Towards a New Theory of Modal Fictionalism

Dombrovszki, A.: Towards a New Theory of Modal Fictionalism. In: Ostium, roč. 13, 2017, č. 4.

Abstract: In our everyday discourse, most of us use modal statements to express possibility, necessity, or contingency. Logicians, linguists, and philosophers of language tend to use the possible world discourse to analyse the semantics of this kind of sentences. There is a disadvantage of this method: in the usual Quinean meta-ontology it commits the users to the existence of possible worlds. Even though there are many theories on metaphysics of these possible worlds, I will focus on the fictionalist approach, which aims to suspend the undesired ontological commitment. Therefore it is an anti-realist theory. The fictionalist strategy is based on the idea that according to our common-sense ontology, fictions and fictional entities do not exist, and one can treat the possible world discourse as a fiction, and the possible worlds as fictional entities. However, in the last two decades, realist theories about fictions and fictional entities have arisen, especially the abstract artefact theory. In this paper I try to revise the fictionalist approach to modality by keeping the main idea that possible worlds are fictional entities, while accepting that fictional entities are abstract artefacts. The result is a new realist theory of modal fictionalism.

Keywords: analytic philosophy, metaphysics, philosophy of language, fictionalism, modality, possible worlds

The aim of this paper is to provide a revised version of the so-called fictionalist approach to modality.[1] Proponents of this view hold that realist theories about possible worlds, and therefore the whole possible world discourse is just a useful fiction, and – at least I suppose that – possible worlds are fictional entities. It seems that with this strategy fictionalists can use the possible world discourse without the undesired ontological commitment to the worlds. However, according to a widespread analysis of fictional discourse, fictions, as well as fictional entities, are actually existing abstract artefacts, despite our common-sense intuitions about the non-existence of them. If this analysis is correct, then the modal fictionalist can’t maintain her antirealist position towards possible worlds: she must accept that possible worlds are abstract artefacts too. But this is certainly not a problem for her: I will argue in favour of this new, realist modal fictionalism.

In Section 1.1, I introduce a set of background assumptions from the field of philosophy of language and ontology, to restrict the domain of my investigation. I call this sketched meta-theory the “Kripkean framework”, assuming that most followers of Kripke share these assumptions. In Section 1.2, I point out one major problem for the Kripkean framework, which is the improper treatment of sentences containing empty names. Probably the best solution for this problem is the complete elimination of empty names, introducing artificial reference-surrogates for every proper name. It means that, accepting this solution, fictional characters – like Emma Woodhouse or Sherlock Holmes – are existing abstract artefacts. In Section 2, I put forth an analogue between the analysis of fictional discourse and the fictionalist treatment of the possible world discourse, concluding that if one accepts that Jane Austen’s Emma, and the fictional character Emma Woodhouse in it exist, then she must accept that David Lewis’s On the Plurality of Worlds[2], and the possible worlds in it also exist. This analogue is a key component in my argument. In Section 3, I try to elaborate a realist modal fictionalist theory in detail, contrasting it with extant variations in the modal fictionalist theory.

1.1 The Kripkean Framework
In the following, I am going to define a theoretical framework for our current investigations, called the Kripkean Framework (KF).[3] I suppose that most followers of Kripke accept this set of beliefs, and that its assumptions are widely hold in contemporary analytic philosophy. KF is going to be summarised in six points, from (a) to (f).

(a) The Quinean meta-ontology.[4] Instead of something like the Meinongian approach to metaphysics, I will use a Quinean meta-theory, relying on its two basic features. Firstly, the ontological system is not hierarchical, every entity exists on the very same level, e.g. if one says that abstract entities exist, she means that they exist in the same sense as chairs, tables, trees, mountains and so on, except that abstract entities lacks spatial or spatiotemporal extension. Secondly, one’s ontological commitments are indicated by the bounded variables of the existential quantifier, and nothing else.

(b) The neo-Russellian theory of propositions.[5] Expressing a proposition is a necessary condition for a sentence to be meaningful. I consider propositions to be structured entities, containing different types of entities. For example, in KF, the sentence “Peter is taller than David” expresses a proposition, containing Peter and David as physical entities, and the “x is taller than y” binary relation which can be considered a universal. Although the metaphysical nature of these propositions is an important question, especially when their constituents include physical and abstract objects together, it won’t be elaborated here, because this problem does not affect the current aims of this paper.

(c) The Millian semantics of proper names. As J. S. Mill put forth, proper names have only denotation, without any connotation.[6] It means that, contrasted to a descriptivist approach,[7] the only semantic value of a given proper name is its referent. Think of them as tags. From this theory, (d) and (e) follows.[8]

(d) The direct-reference theory (DR). According to the DR, a proper name’s only semantic contribution to a given proposition – expressed by a sentence containing that name – is its reference. While there is a certain similarity between (c) and (d), they are not the same: the Millian account is a semantic theory about proper names, while DR is about a relation between proper names and propositions. These two theories together need not give an exhaustive semantics for proper names.

(e) Kripke’s rigid designation thesis.[9]A rigid designator refers to the very same individual in every possible world (PW) where the individual exists, and worlds where the individual doesn’t exist, the name fails to refer. According to Kripke, every proper name is a rigid designator. The concept of PW and the PW discourse itself are important notions for this paper, so further clarification follows.

(f) Analysis in terms of the PW discourse. During our ordinary discourse, the most of us are use modal statements to express possibility, or necessity. See for example the following statement:

(1) It is possible that Richard Nixon had a green hair.

To analyse these kind of statements, logicians, linguists, and philosophers tend to use the PW discourse, which aims to paraphrase modal sentences. In this particular case, from (1) we can get

(2) There is a possible world where Richard Nixon had a green hair.

Technical advantages of this method would be the topic of another paper, what is important here is that PW discourse is a useful tool for semantic investigation.

1.2 The Problem of Empty Names
While KF works well with the usual cases of our ordinary discourse, sentences containing empty names raise a challenge to the view. Empty names – proper names without reference – have three types: names of mythical entities (e.g. “Zeus”), names of hypothetical entities (e.g. the planet “Vulcan”), and names of fictional characters (e.g. “Emma Woodhouse”). To illustrate the problem of empty names, consider the following sentence:

(3) Emma Woodhouse lives in Hartfield.

The problem is that, according to (d), the contribution of the proper name “Emma Woodhouse” to the proposition expressed by (3) would be the reference of the name – Emma herself –, who doesn’t exist. So, a full-fledged proposition cannot be expressed, and taking (b), one can conclude that (3) is meaningless and therefore without any truth value in the KF. But (3) is considered as true, and can be understood by any competent speaker of English – so something is wrong in the analysis. This problem raises a trilemma for the proponents of KF.

Firstly, one could solve the problem by denying the universality of the framework. It means that although we analyse the semantics of our ordinary discourse within the KF, for fictional discourse, and for other sentences containing empty names, we need a different theory with the assumption that fictional names are not Millian. However, giving up the universality of KF wouldn’t be the best choice to solve this problem since having one general theory for describing the very same phenomena is always better than having two.

Secondly, one might assume that sentences containing empty names express gappy propositions.[10] The gappy proposition view is indeed compatible with (b) and (d), but it has its own disadvantages. One major problem is the evaluation of gappy propositions: how can one assign truth-value to these structures? Moreover, some philosophers argue that gappy propositions are not propositions at all. Here, I’m not going to argue against this view,[11] from now on I just assume that propositions must be full-fledged.

Thirdly, the remaining option is to eliminate all empty names from KF. It means that artificial reference-surrogates are introduced into our ontology: fictional characters, mythical entities and hypothetical entities are now considered as actual existing objects. This includes the existence of fictions and theories in which these entities appear. I think this is the best option which one can choose to save KF, but this one also comes with a compromise: the cost is the extended ontology.

If one can accept the existence of fictional characters, then she should ask in what metaphysical category they fit. There are three options which are compatible with the meta-ontological presumption (a) in KF. These are Platonism, Lewisian concretism, and abstract artefact theory. The latter is the most popular version of realism about fictional characters, assuming that, in agreement with our common-sense intuitions, fictional characters are made by their authors. With this intuitive assumption, most of the problems of Platonism and concretism can be avoided.[12] So, from now on, I suppose that fictions and fictional characters are existing abstract artefacts.[13]

2. The Analysis of Fictional Discourse and Modal Fictionalism: An Analogue
In this section, I put forth an analogy between the usual analysis of fictional discourse in KF, and the fictionalist approach to modality. I think from this analogy follows an important conclusion, namely, if one treats the theory of modality as fiction, and the possible worlds in this fiction as fictional entities, then the ontological categorization must be the same as in the case of “regular” fictions. It means that from a fictionalist point of view, possible words are existing abstract artefacts, despite the original goal, which is antirealism.

Let’s see a typical analysis of fictional discourse by Nathan Salmon.[14] This approach is based on two important distinctions. The first is about the interpretation and truthmaker of sentences. Consider again sentence (3). It is obvious that taking literally, as it’s about our actual world, (3) is a false sentence, because Emma is just a fictional character, an abstract entity, which lives nowhere. Call this interpretation the external reading of (3). But we have a strong intuition that in some contexts we rather take (3) as a true sentence. The reason is that in the book Emma, we can find information about the protagonist’s home: she lives in Hartfield indeed. When we rely on the fiction evaluating the sentence, we are in the internal reading.

So, sentences of the fictional discourse are ambiguous. A convenient tool to explain this ambiguity is the so-called story operator: “according to fiction f”. This leads us to a distinction between object-fictional and meta-fictional sentences. Object-fictional sentences are the regular sentences of fictional discourse, like (3). But one can compose a meta-fictional pair of every object-fictional sentence, by taking it into the scope of the story operator. Thus, the meta-fictional pair of (3) looks like the following:

(4) According to Emma, Emma Woodhouse lives in Hartfield.

Using meta-fictional sentences, we can explain our intuitions about the truth-value of the internal reading of (3). According to this analysis, in the internal contexts we use the object-fictional sentences as a shorthand for their meta-fictional pairs. This provides a plausible explanation of the ambiguity of the sentences about fictional entities.

After this very brief introduction, I’m going to sketch the main idea behind the fictionalist approach to modality.[15] The starting point is sentence (2). Remember that, according to (a), we accept a Quinean meta-ontology, which means that (2) is an existentially committing sentence, because the term “possible world” is in the scope of the existential quantifier “is”. Of course, there are a lot of philosophers who not satisfied with the idea that using the PW discourse they must accept the existence of possible worlds. The main aim of modal fictionalism is to provide a theory which allows them to use the PW discourse, without undesired ontological commitments.

This aim can be reached by considering the PW discourse as a fictional discourse, and the possible worlds as fictional entities in this “story”. The account is driven by the intuition that fictions and fictional entities do not exist in our actual world.

Fictionalists also appeal to the story operator. In fact, the fictionalist treats sentences of PW discourse, like (2), literally false in the external interpretation, while in the internal contexts she considers it an abbreviation of the sentence:

(5) According to Lewis 1986, there is a possible world where Richard Nixon had a green hair.

It may look arbitrary to choose David Lewis’s theory as the fiction, on which the fictionalist relies on, but this is just a practical choice: Lewis had the most detailed elaboration of the theory of possible worlds.

This approach seems to work at first sight: putting the existential quantifier behind the scope of the intensional story operator leads to the suspension of the undesired ontological commitments.

My concern here is that in KF the existence of fiction and fictional characters are taken for granted – without these abstract artefacts KF is not able to handle the problem of empty names. For those who choose KF as the general framework in philosophy of language, modal fictionalism also can’t provide an antirealist account. The reason is if one accepts the existence of the fiction Emma, and of the fictional entity Emma Woodhouse in this fiction, then she must accept the existence of the “fiction” Lewis 1986, and the fictional entities in it. It would be an arbitrary choice to say that some fictions and fictional entities exist but others do not. So, for advocates of KF, classic modal fictionalism seems an unsuccessful theory.

One question arise here: what’s the point of this whole investigation? One can immediately reply to my critique: “those who have already accepted the existence of Platonic objects (propositions are assumed in KF), moreover, the existence of abstract artefacts, are just not the ones who would willingly accept the fictionalist approach to modality. They don’t want to be antirealist about possible worlds, because they are already working with a broad ontology.” It looks like my objection misses the target, no actual philosopher would ever meet such a problem.

Here comes the positive part of my project: I think the objection I raised above leads us to a new theory of modal fictionalism, which is realist, and considers possible worlds as abstract artefacts. I also think that this realist version of modal fictionalism is a better choice than other variations, especially for those who has already accepted the existence of abstract artefacts. In the final section I will develop a detailed version of this account, called the hybrid modal fictionalism (HMF).

3. Hybrid Modal Fictionalism
Let the variations of “classical” modal fictionalism be our starting point to differentiate between them and HMF. Two questions will help to do this job.

The first question is what makes our everyday modal sentences like (1) true? Or more accurately: what are the truthmakers of these sentences? Advocates of timid modal fictionalism are agnostic about this question. They don’t aim to build a robust metaphysical theory about modality, all they want is a “permission” to use the PW discourse without the undesired ontological commitments. On the other hand, strong modal fictionalists (SMF) assume that the truthmaker of our ordinary modal sentences is the fiction itself, e.g. the truthmaker of (1) is Lewis 1986.[16]

The second question: what is the ontology of the fiction on which the fictionalist relies, in this case Lewis 1986? There are also two answers here. According to the Platonist version of modal fictionalism, Lewis 1986 is a Platonic entity, discovered by David Lewis, while the artefactualist view considers the theory as a human-made entity. Of course, the possible worlds described in the stories don’t exist, because the classic versions of modal fictionalism are antirealist views.

To sum up, there are three versions of modal fictionalism, the strong Platonist, the strong artefactualist and the timid modal fictionalism. Advocates of the timid version are not concerned with metaphysical issues, even though it would be possible to distinguish between them regarding the other aspect too.

In the previous section, I presupposed that if fictions and fictional characters exist, then they should be treated as abstract artefacts. Because of this assumption, I am not going to consider the Platonist version from now on, which means that only two options remained to keep the possible world discourse without the undesired ontological commitments.

But a certain doubt arises about the success of this kind of antirealism. The main idea behind the fictionalist’s programme is that fictional entities, like Emma Woodhouse or Sherlock Holmes don’t exist, and possible worlds can be considered as fictional entities. But in KF, there are good reasons to treat fictional entities as existing, human-made abstract objects. Otherwise, the theory cannot explain the semantics of fictional discourse.

My main thesis is that if a KF theorist chooses the fictionalist view about possible worlds, she chooses a realist option, because it would be an ad hoc decision to accept the existence of one type of fictional entities (the fictional characters), but deny another type, e.g. the existence of hypothetical entities (like the possible worlds). So, for those who accept KF, modal fictionalism can’t be the antirealist position which it aims to be.

But does it mean that modal fictionalism is not a feasible choice for KF theorists? Not at all. My positive project aims to develop a realist account of modal fictionalism, which considers possible worlds as abstract artefacts.

Someone can raise certain doubts about the significance of this project. There are already two fine options to be realist about possible worlds: genuine realism and ersatz realism, both developed in detail, and can explain the metaphysics of modality. Moreover, realist modal fictionalism seems like a twisted idea: the aim of modal fictionalism was to provide an antirealist account about possible worlds. The account which I would like to develop just seems at least suspicious, so let’s see what we can say in favour of such a realist modal fictionalism.

To motivate the idea, I will sketch a descriptive picture of the early history of modality in a nutshell. I do this for strengthen certain intuitions, and to show that there is a place for a theory which considers possible worlds as abstract artefacts.

The idea to analyse modal sentences in our ordinary discourse is not a new one: Aristotle wrote about it already. However, the rapid expansion of the topic was in the mid 20^th century, when modal logic arose, and different systems of it were developed. It is essential to consider these systems as a creation of logicians, which I think matches our pre-philosophical intuitions. So, logicians tried to model these kind of sentences, and for the formal models they coined the concept of „possible world”, to help us understand the semantics by getting an intuitive picture of what’s going on in these mathematical structures. Unfortunately, they weren’t careful enough, and curious philosophers asked the question: what are the possible worlds? This innocent question led to a whole segment of mainstream analytic philosophy, modal metaphysics – leading to some wild theories, e.g. genuine realism, which assumes that possible worlds represent possibilia, and they are natural entities like mountains, rivers or forests.

According to the account presented here, keeping in mind this description from an external perspective of what happened, possible worlds are not representing possibilia. They do not exist like mountains, rivers or other natural, mind-independent objects. Possible worlds are creations of logicians and philosophers, they use them to solve semantic problems. I think this approach has a greater intuitive power than ersatz realism.

Maybe one can worry about the inflation of ontology: isn’t it the case that ersatz realism is more parsimonious then HMF? The answer is no, we already accepted the existence of abstract artefacts to solve the problem of empty names. Accepting that possible worlds are also abstract artefacts, doesn’t inflate further the ontology.[17]

Now, let’s see the general features of HMF. In the following, I will elaborate the theory, summarising five important properties of it.

(i) Realism about possible worlds, and (ii) possible worlds are abstract artefacts. These points are already explained by the analogy between the novel Emma, the fictional character Emma Woodhouse, and the fiction of Lewis 1986, the fictional entities, which are the possible worlds. It’s worth noting that while “Emma Woodhouse” is a singular referring term, “possible world” refers to a natural kind. We can distinguish a weaker and a stronger argument supporting the analogy sketched above. According to the weaker argument, we don’t suppose that the semantics of proper names and natural kind terms are the same. In KF, using the term “possible world” does not necessitate the existence of possible worlds as we saw in the case of proper names. But according to the fictionalist’s viewpoint, we put these terms into the very same category, both refers to a fictional entity, or a kind of fictional entities. It would be an ad hoc decision to consider one of them an abstract artefact, while the other a Platonist or even a non-existent entity. The stronger version of this argument is when we suppose that singular terms and natural kind terms have the same semantics, i.e. both are directly referring terms.[18] If this stronger claim is true, then the existence of possible worlds is a consequence of KF.

(iii) Paraphrasing ordinary language sentences using PW discourse can be considered as an explication in the Carnapian sense.[19] By paraphrasing I mean the translation of (1) to (2). To understand the concept of explication, it is useful to draw a distinction between ordinary language and scientific language. Expressions of ordinary language are imprecise, ambiguous, subjective and qualitative – expressions with these properties are inappropriate for scientific purposes. If one examines the expressions of scientific language, she will find precise, univocal, objective and quantitative expressions, because these are required for scientific investigations.

Explication is a process, in which we substitute the explicandum, which is usually an ordinary language expression, or an outdated expression of scientific language, with the explicans: an adequate expression for scientific usage. For example, in our ordinary discourse, all of us say things like “x is hotter than y”, “the Earth is round”, or “x is red”; things which we perfectly understand. However, these expressions are rough for the sophisticated scientific discourse. Scientists usually express these properties as “x has a higher temperature than y”, “the Earth has a geoid shape”, “x reflects a light with a wavelength between 620 and 740 nanometres”. These scientific expressions describe the same phenomena, but in a more precise manner, better suited for scientific discourse.

We can also think about the PW discourse as an explication.[20] In this case, the explicanda are the modal terms of our ordinary language: “necessarily p”, “possibly p”, and “impossibly p”, and the explicantia are “according to Lewis 1986, it is true in every possible world, that p”, “according to Lewis 1986, there is at least one possible world where it is true, that p”, and “according to Lewis 1986, there isn’t any possible world where it is true, that p”. While our modal intuitions can be different about what is the meaning of the word “necessary”, logicians adequately defined it as “is true in every possible world”.

(iv) HMF is agnostic about the truthmakers of modal sentences used in ordinary language. Practically, it means that we simply don’t apply the story operator to these statements. We take them literally, and assign a truth-value to them, but when it comes to the truthmakers – just like the timid fictionalists – we are silent about the issue. If we recall, proponents of SMF think in a different way: the fiction – Lewis 1986 – is the truthmaker of sentences like (1). This is a very strong claim, which also not a plausible one, if we consider the consequences of such a view. Let’s see the two main objections against SMF.

The first one is Daniel Nolan’s argument about artificiality.[21] Usually, we think about modalities as natural, objective properties of things, which are rather independent of what we think about them. The fact that the table in front of me could have been painted red, doesn’t depend on people’s activities. According to Nolan, if SMF is true, then modalities must be some artificial properties, which only exist because David Lewis wrote a book in 1986. If he hadn’t existed, and nobody else thought up a theory of possible worlds instead of him, then nothing would have been possible. In addition to this, before 1986, there weren’t any modalities at all, because the truthmaker wasn’t available back then. This seems to be false. Moreover, there is a modal argument in the same manner, according to which it is a contingent fact that Lewis wrote his book in 1986. But it seems that the existence of our objective modal truths cannot depend on contingent facts.[22]

The second one is an epistemological issue. If Lewis 1986 underpins the semantic value of our ordinary modal claims, then epistemic access to this theory – or fiction – is required to understand sentences like (1). It means that, if someone is not familiar with Lewis’s theory, then she is not able to evaluate the truth-value of (1), and maybe she can’t understand it at all. Given the fact that only very few people – probably philosophers – are familiar with genuine PW realism, most speakers of language are in trouble – who otherwise are considered as competent language users. This consequence doesn’t seem like an acceptable one.

These two roughly sketched arguments aim to show that SMF is in trouble: the idea that Lewis 1986 has any role in our ordinary modal discourse is not a viable one. The objections can arise because of this broader application of the story operator. Since HMF doesn’t extend it to the ordinary language, it can avoid the problems. But I think the story operator certainly has a role in PW discourse, in fact, the story itself is the truthmaker of its sentences, like (2). This leads us to the last property of HMF.

(v) The truthmaker of the PW discourse is Lewis 1986. Unlike in the timid version of modal fictionalism, I admit that, there is some semantic role for the story operator “according to Lewis 1986”; but I don’t think this role is as broad as in SMF. The assumption that Lewis 1986 is the truthmaker of the sentences of the PW discourse, analytically follows from the semantics story operator. If we look at sentence (5), it exactly means that the sentence is true according to the fiction, basically because of Lewis 1986. Another motivation for this assumption can be reached if one takes the “external viewpoint”,[23] which was introduced in the descriptive picture of modality above. Users of the PW discourse created the theory to get a model of our ordinary language discourse, to use it for semantic analysis. Because PW discourse is an artificial linguistic structure with semantic rules given by logicians and linguists, it is natural to see in a way that these rules are chosen by a theory, a “fiction” which they consider as a good model. Of course, one can decide that Lewis 1986 is not the best story which can be applied as a theory of modality, but this is not the point here – it can be replaced with any other fiction.

Note that the arguments against SMF do not hold for HMF. Nolan’s artificiality problem can’t be raised, because PW discourse is admittedly an artificial linguistic tool, which wouldn’t exist without the activity of philosophers like David Lewis. Similar things can be said about Sauchelli’s epistemic problem: if one is not familiar with Lewis’s work, or at least she doesn’t have a grasp of what possible worlds are, she won’t be able to evaluate the truth-value of (5) – maybe she wouldn’t understand what this sentence says at all. One interesting thing about HMF is that it not just avoids these problems, but I think these two objections even strengthen the account, showing that it matches our intuitions.

So, according to HMF, in contrast with the weak version, Lewis 1986 has an important role as a truthmaker, but in contrast with the strong version, this role extends only to PW discourse, the ordinary modal sentences are not involved in this picture. This is why Hybrid Modal Fictionalism got its name: the account is between the two versions of modal fictionalism with respect to this semantic aspect. But from a metaphysical point of view, there is another difference: while the fictionalist’s main project is to keep the PW discourse without the ontological commitments, HMF is a realist theory. Accepting that possible worlds are abstract artefacts doesn’t mean that we inflate our ontology since its already accepted in KF to provide a solution for the problem of empty names.

4. Summary
In this paper, I introduced a new approach to modal metaphysics, which I called Hybrid Modal Fictionalism. The main idea was presented by an analogy between the analysis of fictional discourse in the Kripkean Framework, and the fictionalists’ usual treatment of modality.

In section 1.1, I sketched a general framework for the further investigations, including the neo-Russellian theory of propositions and the direct reference theory. These assumptions led to a problem (section 1.2) with empty names, which can be eliminated by extending the ontology, providing reference-candidates for empty names. In section 2, the aforementioned analogy was sketched between the fiction Emma and Lewis 1986, and between the fictional character Emma Woodhouse and the fictional entities, the possible worlds. If one accepts the existence of the first kind, she should accept the existence of the second also. Finally, in section 3, I tried to elaborate further the idea of HMF. The finer details are still missing, but it seems that there is a good prospect for extending the classical taxonomy of modal metaphysics to include the new option presented in this paper.

B i b l i o g r a p h y
CARNAP, R.: Meaning and Necessity. University of Chicago Press: Chicago, 1947.
CARNAP, R.: Logical Foundations of Probability. University of Chicago Press: Chicago, 1950.
EVERETT, A.: Empty Names and ‘Gappy’ Propositions. In: Philosophical Studies, Vol. 116, No. 1, 2003, pp. 1–36.
FITCH, G. – NELSON, M.: Singular Propositions. In: ZALTA E. N. (ed.): Stanford Encyclopedia of Philosophy. Winter 2016 Edition: https://plato.stanford.edu/archives/win2016/entries/propositions-singular.
FREGE G.: Sense and Reference. In: The Philosophical Review, Vol. 57, 1948, pp. 209–230.
KAPLAN, D.: Demonstratives. In: ALMOG, J. (ed.): Themes from Kaplan. Oxford University Press: Oxford, 1989, pp. 481–563.
KIM, S.: Modal Fictionalism and Analysis. In: KALDERON, M. E. (ed.): Fictionalism in Metaphysics. Oxford University Press: Oxford, 2005, pp. 116–133.
KRIPKE, S.: Naming and Necessity. Harvard University Press: Cambridge, 1980.
KRIPKE, S.: Reference and Existence. Oxford University Press: New York, 2013.
LEWIS, D.: On the Plurality of Worlds. Basil Blackwell: Oxford, 1986.
MARTÍ, G.: The Question of Rigidity in the New Theories of Reference. In: Nous, Vol. 37, No. 1, 2003, pp. 161–179.
MILL J. S.: A System of Logic. John W. Parker: London, 1893.
MOUSAVIAN, S. N.: Gappy Propositions. In: Canadian Journal of Philosophy, Vol. 41, No. 1, 2011, pp. 125–158.
NOLAN, D.: Three Problems for ‘Strong’ Modal Fictionalism. In: Philosophical Studies, Vol. 87, No. 3, 1997, pp. 259–275.
ROSEN, G.: Modal Fictionalism. In: Mind, Vol. 99, No. 395, 1990, pp. 327–354.
SALMON, N.: Nonexistence. In: SALMON, N. (ed.): Metaphysics, Mathematics and Meaning. Oxford University Press: New York, 2005, pp. 50–90.
SEARLE J.: Proper Names. In: Mind, Vol. 67, 1958, pp. 166–173.
SEARLE, J.: The Logical Status of Fictional Discourse. In: SEARLE, J. (ed.): Expression and Meaning. Cambridge University Press: New York, 1979, 58–75.
SOAMES, S.: Beyond Rigidity. Oxford University Press: New York, 2002.
THOMASSON, A. L.: Fiction and Metaphysics. Cambridge University Press: Cambridge, 1999.
VAN INWAGEN, P.: Meta-Ontology. In: Erkenntnis, Vol. 48, No. 2, 1998, pp. 233–250.
WOODWARD, R. Is Modal Fictionalism Artificial? In: Pacific Philosophical Quarterly, Vol. 92, No. 4, 2011, pp. 541–548.
ZVOLENSZKY, Zs.: Against Sainsbury’s Irrealism About Fictional Characters. In: Magyar Filozófiai Szemle, Vol. 54, No. 4, 2012, pp. 83–109.
ZVOLENSZKY Zs.: Artifactualism and Authorial Creation. In: Proceedings of the European Society for Aesthetics, Vol. 6, No. 1, 2014, pp. 457–469.
ZVOLENSZKY Zs.: An Argument for Authorial Creation. In: Organon F, Vol. 22, No. 4, 2015, pp. 461–487.

N o t e s
[1] I am grateful to Andrea Komlosi, Zsofia Zvolenszky, and the participants of the Characters of Current Philosophy and Its Methods conference in March 2017 for discussions of the issues raised in this paper. This research has been supported by the UNKP–17–2 New National Excellence Program of the Ministry of Human Capacities.
[2] LEWIS, D.: On the Plurality of Worlds. Basil Blackwell: Oxford, 1986.
[3] This tag doesn’t mean that KF actually represents Kripke’s standpoint.
[4] In the way as its presented in VAN INWAGEN, P.: Meta-Ontology. In: Erkenntnis, Vol. 48, No. 2, 1998, pp. 233–250.
[5] See for an introduction to proposition theories: FITCH, G. – NELSON, M.: Singular Propositions. In ZALTA E. N. (ed.): Stanford Encyclopedia of Philosophy, Winter 2016 Edition: https://plato.stanford.edu/archives/win2016/entries/propositions-singular.
[6] MILL J. S.: A System of Logic. John W. Parker: London, 1893, pp. 34–41.
[7] According to the descriptivist picture, proper names has another semantic value, identical with a description (FREGE G.: Sense and Reference. In: The Philosophical Review, Vol. 57, 1948, pp. 209–230.), or a set of descriptions (SEARLE J.: Proper Names. In: Mind Vol. 67, 1958, pp. 166–173.).
[8] If one thinks that the only semantic value of a proper name is its reference, then she should think that that reference is the only contribution to the proposition. The other choice is to think that proper names doesn’t contribute to propositions, which is not a plausible option. In the case of (e), see MARTÍ, G.: The Question of Rigidity in the New Theories of Reference. In: Nous, Vol. 37, No. 1, 2003, pp. 161–179.
[9] KRIPKE, S.: Naming and Necessity. Harvard University Press: Cambridge, 1980.
[10] The idea of gappy propositions comes from KAPLAN, D.: Demonstratives. In: ALMOG, J. (ed.): Themes from Kaplan. Oxford University Press: Oxford, 1989, pp. 496.
[11] Others have done it. See EVERETT, A.: Empty Names and ‘Gappy’ Propositions. In: Philosophical Studies, Vol. 116, No. 1, 2003, pp. 1–36, and MOUSAVIAN, S. N.: Gappy Propositions. In: Canadian Journal of Philosophy, Vol. 41, No. 1, 2011, pp. 125–158.
[12] For the arguments, see ZVOLENSZKY, Zs.: Against Sainsbury’s Irrealism About Fictional Characters. In: Magyar Filozófiai Szemle, Vol. 54, No. 4, 2012, pp. 83–109; ZVOLENSZKY Zs.: Artifactualism and Authorial Creation. In: Proceedings of the European Society for Aesthetics, Vol. 6, No. 1, 2014, pp. 457–469; ZVOLENSZKY Zs.: An Argument for Authorial Creation. In: Organon F, Vol. 22, No. 4, 2015, pp. 461–487.
[13] Most important advocates of the abstract artefact theory are KRIPKE, S.: Reference and Existence. Oxford University Press: New York, 2013; SALMON, N.: Nonexistence. In: SALMON, N. (ed.): Metaphysics, Mathematics and Meaning. Oxford University Press: New York, 2005, pp. 50–90.; SEARLE, J.: The Logical Status of Fictional Discourse. In: SEARLE, J. (ed.): Expression and Meaning. Cambridge University Press: New York, 1979, 58–75.; and THOMASSON, A. L.: Fiction and Metaphysics. Cambridge University Press: Cambridge, 1999.
[14] SALMON, N.: Nonexistence.
[15] Originally introduced by ROSEN, G.: Modal Fictionalism. In: Mind, Vol. 99, No. 395, 1990, pp. 327–354.
[16] These versions are differentiated by NOLAN, D.: Three Problems for ‘Strong’ Modal Fictionalism. In: Philosophical Studies, Vol. 87, No. 3, 1997, pp. 261–264.
[17] For the explanation of this idea, and the case of the false ontological parsimony, see THOMASSON A. L.: Fiction and Metaphysics, pp. 137–145.
[18] There are good arguments for this, see for example the last three chapters of SOAMES, S.: Beyond Rigidity. Oxford University Press: New York, 2002.
[19] See CARNAP, R.: Logical Foundations of Probability. University of Chicago Press: Chicago, 1950, pp. 1–18.
[20] This viewpoint is not my own, Carnap also thought it as an explication process, however, he hasn’t use the term “possible world” back then. See: CARNAP, R.: Meaning and Necessity. University of Chicago Press: Chicago, 1947.
[21] NOLAN, D.: Three Problems for ‘Strong’ Modal Fictionalism, pp. 264–266.
[22] Contingency problem is discussed further in WOODWARD, R.: Is Modal Fictionalism Artificial? In: Pacific Philosophical Quarterly, Vol. 92, No. 4, 2011, pp. 541–548; and in KIM, S.: Modal Fictionalism and Analysis. In: KALDERON, M. E. (ed.): Fictionalism in Metaphysics. Oxford University Press: Oxford, 2005, pp. 116–133.
[23] This „external viewpoint” is somehow essential for HMF: I don’t consider possible worlds as representations of counterfactual situations, I consider them as merely theoretical tools of linguists, logicians and philosophers. The easiest way to understand this is to look at they discourses, they papers, they works, and describe your experiences. Without this changed viewpoint, it would be hard to accept for example the idea that possible worlds are abstract artefacts, which explains why this option haven’t arise yet. Consider an analogy with the hammer: for the users, the hammer is an instrument with a function, it can be used to drive a nail into the wall, like the possible worlds are representations of counterfactual situations for a philosopher. On the other hand, from an “external viewpoint”, e.g. for a physicist, the hammer is not an instrument, just a physical thing, a construction of particles, as I think possible worlds are just artificial abstract artefacts. In this project, I try to regard possible worlds as a physicist regards an ordinary object.

Áron Dombrovszki
Department of Logic
Institute of Philosophy
Eötvös Loránd University
Budapest
e-mail: d_aron@outlook.com