenever the author of a fiction is producing a non-fictive utterance, it cannot be taken as furnishing a fictional world.
This argument can be taken as a reductio ad absurdum of Currie's account.
Indeed, authors routinely use what counts for Currie as non-fictive utterances to furnish fictional worlds. Currie's view, however, makes this fact either impossible (as the above argument has it) or at least quite hard to grasp (as Stock's own formulation has it). In either case, we cannot rest content with Currie's account as it stands.
I shall assume that this argument is sound. Hence, Currie's account cannot remain as it stands. But which part exactly should go? Premises (1), (4), and (5) don't seem to be plausible candidates for rejection, since they come very close to being stipulations. I shall henceforth leave them aside. Premise (2) is equally non-substantial, since the notion of fictional world it appeals to is very minimal and merely elliptically refers to what is true in a fiction. This leaves us with options that we can appeal to only two kinds of utterances, fictive and non-fictive ones, this fact neatly explains the need for the claim that a non-fictive utterance be allowed, in its competence as a non-fictive utterance, also to furnish a fictional work. Indeed, this claim certainly is more plausible than the one that a fictive utterance can aim at depicting the real world. Hence, premise (3) may look bizarre but, for the time being, rejecting it would be even worse, since it would be tantamount to failing to meet the second of Currie's three desiderata.
What about premise (6)? Generally speaking it doesn't seem impossible for an utterer to use a single utterance to perform more than one communicative act.
Take the following case. A firm is hiring and examining candidates. About one of them, an examiner says: 'He is a hard worker.' It is a quite mundane claim that making such an utterance can realize more than one communicative act. First, it can convey the examiner's belief that the candidate is a hard worker and, second, her conviction that there isn't much more to expect from this candidate than hard work. Why wouldn't it be the case that the same utterance can both furnish a fictional world and aim at conveying a truth about the real world? As Stock puts it, It is […] possible to have different, non-competing intentions fulfilled by the same utterance. An utterance in code can be intended to convey two distinct true claims, for instance. Likewise, in principle there seems no impediment to a single utterance's being seriously intended both to say something true and to be recognized as such, and seriously intended to be imagined and recognized as such, simultaneously, assuming that the right sort of context can be generated such that the relevant intentions are communicable by the form and context of the utterance. 11 In substance, premise (6) claims that the nature of the utterances that compose a work of fiction are somewhat special in that they cannot convey more than one communicative act: they are either fictive or non-fictive but cannot be both. The rejection of premise (6) allows us to paint a picture very different from Currie's -namely, works of fiction are not patchworks of fictive and non-fictive utterances. Rather, works of fiction are composed entirely of fictive utterances and, in addition, at least some such utterances convey a subsidiary, non-fictive, communicative act. Moreover, if this is the case, then it becomes immediately obvious that premise (3) is superfluous. If a same utterance can be both fictive and non-fictive, then there is no need to endorse the weird claim that a non-fictive utterance can furnish a fictional work. Hence, we can solve the first patchwork puzzle by getting rid of both (3) and (6).
Interestingly, Currie comes very close to claim just this, revealing some tension in his original account. Indeed, he writes: The reader of fiction is invited by the author to engage in a game of make-believe, the structure of the game being in part dictated by the text of the author's work. What is said in the text, together with certain background assumptions, generates a set of fictional truths: those things that are true in the fiction. Anything that is true in the fiction is available for the reader to make-believe. 13 Assuming that 'what is said in the text' covers both fictive and non-fictive utterances, then Currie comes very close to affirming that we can in fact makebelieve the content of utterances whose content we are prescribed to believe.
As a result, there is a tension between, on the one side, what Currie says about the nature of fictive and non-fictive utterances and, on the other, his claim that everything that is true in a work of fiction is available for the reader to makebelieve. It is reasonable to conclude one of two things from this: either Currie's account is inconsistent or, more charitably, he indeed believes that the content of non-fictive utterances can be make-believed. But if the latter is the case, then he leaves us completely in the dark as to how non-fictive utterances can prescribe us to do this. This possibility has now been made explicit.
An important consequence of our rejection of (3) and (6) is that, contrary to what Currie claims, a fictive utterance can successfully prescribe its content for make-believe and yet its content can be non-accidentally true. Accordingly, even though a fiction cannot be a patchwork of exclusively fictive and non-fictive utterances, it can be composed of two kinds of fictive utterances: those with a content which meets the 'accidentality of truth' condition and those with a content that fails to meet it. Of course, this does not entail that a work of fiction can be composed entirely of fictive utterances whose contents fail to meet the accidentality of truth condition, as would be the case in Smith's story. This 13 Currie, Nature of Fiction, 70; emphasis added.

Patrik Engisch
Estetika: The Central European Journal of Aesthetics, LVI/XII, 2019, No. 1, 0-00 seems good enough when it comes to accommodating the intuition that at least some of the content of a fiction must be made up, as endorsed by Currie. It can still be a necessary condition upon a work being fictional that it must contain at least some contents which, if true, are at most accidentally so. Taking the claim that works of fiction are patchworks of utterances and replacing it with this duality of fictive utterances does not conflict with this intuition; indeed, it preserves it.
Nevertheless, it must be said that important differences exist between these two kinds of fictive utterance. First, there is the question of what composes a fiction. My claim here is that works of fiction are entirely composed of fictive utterances which come in two kinds, differentiated by the nature of their content.
That is, the relation of composing a work of fiction that obtains between a set of utterances and a fictional work will obtain between the set of all the fictive utterances and the work they compose. This matter must be distinguished from a second one -namely, the question of why something turns out to be a fiction.
Here, several options are open.
A first possibility would be to rely again on the notion of fictive utterance. This, however, turns out to be unattractive insofar as it would not permit those who, like Currie, share the intuition that a story like Smith's should not count as fiction to defend their intuition. A better option is to preserve some asymmetry amongst the constituents of a work of fiction. That is to say that, although all fictive utterances compose a work of fiction, only a sub-set of them grounds its fictional nature -namely, those fictive utterances which, if true, are at most accidentally so. In other words, grounding a work of fiction is a relation that obtains only between a sub-set of fictive utterances and a work. In line with the intuition behind Currie's Smith example, a work full of fictive utterances with non-fictional content would not yet be a fiction.
Moreover, there is some difference with respect to the normative relation that obtains between fictive utterances and their content. Generally speaking, an utterance of type t alone rarely constitutes a conclusive reason for its audience to entertain the prescribed attitude towards its content. For instance, the fact that p is asserted by speaker S does not alone constitute a conclusive reason for one to believe that p. Some minimal condition must be added, for example, that one be disposed to judge that S constitutes a reliable epistemic source. In the parallel case of fictive utterances, Currie's Smith example demonstrates that authorial intention doesn't constitute a conclusive reason for a reader to make-believe some content. Smith produces a work with the intention that her readership should make-believe its content but, as Currie points out, this is not sufficient for her readership to possess conclusive reasons to make-believe the content.

Patchwork Puzzles and the Nature of Fiction
In addition, the content submitted by an author for make-believe must be of the right kind.
This point is weakened, however, by Currie's introduction of the category of 'pseudo-fictions' . 14 These are works that have not been produced using fictive utterances yet we may treat them as if they were fictional. But not any kind of work can be thus treated. According to Currie, a first condition is that their content must be false so that there is no point in believing them. However, we do not possess a reason to make-believe the content of a non-fictional work merely because its content is false. In addition, Currie says, we must possess a good reason to do so, for example, pleasure or entertainment. His paradigmatic example is Defoe's Robinson Crusoe. The work was initially presented by its author as non-fiction and its content prescribed for belief. But the falsity of the content was later revealed, and the readers thereafter lacked any reason to believe its content. Yet, as Currie points out, treating the work as if it were fiction and makebelieving its content makes for a pleasurable experience. Hence, we have a reason to do so.
The point I want to make with this notion of pseudo-fiction is that Currie allows for two different kinds of reason to make-believe the content of an utterance. On the one hand, there are fictive utterances such that we have an intrinsic mandate to believe their content. These are utterances which come with a prescription that we make-believe their content and which are at most accidentally true.
On the other hand, there are fictive utterances such that we have an extrinsic mandate to make-believe their content, that is, utterances for which there may be an operative prescription that we make-believe their content though this prescription doesn't stem from the fact that these utterances meet the conditions Currie sets for fictive utterances to be fictional statements.
I now suggest pushing this distinction further and distinguishing between two kinds of fictive utterance whose content we may have an extrinsic mandate to make-believe. First, there are utterances in the context of pseudo-fictions.
Granted that we have a prescription to make-believe their content, it has nothing to do with intrinsic properties of the work. It is because our psychological nature is such that treating Crusoe as fiction makes for a pleasurable experience that we have a reason to make-believe its content. Were our psychological nature altered such that we would no longer be able to derive pleasure from treating Crusoe as fiction, any reason to make-believe its content would simply disappear.
14 Ibid, 37. Second, there are, in the right context, the contents of fictive utterances which fail to meet Currie's condition for being fictional statements. These utterances are a mixed bag. Like fictional statements, they too are fictive utterances and, hence, come with a prescription that we make-believe their content. But, as we saw in the discussion of Currie's Smith's case, this prescription does not constitute a conclusive reason to make-believe their content. The content of nonaccidentally true utterances should be believed, not make-believed. But, and here lies the truly important claim, when such utterances are constitutive of a fiction we have reason to make-believe their content. This reason counts as extrinsic, not intrinsic: it is because fictive utterances furnish a fictional world in collaboration with fictional statements that we have a conclusive reason to make-believe their content, not, as with fictional statements, merely in virtue of properties of the utterances themselves.
With these elements in place, I may now formulate my full answer to the first patchwork puzzle: there is no puzzle at the level of utterances because all the utterances that compose a work of fiction are fictive. But fictive utterances are of two kinds, those that meet Currie's criteria for being fictional statements, and those which do not meet these criteria but for which they might, in the right context, be a prescription that we make-believe their content. All fictive utterances prescribe make-believe, but only the first kind do so intrinsically.
Finally, all fictive utterances compose a work of fiction, but only fictional statements ground its fictional nature.
Let us now turn to the second puzzle. If a work of fiction is entirely composed of fictive utterances, then there is no reason to believe that to engage with a work of fiction qua work of fiction results in a puzzle at the level of attitudes. Our answer to the first puzzle entails the dissolution of the second one.
Yet, this is obviously moving a bit too quickly. The initial puzzle was that when engaging with a fiction a competent reader would make-believe some part and believe some other, resulting in a 'schizophrenic' picture of engaging with works of fiction. We now have a rather different situation: since the whole work of fiction is composed of fictive utterances, the competent reader would make-believe the whole work. But if works of fiction are also to be vehicles for genuine truths, then we should also be able to believe some of its parts, in addition to makebelieving them. And here is the rub. Can we really simultaneously believe and make-believe the same proposition? This will turn out to be crucial not only for our solving of the second puzzle but also for our solving of the first one. Indeed, we rejected premise (6), claiming that a single utterance could be both fictive and non-fictive. At the level of attitudes, this means that the very same proposition can simultaneously be both make-believed and believed. But, as Stock remarks, Patchwork Puzzles and the Nature of Fiction 'if it can't be done, it can't be seriously prescribed' . 15 I shall now, at last, account for this possibility.
Belief is standardly treated as the attitude we take towards things we consider to be true and, as such, it possesses a specific link to action. It is because we believe that there is beer in the fridge that, if we desire to have a beer, we form the intention to reach for the fridge. Make-believe, on the other hand, is often treated as a mental state whose role is to substitute for belief when belief is unavailable. For instance, if one make-believes that there is beer in the fridge, one may, in conjunction with a desire to have a beer, experience the formation of something like the intention to reach for the fridge. But there is a strict difference between belief and make-believe: though both represent something as being true, make-believe is divorced from any behavioural output.
Is it possible both to believe and to make-believe that there is beer in the fridge? Admittedly, the real question is vacuity, not compatibility. Indeed, if one already believes that there is beer in the fridge, then it would be superfluous also to make-believe it. Similarly, if one make-believes that p, then this state would become pointless as soon as one were to come to believe that p. These remarks fall short of establishing incompatibility. At best, they establish a certain oddity which, as we shall now see, does not turn out to be really puzzling once an adequate picture of what is going on is in place.
First, some interesting examples discussed in the literature suggest that there are genuine cases where one simultaneously believes and make-believes the very same proposition, such as in Leslie's cup scenario. 16 A child is handed an empty cup by an adult and believes that it is empty. As part of a game of make-believe, the child is then asked to pretend that the cup has been filled up before emptying it. As a result, Leslie claims, the child both believes and make-believes the same proposition: that the cup is empty.
It is important to note that the child's resulting two attitudes cannot be conflated. Indeed, the child would certainly be ready to conjoin her make-belief that the cup is empty with the make-believe that the cup was full seconds ago while refusing similarly to conjoin her belief. Even though the child both believes and make-believes that the cup is empty, these two attitudes are distinct in so far as the child is inferentially disposed towards them in different ways.
What Leslie's cup scenario shows is that it is a mistake to claim that makebelieve is a strict substitute for belief. There are at least some cases, such as this one, where the relation between these attitudes is different. The question, then, 15 Stock, 'Fictive Utterance and Imagining' , 149. is what underlies this possibility? Here, I submit the introduction of a new notion to account for this fact: a quasi-factual mental project (QMP). By 'quasi-factual' I mean a kind of mental attitude that represents its content as being the case. 17 To represent some content as being the case, however, must be distinguished from two things. First, being represented as factive doesn't entail being factive; second, believing that p is one way among others to represent p as being the case. Second, by 'project' I mean an endeavour extended over time, though not necessarily continuous. A QMP, then, is an endeavour extended over time, possibly discontinuous, with a particular aim, that is, forming quasi-factual attitudes with respect to subject matter.
Imagine, for instance, that one grows fond of cheese. As a result, one engages in a particular QMP, extended over time and possibly discontinuous: forming all kinds of beliefs about cheese. Here is another example: engaging with fiction.
Such an engagement can also reasonably be characterized as a QMP, extended over time and possibly discontinuous, of forming make-believes about the world of a fiction. Both kinds of QMP are similar: they are a matter of determining how things are with respect to some subject matter. This subject matter can be either real or fictional. If it is real, the QMP involves forming beliefs; if it is fictional, the QMP involves forming make-believes.
A decisive claim about QMPs is that we can conduct more than one such episode at a time. This is obvious with QMPs that aim at forming beliefs.
Presumably, all such kinds of QMPs are sub-projects of a bigger one: finding out what our world is like. It is also obvious with respect to other kinds of QMP. Indeed, the quasi-factive attitudes we form towards our world and the world of a fiction are not in competition with each other and are therefore compatible. Take the cup case again. On the one hand, the child is engaged in a continuous QMP of forming beliefs about her environment and believes that the cup is empty.
On the other hand, she simultaneously takes part in a game of make-believe and, as a result, engages in a second QMP of forming make-believes about the world of the game, where a cup is filled and then emptied. The two projects are not in competition with each other and, because they have radically different aims, are compartmentalized.
We can now have a fresh look at the beer example with these new elements in place. It seems odd, I suggest, that one simultaneously believes and makebelieves that there is beer in the fridge as long as these attitudes occur as part of the same QMP, that is, determining the availability of drinks in one's environment.
But a subject could well engage in two different QMPs with two different aims Patchwork Puzzles and the Nature of Fiction 17 See Stock, 'Fictive Utterance and Imagining'; Only Imagine. and, consequently, both believe and make-believe that there is beer in the fridge.
As long as the aim of the two QMPs are distinct, there is no reason to regard such a combination of attitudes as odd.
This fits very well with what is going on in the case of fiction. An author can, by means of writing a single sentence, aim at realizing two intentions: prescribing both believing and make-believing the very same proposition. These prescriptions offer guidance that pertains to two different QMPs on the reader's side: forming make-believes about the world of the fiction and forming beliefs about the world. If we return to Tolstoy's example, someone reading the opening sentence of Anna Karenina can, first, make-believe its content as part of an ongoing QMP that aims at finding out what is the case in the world of the fiction; and, second, believe it as part of an ongoing QMP that aims at forming beliefs about families.
This also fits well with the phenomenology of engaging with a work of fiction.
To engage with a work of fiction is to be immersed in a fictional world. Yet it often happens that elements of a work of fiction strike us as being not only true in the work of fiction but also genuinely true. In such cases, we don't trade immersion in a fictional world for engagement with the real world. On the contrary, we keep both doors open, so to speak, and add that truth both as a component of our QMP that pertains to the fictional world of the work and to another QMP that pertains to the real world such as, recalling the Tolstoy example, our forming beliefs about the general functioning of families.
It is time to sum up the main theses of this section. First, works of fiction are entirely composed of fictive utterances. Hence, there is no puzzle at the level of utterances. Second, since all fictive utterances prescribe make-believe, there is also no puzzle at the level of our engagement with works of fiction: a work of fiction is entirely imagined. But in order to account for the claim that fiction can also be a vehicle of genuine truths, it must be recognized that at least parts of works of fiction can also simultaneously be believed. This is accounted for by the notion of a QMP: one can competently engage with a fiction while both imagining and believing some part of it as long as each attitude is part of a QMP.
In the next section, I compare this account to the one recently proposed by Stock, and argue that there are some reasons to prefer the account I have just sketched out.

IV. STOCK'S SOLUTION TO THE PUZZLES
Having identified the patchwork puzzles, Stock sets out to dissolve them. But her proposal is very different from the one I have just sketched. In this section, I shall offer three reasons to prefer my account over hers. First, I present some doubts Stock characterizes the attitude of F-imagining as follows. First, it is propositional. Second, it is quasi-factive, in other words, it is an attitude that presents its content as being the case and not, contrary to desire, as having to be made the case. Third, it is characterized by two central principles: Connect 1: Necessarily, a thinker T who imagines that p is disposed to connect her thought that p is the case with other propositional thoughts about what is the case. 18 Connect 2: Necessarily, where a thinker T imagines that p at time t, either T does not believe that p or T is disposed to connect her thought that p is the case to some further proposition(s) about what is the case, whose content is not reduplicated by any belief of hers at t. 19 What is essential about these two claims is their dispositional nature. That is, to F-imagine that p is to be disposed to do certain things with some propositions whose content one takes as being the case. The first general disposition is that one should be disposed to connect a proposition p that one imagines to other propositions one is also thinking. This first requirement is pretty weak and seems to apply to any kind of thought, not merely imaginative ones. Generally, to think is to connect thoughts and to engage in inferential activities.
Connect 2 is much more substantial since it is supposed to account for the specific dispositional profile of states of imaginings. The idea is as follows.
If one F-imagines, either one doesn't believe that p but meets the criteria for F-imagining or one does believe that p but is disposed to connect the content of one's beliefs to thoughts one does not believe -a functional profile much different from the one of belief, where one should be disposed to connect what one believes only to other thoughts one believes. Take the cup example again.
According to Stock, the child F-imagines that the cup is empty in so far as she is disposed, even while believing that the cup is empty, to connect this thought to ones she does not believe, that is, that the cup was full thirty seconds ago.
Stock then proposes to define works of fiction in terms of her notion of F-imagining. A work W is a fiction if, and only if, it is produced by an author with the reflexive intention that its audience should F-imagine its content. As she puts it: 'A fiction should be understood as a collection of utterances of a certain kind: utterances intended by their author to produce conjoined F-imagining in readers Finally, Stock claims that this is sufficient to dissolve the two patchwork puzzles.
First, since we are supposed to F-imagine each utterance contained in a work of fiction and, moreover, since an utterance is fictive only when there is a prescription that we imagine it, then a work of fiction is entirely composed of fictive utterances. Second, the content of fictive utterances can overlap with the content of what one believes, and since to F-imagine is characterized in terms of a disjunctive claim whose second disjunct explicitly mentions that it is possible to believe that p and yet to imagine that p, then the second puzzle also disappears.
However, as I shall now suggest, there are good reasons to prefer my account over hers. First, imagine that an author produces a work with the intention that her readership should F-imagine its content though she intentionally does not introduce any content that her readership does not believe. According to Stock, such a scenario would be impossible. Indeed, she stipulates the nature of F-imagining in such a way that one cannot F-imagine only what one believes. But what is the rationale for such a requirement? After all, she spells out the nature of F-imagining in dispositional terms: if one both believes and F-imagines that p, then one should at least be disposed to connect one's thought that p to other thoughts that one does not believe. But one might well be so disposed without the disposition actually being triggered. In other words, there is a suspicion that this supplementary clause is merely ad hoc. It does not really pertain to the nature 20 Stock, Only Imagine, 150.

Patrik Engisch
Estetika: The Central European Journal of Aesthetics, LVI/XII, 2019, No. 1, 0-00 of F-imagining but, rather, is introduced to prevent the possibility of problematic cases such as the one just described.
Second, Stock claims that her notion of F-imagining also provides us with an explanation of what is going on in cases like the cup scenario. Indeed, such a scenario admittedly demonstrates that one can both simultaneously believe and imagine the same proposition. Yet it doesn't tell us what it is to imagine a proposition that one also believes. A task for philosophy is to clarify conceptually the phenomenon of imagining so as to offer some notion of imagining that is compatible with simultaneously believing what one imagines. And, as we have seen, Stock claims that the notion of F-imagining is up to this task.
But here is the rub. The solution only works if the belief remains dispositional.
If both dispositions were to be manifested simultaneously, then the child would find herself in an impossible position. On the one hand, as part of the dispositional profile of belief, she would have to be disposed to connect her thought only to what she already believes. On the other, as part of the dispositional profile of F-imagining, she would have to be disposed to connect her thought to other thoughts she doesn't believe. And, surely, one cannot be both disposed and not disposed to connect what one believes to what one does not believe.
But is there a reason to believe that we can simultaneously and occurrently both believe and imagine the same proposition? It is certainly false that two different attitudes, believing that p and imagining that p, can both be at the centre of one's attention at the same time. However, there is no reason to deny that both can be occurrent though only one can occupy the centre of one's attention at any one time. For instance, we could well imagine the child reflecting on her situation and thinking: 'How odd it is that I am currently imagining a proposition I already believe. ' Hence, after all, Stock's notion of F-imagine is, contrary to my solution, not well suited to describe a case like Leslie's cup example.
Finally, imagine the following case. At a time t, an author writes a work of fiction F in which two sets of propositions intermingle: S 1 , containing propositions that everyone believed at time t; S 2 , containing propositions that no one believed at time t.
Yet, at time t+n, beliefs have changed in such a way that everyone both believes the contents of S 1 and S 2 . As a result, at time t+n, readers cannot comply anymore with the author's intention: whatever they imagine they also believe. How should we treat such works? Is F fiction at time t but non-fiction at time t+n? Is it still fiction at time t+n but only if one reads it with the perspective of someone engaging with it at time t? These are weird questions. Moreover, these are questions that pop up only against the background of Stock's own theory. No one Patchwork Puzzles and the Nature of Fiction that our Currie-inspired account provides a straightforward answer to the problem. The work would still count as fiction at t+n because its content, if true, would at most be non-accidentally so.
On the basis of these three elements, I then conclude that my solution to the patchwork puzzles should be preferred over Stock's.

V. CONCLUSION
The aim of this paper was to discuss two patchwork puzzles that Stock claims beset Currie's theory of fiction. The first puzzle occurs at the level of occurrences, the second at the level of attitude. I started by presenting Currie's account and then identified precisely the source of the puzzle -namely, the two claims that, first, works of fiction can be furnished both by fictive and non-fictive utterances and, second, that some utterances cannot realize more than one intention. I then identified the reasons that push Currie to make such claims and replaced them with suitable alternative theses. Next, I showed that adopting them is compatible with Currie's main aims in formulating his theory. As such, my account constitutes a suitable alternative to Currie's. I then discussed an alternative solution proposed by Stock, before presenting three objections to it, and concluding that there are good reasons to prefer my account to hers.