Is There a Part of Snow White Instantiated While the Story is Told? A Novel Artefactualist Theory of Fictitious Objects [Alberto Voltolini, How Ficta Follow Fiction. A Syncretistic Account of Fictional Entities, 2006.]

Tom Kindt

Maria Elisabeth Reicher

Is There a Part of Snow White Instantiated While the Story Is Told?

A Novel Artefactualist Theory of Fictitious Objects

Alberto Voltolini, How Ficta Follow Fiction. A Syncretistic Account of Fictional Entities. Dordrecht: Springer 2006. xxiv, 273 S. (Preis: EUR 106,95). ISBN: 978-1-4020-5146-3.

1. The Main Questions of the Book

There are two distinct, though interrelated, questions which Alberto Voltolini answers in this book. First, what kind of entities are fictional entities (if there are any)? Voltolini calls this »the metaphysical question«. The second question is whether there are fictional entities, which Voltolini calls »the ontological question« (127f.).

Fictional entities (ficta) are characters, things, places and events etc. of fictional works (fictions), e.g., Hamlet, the main character of Shakespeare’s play Hamlet, Pegasus, the flying horse of Greek mythology, or Leopold Bloom’s eating of inner organs of beasts and fowls in James Joyce’s Ulysses. As the given examples suggest, Voltolini’s focus is on literary fictions.

The first half of the book (Part I, entitled »The Metaphysical Side«, consisting of Chapters 1-4) is dedicated to the »metaphysical« question (»What kind of entities are fictional entities?«). Part II (»The Semantic Side«), consisting of Chapters 5 and 6, and Part III (»The Ontological Side«), consisting only of Chapter 7, deal with the »ontological question« of whether there are fictional entities.

2. Voltolini’s Theoretical Background

Voltolini develops his own theory out of a careful critical examination of a number of important alternative theories which he discusses in detail. These theories fall into two camps, the camps of »committal theories« and »noncommittal theories«. Committal (or realist) theories are those according to which there are fictional entities. Under the heading »Committal Theories«, Voltolini discusses »Meinongian theories« (Chapter 1, Section 2), »the possibilist conception« (Chapter 1, Section 3), »the Neo-Meinongian abstractionist conception« (Chapter 1, Section 4), and »the artefactualist abstractionist conception« (Chapter 2). Noncommittal (or antirealist) theories are those »that attempt to dispense with fictional objects by following a semantic path, that is, by maintaining that the best truthconditional account of sentences apparently about ficta does not involve such entities« (127). Voltolini ponders the noncommittal theories of Frege and Russell (Chapter 5, Sections 3 and 4) and »the intensionalist theories« (Chapter 5, Sections 5.1 and 5.2). I shall now consider one by one Voltolini’s presentation and discussion of both the committal and the noncommittal theories which together form his theoretical background.

2.1. »Meinongian Theories«

The term »Meinongian theories« of course refers to Alexius Meinong, but what is called a »Meinongian theory« in this book is not necessarily something that Meinong had ever held in reality (as Voltolini himself admits). One of the main distinguishing features of »Meinongian theories« is the assumption that there are objects that do not exist (and have no kind of being whatsoever). They only have what Meinong calls »Außersein«. According to »Meinongian theories«, fictitious objects are a kind of nonexistent, yet (usually) concrete (in contrast to abstract) objects. That is to say that fictitious objects literally have all the properties they have »in« their respective stories. Pegasus, for instance, literally is a horse-like animal with wings.

Voltolini distinguishes two kinds of »Meinongian theories«: a phenomenological Meinongian theory and a Platonist Meinongian theory. According to the phenomenological version, nonexistent objects are created by intentional acts (thoughts, imaginations, dreams etc.). Thus, for instance, if a person imagines a flying horse, she thereby creates a flying horse (though, of course, a nonexistent one). According to the Platonist version, by contrast, nonexistent objects are not created at all but rather have their Außersein necessarily and thus independently of any intentional acts. According to this theory, to every description whatsoever, there is necessarily an object that fulfils this description, if not an existent then a nonexistent one. Thus, there is, for instance, a golden mountain, a round square, a golden round square, etc. – irrespective of whether anybody has ever thought of or imagined such things.

Voltolini objects to the Platonist version of Meinongianism that it does not provide us with an explanation of the particular status of fictitious objects among the huge realm of nonexistents. Whether there really are as many nonexistent objects as the Platonist Meinongian thinks is a question that does not need to be settled within the context of a theory of ficta. However, for a Platonist Meinongian theory of fictitious objects to be adequate, the Platonist Meinongian must provide an explanation of what distinguishes an »ordinary« nonexistent object from a fictum.

The phenomenological version of Meinongianism runs into a different, though structurally similar, problem. According to this theory, every intentional act whatsoever (thought, imagination, dream etc.) yields a corresponding intentional object. Yet not every intentional object is a fictitious object. There is a difference between »ordinary« intentional objects (those that arise »automatically« whenever something is thought of or imagined) and fictitious objects in the sense explicated above. This is actually one of the recurrent themes in Voltolini’s book: Ficta are not just intentional objects; they are not brought into the world through arbitrary intentional acts (at least not through such acts alone).

2.2. »The Possibilist Conception«

According to the possibilist conception, fictitious objects do not exist in the actual world but in other, non-actual, merely possible worlds, i.e., they are possibilia. Voltolini objects to this that a fictitious object actually has those properties that are ascribed to it in the story (though in a particular sense which will be explicated below), whereas a merely possible object does not have these properties actually but merely possibly. Moreover (this is a standard argument against the possibilist conception of fictitious entities), fictional entities may have contradictory properties and thus may be impossible objects.

2.3. »The Neo-Meinongian Abstractionist Conception«

According to this theory, Meinongian objects in general (such as the golden mountain and the round square) and fictitious objects in particular (such as Hamlet and Pegasus) are abstract objects and as such they exist (in the actual world).

Voltolini distinguishes two versions of Neo-Meinongian abstractionist theories. According to the first one, Meinongian objects are sets of properties (or »correlates« of such sets). According to this theory, Pegasus, for instance, is just the set of all the properties Pegasus has in the Pegasus story in Greek mythology, i.e., being a horse-like animal, having wings etc. According to the second version of Neo-Meinongianism, fictitious objects are universals, Platonist types, i.e., entities that may be instantiated in (concrete) individuals. According to this theory, one might say that Pegasus would be the type winged horse and thus it could be instantiated in »real« winged horses (if there were any), just as the type rectangular triangle can be instantiated in all concrete rectangular triangles. (This way of putting things is, of course, a bit simplifying, because Pegasus has more properties in his story than just being a horse and being winged; therefore, if Pegasus is a type, it is a more complex one than the type winged horse.) (cf. 16f.)

One important feature of Neo-Meinongian theories of all kinds is that they acknowledge that, with respect to Meinongian objects (and thus also with respect to fictitious objects), one has to distinguish two kinds of predication. Consider, for instance, the following two sentences: (1) Pegasus is a winged horse. (2) Pegasus is a character of Greek mythology. (1) is, as Voltolini puts it, a case of internal predication, (2) is a case of external predication. An advocate of the property set version of Neo-Meinongianism may explain the distinction as follows: A predication is internal if, and only if, the property ascribed to the object in question is among those properties that are members of the property set that constitute the object. Consequently, a predication is external if, and only if, the property ascribed to the object in question is not among the properties in the set that constitute the object. With respect to fictitious objects, one might explain the distinction as follows: A predication is internal if, and only if, the property ascribed to the object is among the properties which the object has in the story. Consequently, a predication is external if, and only if, the property ascribed to the object is not among the properties which the object has in the story.

The mode of predication distinction is an indispensable feature of Neo-Meinongian theories. It is the remedy against inconsistencies which otherwise would infest these theories, and in its application to fictitious objects it has an obvious intuitive appeal.

Voltolini raises two objections to the Neo-Meinongian abstractionist conception. He calls them »the many-ficta problem« and »the no-ficta problem«, respectively (31-36). The latter (the no-ficta problem) has already been raised to »Meinongian theories«. Neo-Meinongian theories have difficulties to account for the difference between fictitious objects and »ordinary« Meinongian objects.

Though this problem recurs several times in Voltolini’s book and seems to be among the most important driving forces for the development of his own theory, it seems that Voltolini takes the many-ficta problem to be even the more serious one of the two. In Voltolini’s lights, there may be distinct fictitious objects that have exactly the same (internal) properties. To make this claim plausible, Voltolini makes use of the well-known »Menard case«. This prominent example goes back to a (fictional) short story by the Argentinean writer Jorge Luis Borges. It is the story of a man, Menard (a French intellectual of the 20th century), who sets out to re-write Cervantes’ Don Quixote. However, he does not produce a new text (a »modernized« version of the original or a sequel) but is writing down the original text word by word. What has struck (and obviously still strikes) many theorists of literature as a deep insight is the claim (made by the narrator of Borges’ story) that Menard thereby has produced a new work, simply because the same utterances made in a different cultural-historical context gain new connotations.

However, Voltolini goes even a step farther. He »idealizes« the case in that he construes Menard not as a 20th century intellectual but as a contemporary of Cervantes, even as a neighbour of him, i.e., as somebody who shares Cervantes’ cultural-historical background. In Voltolini’s case (which I shall call in what follows »the idealized Menard case«), Menard writes down exactly the sentences of Cervantes’ Don Quixote, but without knowing anything about Cervantes’ work (32f.). Thus, it is a case, where, by sheer coincidence, two distinct authors in the same cultural-historical context produce the same text (assumingly with the same communicative intentions, due to their common background).

Voltolini claims that even in such a case the work of the one author is not identical with the work of the other (228-241, especially 235), a claim that plays a crucial role in his »ontological argument« in favour of fictitious objects put forward in the last chapter. Closely related to this, he claims that, in the idealized Menard case, the character Quixote of Menard’s story is not identical with the character Quixote of Cervantes’ story, although, by hypothesis, Menard’s Quixote and Cervantes’ Quixote share all of their internal properties (33). Therefore, Voltolini claims, a character’s identity cannot be defined by reference to its internal properties alone. Therefore, Neo-Meinongian theories fail to provide adequate identity criteria for fictional characters. This is the »many-ficta problem«.

2.4. »The Artefactualist Abstractionist Conception«

According to the artefactual theory (as I shall call it briefly), fictitious objects are abstract objects which have been created by their authors. As such, they are artefacts that exist in the actual world.

Voltolini discusses exclusively Amie Thomasson’s version of the artefactual theory. Therefore, a number of his objections are not really objections to the artefactual conception per se (although the reader might sometimes get the contrary impression) but rather to Thomasson’s variety of it.

According to Thomasson’s version of the artefactual theory, fictitious objects are a particular kind of intentional objects. That is to say, fictitious objects come into existence through intentional acts (thoughts, imaginations, hallucinations, dreams etc.). At this point, however, Voltolini raises an objection already raised to the phenomenological version of »Meinongian theories«: Even if we accept that fictitious objects come into existence through intentional acts (and thus are a kind of intentional objects), it is plain that not every intentional object (i.e., the objectual correlate of an arbitrary intentional act) is a fictitious object. An objectual correlate of an arbitrary intentional act cannot rightly be called an »artefact«. Thus, Thomasson needs to give an account for the »artefactuality« of fictitious objects (49-52).

As a matter of fact, Thomasson tries to give such an account. Whereas »purely intentional objects« (i.e., objectual correlates of intentional acts that are not fictitious objects) perish as soon as their correlated intentional acts have come to an end, fictitious objects are less ephemeral entities. They may go on existing long after the intentional acts which originally brought them into existence have perished – provided that there are literary works in which these objects occur. Thus, one might think »that what gives a fictum its artefactual character is not its origin but rather its protraction – its life in some work or other; in Thomasson’s terms, its constant generic dependence on literary works« (52).

Voltolini, however, thinks that this account does not work:

On the one hand it seems that, if a fictum is an artifact, it must be such from its very beginning; hence, whatever accounts for its protraction does not account for its artefactuality. On the other hand, it seems that a fictum cannot begin its life as an artifact as we have seen that if a fictum originates in the same way as a purely intentional object, at its beginning it cannot be an artifact. The question therefore remains: if ficta are artifacts whereas pure intentionalia are not, how is this to be explained? (53)

I turn now to the »noncommittal theories«.

2.5. Frege’s Theory

For Frege, fictional names are just special cases of empty names, i.e., names that lack reference (though they may have sense). Voltolini points out that Frege’s theory is fine for what he calls »the conniving use« of fictional names but that it does not tell us anything about the »nonconniving« use. The conniving use of names and sentences is its use in »fictional discourse« (to use a familiar term). Fictional discourse occurs whenever somebody is telling a fictional story (irrespective of whether the story is created by that particular speech acts or whether the respective speech acts constitute a »re-telling« of the story in question). The distinguishing feature of fictional discourse is that the speaker does not claim truth for what she says, even if she utters assertive sentences. In Voltolini’s words, conniving use is the »use of a fictional sentence […] in order to say make-believedly that something is the case« (xxi). Fictional sentences are »sentences occurring in fictional texts […] that may be used both connivingly and nonconnivingly« (xxii). Nonconniving use is the »use of a fictional sentence in order to say that something is the case« (xxiii). What Voltolini calls »the nonconniving use of fictional sentences« is part of what is often called »discourse about fiction«. Discourse about fiction occurs whenever somebody makes serious (i.e., not make-believe) assertions about fictitious objects, for instance in the context of an examination in literary history. Since Frege’s theory concerns exclusively the conniving use of fictional names, it is perfectly consistent with a committal theory of fictitious objects.

2.6. Russell’s Theory

The theory in question is, of course, Russell’s theory of definite descriptions, together with Russell’s description theory of proper names. According to the latter, every proper name is a disguised definite description (i.e., a locution of the form »the F«). According to the former, subject-predicate sentences containing definite descriptions in subject position are to be analyzed in such a way that the analysans is an existential quantification that contains only general terms plus an existential quantifier, variables and a »uniqueness condition« instead of the definite description. Thus, to use Russell’s famous example, the sentence »The present king of France is bald« is to be analyzed as »There is exactly one x, such that x is a present king of France and x is bald«.

Somebody who does not wish to commit himself to fictitious entities might make use of Russell’s theory in the following way: Take, for instance, the sentence (1) Pegasus was ridden by Bellerophon. Prima facie, this seems to be a true sentence. Thus, given that the truth of this sentence entails the existence of its subject, it threatens to commit one to the assumption that Pegasus exists. According to Russell’s theory of proper names, »Pegasus« is a disguised definite description, say, »the winged horse«. Thus, (1) is to be analysed as (1') The winged horse was ridden by Bellerophon. To this, in turn, one may apply Russell’s theory of definite descriptions, and thus one gets: (1'') There is exactly one x such that x is a winged horse and x was ridden by Bellerophon. Clearly, however, (1'') is a false sentence, since the first part of the conjunction is false (there is no such thing as a winged horse in the world). Therefore, an analysis along Russellian lines shows us that (contrary to the first impression) sentence (1) is not to be accepted as true; and therefore, there is no need to commit oneself to the existence of Pegasus. Given that this method of analysis can be applied to all sentences that are allegedly »about« fictitious objects, such sentences do not force us to commit ourselves to the existence of fictitious entities.

Voltolini is, of course, aware of the fact that one might question Russell’s theory for several reasons, but he argues that one does not need to do so in order to defend a committal theory of fictitious objects. As he points out, Russell’s theory, just as Frege’s, is compatible with a committal theory of fictitious objects. All depends on exactly what definite description is taken as the sense of a given fictitious name. The sort of definite descriptions Voltolini considers to be the senses of fictional names involve sets of properties and make-believe processes (139). (I shall come to this shortly.) But to get the idea one just needs to consider, for instance, the definite description »the character from Greek mythology that is a winged horse« instead of »the winged horse«. If one replaces »Pegasus« by this definite description, then Russell’s analysis yields the following: (1''') There is exactly one x such that x is a character from Greek mythology and x is a winged horse and x was ridden by Bellerophon. In contrast to (1''), this sentence seems to be intuitively true. (The predicates in (1''') are to be understood in the internal sense, of course.) Thus, not even acceptance of Russell’s theory of proper names together with Russell’s theory of definite descriptions forces one to accept a noncommittal theory of fictitious objects.

2.7. »The Intentionalist Theories«

What Voltolini calls »intensionalist theories« are theories which make use of story operators. A story operator is a locution of the form »In the story S« or »According to the story S«. By means of such story operators, fictional sentences nonconnivingly used or »parafictional sentences« are paraphrased in such a way that the ontological commitment to fictitious objects (allegedly entailed by the nonconnivingly used or parafictional sentences) seemingly can be avoided. A »parafictional sentence« is a sentence that does not, literally, occur within a story (and thus is not a fictional sentence) but is equivalent to a fictional sentence (such sentences are explicit parafictional sentences) or a sentence that is entailed by an explicit parafictional sentence (such sentences are implicit parafictional sentences) (xxiii). Given that nowhere in Shakespeare’s play Hamlet the sentence »Hamlet is a prince« occurs, »Hamlet is a prince« is not a fictional but a parafictional sentence (probably an implicit parafictional sentence). Simply put, a person who describes the characters and events of a fictional story in her own words utters parafictional sentences.

The parafictional sentence »Hamlet is a prince« seems to entail an ontological commitment to Hamlet. However, if we put a story operator in front of it, we get: »According to Hamlet, Hamlet is a prince«, which does not seem to entail an ontological commitment to Hamlet. Thus, one might think that the use of story operators, in general, is a means to avoid the ontological commitment to fictitious entities.

The standard objection to this line of reasoning is that story operator paraphrase strategies are applicable only to parafictional sentences and fictional sentences nonconnivingly used, but not to external metafictional sentences like »Pegasus is a mythical character«. For if we put a story operator in front of the latter, the resulting sentences have the wrong truth-value. While »Pegasus is a mythical character« is obviously true, »According to Greek mythology, Pegasus is a mythical character« is false, since, according to Greek mythology, Pegasus is a being of flesh and blood, as real and concrete as can be. But if intensionalist theories fail to provide adequate paraphrases for external metafictional sentences, the whole eliminative project is bound to fail (145f.).

Voltolini, however, is dissatisfied with this line of reasoning (although he holds the conclusion, i.e., that eliminativism fails, to be true):

Although, as a realist on ficta, I obviously believe that external metafictional sentences commit us to such entities, I must confess that I find this line of reply disconcerting. It is not only because, if the problem with external metafictional sentences were simply that intensionalist paraphrases do not work in their case, it might well be true that other paraphrases would […]. […] But also and above all, it would be curious if our ontological inventory allowed for fictional entities only because we seemingly fail to account noncommittally for a rather limited and marginal number of sentences, the external metafictional sentences. Indeed, the sentences in which we seem commonly to discuss ficta are the parafictional sentences. Not only are they greater in number than the external metafictional sentences; they are also more important. For in them we discuss the properties that are assumed to characterize fictional entities: Holmes as a clever detective, Othello as a jealous man, Roland as an insane paladin. (146)

Therefore, Voltolini tries to show that intensionalist theories do not even work for fictional and parafictional sentences (see section 5.2). Among other things, problems seem to arise with sentences that contain indexicals. Consider, for instance, the sentence (1) For a long time I used to go to bed early (the first sentence of Proust’s In Search of Lost Time). According to intensionalist theories, we should expect that this sentence, nonconnivingly used, says the same as (1') In In Search of Lost Time, for a long time I used to go to bed early (166f.). Voltolini comments on this:

[T]his equivalence clearly does not work. For, supposing that [1'] is uttered by me, what it says is that the real utterer of »I« in [1'], namely myself, is such that in the imaginary »world« of Proust’s Recherche he used to go to bed early for a long time. This definitely is false, for such a »world,« whatever it is, definitely does not contain me (as having certain properties there). But we would expect that if [1] had a real truth value, this would be the True, not the False. (167)

3. Critical Assessment of Voltolini’s Discussion of Alternative Theories

All in all, Voltolini provides a quite comprehensive and well-organized overview of theories of fictitious entities, which makes his book not only a worthwhile reading for theorists specialized in this rather particular field but also a good introduction. He takes into account an impressive amount of literature. To reproach him for nevertheless having disregarded some interesting authors and works would be unfair, given the plethora of relevant literature on this subject. Also, that the choice of particular authors and works discussed is partly a matter of personal taste and other contingencies is quite natural and an author’s right. However, it is an unnecessary flaw that the discussion of the »artefactualist conception« is focused exclusively on Amie Thomasson’s theory, for this gives a distorted picture of this conception to the reader and it prompts Voltolini to use quite some space for the discussion of idiosyncratic difficulties of Thomasson’s theory which are not really essential for the artefactualist conception per se, at the expense of other important artefactualists, like, for instance, Roman Ingarden and Peter van Inwagen.

As far as the details of Voltolini’s discussion of alternative (or, in some cases, rather complementing) theories are concerned, there is hardly any fault to find with it. I shall confine myself to two small comments on Voltolini’s discussion of »intensionalist theories«. First, I do not agree with Voltolini’s claim that reference to (external) metafictional sentences is a weak basis for realism with respect to fictitious objects. Voltolini argues that »it would be curious if our ontological inventory allowed for fictional entities only because we seemingly fail to account noncommittally for a rather limited and marginal number of sentences, the external metafictional sentences« (146; for the context of this quotation see the end of the former section). It may be that parafictional sentences outweigh external metafictional sentences in number, but this fact (if it is a fact) does not weaken arguments that rely on metafictional sentences.

Second, Voltolini’s above-quoted argument to the conclusion that story operator strategies do not even work for parafictional sentences strikes me as inconclusive. Remember that Voltolini argues that the story operator strategy does not work for the sentence »For a long time, I used to go to bed early« (the first sentence of In Search of Lost Time). For, as Voltolini points out, if we used this sentence nonconnivingly, according to the intensionalist it would have to be read as »In In Search of Lost Time, for a long time I used to go to bed early«, which obviously yields the wrong truth value. Of course, the source of the problem is that the sentence contains the indexical »I«. So far I agree, but I think the reason for this is simply that this sentence (and other sentences containing indexicals) cannot be used nonconnivingly. Giving a description of the contents of In Search of Lost Time, I simply cannot say that, for a long time, I used to go to bed early. If I want to convey the beginning, I have to say something like: »The narrator tells us that, for a long time, he used to go to bed early.«

4. Voltolini’s Theory

Voltolini dubs his own theory »the syncretistic theory«, since, as he says, it »firmly acknowledges that the various other theories already developed on this subject have great merits« and since it »[integrates these other theories] into a single theory that aims both to maintain their positive results and to overcome their defects« (xiii).

Voltolini sees his theory as »syncretistic« in a twofold manner: He strives to combine, on the one hand, committal and noncommittal theories and, on the other hand, Platonist and artefactualist theories.

The reader might wonder how one can integrate these obviously incompatible theories into a single theory without inconsistency. In fact, Voltolini’s »diplomatic« characterization is a bit distorting. First, as he himself makes clear at the very outset, his theory is committal. According to the »syncretistic theory«, ficta are actually existing abstract objects. The noncommittal element comes in only insofar as Voltolini acknowledges that there are noncommittal uses of fictional names and sentences, namely the uses of fictional discourse. This, however, seems to be fairly undisputed even among advocates of those kinds of committal theories Voltolini discusses in Part I of the book (with the exception of Amie Thomasson). Second, as far as Voltolini’s alleged »syncretism« with respect to Platonist and artefactualist theories is concerned, Voltolini’s theory is clearly artefactualist. For Voltolini, fictitious objects are contingent beings that are created by the authors of fictional stories through acts of storytelling and thus come into being at a certain moment in history. The Platonist element of Voltolini’s theory consists in the assumption that fictitious objects contain sets of properties as parts. In Voltolini’s words,

a fictum is a compound entity composed of a pretense-theoretical and a set-theoretical element. These elements are, on the one hand, the make-believe process-type in which it is pretended that there is a (typically concrete) individual that has certain properties and, on the other, a set of properties. (65)

As shall be seen shortly, however, this is not yet the full picture. But since the additional element of the full picture is motivated by a particular difficulty that Voltolini sees with this provisional characterization, I shall stick to the latter one for a moment. Voltolini continues the outline of his theory as follows:

[T]he syncretistic theory accepts the Neo-Meinongians’ claim that a certain property set yields a necessary condition for the identity of a fictional entity. This is the set of the properties corresponding to the properties that are directly or indirectly mobilized in a certain process of make-believe, that is, the process of storytelling in which one makes believe that a certain, typically concrete, individual explicitly or implicitly possesses precisely the properties in question. (66)

But identity of constituting property sets is not a sufficient condition for the identity of fictitious objects. That is, there may be distinct characters that share exactly the same internal properties. Furthermore, the existence of a particular set of properties is a necessary but not a sufficient condition for the existence of a fictitious object that has this set as one of its constituents. In addition, it needs a storytelling [=make-believe] process type which must be instantiated. Therefore, storytellers can rightly be considered to be creators of fictitious objects. Storytelling processes involve mental acts but are not reducible to these (69-72). It should be emphasized that the storytelling process type is not just a necessary condition for a fictitious object’s coming into being, but literally a constituent of the fictitious object (75).

Another crucial claim of Voltolini’s theory is the following: The existence of a (instantiated) storytelling process type is a necessary but not a sufficient condition for the existence of a corresponding fictitious object. Assume that a storyteller tells a certain story that has never been told by anyone before, and immediately after the end of the storytelling process passes away. By Voltolini’s lights, in such a case the storyteller has not created a fictitious object! This contention is closely related, once again, to Voltolini’s conviction that fictitious objects have to be distinguished from mere intentional correlates of intentional acts (75-78). Consider the following passages that throw some light on this line of thought:

As I remarked in the previous chapter, if the participant(s) in one such [storytelling] process ceased to exist as soon as that process came to an end, so that no trace of the process remained, no fictional entity would emerge from it. This does not depend on the fact that such a process is private for a make-believe game may well be (in fact, typically is) intersubjective. Rather, it depends on the fact that pretending is just pretending. That is, pretending is an activity in which one may make believe that there are plenty of (typically concrete) individuals. Yet pretending that there are such individuals does not mean that there really are such things, nor that there really are abstract entities related to those individuals in some way. Therefore, by merely pretending that there is a (typically concrete) individual, no fictional individual comes into being. (76)

In this respect, an existentially creative make-believe process is like a dream. In an existentially creative dream, one imagines that there are many concrete individuals; yet these individuals have no being at all outside the context of the dream itself. Over and above the oneiric fantasy that there are such concrete individuals, there are no further phantasmic entities that the dream is concerned with. That is, dreams have no magical power to generate phantasmic entities. Likewise, neither do make-believe processes have any magical power to generate fictional entities. Admittedly, dreams are intrasubjective processes whereas make-believe games are intersubjective but, if one wishes to do so, it [is] possible to treat them as collective dreams. (77f.)

If the existence of a certain make-believe process type is not a sufficient condition for the existence of a corresponding fictitious object, a problem arises for Voltolini’s theory: Since the existence of a certain property set is not a sufficient condition for a corresponding fictitious object either, it seems that neither the existence of a property set and the existence of an instantiated storytelling process type taken together can be a sufficient condition for a fictitious object’s existence (87f.). Since it is one of Voltolini’s main criticisms of alternative theories that they do not provide appropriate existence conditions for fictitious objects, this is a problem that he has to overcome.

Therefore, he postulates an additional condition, namely that »the [storytelling] process is in fact to be taken as pertaining to that [property] set« (88). This locution occurs several times (see, for instance, also 86) but it is nowhere defined. An alternative formulation for the alleged relationship between property set and storytelling process is contained in the following:

[A] fictional individual is a compound entity consisting of both a make-believe process-type and the set of properties corresponding to the properties mobilized in that process, as a result of seeing that process-type as regarding that set. (89) [My italics.]

Let me summarize some further interesting and important features of Voltolini’s theory:

As many theorists before him (both Neo-Meinongians and »artefactualists«), Voltolini uses the distinction between different modes of predication to explain the so-called »incompleteness« of fictitious objects. »Incompleteness« arises out of the fact that authors do not and cannot determine for every property whether a given character has or lacks this property within the story. The modes of predication distinction allows giving an account of this particular feature of fictitious objects that is compatible with the principle of excluded middle.
Although fictitious objects come into being at a certain moment in history, once they have been created, they cannot cease to exist. (See 92f. but in particular note 32 on p. 236.)
In a series of stories, a character of episode 1 cannot be identical with a character of episode 2 or any other episode. This is ruled out by Voltolini’s assertion that property set identity is a necessary condition for character identity. However, Voltolini admits that sometimes we refer to a character without having a particular episode in mind. We may refer, for instance, to Sherlock Holmes, without having any particular Holmes story in mind. In this case, Voltolini argues, we refer to a »general character«, where the »general character« Holmes

is larger than the one generated by virtue of the initial stage of the relevant make-believe process, and also larger than any of these particular Holmeses. As I just said, this general Holmes consists of the protracted make-believe process-type occurring throughout the storytelling of the whole cycle of the Holmes stories, together with the set of all the properties corresponding to those invoked in that protracted process. (110)
There are no »real« immigrant objects in fictions, i.e., the Napoleon in War and Peace is not identical with the real Napoleon, the London of the Holmes stories is not the real London etc. (101)
Fictitious objects are constituents of stories. Stories are sets of states of affairs. At this point, it should be mentioned that Voltolini mostly uses the term »proposition« instead of »state of affairs«. It is plain, however, that what he has in mind are not Fregean »thoughts« (which are nowadays usually called »propositions«) but rather the truthmakers of Fregean thoughts/propositions. (See 191-194, in particular note 107 on p. 194.) Therefore, his use of »proposition« is potentially misleading. For this reason, I deviate throughout this whole review from Voltolini’s terminology in this respect and use »states of affairs« instead of »propositions«.

For instance, the story of Shakespeare’s play Hamlet contains, among others, the state of affairs that Hamlet is (internally) a prince (195). It must be emphasized that the states of affairs that constitute a fictional story are of the form that a is internally F, as Voltolini makes clear on 212: »If ›the story S‹ designates a propositional set, then the fictional individuals existing ›in‹ this set possess the properties figuring in the propositions of this set internally, not externally.« At any rate, the state of affairs that Hamlet is (internally) a prince contains Hamlet, the fictitious character, as a constituent.

The claim that fictional stories consist of states of affairs and that these states of affairs consist (among other things) of fictional objects is important for what Voltolini calls »the ontological argument in favour of fictional objects«:

It is quite a simple argument. If we admit a certain kind of entity, we cannot but admit all the other kinds of entities that figure in the identity conditions of such an entity. We admit fictional works; so we cannot but also admit fictional objects because they figure in the identity conditions of fictional works. (241)

I shall come back to this argument at the end of the next section.

5. Critical Assessment of Voltolini’s Theory

I agree with most of what is said in this book. In particular, I have no objections against Voltolini’s realism with respect to fictitious objects. I also wholeheartedly agree with the abstract artefactuality thesis, i.e., the claim that fictitious objects are abstract, yet contingent entities that come to existence at a certain moment in history and are created through particular human acts. [1] Thus, the following critical remarks concern rather some details than central tenets of the theory.

1. The most innovative aspect of Voltolini’s theory is his proposal to consider fictitious objects as composita consisting of storytelling process types and sets of properties. This idea is prima facie attractive because it provides an account of the »artefactuality« of fictitious entities as well as a definition for internal predication: »[I]nternal predication is just set-membership: a property is possessed internally by a fictum iff it belongs to the property set that constitutes that fictum« (31). Especially the latter is an advantage of Voltolini’s theory.

Nevertheless, there are reasons for reservations: First, is the particular double structure that ficta have according to Voltolini’s theory (i.e., property set plus process type) really intuitively plausible? It seems to me that it is not. As far as I can see, it is nowhere reflected in our ordinary thinking and speaking about fictitious characters; rather to the contrary. It is of course not the complexity per se that makes Voltolini’s ficta bewildering, but rather the fact that the alleged constituents are so different. We may think of fictitious characters as quite complex entities, but do we think of them as something that consists of entities that belong to different categories? This is doubtful. By contrast, it seems natural to consider fictitious characters as types (of persons, animals etc.). This seems to be something that is implicit in our ordinary thinking about fictitious characters.

The following may highlight the point of counterintuitiveness: It seems that one can (truthfully) apply certain predicates to a storytelling process type which one cannot (truthfully) apply to a fictitious character. For instance, one might (truthfully) say of a storytelling process type that an instantiation of it takes at least three hours. To say of a character that an instantiation of it takes at least three hours, however, is at best wrong and at worst nonsensical. Of course, in principle, a part of a composite whole may have properties that the whole lacks. But it seems also odd to say that a part of a character is such that its instantiation takes at least three hours.

Moreover, storytelling process types may be instantiated. Thus, according to Voltolini’s theory, if, for instance, the Pegasus story is told, a part of Pegasus, the mythical character, is instantiated. This, again, seems odd to say. It is not odd to say, of course, that Pegasus, the mythical character, can be instantiated. It is just odd to say that storytelling processes are instantiations of (parts of) Pegasus. A (wingless) horse seems to be a better candidate for an instantiation of a part of Pegasus than a storytelling process.

As has been said in the previous paragraph, characters are instantiable. But, intuitively, they are instantiable as a whole (unless they have contradictory internal properties). According to Voltolini’s theory, only parts of characters (namely the process parts) are instantiable (sets are not instantiable).

2. In comparison to other metaphysical theories of fictitious objects, Voltolini’s theory is ontologically abundant. As it stands, the theory entails a commitment to properties, sets and types – although at one point, Voltolini seems to try to circumvent the commitment to types by claiming that one could interpret storytelling process types as sets of storytelling process tokens (see note 22 on p. 75). However, the claim that fictitious objects are property sets plus storytelling process types is obviously not equivalent with the claim that fictitious objects are property sets plus sets of storytelling process tokens. Thus, a theory that implies the former claim is substantially different from one that implies the latter; i.e., the difference is not just a terminological one.

As a matter of fact, a »nominalistic« version of Voltolini’s theory (i.e., one which assumes sets of storytelling process tokens instead of storytelling process types) would run into difficulties of its own. Suppose a particular story S, »home« of a particular character C, is not told anywhere in the world during a particular period of time t1–tn. Given that the elements of a set must be actually existing entities, the set of storytelling processes which is a constituent of C would be the empty set during t1–tn. But the empty set is clearly distinct from any non-empty set. Moreover, a set that has one element is distinct from a set that has two elements etc. Consequently, according to a »nominalistic« version of Voltolini’s theory, whenever a storytelling process were terminated or a new storytelling process were started, the nature of the characters of the respective story would be substantially altered – which seems to be a weird consequence.

Furthermore, the »nominalistic« version of the theory would raise a problem for Voltolini’s identity conditions for fictitious objects. As was said above, Voltolini holds that there may be distinct fictitious objects that share exactly the same internal properties. Accordingly, in terms of Voltolini’s theory, sameness of property sets is not a sufficient condition for character identity. For instance, in the idealized Menard case, Cervantes’ Don Quixote and Menard’s Don Quixote are distinct, although they share the same property set, because of their distinct storytelling process constituents. But if neither Cervantes’ story nor Menard’s story were told, both Cervantes’ Quixote and Menard’s Quixote would contain the empty set as second constituent (in addition to the property set, which is, ex hypothesi, also the same in both of them). How, then, could Voltolini account for the alleged difference of Cervantes’ Quixote and Menard’s Quixote?

Voltolini could avoid these difficulties by stating that the set of storytelling processes does not contain just the actually occurring storytelling process tokens but the storytelling process tokens of all times (i.e., present, past and future ones). In this case, however, his theory would imply either that past and future objects exist or that sets may contain entities that do not exist. Since Voltolini does not address this problem, it is unclear whether he would be willing to accept any of these consequences. In my opinion, he is better off with a commitment to storytelling process types.

3. A third potentially problematic aspect of the theory concerns the postulated link between the property set and the storytelling process. As was said above, Voltolini states that a fictitious object cannot be brought into existence merely by a storytelling process, since, as he sees it, a storytelling process is on a par with a dream or another series of intentional acts of some kind; and it is one of his main concerns to distinguish sharply between fictitious objects and mere objectual correlates of intentional states. (See the passage from p. 77f. quoted above.) Therefore, an additional condition for the creation of a fictum must be fulfilled: the storytelling process must »be taken as pertaining to the property set« in question; or, alternatively, the process type must »be seen as regarding that set«. (See 88f.)

Neither of these formulations is very clear. But it seems that Voltolini has in mind a particular intentional act over and above the intentional acts that constitute the storytelling process itself. Voltolini states at several places that this particular act can take place only after the storytelling process is finished. Unfortunately, the particular nature of this intentional act remains mysterious.

I share with Voltolini the conviction that not every act of imagination yields fictitious objects. I agree that one has to distinguish sharply between fictitious objects and arbitrary objectual correlates of intentional acts. It seems to me, however, that he is mistaken in thinking that processes of storytelling (or the respective intentional aspects of such processes) are on a par with dreaming or other processes of imagination. To the contrary, making up a story is an intentional process of a particular kind. What distinguishes it from dreaming etc. is the aspect of conscious decision. The author determines the traits of her characters, the outcome of the events etc. The dreamer is the (more or less) passive subject of a flow of imagination. The author actively takes decisions – usually with a communicative intention. Accordingly, it is the particular nature of the storytelling process itself that accounts for the difference between mere intentionalia on the one hand (if there were such things) and fictitious objects on the other. There is no need to postulate a mysterious additional act that takes place after the storytelling itself is finished.

4. Voltolini’s conception of a general character as the sum of the particular characters of various episodes of a series (e.g., the general Holmes character as the sum of all the »Holmeses« of the particular Holmes stories) is adequate for some cases of transfictional reference, but not for all of them. By »transfictional reference« I mean the reference to a character that occurs in more than one story without (implicit or explicit) reference to a particular story. Voltolini’s conception works well in certain cases of reference to serialized novels and the like, i.e., in cases where we have episodes that may be considered as parts of one large continuous story. However, it is not adequate for cases where (as we use to say) one and the same character occurs in distinct stories that are not episodes of one large story but rather different versions or variations of one and the same story. Paradigm instances of such characters are the Faust character and the Don Juan character.

Of course, sometimes we refer, for instance, to Goethe’s Faust and to Christopher Marlowe’s Doctor Faustus, and it is clear that the former is not identical with the latter. However, in some cases, we wish to refer to the Faust character, which somehow occurs both in Goethe’s and in Marlowe’s Faust story, yet is not strictly identical either with Goethe’s Faust or with Marlowe’s Faustus. It will not do to say that the Faust character is just the sum of all the particular Fausts in all the particular Faust stories (i.e., the general Faust character in Voltolini’s sense). For in some Faust stories, Faust goes to hell, and in others he is saved by God’s grace. Thus, Voltolini’s general Faust character would have incompatible properties. It seems that Voltolini is aware of this consequence (see his example on p. 196) but that he does not consider it to be a problem for his theory. Yet I claim that it is simply not true that we always (or even usually) refer to an inconsistent general character when we refer, for instance, to the Faust character. Rather, we refer to a character that is included in all (or at least most of) the particular Faust characters of the particular Faust stories. This character is not the sum of all particular Faust characters but rather a part that all these particular characters have in common.

5. Finally, I come back to the »ontological argument« in favour of fictitious entities. The argument contains three premises:

If we admit a certain kind of entity, we cannot but admit all the other kinds of entities that figure in the identity conditions of such an entity.
We admit fictional works.
Fictional objects figure in the identity conditions of fictional works. Therefore: We cannot but also admit fictional objects. (241)

I grant premises 1 and 2, but not premise 3. Let’s look how Voltolini justifies this premise. If I have understood it correctly, the basis of Voltolini’s argument in favour of premise 3 is his intuition that in the idealized Menard case, the fictional work by Cervantes is not identical to the fictional work by Menard. But, the argument proceeds, the distinctness of Cervantes’ Quixote and Menard’s Quixote can be grounded neither in the syntactic structure nor in the meaning. For both syntax and semantics are exactly the same in these two works. Therefore, the distinctness of these two works must be grounded in the distinctness of their characters. That is, Menard’s work is distinct from Cervantes’ work because the Quixote character of Menard’s work is distinct from the Quixote character of Cervantes’ work (229-238, esp. 235). Thus, Voltolini argues, fictional characters figure in the identity conditions of fictional works.

I deny what I take to be the basis of this whole line of reasoning, i.e., the claim that in the idealized Menard case the work by Cervantes is distinct from Menard’s work. I see no reason to accept this claim. Obviously, Voltolini has a very firm intuition in favour of it, but I failed to find out the ground for this intuition. (Incidentally, I have a firm intuition to the contrary.) If Voltolini’s »ontological argument« were in fact the strongest argument in favour of an ontological commitment to fictitious objects (as Voltolini seems to think), the case for fictitious objects would be very weak indeed.

However, there are stronger arguments in favour of fictitious objects, namely those arguments that Voltolini calls »semantic«. These arguments are of the following structure: 1. »p« is true (where »p« is a sentence that is, on the surface at least, about one or more fictitious objects). 2. If »p« is true, then fictitious objects exist. 3. Thus, fictitious objects exist.

Voltolini finds this line of reasoning insufficient for the following reason: Antirealists use to argue that all those sentences which seem to be »about« fictitious objects can be paraphrased in such a way that the alleged reference to fictitious objects disappears. Realists, in their turn, have to show that at least some of the paraphrases proposed by the antirealists are inadequate. But

it is clearly not sufficient for the syncretist to be able to show that none of the antirealist paraphrases of those sentences in nonconniving uses which have hitherto been provided actually work. It cannot be excluded that in the future a new antirealist paraphrase will be thought up that overcomes all the putative counterexamples which previous antirealist paraphrases allegedly failed to deal with. (225)

Even more important, however, is the point that paraphrases are in principle »ontologically neutral«, because

a paraphrase and the sentence it paraphrases are merely same-sayers. Insofar as this is the case, the sentence to be paraphrased can be read in terms of its paraphrase as well as the other way round. So, if one paraphrases a sentence in apparently noncommittal terms, it is also true that one can vice versa read the paraphrase in terms of the apparently committal sentence. (226)

If this really were the case, then, of course, paraphrase strategies would be utterly useless for the purpose of settling questions of ontological commitment. However, Voltolini’s objection rests on an erroneous conception of adequacy conditions for paraphrases. As he sees it,

any paraphrase of a given sentence must share with it not only its actual, but also its possible truth evaluations: in order for a sentence to say the same as another sentence, these sentences must share their modal content, that is, they must obtain the same evaluation at all possible worlds. (178f.)

This equivalence condition for the adequacy of paraphrases indeed leads into the dilemma that a paraphrase is either inadequate (because it is not equivalent with the paraphrased sentence) or useless (because it works in the other direction equally well). But in fact, an adequate paraphrase need not and in many cases must not be equivalent with the paraphrased sentence. When philosophers set out to find a paraphrase for a sentence, usually they take the original sentence, if understood in its literal sense, either to be plainly false or to be ambiguous (where at least one of the possible readings of the ambiguous sentence is false). Often (though not always) philosophers offering paraphrases implicitly or explicitly assume that the original sentences are just misleading façons de parler, i.e., that the original sentences (if taken literally) do not really express what the speakers intend to express by using them. In such cases, the task of paraphrasing is to replace a sentence that is literally false and whose literal meaning is not what the speakers intend to communicate with it with a sentence that is literally true and whose literal meaning is exactly what the speakers intend to communicate with it. Clearly, such a paraphrase cannot have the same modal content as the original sentence, but this does not make the paraphrase inadequate – quite to the contrary. A paraphrase is adequate if its literal meaning is the meaning that speakers intend to communicate with the original sentence (although, perhaps, in a misleading way). Given this adequacy condition for paraphrases, paraphrase strategies are, in general, not vulnerable to Voltolini’s above mentioned objections.

Univ.-Doz. Dr. Maria Elisabeth Reicher

Karl-Franzens-Universität Graz

Institut für Philosophie

Notes

[1] Voltolini erroneously suggests that I am a proponent of Neo-Meinongianism. (See note 10 on p. 108.) In fact, I have always (i.e., from the mid 90ies onwards) advocated a version of artefactualism. See, for instance, my Fiktive Gegenstände als abstrakte Individuen, in: Kjell S. Johannessen/Tore Nordenstam (eds.), Culture and Value. Philosophy and the Cultural Sciences, Kirchberg am Wechsel 1995, 233-240; Referenz, Quantifikation und ontologische Festlegung, Frankfurt/Main 2005 (Part IV, Chapter 1, 226-252); Two Interpretations of »According to a Story«, in: Andrea Bottani/Richard Davies (eds.), Modes of Existence. Papers in Ontology and Philosophical Logic, Frankfurt/Main 2006, 153-172. [zurück]

2008-09-25

JLTonline ISSN 1862-8990

This work may be copied for non-profit educational use if proper credit is given to the author and JLTonline.

For other permission, please contact JLTonline.

How to cite this item:

Maria Elisabeth Reicher, Is There a Part of Snow White Instantiated While the Story Is Told?. A Novel Artefactualist Theory of Fictitious Objects. (Review of: Alberto Voltolini, How Ficta Follow Fiction. A Syncretistic Account of Fictional Entities, Dordrecht: Springer 2006.)

In: JLTonline (25.09.2008)

URL: http://www.jltonline.de/index.php/reviews/article/view/31/187

A Persistent Identifier can be found in the PDF-Version of this article.