Fictional Creations

Many people assume that fictional entities are encapsulated in the world of fiction. I show that this cannot be right. Some works of fiction tell us about pieces of poetry, music, or theatre written by fictional characters. Such creations are fictional creations, as I call them. Their authors do not exist. But that does not take away that we can perform, recite, or otherwise generate actual instances of such works. This means we can bring such individuals actually into existence, as the works they are. I conclude that the assumption about encapsulation is untenable, unless an exception is made for types. philosophical claim, some philosophers entity can actual I will show that both claims are false. To say that fictional entities do not exist, or to claim that they could not exist, is mistaken, I argue, for reasons that have little to do with the distinction between fiction and reality.

It is commonly assumed that fictional entities do not exist, at least not actually. "There is no Dracula!" we tell the children before bedtime, "It's only a story. " That no fictional entity exists is however a substantive philosophical claim, and some philosophers have even gone so far to claim that no fictional entity can exist in the actual world. I will show that both claims are false. To say that fictional entities do not exist, or to claim that they could not exist, is mistaken, I argue, for reasons that have little to do with the distinction between fiction and reality.

I. FICTION AND ENCAPSUL ATION
It is undeniable that some works of fiction tell us about and refer to specific paintings, poems, or plays painted, written, or scripted by merely fictional characters. The acts of writing, painting, composition, and so forth that (according to the fiction) brought these works into being, are merely fictional creative acts. I call such fictional works of art fictional creations.
What makes fictional creations special, compared to ordinary creations such as Vermeer's View of Delft or Donna Tartt's The Goldfinch, is that fictional creations find their origin in a merely fictional creative act. The picture of Dorian Gray is the result of a fictional act of painting, and consequently it has never actually been painted. Because fictional creations are created merely "according to a story," fictional creations are fictional entities, on a par with Sherlock Holmes and the titular antagonist of Stephen King's It. Yet, unlike Holmes and It, some fictional creations actually exist as the works of art they are. I will elaborate one example in detail.
In W.H. Mallock's satire The New Republic (1877) we read about the meeting of a party of friends, who end up discussing increasingly philosophical matters. Eventually they become gripped by a Platonic question, what is the essence of a good society? Many of the novel's characters satirize famous aesthetes in Victorian society, such as Walter Pater, Violet Fane, and Matthew Arnold. Mallock portrays his characters often as oversensitive and as seeping with poetry. Their conversations are littered with remembered lines of Tennyson, Browning, and Goethe. However, some of the poems, letters, and sermons that are printed in the novel are entirely fictional. Mr. Rose, whom Mallock intended to represent Pater, explains that the ideal society shall be quite without prejudice or bigotry. It will allow for a variety of styles, he says, "whether they be pagan or Catholic, classical or mediaeval. " To illustrate his point, Mr. Rose produces a piece of paper from his pocket. "These lines," he says, "were written by a boy of eighteen -a youth of extraordinary promise, I think, whose education I may myself claim to have had some share in directing. Listen," Three visions in the watches of one night Made sweet my sleep -almost too sweet to tell. One was Narcissus by a woodside well, And on the moss his limbs and feet were white; And one, Queen Venus, blown for my delight Across the blue sea in a rosy shell; And one, a lean Aquinas in his cell, Kneeling, his pen in hand, with aching sight Strained towards a carven Christ; and of these three I know not which was fairest. First I turned Towards that soft boy, who laughed and fled from me; Towards Venus then; and she smiled once, and she Fled also. Then with teeming heart I yearned, O Angel of the Schools, towards Christ with thee! (Mallock 1877, 171-2) When Mr. Rose puts away the paper he expresses his approval of the sonnet to Miss Merton, one of his interlocutors. It captures the true spirit of modern aestheticism, he says-the spirit in which the architects of the ideal state will set to work.
The eighteen-year-old poet who composed the sonnet for Mr. Rose remains unnamed in the novel, and it is also unclear if the poem has a title. I refer to the poem as Three Visions. Mr. Rose's gifted student does not actually exist. Nor could he actually exist, it seems, for the undergraduate author of Three Visions is a purely fictional person, and fictional people cannot exist actually. Also, the piece of paper on which the sonnet was drafted, and that was handed to Mr. Rose, probably as a gesture of fondness, is a merely fictional piece of paper. The composition of the poem, an event that is part of Mr. Rose's backstory, inherits its fictionality from the boy of eighteen who carried it out (cf. Nolan and Sandgren 2014, 616). Therefore, Three Visions is a fictional creation. But does it follow that it is necessarily a merely fictional entity, or can Three Visions exist actually?
Many assume that individuals that find their origin in a fiction are "encapsulated" in that fiction, or at least in the world of fiction. Such originally fictional individuals have not migrated from the real world into a fiction (as can be said of Napoleon, who figured in Tolstoy's fiction War and Peace), but are pure inventions of the literary imagination. When we tell children that Dracula is only a story, we build our words of comfort on the assumption that what is purely imaginary is not real. Even stronger, it is not just that Count Dracula does not happen to exist actually, it is commonly assumed that he cannot actually exist because he is an originally fictional individual. The credo is that what happens in the fiction, stays in the fiction. This Encapsulation Thesis underpins many of our beliefs and theories about fiction, and I think it is a philosophically interesting thesis.
Encapsulation Thesis: originally fictional individuals cannot exist in the actual world The possibility of fictional creations such as Three Visions presents us with potential counterexamples to the Encapsulation Thesis.
Note, the Encapsulation Thesis, as it underpins beliefs and theories, does not rule out that Dracula actually exists as a fictional character. Some authors have wished to claim that fictional characters actually exist, as some sort of cultural construction (e.g. Thomasson 1999). Such realism about fictional entities is compatible with the idea that Dracula is harmless to us, because according to the realist Dracula can actually exist as a fictional character. The Encapsulation Thesis should be read as denying that Dracula only actually exists as a blood-guzzling monster. And it is only things like monsters that we should be afraid of, not fictional characters.
Two other points of clarification are in place. First, fictional creations are works written, created or invented within a fiction. We should not confuse the notion of a fictional creation with the notion of a work of fiction. The latter denotes stories or narratives that, for aesthetic effect, do not conform to how things are. Works of fiction belong to the genre category "fiction" (for a good discussion of the concept of a work of fiction, see Friend 2012).
Second, it is also not the case that all fictional creations are works of fiction. In Mallock's novel, a certain Doctor Jenkinson delivers an improvised sermon to a flamboyant congregation. We should count that sermon as a fictional creation, assuming that Jenkinson wrote it, according to the story, and that Jenkinson himself is fictional. But given that it is a sermon, what Jenkinson wrote is not fiction. Furthermore, it is also not true that all fictions that occur within fictions are fictional creations. Robin Le Poidevin uses the term 'embedded fiction' to refer to fictions within a fiction (1995,227). As an example, he mentions Ronald Harwood's The Dresser, which is a play containing some scenes from Shakespeare. Those scenes are fictions also according to the story of the play, but they are not presented as written by someone who just belongs to the story-they are presented as being written by Shakespeare. They are embedded fictions, but not fictional creations.
With these clarifications in mind, we should accept that some embedded fictions are fictional creations. Examples are easy to come by. Think of the stories told by the fictional protagonists of Arabian Nights or of the play The Murder of Gonzago (or The Mouse Trap) that is staged in Shakespeare's Hamlet. But other works of fiction, such as Shakespeare's Hamlet itself or The New Republic written by Mallock, are clearly not fictional creations. They are actually authored novels or plays.
I claimed that the possibility of fictional creations presents us with potential counterexamples to the Encapsulation Thesis. It seems obvious that works written or composed in a fiction can be printed, recited, or performed inside that fiction. Many stories are about people performing plays, or singing songs, where those plays or songs find their origin in the fictional stories themselves. I have already mentioned Three Visions (a poem fictionally recited) and The Murder of Gonzago (a play fictionally staged), but additional examples are easily found. Think for example of the Vinteuil Sonata, a fictional musical work that is performed in Marcel Proust's In Search of Lost Time. There does not seem anything strange or incoherent about imagining an entirely fictional person composing a sonata and playing it to an audience, or conceiving of merely possible Oxford undergraduate students writing sonnets for their tutor (see also Hayaki 2009).
To my mind, it seems just as obvious that such works can be performed to an audience outside the story. And here I do not mean to an audience in another fiction, as when a fictional entity that originates in one corpus makes a cameo in another fictional work, as it were. No, I mean that a fictional poem, play, or piece of music could just as well be recited or performed by actual people, to an actual audience, at an actual poetry night, say, on the Monday evening of March 2nd, 2021.
On the face of it, Three Visions seemed a merely fictional poem, just as Dracula, a vampire, is not a "real" vampire. Yet if we saw someone on a stage reciting Three Visions from memory, giving voice to its lines and expression to its sentiments, would we not be willing to accept that we were listening to a reading of Three Visions? Or, closer to your present situation, would you not accept that the sonnet I reproduced a few paragraphs back was Three Visions? And is this not just the poem composed by that unknown student of Mr. Rose in Mallock's story? I think the answers to these questions should be affirmative. But this means that a fictional individual can, and in fact has, transferred into the actual world.

II. AUTHOR SHIP IN FICTION
At this point a critic could object that we have been too quick in our attributions. No eighteen-yearold Oxford undergraduate wrote Three Visions. Mallock did. Of course, this critic-is it you?-will say, "It is true that in the fiction that Oxford undergraduate composed the sonnet, but in reality it was Mallock who did so. " If what this critic claims were true, then Three Visions would have originated in an actual act of composition, one undertaken by Mallock. And that would mean Three Visions is not a fictional creation at all. It would be Mallock's poem, a poem actually written by an actual writer, only to be embedded in a fiction. And of course, reproducing or reciting a poem that was actually written by an actual writer does not count as a counterexample to the Encapsulation Thesis. This objection needs to be addressed.
It is true that the thought that Mallock composed Three Visions may seem attractive at first. But any such attractiveness disappears once we pull apart some ideas that are easily confused.
To begin, it is no general truth that in order to write a story about a work, you also need to author that work. True enough, all poems have some author, just as all paintings were painted by someone. But sometimes these authors or painters are merely fictional. Just as someone can paint a picture of a book without writing a book, a writer can write a story about a painting without ever touching a brush. This is just to reiterate that there can be fictional creations: that some stories can be about originally fictional works of art, where an originally fictional work of art is a work of art that was not actually created. That there can be fictional creations in this sense is uncontroversial.
Accordingly, the critic must have something more specific in mind. What the critic must have in mind is that Mallock at some point sat down to arrange and write down the specific collection of words that constitutes Three Visions. Mallock's novel does not only refer to and describe Three Visions, the novel incorporates the very words and lines that make up the poem. Therefore, the critic urges, Mallock's act of writing those lines could not have been anything else than the act of authoring Three Visions.
But this is not true either. First of all, there is no general principle that says that whenever someone arranges and writes down a collection of words that form a poem, they are also and thereby the author of that poem. If I write a biography of Oscar Wilde and I decide to write about his poem Ravenna, perhaps letting one of Wilde's close friends recite portions of it in my narrative, I may also find myself writing lines of verse while drafting my manuscript, in much the same way Mallock did. Yet it is clear that I am not the author of Ravenna.
Of course, the critic might say that this again misses the point. Wilde's poetry already existed well before biographies about him were written, and so Wilde's biographers solely reproduce his poems when they overtly incorporate them in their biographies. In the case of Three Visions, the critic insists, with Mallock's act of composition of those lines the poem was authored, because prior to Mallock's writing those lines, Three Visions simply did not exist.
This response brings us to the heart of the objection. Ultimately, the critic relies on an illegitimate move from what is actually true to what is true according to a fiction. This is because the critic must assume that when Mallock wrote the lines that make up Three Visions while drafting his novel, he created the poem because no one else had yet done so. But that no one else had done so is only partly true.
It is true that no one had actually composed the poem Mallock wrote down. This should have been obvious all along, given that Three Visions so clearly seems to be a fictional creation. However, what the critic must assume as well, and without good reason, is that, if no one had in actuality composed the poem, then no one had composed that poem according to the fiction Mallock was working on. This is clearly false. The poem Mallock included in his manuscript had been composed by one of Mr. Rose's students (at least according to the story), and this student had done this (again, according to the story) long before Mr. Rose produces the piece of paper and recites Three Visions to Miss Merton. It is perfectly consistent to suppose that by the time Mallock came to write Mr. Rose's recitation of the poem, it was true according to Mallock's story that this poem was written by one of Mr. Rose's undergraduate students. What Mallock did was commit to paper a poem that, according to his story, had already been composed by Mr. Rose's undergraduate student.
We may suppose that all poems have some author and may suppose that according to the fiction one of Mr. Rose's students wrote Three Visions. We may also suppose that no one actually wrote Three Visions before Mallock wrote down the lines that make up Three Visions. But from these assumptions it is logically invalid to infer that Mallock himself was the actual author of Three Visions. For all we know no one in actuality wrote the poem-something that should not even be surprising, given that Three Visions is a fictional creation.
Of course, this does not mean that Mallock had no role to play in the story's fictional creation. He obviously did. It was Mallock who imagined The New Republic's story about this group of aesthetes in the first place. So it should be clear that Mallock's writing the lines that make up Three Visions was only possible because of an overarching (and aesthetically impressive) imaginative project. And with regard to the poem specifically, Mallock's creative contribution was to write a story that was in part about Three Visions. In writing this story, he ended up writing about the poem too, and by reproducing the poem when drafting Mr. Rose's speech to Miss Merton, even made Three Visions quite explicitly part of a fictional world.

Fictional Creations • 483
It is only in this broader sense that Three Visions originates with Mallock, as part of a novel he wrote. And it is for this reason that in many situations we would be right to give Mallock credit for it. Ultimately, it was a fictional world that sprung from his imagination. But this does not make Mallock the author of Three Visions. As it happens, Otho Laurence, who in the novel represents Mallock himself, at some point even says, "I don't think I ever wrote any original poetry" (1877,203). (This happens when Lady Ambrose is reading out one of Laurence's own published poems, which Mr. Rose points out is in fact a translation of Euripides's address to Artemis of Hippolytus.) Where does that leave us? I have suggested we can actually recite or print a fictional creation such as Three Visions. If one can actually reproduce a fictional creation-it does not matter in what formthen a fictional individual can exist actually. But the Encapsulation Thesis entails that no fictional individual can exist actually. All ficta are supposed to be encapsulated in the world of fiction.
No doubt it remains prudent to get children to believe the Encapsulation Thesis, as it certainly makes them sleep better. But as a philosophical thesis about fiction, the thesis is untenable.

III. INDIVIDUALS, REPE ATABLES, SCRIPTS
Is that the end of the Encapsulation Thesis? There may seem to be another way out. Let us consider this for a moment.
Someone critical of what I just said could object that all I have identified is the appearance of fictionto-reality transfer, and not the real thing. They think that a poem we actually recite can at best resemble the poem Mr. Rose's student wrote, even resemble it very closely indeed. But resemblance is not yet identity. Despite appearances, it may not be the same poem. We may of course call the actual poem by the same name as the fictional poem, and even believe that it is the same poem, yet the critic I here foresee brings in that, nonetheless, it simply cannot be the same poem.
How can they make this clearly metaphysical conjecture with confidence? This may be because we readily conceive of plays, of works of art more generally, as individuals. Indeed, I assumed that Three Visions, as a work of art, is an individual, be it one that finds its origin in a fictional world.
It would not be entirely misplaced to emphasize that fictional creations are individuals. Any individual is an entity uniquely placed at a point in time and space. It can only travel, grow, be moved, or reproduce itself into places and times connected with a point it occupied. Assuming that each fiction, or perhaps the world of fiction as such, comprises an entirely separate historical setting in its own right, it seems impossible for originally fictional individuals to leap from fiction into reality. This is because there is no continuity between these two realms. Because a purely fictional narrative unfolds in a distinct universe, so to say, there are no spatio-temporal paths between such a fictional universe and the actual one, and so the history of a fictional individual can never extend into the actual world.
These metaphysical presuppositions about the individual might explain the critic's confidence. And there is something to this. But the observation can only be confidently developed as an objection to my argument if an overgeneralization is made. And I see no reason to accept that overgeneralization. Let me explain this.
It is worth bringing out what is right about the observation. Individuals, at least as they are commonly conceived, are intimately tied to their historical setting. Think of buildings. Buildings are built at some point in time, and at specific locations. Such spatiotemporal facts about a building are essential to it. If we know of a building x that it is located at a time and place in which building y is not located, we can safely conclude that building x and building y are distinct. The Japanese Hausu Ten Bosu is a theme park known for its replicas of Dutch houses. But no matter how hard the park's designers try, they will never manage to build one of the buildings erected in Holland during the Golden Age. Any degree of resemblance between them is simply irrelevant to questions of numerical identity. And there is a symmetry here: for the same reason, an originally fictional building cannot ever exist in the actual world.
However, a play is not just any sort of individual. It is a repeatable individual or, in P.F. Strawson's way, a type (1959,231). What is philosophically significant about types is that multiple "tokens" of the same type can exist simultaneously in entirely separate locations without any difficulty. When this happens, we may say that a single thing (the type) is found in many places (the locations of its tokens).
Think of how both of us can at the same time have the £5 note in our wallets. Here we have one individual (i.e. a type, the £5 pound note), existing simultaneously in entirely separate locations.
Given that poems and plays are types, it could be thought that their repeatability liberates them from their historical situation. Put differently, it could be thought that the objection presented above can be defused by pointing out that poems and plays are not just individuals, but repeatable individuals, and so it simply lies within their nature that they can exist both in and outside a story. Yet I think this response would be too quick. The current objection nicely brings out that point. For, even types, as repeatable individuals, seem to be confined to their historical setting in a way that still seems to confirm the Encapsulation Thesis.
Take an etching, for instance. Etchings are types, repeatable individuals. Two museums can have the same self-portrait by Rembrandt in their print collections, because both may have been able to get hold of different impressions (tokens) made from the original copper plate. Yet, there is no way you can hang a print in your house pulled from an etching done by a merely fictional character. Why not? Well, for it to be such a print, the paper on your wall would have had to have touched a merely fictional etching plate. Impossible! Actual touching requires just that kind of historical path that is unavailable to a creature of fiction. And this I think is the grain of truth in the Encapsulation Thesis.
So, the critic does seem to have a point. Earlier I suggested that there seems to be no good reason why some actual poetry lovers could not burst out and recite Three Visions, or include it in their collection of Victorian sonnets. The critic now thinks to have such a reason: individuals, even repeatable ones, are tied to their own historical setting in a way that forever confines them to this setting. It seems that individuality trumps repeatability. This is almost entirely correct. Yet it overlooks the role of notation or scoring. At least on a plausible and widespread understanding of repeatable individuals, not all instances of them depend on such moments of physical touching, on a spatiotemporal path from original to print, cast, or reproduction. As we typically conceive of them, all repeatable works allow for further instances of those works to be generated either by mechanically copying from existing instances (e.g., from an original or mold), or else by producing instances by following a set of guidelines or recipe (cf. Walters 2013). Accordingly, even if we cannot mechanically copy a fictional creation, we may still be able to generate instances of it by following a set of guidelines or recipe, if these are available. (I should note that this does not require assuming a Platonism about types, according to which types are a kind of universals, such as developed by Julian Dodd (2007). Universals also allow their instances to be spatiotemporally isolated, but in addition they exist without being created. Types, as I conceive of them, are created individuals.) It is important to keep sight of the fact that notation or scoring makes it that created types such as plays, poems, sonatas and the like are a special sort of repeatable individuals. They are allographic, to use Nelson Goodman's term (but without accepting the specific theory of work identity he builds on top of it). Even when fictionally created, allographic repeatables can be printed, performed, or recited in real life, as long as their fictional scores or scripts are accessible. And for some fictional creations, such scores or scripts may be accessible. An author, for example, can write about them in a novel. Ultimately, then, notation trumps individuality.

IV. RE AL-WORLD IMMIGR ANTS
I have shown that the Encapsulation Thesis faces counterexamples. This is not because fictional creations can be repeatable individuals, but, more specifically, because they can be repeatable individuals of the allographic sort. That is, they can be individuals created at a specific point in a fictional history that allow for instances to be generated in light of a set of rules or following a procedure. This liberates them, at least potentially, from their original fictional setting. All that is required is that someone (fictional or not) pick up the task of following the rules or instructions, or of executing the relevant procedures. This entails that fictional creations can break free from their spatio-temporal setting in a way that is impossible for other sorts of individuals, such as people, buildings, and etchings.
If this is right, then potentially some of the individuals we can find in the world around us may have originated in a merely fictional world. To stretch somewhat the terminology suggested by Terence Parsons, it is possible for the actual world to contain "immigrants," individuals that came into existence due to what happened in a merely fictional story (1980, 51-2). Native to some fiction or other, such immigrants would have found their way into the real world because someone generated an actual instance of them, for example by performing a fictional play, or reproducing a fictional poem.
I should emphasize, the range of possible real-world immigrants is restricted. At least on the basis of what I have said so far, we have only reason to point to possible candidates that fall into the specific allographic class of repeatable individual. My argument does not set Dracula free, for instance, and neither the portrait painted of Dorian Gray. In the next section I want to consider a way of pushing the argument one step further.

V. EMBODIMENT AND T YPE E XISTENCE
As I formulated the Encapsulation Thesis, it maintained that fictional individuals do not and cannot exist in the actual world. This is untenable as a general thesis, because, as I have shown, we can find counterexamples. My argument so far thus only builds on examples of such real-world immigrantsby reproducing Three Visions here I have released the poem into the actual world, thereby establishing the point. This is enough to undermine the Encapsulation Thesis, which denies the possibility of realworld immigrants.
Yet if we replace one of my argument's premises with a stronger version, we can use the argument to establish that some fictional creations actually exist regardless of whether anyone actually reproduces them or not. In other words, it would show that simply by figuring in a fictional work in a certain way is enough for a fictional creation to exist as a work in the actual world.
This may seem surprising, but the point is easily illustrated with an example. The Murder of Gonzago (or The Mouse Trap) is a fictional play that is performed in Hamlet. Moreover, we may assume that the script of Hamlet gives enough information about The Murder of Gonzago for an actual theatre company to stage the play in an actual theatre. Actually performing the play would be enough for the play actually to exist: a repeatable work actually exists if some instance of that work actually exists. But could it be that it existed even before that, just in virtue of the completion of Hamlet, the play that contains the play? Although it may seem far-fetched at first, I think there is some reason to take also this seriously. To see this, it is helpful to look more closely at the premises of my earlier argument, as it applies to a play like The Murder of Gonzago: 1. The Murder of Gonzago is a fictional individual. 2. The Murder of Gonzago, a fictional individual, is a play. 3. Plays are repeatable individuals. 4. If one can actually reproduce a repeatable individual, then that individual can actually exist. 5. We can actually reproduce The Murder of Gonzago, a fictional individual. 6. The Murder of Gonzago, a fictional individual, can actually exist.
Recall, the Encapsulation Thesis implies that, say, Sherlock Holmes, because he is a fictional detective, cannot actually exist, at least not as a detective. Similarly, it implies that The Murder of Gonzago, because it is a fictional play, cannot actually exist, at least not as a play. And so earlier I used (6) to undermine the Encapsulation Thesis: 6. The Murder of Gonzago, a fictional individual, can actually exist. 7. If the Encapsulation Thesis is right, then a fictional individual cannot actually exist. 8. Therefore, the Encapsulation Thesis is not right.
The reasoning here is straightforward. But if we could replace the fourth premise with a conditional involving actual existence in its consequent, instead of merely possible existence, we would be able to use the argument from (1) to (6) to conclude that The Murder of Gonzago exists, period. The amended argument would then run as follows: 1. The Murder of Gonzago is a fictional individual. 2. The Murder of Gonzago is a play. 3. Plays are repeatable individuals. 4*. If one can actually reproduce a repeatable individual, then that individual does actually exist. 5. We can actually reproduce The Murder of Gonzago. 6' . Therefore, The Murder of Gonzago does actually exist.
So potentially the argument leads to an even bolder conclusion. It requires us to make the stronger assumption (4*), and this may seem incredible at first. Could the sheer possibility for performance really bring repeatable works into existence? I think there is something to be said for this assumption. It turns on the interpretation of "can reproduce. " If by "can reproduce" we mean that the materials are available to us, that we have access to an exemplar or script or score, or can reasonably believe we have reconstructed (part of) that script or score, then the assumption is acceptable. By such lights, The Murder of Gonzago can be reproduced. For those in possession of a copy of Hamlet do have access to at least part of the play's stage directions and script (and, we may assume, to enough of it).
Some authors seem to be sympathetic to this way of strengthening the argument. Lee Walters, for example, writes that "types, even created types, can exist when they have no tokens," (2013, 476) and goes on to suggest that all that is needed is that there is an "embodiment" of the type: It is the existence of the embodiment of a type that enables the production of (further) tokens of the type. What's important for the existence of a type is, then, the possibility of generating tokens of that type-that is, that some embodiment of the type exists. (463) An embodiment, as Walters understands it, is not itself a token, but some individual that counts as standard or rule for the generation of instances or tokens of a repeatable work. As examples he mentions both pre-existing exemplars, and recipes for producing tokens of the type, such as musical scores. Embodiments can be stored in various ways: in museum storage, on film, or in human or computer memory (462).
Together with Walters I think that the idea behind (4*) captures well the persistence conditions of repeatable artworks as experts and laypeople alike conceive of them. Whether a piece of music exists or not depends on the availability of its score, regardless of whether the work in question is ever performed. Whether such a score, or a theatrical script, is accessible clearly does not turn on the actual existence of any token performance or staging.
If this is right, then given that Shakespeare's Hamlet exists, and given that Hamlet contains an embodiment of The Murder of Gonzago, we must conclude that The Murder of Gonzago already exists as a work of art.

VI. KRIPKE AND GONZ AGO'S MURDER
Before concluding I want to bring out how my conclusions contradict a point emphasized by Saul Kripke. In his John Locke Lectures from 1973, Saul Kripke maintains that the characters in a fictional creation such as Three Visions or The Murder of Gonzago at best exist as fictional fictional characters; they are not or perhaps cannot exist as fictional characters. A fictional character, according to Kripke, is actually a socially constructed entity, even though such entities can according to a story be, say, human beings or tropical islands. Kripke's point is that Gonzago (a main character of The Murder of Gonzago) is only a fictional character according to the story of Hamlet. And so Gonzago does not actually exist as the socially constructed entity a fictional character would be, but only according to a story exists as such a socially constructed entity. He writes the following: Only in the play Hamlet, or let's suppose so, is it said that there is such a play as The Murder of Gonzago. If so, we can say that there is no such fictional character as Gonzago. Here we are not reporting on what is in the play, because the play does say that there is such a fictional character as Gonzago. We

Fictional Creations • 487
are speaking now about the real world. There is in fact no such fictional character as Gonzago, though the play pretends that there is. There is, however, a fictional fictional character called 'Gonzago' . This is true in virtue of the existence of the play Hamlet. (Kripke 2013, 72-3) Kripke thinks that the mere fact that Hamlet contains a fictional play that tells us about Gonzago does not entail that Gonzago actually is a fictional character. Stronger, even, he seems to think the mere fact that according to Hamlet there is a play about Gonzago rules out that "Gonzago" can refer to any actual fictional character. Just as there is no possible blood-guzzling creature named "Dracula," Kripke thinks that there is no possible fictional character which we call "Gonzago. " I think that here Kripke falls prey to the kind of overgeneralization I discussed earlier. He assumes that because The Murder of Gonzago is a fictional individual, it cannot exist in the actual world. But as I have brought out, The Murder of Gonzago is not just any fictional individual. It is a repeatable, one that allows instances of it to be generated via a script and stage directions. Assuming that Hamlet makes that script and directions accessible to a real-world performer, The Murder of Gonzago surely could exist as an actual work of theatre, or, if we embrace the stronger premise (4*), in fact already so exists.
Even by Kripke's own lights, characters of actually existing plays do themselves actually exist as fictional characters. And I suppose Kripke would agree that from this it follows that characters of possibly existing works of theatre possibly exist as fictional characters. And so, even if the existence of Hamlet, as a work of fiction, does not entail the actual existence of The Murder of Gonzago, it at least entails the possibility of that play's actual existence. This implies that there is a possible fictional character named "Gonzago," contrary to what Kripke insists.

VII. CONCLUSION
I have wanted to draw attention to a specific puzzle about fictional creations. I argued that, given a plausible understanding of the nature of created types, the possibility of fictional creations has ramifications for our conception of fictional entities. Instead of being necessarily encapsulated in the world of fiction, some fictional individuals can and perhaps even do exist as real-world immigrants in the actual world. The examples of such immigrants I have discussed are all allographic and repeatable fictional works of art.
Although it is difficult to draw a sharp line, it is clear that there are more potential real-world immigrants. Some of these will be marginal. Others may come to figure prominently in our lives. Think of types of artifacts first invented in works of science fiction, and only decades later assembled in real life. When people say of something that it is as if it came straight out of a story, they could in fact be right.