How are fictions given? Conjoining the ‘artifactual theory’ and the ‘imaginary-object theory’

According to the so-called ‘artifactual theory’ of fiction, fictional objects are to be considered as abstract artifacts. Within this framework, fictional objects are defined on the basis of their complex dependence on literary works, authors, and readership. This theory is explicitly distinguished from other approaches to fictions, notably from the imaginary-object theory. In this article, I argue that the two approaches are not mutually exclusive but can and should be integrated. In particular, the ontology of fiction can be fruitfully supplemented by a phenomenological analysis, which allows us to clarify the defining modes of givenness of fictional objects. Likewise, based on the results of the artifactual theory, some assumptions in the imaginary-object theory, which are liable to be interpreted as laying the ground to phenomenalism, can be corrected.


Introduction
Do fictional objects, such as Little Red Riding Hood, Wonderland, or Humbert Humbert, exist? And, if they do, what is their ontological status? Is there a relation between the mode of existence and the mode of givenness of fictional objects? These and similar questions have been largely debated in the philosophy of fiction. Roughly speaking, the main divide is between eliminativist and non-eliminativist approaches. Eliminativist approaches deny the existence of fictional objects toutcourt. Eliminativists are mostly interested in problems related to the truth value of 1 3 fictional statements, the clarification of which would not profit from attributing some kind of existence to fictional objects. Fictional statements, in fact, can be understood as intensional contexts obtained by means of appropriate operators (such as the narrative operator "it is true in the fiction that…"), and accordingly they do not require what is represented in the statement to exist (cf. Beyer, 2004;Künne, 2007;Sainsbury, 2010). 1 Non-eliminativist approaches, instead, generally attribute some kind of existence to fictional objects, but positions diverge when it comes to determining what kind of existence. 2 Eliminativism shows that it is possible to envisage and even to construct intelligible and coherent theories of fiction based on linguistic analysis without assuming that fictional objects have any kind of existence. However, ficta have an impact on our experiences, in a way that goes beyond the simple understanding of fictional discourse. We are moved by fictions in ways that often impinge on our real life (Summa, 2018(Summa, , 2019a. We refer to fictions as something that is part of a culturally shared heritage. We either agree or disagree with others in a social context with respect to fictions. All this seems to imply that we not only have an understanding of the meaning of fictional propositions, but that we intend something in and through these propositions, and that this intention cannot be empty. Additionally, there seem to be differences between objects that do not exist in reality but that we can fictionally conceive and objects that we cannot conceive, such as a round triangle, and, among the objects we can conceive, between the ones that we intend without any stability in intermittent acts of imagining and the ones that are part of a shared cultural heritage. 1 In this regard, the category of fictional discourse should be further specified. Notably, following Künne (2007), we should distinguish: transfictional discourse (assertions such as "I have an image of Little Red Riding Hood" or "Little Red Riding Hood is smaller than I am", in which only one term is fictional and is embedded in a context that refers to reality); intrafictional discourse (assertions about fictional objects from within the fictional context, such as "Little Red Riding Hood encounters a wolf"), interfictional discourse (assertions about fictional objects that go beyond the fiction, for instance involving objects of other fictions, such as "Little Red Riding Hood is younger than Rapunzel"); and status assertions or metafictional discourse (assertions about fictional objects, from a meta-perspective, such as: "Little Red Riding Hood does not exist"). Cf. Künne (2007, pp. 279 f.); Beyer (2004). On this topic, see also Erhard (2014, pp. 443 f.), who discusses how the different kinds of fictional discourse can be understood on the basis of a non-eliminativist ontology, such as the one of the later Husserl. 2 A non-exhaustive list of non-eliminativist positions includes: (a) possibilist theories, according to which fictional objects are part of possible words and fictional truth should also be considered on the basis of the ontology of possible words (e.g., Lewis 1978); (b) neo-Meinongian theories, according to which fictional objects are 'Meinongian objects', that is, objects we can predicate about, even if they are not existing in space and time, being only characterized by 'mere subsistence' (Bestehen) (e.g., Parsons 1980); (c) abstract-object theories, contending that fictional entities are abstract entities (Zalta 1983) or "theoretical entities of literary criticism" (van Inwagen 1977); (d) pretense theories, according to which fictional entities do not exist in reality but are the products of make-believe (e.g., Walton 1990); (e) abstract-artifact theories, according to which, as I show more extensively, fictions are abstract artifacts; (f) intentionalist theories, according to which fictional objects are intentionally constituted by the creative acts of the author and entail the properties the author intends the reader to imagine. While the latter view has some affinities with the phenomenological approach (e.g., Stock 2017), it defends a narrower account of intentionality. In this paper, I will only concentrate on the abstract-artifact theory and an intentionalist approach based on the phenomenological theory of intentionality. Cf. Kroon and Voltolini (2018) for an overview.
For these reasons, I believe a theory is required that is able to extend our ontology to embrace fictional entities in their specific mode of existence.
Among the non-eliminativist approaches, Amy Thomasson's (1999) stands out for developing an ontological theory of fictional objects as abstract artifacts. According to this theory, fictional objects are notably irreducible to the standard dichotomy between real and ideal objects and should be rather assimilated to cultural objects as abstract artifacts and socially constituted entities. As she introduces her view, Thomasson distinguishes the artifactual theory from other approaches to the ontology of fiction, including the so-called 'imaginary-object' theory, the main proponent of which is Jean-Paul Sartre (1966Sartre ( , 2004. Thomasson contends that, although the artifactual theory and the imaginary-object theory have "certain similarity in spirit," they eventually "differ in substance" (Thomasson, 1999, p. 21). The two theories are similar insofar as they share a noneliminativist view on fictional entities; they consider the dichotomy between the real and the ideal as inadequate to capture the ontological status of fictional entities; and they claim that fictional entities are dependent entities. The apparent differences between the two approaches are instead connected, not without importance, to the role ascribed to the imagination in making the givenness of fictional objects possible. Sartre's ontological claims regarding fictional objects as irreal or imaginary is namely based on his intentional theory of the givenness and constitution of ficta as products of spontaneous imaginative acts, which are irrealizing insofar as they establish a discontinuity with what is experienced as spatiotemporally real. Differently, Thomasson does not focus on the specific intentional imaginative accomplishments that make the constitution of ficta possible, and rather develops her position within a strictly ontological and semantical framework.
The theoretical question raised by Thomasson's demarcation between her own approach to the ontology of fiction and Sartre's is whether considering fictional objects as imaginatively given and constituted means to rule out their existence as abstract artifacts in our real world. In a broader and more radical sense, the question is whether ontology-and the ontology of fiction in particular-can and should be discussed independently of the phenomenology of experiential givenness and constitution. In this article, I wish to argue that, although such an ontology is internally consistent, it can be fruitfully supplemented by investigations into the givenness and constitution of fictional objects. Givenness and constitution are the two sides of the phenomenological program. 'Givenness' refers to the direct mode of presentation of intentional objects, in our case of fictional objects, to experiencing subjects. 'Constitution' is a broader term referring to the intentional structures that make the givenness of objects possible. Concerning our topic, we are to take into account the intentional structures characterizing the acts of both producers and receivers (readers or spectators) of fictional works. The concept of intentional constitution can be understood in a general and in a specific sense. In the general sense, constitution means the apprehension of something as something, for instance the apprehension of a unitary object on the basis of its plural modes of appearance. This is the phenomenological meaning of constitution, which holds for every intentional object. Assuming thus that all intentional objects are constituted in this general sense, there are 1 3 specific modes of constitution in each ontological region. In particular, one should distinguish the constitution of real objects from the constitution of imaginary objects and of fictional objects. With regard to the latter, hence in a specific sense, constitution also means ontological creation.
I wish to argue that the givenness and constitution of fictional objects rely on imaginative intentional acts and that the ontological claims resulting from the phenomenology of constitution of fictional objects are not at variance with the artifactual theory, but rather complement it by highlighting what the conditions for the experience of fictional objects are. Furthermore, reassessing the imaginative constitution of fictional objects in the light of the results of the artifactual theory allows us to correct some controversial claims made by Sartre, which are liable to be interpreted as laying the ground to phenomenalism.
Let me try to be more precise on this latter point. What Thomasson explicitly criticizes in Sartre's view is the discernible phenomenalist implications of the claim that imaginary objects exist only as long as they are imaginatively actualized or "as long as someone is thinking of them" (Thomasson, 1999, p. 22). This is certainly a serious concern raised by some of Sartre's assumptions. But is such phenomenalism the unavoidable consequence of considering fictional entities as given in and through acts of imagination? I contend that this is not the case. In order to support this claim, in what follows, I elaborate on the connection between the ontology of fictional objects and the phenomenological analysis of their givenness and constitution. I first address the ontology of fictional objects as dependent (Sect. 2). Subsequently, I discuss in what sense the ontology of fiction needs to be supplemented with an inquiry into the givenness and constitution of fictional objects (Sect. 3). And finally, I discuss in what ways imaginative activity is required in order to account for the specific modes of givenness and constitution of fictional objects (Sect. 4).

Fictional entities as abstract artifacts: an ontology of dependence
According to the artifactual theory, fictional objects are to be considered as abstract artifacts. As artifacts, fictional objects owe their existence to the creative activity of someone who produces them. As abstract, these artifacts are distinguished from material or concrete ones, such as chairs and tables, being rather assimilated to cultural and social institutions, such as laws, governments, marriages, etc. (Thomasson, 1999, pp. xi, 12). In fact, the affinity with social and cultural objects is considered to be one of the main reasons why fictional entities are philosophically interesting: "It is […] their place as cultural artifacts that makes fictional objects of broader philosophical interest, for the ontology of fiction can thus serve as a model for the ontology of other social and cultural objects in the everyday world" (Thomasson, 1999, p. 14).
Against this background, ficta should be assessed on the basis of two-level dependence relations. 3 Proceeding top-down: fictional entities depend on both the creative activities of the authors of fictional works, like novels, tales, etc. and on the fictional works themselves (first level). Yet, fictional works are also dependent entities: they depend on material and meaningful signs and on competent readership (second level). Since 'dependence' is a transitive relation, and since there is a juncture between these two levels, fictional entities also depend on material and meaningful signs and on competent readership. Thomasson (1999, pp. 24 f.) further specifies these dependence relations by distinguishing (a) between constant and historical dependence, and (b) between rigid and generic dependence. As to the former distinction (a), if x constantly depends on y, then, necessarily, whenever x exists, y also exists (Thomasson, 1999, pp. 30-31). Instead, if the x historically depends on y, then x needs y in order to come to existence, but x may continue to exist independently of whether y continues to exist or not (Thomasson, 1999, pp. 31-32). As to the latter distinction (b) x rigidly depends on y, if y is a determinate and irreplaceable individual; instead, x generically depends on y, if y is a token of a particular type and can in principle be substituted by another token (Thomasson, 1999, p. 27).
Based on these distinctions, on the first level, fictional entities are rigidly and historically dependent on the creative activity of their authors. Such dependence is rigid, because the author is a singular and not interchangeable individual; also, it is historical, because, while fictional entities depend on the authors of fictions in order to come into existence, they can continue to exist independently of them. Moreover, fictional entities are generically and constantly dependent on the work of fiction itself. Such dependence is generic, because fictional entities do not depend on one individual and not-interchangeable copy of a book; they do constantly depend on the existence of some material copies or even only on orally transmittable memories of the work, but these material supports are interchangeable.
The second level of dependence concerns fictional works. In this respect, Thomasson follows Ingarden in considering fictional works as being themselves dependent entities, notably dependent on material and meaningful signs and on competent readership or competent recipients. The reference to meaningful signs should be clarified in relation to Ingarden's account of the stratified structure of fictional works. On the first layer, fictional works are made of 'linguistic sound formations,' including material-written or oral-signs and/or sounds, the metric and prosodic components, rhythm, melody, the structuration of the text in paragraphs, etc. (Ingarden, 1973b, pp. 34 f.). Written and oral material signs signify something. What is signified by those signs-the socalled 'meaning units'-makes up the second layer of fictional works. This includes individuals (meanings of proper names) as well as more complex states of affairs or processes in temporal development (meanings of judgments, of entire paragraphs or chapters) (Ingarden, 1973b, pp. 62 f.). Meaning units, then, refer to something, and the 'objects' they refer to form the third layer of fictional works. This includes all the entities (characters, episodes, events, states of affairs, etc.) that are presented in the fiction (Ingarden, 1973b, pp. 217 f.). Finally, expressions always refer to something in a certain way, that is to say, they refer to their objects in and through certain descriptions. Each description includes only particular aspects of the object, which is presented from one particular perspective. And a reference is established by means of these particular descriptions alone (Ingarden, 1973b, pp. 255 f.). These latter coincide with the aspects and the modes of appearance in and through which characters, states of affairs, situations, stories, etc. are presented in a literary work. They make up what Ingarden calls the layer of 'schematized aspects'.
In short, fictional works are dependent entities insofar as they depend: (i) on the just mentioned strata (including both material and meaningful components); (ii) on authors who produce such material and meaningful components, and (iii) on competent readership (Thomasson, 1999, p. 36). The claim concerning competent readership derives from the rather straightforward observation that no fictional entity would be produced if there were no readers capable of understanding what is written and signified in a fictional work. The dependence is here generic, because it is not the case that fictional works depend on singular, non-interchangeable readers; and it is constant, not because fictional works need to be constantly read, but because they need to be constantly readable. Accordingly, fictional entities, like characters in a novel, depend immediately on literary works and authors; while the literary work itself immediately depends on authors, copies or memories, and a competent readership (Thomasson, 1999, p. 36). This is a complete description, which exhausts the dependence relationships that define a fictional entity. Also, given that dependence is a transitive relation, fictional entities are also mediately dependent on what fictional work are dependent on: authorship, material copies, and readership. Schematically:

The ontology of fiction and the problem of givenness and constitution
While subscribing to the artifactual view, in what follows, I wish to argue that the ontology of fiction cannot do without considerations concerning how fictional entities are imaginatively given and constituted. In order to understand the import of this claim, we should first address what I consider to be a kind of tension between realism and idealism about fiction. While assessing what is meant by saying that fictional entities are 'real' or 'ideal,' I argue that this tension is-to a large extentonly superficial and that the actual controversy touches on questions concerning the constitution and givenness of fictional entities. Thomasson characterizes her approach to fiction as 'realist' (cf. Thomasson, 2010). Consistent with a generally accepted understanding of 'realism'-for instance in debates concerning realism about mathematical objects or moral factsthis does not mean that she believes fictional entities exist as physical entities in the spatiotemporal world, nor that they exist in some other world. More specifically, being realist about fictional entities, for Thomasson, means assuming that these entities are abstract artifacts that exist as such (that is to say, as abstract artifacts) in the real world. 4 In order to define the ontological status of ficta, a system of categorization is required (Thomasson, 1999, pp. 115 f.), which is not restricted to the binary division of 'real vs. ideal' or of 'material vs. mental,' since "[t]he world seems to be far more varied than these divisions would allow" (Thomasson, 1999, p. 135). This system of categorization embraces "unfamiliar categories" between those extremes, that is to say, between the material and the mental (e.g., social objects, such as mortgages, money, and governments) and between the real and the ideal (e.g., several kinds of art objects, such as works of music, literature, and visual design). Fictional entities also fall under these categories (Thomasson, 1999, p. 127, cf. ibid. p. xii).
In this respect, Thomasson agrees with Ingarden, who argues that the ontology of fictional objects actually breaks the dichotomy between the real and ideal and requires a more differentiated ontology. Ingarden thematizes the specific ontological status of fictional objects based on four criteria: (a) dependence on mental activity in order to come to existence; (b) origin in time; (c) identity criteria; (d) spatial and temporal constraints (Ingarden, 1973b, pp. 10 f.).
(a) Different from both real and ideal entities, fictional entities are dependent on someone's mental activity in order to come to existence: without Goethe's creative acts, there would be no Faust. (b) Different from ideal entities, which are a-temporal or omnitemporal, and like real entities, fictional entities come into existence at a certain moment in objective/historical time. One can thus say that they have an origin in time: fictional entities begin to exist when they are produced. (c) Fictional entities are subject to specific identity criteria, which are different from the identity criteria of real entities. For instance, even though Goethe's Faust had been subject to changes, the identity of the character Faust would be maintained if the changes were made or approved by Goethe himself; conversely, we cannot say that Dafoe's and Tournier's Robinson are one and the same character (cf. Ingarden, 1973b, p. 11;Thomasson, 1999, pp. 55 f.). (d) Finally, different from real-and more precisely, physical-entities, fictional entities have no objective spatiotemporal being and are not subject to causal determinations of material nature. However, unlike ideal beings, they cannot be said to be a-temporal either, not only because they have an origin in time, but also because the narrative that produces them is temporal. Although this might be only a terminological difference regarding what is included under the concept of reality, at variance with Thomasson, Ingarden does not present his own view as a 'realist' theory of fiction, but rather supports existential pluralism by claiming that, besides real and ideal beings, we should also recognize 'purely intentional' beings or formations (Gebilde). 5 Besides dependence relations, which can be established by means of ontological considerations, there are further characteristics that pertain to fictional objects precisely insofar as they are intentionally constituted. 6 In particular, in the experience of fiction, recipients are confronted with a unique kind of indeterminacy in the mode of givenness of fictional objects. This mode of givenness marks a crucial difference between fictional objects and real (material or physical) objects. Fictional entities are always and necessarily presented in and through schematic aspects-which means not only that descriptions are partial and perspectival, but also that they are necessarily characterized by empty or indeterminate spots (Leerstellen; Unbestimmtheitsstellen). Accordingly, fictional objects are in a specific way constitutively indeterminate. This indeterminacy is different from the incidental indeterminacy that can characterize real objects. Even if there are aspects or properties of a real object that are not determinate for us at a given moment, these aspects or properties can be determined. We can thus speak in this respect of a determinable indeterminacy, while in the case of fictional entities we should speak of indeterminable indeterminacy. The spots of indeterminacy of fictional entities are such that they cannot in principle be determined; they are indeterminable tout court (Ingarden, 1973b, p. 246).
This approach to fiction can be said to be realist only if one takes a broad concept of reality, in other words, only if one assumes that fictional objects exist in the world as intentionally created and intentionally dependent. It is not a realist account if one considers real being as material/physical being, that is to say, as independent of mental activity (as we saw, fictional objects are historically dependent thereupon); as spatio-temporally localizable and identifiable in the physical world; and as causally determined. The qualification of indeterminacy also touches on the problem of the reference of simple everyday empirical concepts in fiction. If a novel describes a scene with a There is an 'empty' spot here, a 'spot of indeterminacy' […] such empty spots are impossible in the case of a real object. At most, the material may, for example, be unknown" (Ingarden, 1973b, p. 249). When saying that fictional objects as schematic formations have spots of indeterminable indeterminacy in their content, I am thus implying that recipients have no means to determine whether they have the property P, if this property is not attributed to them by the author (or can be inferred by other properties attributed to them by the authors). 7 Saying that Hamlet's blood type marks a spot of indeterminacy in content means that according to the story one has no grounds to claim the truth of a judgment on Hamlet's blood type in the story.
I will argue in the next section that this indeterminacy of fictional entities is the correlate of a specific kind of imaginative activity which is constitutive of fictional entities in both the general and the specific sense. Before doing so, I wish to return to the apparent tension between idealism and realism concerning fiction by further discussing the meaning of 'correlation' and 'constitution,' which I understand along the lines of Husserl's and Sartre's phenomenological accounts of fiction. Both philosophers characterize fictional objects as irreal-hence none of them can be labeled strictly speaking as a realist when it comes to fictions-and yet they both claim that they exist as constituted entities. Sartre is rather clear in emphasizing that we should not confuse 'reality' with 'existence,' since both 'reality' and 'irreality' should be considered as qualifications of 'existence'. Accordingly, fictional objects are in fact existing in the mode of irreality: "[i]n reading as in the theatre, we are in the presence of a world and we attribute to that world just as much existence as we do to that of the theatre; that is to say, a complete existence in the irreal" (Sartre, 2004, p. 63).
For his part, Husserl claims that fictional entities exist and belong-together with purely ideal objects, like numbers-to the domain of 'irrealities'. 8 In fact, fictional entities are ideal objects, but, unlike pure idealities, they are 'bound'. One may wonder whether this account is not deeply at odds with the realist view on fictional entities previously outlined. How can we claim that the phenomenology of fictional entities as bound idealities is consistent with the realist account of the artifactual theory? In order to answer these questions and better clarify the relation between 7 Cf. Ingarden (1973a, pp. 50 f.;1973b, pp. 246 f.). It is controversial whether indeterminacy also means ontological incompleteness, whereby 'ontological incompleteness' means that, for a given property P, an incomplete object x neither has P nor has non-P. Thomasson (1999, pp. 105 f.) distinguishes indeterminacy from incompleteness, notably discussing Parsons' Neo-Meinongian approach. Künne (2007, pp. 304-5) seems instead to suggest that indeterminacy of content implies incompleteness. 8 On Husserl's late ontology of fiction, see Erhard (2014, 443 f.), who focuses on further aspects of the constitution of bound idealities and perceptual phantasies, particularly on the associative syntheses that are required for such a constitution. Erhard also interprets fictional language in Husserl with reference to Künne's distinction between intra-, inter-, and trans-fictional discourse. As a result, he consolidates Husserl's distinction between judging about what is phantasized and judging "as if" one lived in the phantasy (Husserl 2005, p. 447). Considering the truth value of fictional statements in relation to the context in which these are meant to be true or false, Husserl's position is compatible with Thomasson's, who argues that "[t]he problems surrounding fictional discourse may be resolved by recognizing that these problems stem from differences of contexts, nor of objects" (Thomasson 1999, p. 105).

3
phenomenological givenness and the ontological status of fiction, it is worth emphasizing that the ontological difference between real and ideal entities should be assessed in relation to the differences concerning their respective constitution and temporality. This is so notably inasmuch as ideal entities-both pure and boundare not simply separated from the real world. Epistemically, such entities can be cognized only insofar as they appear (auftreten) in the real world in and through temporal determinations (Husserl, 1973, pp. 258 f.). This is the case for pure idealities: ontologically, Pythagoras' theorem is not temporally individuated, although it was discovered and demonstrated by a real individual. It is "everywhere and nowhere" (Husserl, 1973, p. 259) and would exist as omnitemporal even if nobody had ever discovered it. In this latter case, however, it would not have any epistemic meaning for us. Accordingly, if we are to understand how it is that irreal entities have a meaning for us, we should consider them in terms of givenness and constitution (Husserl, 1973, pp. 259-260). Saying that ideal objects-and for the moment I am only referring to pure idealities, like those of mathematics-are omnitemporal does not exclude but rather requires that such objects are given to us in the temporal form of experience. In the case of pure idealities, the temporal concretization is contingent from an ontological point of view, but it is necessary from an epistemic point of view: it may well be that an ideal truth is never 'given' to us or is never concretely expressed in an utterance-and it would still be an ideal truth; and yet, such truth would have a meaning for us only if it makes its appearance (auftreten) in the world through an actual utterance, that is, only if it is constituted in and through intentional meaning-giving acts. This is the reason why, while providing a transcendental account tracing the constituted ontology of ideal entities back to the constituting accomplishments in and through which ideal entities are constituted, Husserl eventually contends that omnitemporality is after all still a mode of temporality (Husserl, 1973, p. 261).
This brief detour through the ontology and constitution of pure idealities as irreal entities is meant to show why the investigation of the ontological status of fictional entities can significantly profit from a phenomenological analysis of their constitution and mode of givenness. In fact, what holds true for pure idealities also holds for fictional entities as bound idealities, since the latter are even more tied to real constraints.
'Bound idealities' are necessarily embodied in 'real things,' such as artworks, books, cities, etc. These 'real things' are not simply 'physical things,' since, besides having material determinations, they are also bearers of "determinations of significance," which should not be considered as determinations of the object just in and for itself, but rather "in relation to us, to our appraising and willing, according to what it signifies for us" (Husserl, 1973, p. 265). Such determinations of significance depend on their being embodied in real, spatiotemporal beings; yet, they are called 'irreal' since they are not themselves spatiotemporal beings. In this sense, Goethe's Faust depends on some real material support (a book), while nonetheless it is not bound to any singular material support. A bound ideality, thus, "is certainly 'embodied' in the real world, but it is not individualized by this embodiment" (Husserl, 1973, pp. 265-266). Synthese (2021) 199:13749-13769 In short, we can say that the difference between pure and bound idealities consists in the fact that the former are only epistemically dependent on their appearing (Auftreten) in the real world, while the latter are ontologically dependent on such appearing. Pure idealities (like a triangle or Pythagoras' theorem) can also be embodied and need to appear in the world in order to be cognized. Yet, the difference between them and bound idealities like Goethe's Faust is that the latter not only need to be embodied in order to be cognized, but they also need to be embodied in concrete-material exemplars in order to exist. Thus, in the case of pure idealities, the embodiment in the material, physical, and linguistic world has an epistemic significance. Differently, in the case of bound idealities, such embodiment has an ontological significance: cultural idealities-and fictional entities as cultural idealities-would not exist if they were not created and embodied in some material entities or, in Husserl's terms, if they were not bound to some "territories" (Husserl, 1973, p. 267).
Partly different from Thomasson's, this approach to fictional entities as bound idealities allows us to clarify to what extent assumptions concerning entities in between the real and the ideal do not stem from top-down categorial constructions, but rather from criteria generated by the way in which we experience different kinds of objects, or, in other words, from the way in which different kinds of objects are given to us. This holds not only for cultural objects-hence for what Ingarden calls 'purely intentional' objects-but for every object in general. Are these observations necessary to make Thomasson's theory, and in general the ontology of fiction, more intelligible? While the intelligibility and the internal coherence of the abstract-artifact ontological approach does not necessarily require a phenomenological analysis, I believe this latter allows us to successfully tackle some of the issues arising from that approach. This should become clearer if we return to the problem from which we started: Thomasson's critique of the imaginary-object theory. Claiming that fictional objects depend on the creative acts of the producers and the receptive acts of competent readership, Thomasson's work does indeed suggest that there is a relation between the ontology of fictional entities and their givenness and constitution. But on what grounds exactly should we reject that fictional objects are imaginary objects, whereby 'imaginary objects' precisely means 'objects intentionally constituted through imaginative acts'?

Fictional entities and their imaginative givenness
As already mentioned, there are similarities between fictional entities and other cultural objects, such as cities, governments, marriages, etc. In both cases, we have to do with dependent, intentionally constituted, objects (abstract artifacts) that somehow exist in our world (cf. Husserl, 1973, pp. 258 f.; Thomasson, 1999, p. 89). However, are there any differences between those abstract artifacts that make up our social 'reality' and those we recognize as 'fictions'? In general terms, we can say that the former affect our cognitive and practical behavior differently compared to the latter, and this seems already to give us reasons not to consider them as belonging to one and the same class. One way to address this difference points to the 'unserious' nature of fictional entities (cf. Austin, 1962, pp. 21 f.;1958;Ryle, 2009, pp. 245 f.). However, 'unseriousness' seems too vague a characterization: what is the special meaning of this 'unseriousness' and what allows us to characterize something as serious or unserious? I wish to argue that we can elaborate on the distinctive nature of fictional entities if we consider them not only as abstract artifacts, but also as imaginatively constituted objects. Notably, the imaginative constitution of fictional entities should be investigated in relation to the three following moments: (a) the as-if structure of the experience of fiction, (b) the spots of indeterminacies and the appeal to imagine, (c) the sociality of imagination in fiction.

(a) Experiencing as-if: fiction and bound phantasy
While assimilating fictional entities to cultural artifacts, we should not neglect their imaginative constitution. 9 Such imaginative constitution can be investigated by more closely taking into account the rigid and historical dependence of fictional entities on the creative acts of the authors. These are imaginative acts shaped and expressed in a linguistic and narrative form. When the authors sketch a story or come to an idea on how to write an episode, they imagine or phantasize how such a story, the episode, the characters in it, etc. would be, feel, behave. They do neither perceive, nor remember such story, episode, etc. Instead, they linguistically shape phantasies that also have a sensory component (Casey, 2000, pp. 41 f.). In order to understand what is meant by the idea that imagination is experiencing in the as-if mode, it is important to depart from the idea that imagining is 'having mental images,' as well as from the limitation of the quasi-sensory component to the visual field. Imagining is an intuitive act simulating possible experiences (Husserl, 2005;Jansen, 2005Jansen, , 2016Summa, 2021), and the quasi-sensory components include different senses, as well as intersensorial experiences, feelings, and atmospheres. It should be made clear that I am not claiming that imagining is first shaped in a pre-linguistic, sensory way and then transposed into a linguistic form. On the contrary, when it comes to the fictional entities here under consideration, linguistic expressions (but also other kinds of expression, e.g., figurative or performative) shape the imaginative experience. This however does not exclude that there are forms of imagining that are not linguistically structured. My claim is then that fictions are narratively constituted imaginative productions. This also implies that they are socially constituted entities (Ricoeur, 1983(Ricoeur, , 1984(Ricoeur, , 1985Summa, 2019b). Furthermore, not only the authors' acts, but also the readers' and more generally the recipients' acts are imaginative. Recipients of fiction suspend reality and temporarily-even if not completelytranspose themselves in a context that does not belong to what they ordinarily take to be real (cf. Summa, 2018Summa, , 2019a. They read about characters or experience them on stage without properly interacting with them, and yet participate in a specific way in the fiction, for instance empathizing with characters, being genuinely emotionally moved by what happens in an episode of a novel or on stage, etc. The imaginative activity is thereby 'bound,' not only in virtue of the material support we have discussed above, but also because recipients are implicitly asked to follow 'prescriptions' coming from the fiction itself (Husserl, 2005, pp. 617 f.). In this sense, when we read a novel, certain phantasies are forced upon us "by the succession of spoken and written words" (Husserl, 2005, p. 620).
Within Ingarden's theory, taken by Thomasson as a source of inspiration for the artifact theory, imaginative 'simulation'-or experience in the as-if mode-is also a condition for the experience of fiction. Without acts of intuitive, imaginative representation, we would remain on the level of empty signification. The stratum of 'schematized aspects,' thus, guides us through the imaginative givenness of the intended object (Ingarden, 1973b, p. 197).
Unlike pure or free phantasies-which are independent of material supports and not constrained by particular affordances-bound phantasies rely on perceptual objects (like in theater, and also in pretend-play) and/or on linguistic expressions that function as affordances to imagine (cf. Summa, 2020). In most cases, the two bonds are intertwined, so that we have stories and linguistic affordances to imagine, based on perceptual appearances of objects that take up the function of fictional entities or 'figments' (Fikta). Figments share with bound idealities the double nature discussed above: they are irreal, since they are not spatiotemporally and causally determined as real objects, but they depend on the embodiment in real/perceptual objects (Husserl, 2005, pp. 616 f.). Experiencing figments, "[w]e live in neutrality, we do not carry out any actual positing at all with respect to what is intuited" (Husserl, 2005, p. 617). And yet, the figment "presents itself […] in the real thing, when the latter offers us, under given circumstances in its 'genuine' perceptual 'appearance,' precisely that which is 'genuinely perceived' in the figment" (Husserl, 2005, p. 619). In this sense, we can say that we constitute fictional objects by grasping the imaginary as imaginary in and through the real (Husserl, 2009, p. 211).

(b) Spots of indeterminacy and the appellative character of fiction
In the introduction to his The Literary Work of Art, Ingarden departs from the claim that fictional objects can be equated to imaginary objects (Ingarden, 1973b, pp. 16 f.). Thereby, he departs from the assumption that fictional objects are ontologically reducible to mental acts. Yet, this does not mean to exclude that imagination plays a role in the constitution of fictional objects. As previously mentioned, Ingarden recognizes that experiencing fictions means finding oneself in the realm of imagination 1 3 and intuitively fulfilling mental acts (Ingarden, 1973b, pp. 197, 308). Imaginative activity is required for the intuitive concretization and actualization of the work of fiction (Ingarden, 1973a, pp. 50 f.). As part of the above-mentioned intuitive-imaginative fulfillment, Ingarden also introduces a more specific imaginative activity, which, in the concretization of the experience of fiction, consists in filling the gaps of indeterminacy. While maintaining that such indeterminacy cannot be removed by establishing a univocally valid determination, we should recognize that recipients are active in the sense that they also synthetically connect, in an implicit way, what is said and what is not said in the text. Also, without raising claims concerning truth or falseness, they implicitly grasp the unsaid and can make such grasping explicit by means of acts of imagination (Ingarden, 1973b, p. 269).
Concretization is not only the mental actualization of written content, but also what directors do when they turn a screenplay into a movie or a theater performance. Such concretizations, while exemplifying the same literary work, should also be distinguished from the work of fiction itself (Ingarden, 1973b, p. 252). And this distinction hinges precisely upon the fact that each concretization also entails moments that determine-on the basis of the imaginative activity of the recipient-what is indeterminate in the fiction. In certain respects, then, Ingarden's approach to concretization qualifies the idea of repeatability: in each reception and each performance, we have a new and different concretization; yet, what is concretized, the fictional object, is always the same entity. In this sense, the recipient's imaginative activity is as necessary for the constitution of a fictional entity as the activity of the author who presents the object under a certain description. On this ground, as Wolfgang Iser (1975) observes, we can take fictions as appeals to imagine directed to the reader. 10 The imaginative activity that underpins concretization deserves some closer remarks. One can indeed question whether it is phenomenologically adequate to claim that, when reading a novel, one imaginatively figures out what is described and imaginatively fills in what the author does not say. At least to a certain extent this seems not to correspond to our experience of reading fictions: we do not imagine the eye color of all characters, or everything that may have happened between one scene and the other. And nonetheless, we are sometimes disappointed when, watching for instance the film adaptation of a novel we have read, we experience how the director concretizes characters or scenes. This shows that, at least implicitly and vaguely, we do perform some imaginative concretization while reading fictional works. In order to understand what is meant by the claim that recipients are actively involved in the constitution of fictional entities and that this also entails filling in the indeterminate gaps (Ingarden, 1973a, pp. 37 f.), it is useful to pursue the analogy between imaginative and perceptual experience a little further. Imagining means experiencing in the 'as if' mode. Basic kinds of imagining are thus perceptions in the 'as if' mode, perceptual simulations, and in this sense in several respects analogous to perceptions. In particular, they have a quasi-sensory component, that is to say, components of perceptual 'simulation,' which are not reducible to the analogy between 'seeing' and 'seeing with the mind's eye,' but also entail other quasi-sensory modalities, as well as atmospheric components, which we may only vaguely imaginatively re-present.
Besides, one should be careful in distinguishing the imaginative activity that is at stake in the constitution of fictional entities from the awareness of occluded parts in amodal perception. 11 Whenever I perceive a cube, and actually only see three of its faces, I am not imaginatively representing the faces I do not see. Instead, I perceive the cube as a cube precisely because I also perceive its hidden sides as hidden. I have a different experience when I try to actually figure out how the hidden faces may look, for example with regard to their color, smoothness, etc. In the latter case, I am imagining how the faces I do not currently see may look. I simulate the experience in the 'as if' mode, and-if I only imagine that-I do not make any claim concerning the correctness of what I imagine, hence I remain neutral in this respect. Analogously, the distinction between an implicit synthesis and the possibility of making something explicit is also found in imaginative experience, only modified. In short, most of what we imagine by simulating an experience in the 'as if' mode has nothing to do with making explicit how we imagine something that is indeterminate in the text. We implicitly follow the unravelling of descriptions and stories without representing everything that is left unsaid. And yet, we follow the flow of the story, thus implicitly making synthetic connections. We can, however, also make things explicit and, like in perception, seek determinations where we do not have them. This is all the more necessary when the concretization occurs not only in reading, but also in staging. For instance, theatre directors will have to imagine what their Hamlet should look like. In such processes of explicitly filling gaps of indeterminacy-which do not and cannot in principle concern all that which remains unsaid, because this is clearly an indeterminate realm-imagination operates within a context that is already imaginative, so to speak, on a second level. And we speak here of imagination precisely because we remain neutral with regard to the correctness of our concretization. Concerning the constitutively indeterminate gaps, we cannot univocally decide whether one concretization is true or false; we can only consider what is more or less adequate on the basis of consistency or even only on the basis of certain aesthetic criteria.

(c) Fictions as social imaginary objects
Bound imagining-notably when it entails linguistic expression-cannot be reduced to individual experience, as it is rather a social form of imagining. Husserl already hints in this direction when he claims that fictional objects-like all bound idealities-have an "intersubjective existence," and that, accordingly, "descriptive statements, judgments about the characters, about their expected development, and so on, have a kind of objective truth, even though they refer to fictions" (Husserl, 2005, p. 261).
This also has normative implications, connected to our being "forced" (aufgenötigt) to imagine something on the basis of spoken or written words (Husserl, 2005, p. 260). 12 Thus, the constitution of fictional objects, their identity, and their existence as cultural idealities cannot be traced back only to one's private or individual imaginative activity but should rather be seen as a social/collective enterprise. Yet, how should we understand the sociality of imagining in the experience of fiction?
An understanding of fictions as abstract artifacts and as paradigmatic for the class of entities studied in the field of social ontology seems eo ipso to imply the reference to social or collective experience. For instance, it seems that we need to share cultural memories or collective imagination if we are to be able to understand each other when we speak about fictional objects, their stories and their identity. While elaborating on this in connection with the structures of collective experience, however, one is faced with problems concerning the defining criteria for collective imagining. Do acts of collective imagining need to instantiate the same phenomenal content in order to refer to identical objects? Do imagining subjects need to be mutually aware of other subjects' imagining? According to Thomas Szanto (2018), instead of focusing on one exclusive model of collective imagination, we should rather develop a multi-dimensional approach. Such an approach needs to include: (i) the identity of the intentional object as reference (identity condition), 13 (ii) some form of mutual awareness of others participating in the imaginative activity (mutual awareness condition), and in part also (iii) some form of normative joint commitment in the imagining (normativity condition). In the remainder of this section, I wish to show how Sartre's account of the sociality of imagining responds to these three requirements, namely by taking the sociality of imagining not as a kind of sharing of intentions, but rather as a dialectical movement that produces objectivity.
According to Sartre, fictions are products of the imaginative activity of their authors, and in this sense fictional objects depend on such an activity. Imaginative acts are spontaneous-although not necessarily voluntary 14 -intentional acts simulating experience by combining affective and cognitive components. Such intentional acts notably have an irrealizing power: what is imagined is not taken as part of the real world, but rather negates reality (Sartre, 2004, pp. 5 f., 55 f.;cf. Summa, 2019a). However, such an irrealizing power of imaginative acts is not sufficient for the imaginary to become an 'object.' 'Objects' in the strict sense are socially 12 Among other things, these normative implications concern our responses to fiction. For a discussion of these normative implications, see Jansen (2018); Szanto (2018);and Summa (2019a). 13 Szanto speaks of "sameness" in this respect. Yet, in order to avoid possible misunderstandings, I rather speak of the identity of the referent or intentional object. In fact, as I will argue, this condition does not imply that different subjects instantiate the same content. 14 In short, they are spontaneous insofar as they originate within consciousness itself and are not caused by something else. Yet, they might not be the result of an explicitly voluntary act. Cf. Sartre (2004, pp. 14, 18 f.). constituted and for this reason a further movement of negation is required, which makes fictional entities not only the product of my individual acts of imagining, but also objects for others. In other words, Sartre assumes that the constitution of fictional entities should be traced back to the imaginative activity of both producers and recipients. Neither of them, considered alone, is sufficient for the constitution of ficta. Rather, both are required. Furthermore, on both sides, imaginers are aware that the fictional object is not the product of their individual act of imagining alone, but rather requires at least another consciousness in order to come to existence as an object (mutual awareness condition). Yet, how should we understand this mutual awareness and the social constitution of fictional objects?
In The Imaginary, Sartre observes how varied the family of the imaginative acts and their respective correlates is. This family includes such diverse phenomena as the experience of physical images, portraits, faces in the fire, spots on walls, rocks in human forms, hypnagogic images, and mental images. One should also remark that within this list mental images-or our individual phantasies-are at the very end of a series of increasingly immaterial objects or quasi-objects. In the case of mental images, matter disappears, and the imagistic representation is the pure product of the spontaneity of consciousness, in other words, of an act of irrealization (Sartre, 2004, pp. 17 f.). This does not mean that they simply do not exist, for irreality is considered to be the mode of existence of imaginary entities, be they simply imagined or fictional (Sartre, 2004, p. 63). However, since they exist only as correlates of irrealizing acts, mental images are not 'objects' in the strict sense. And this is so for the following reasons: first, correlates of singular imaginative acts are constitutively unstable, intermittent, and elusive; they have a protean character (Husserl, 2005, pp. 63 f.) and as such they do not comply with the identity criteria applying to objects. Secondly, as correlates of my own acts of imagining, they are only accessible to me, and one can even doubt whether they remain accessible to me at different times, whether I can re-identify them without eo ipso changing them or producing new imaginary entities.
For objects, and for fictional objects in particular, criteria of identification and public recognition are instead required. 15 It must be possible for us to be aware that we are referring to an identical object (identity condition). In this respect, while still being products of irrealization, fictional objects do not only stem from my spontaneous imagining; they are also objects for others. The centrality of the role of the other-the reader in the case of literature-is emphasized in What is Literature?, notably in the second essay. As a result, the role of the reader is here not limited to imaginatively concretizing what is transmitted in textual form and to filling spots of indeterminacy. This may well be part of what the reader does, responding to what Sartre calls the "appeal" of the literary work. But most importantly, while doing all this, readers address the product of someone else's spontaneous activity from a perspective that is not the one of the authors; they look at it from the outside, and this is necessary in order to have not only a product of spontaneity, but an object in the proper sense (Sartre, 1966, p. 40).
This view is consistent with the ontological claim according to which the existence of fictional entities is based on the above-mentioned dependence relations. However, there is yet another, and connected, dependence relation: the dependence on the constituting activity of both producer and reader. The "conjoint effort of author and reader" (Sartre, 1966, p. 43) is a form of object constitution, that is to say, the constitution of something publicly available, based on the negation of pure individual spontaneity. 16 Seen from the first-person perspective, the novel one writes is only the expression of one's own spontaneity. It can become an object only by means of the irrealizing activity of readers or recipients, who, with their own imagining, turn one's writings into something more than the product of one's private and spontaneous activity. They turn it into a social and public imaginary object. Accordingly, they somehow expropriate the imaginary, and in so doing co-constitute the fictional object as a social one. In this sense, Sartre argues that "[t]here is no art except for and by others" (Sartre, 1966, p. 43).
One can argue then that reading a work of fiction entails both moments of uncovering, analogous to what we have in perception, and creative moments, which are properly imaginative (Sartre, 1966, pp. 38 f.). Something is in fact given in the literary work: as readers, we first have to uncover meanings we do not produce ourselves but rather find before us; we need to understand words that were written by someone else and shape imaginary scenarios that we have not primarily produced. And this clearly has normative implications: recipients cannot imagine whatever they want or whatever spontaneously comes to their minds. On the contrary, their imagination is bound or constrained by what authors have projected (normativity condition). Yet, this activity of uncovering is supplemented by a creative activity, which also allows us to better qualify the identity condition. Concerning the identity of the intended object, this does not mean that the producer's and the recipient's act have the same phenomenal content. In fact, it is hard to claim that two individuals can have the same imaginary scenarios with the same phenomenal character in mind, and it even seems impossible to find criteria for sameness in this case. The identity of fictional entities is the identity of the intentional object. And this identity is socially constituted by the interaction of the producer's and the recipient's imaginative activity. The recipients, thus, follow the "project" sketched by the producer, but in doing so they at the same time "negate" it, by turning it into something they co-constitute. This is the movement that allows fictional entities, as objects in the public sphere, to exist and to be re-identified as social objects (Sartre, 1966, pp. 40 f.). For this reason, claiming that one only writes for oneself is somehow contradictory. Without an external look on one's own productions, one can in fact only find oneself or one's own spontaneity. As a dialectical movement, the sociality of imagining Sartre envisages is not just a matter of joint activity. Imaginative activities are always activities of negation and estrangement, and these are also part of the described processes: readers negate the spontaneity of writers and, conversely, writers also negate or limit the spontaneity of readers by prescribing what they are to imagine.

Conclusion: toward an integration
Thomasson explicitly criticizes one of Sartre's claims: that imaginary and fictional objects only last as long as the imaginative act is performed. In fact, Sartre does claim that "[i]f we turn ourselves away from them [imaginary objects], they are annihilated; […] they are totally inactive" (Sartre, 2004, p. 126, my emphasis). This claim is embedded in Sartre's overall account of imaginary objects as products of spontaneity, which in fact presupposes not only a constitutive analysis, but also a kind of metaphysics of actualization. However, if one considers the status of merely imagined or phantasized quasi-objects, Sartre's claim seems plausible. These, as we have seen, are merely correlates of private and mostly intermittent acts of imagining, and they should not even be considered as objects. They have no identity that reaches beyond the present act; they are rather ephemeral and intermittent products of our phantasies. One would rather expect a different account of the status of fictional objects in the strict sense-as objects that are not only products of spontaneous imagining, but products of social imaginative constitution. Yet, despite developing a social ontology of fiction, in What is literature? Sartre indeed makes a claim that is very similar to the one he makes in The Imaginary. Although Thomasson does not refer to this text, what it says-unfortunately-confirms her concerns. In the case of literature, with explicit reference to the acts of readers, Sartre concludes that a fictional object can only exist in a dialectical movement, like a spinning top: in order to make it appear "a concrete act called reading is necessary, and it [the literary object] lasts only as long as this act can last. Beyond that, there are only black marks on paper" (Sartre, 1966, pp. 40-41).
Here, the problem indicated by Thomasson comes to the fore. This problem concerns Sartre's assumption that the act of reading and correspondingly the act of imagining need to be constantly actualized in order for the fictional object to exist. One can say that the formulation he chooses is, to say the least, questionable. If we now speak about an episode in Don Quixote, without reading the book, we refer to an entity that 'exists' independently of the fact that we are now reading the book. And if someone attributes to Quixote properties that we know not to belong to him in the fiction, we would correct them, referring to what is written in the book. Should we then give up the claim that fictional objects are the product of imaginative activity? I argue that recognizing the validity of Thomasson's critique does not necessarily mean to deny that fictional objects are imaginary products. Even in the just outlined situation, in which we speak about a fictional character without actually reading the book, we can of course emptily refer to Quixote by appealing to our semantic memory, but this does not mean that we are involved in the fiction, or that we concretely re-present it. If the latter has to occur, then acts of imagining are required.

3
Thomasson's theory of dependence and her account of fictional entities as abstract artifacts can correct the imaginary-object theory, inasmuch as they warn us against any kind of actualism which risks ending up in an extreme form of phenomenalism concerning fictional entities. Conversely, however, the ontology of fiction can be fruitfully complemented by the inquiry into the constitution of fictional entities, and this requires an analysis of the different forms in which imagining-as individual and as social experience-underscores the intuitive givenness of fictional entities, as different from non-fictional social and cultural artifacts.
Funding Open Access funding enabled and organized by Projekt DEAL.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.