Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

There is an obvious plausibility to the claim that videogames are fictions or involve fictive elements. The Elder Scrolls IV: Oblivion represents a world in which players battle goblins, explore ancient ruins, and collect treasure. In reality I have never done any of these things; rather it seems fictional that I have done so. Oblivion appears to be a work of fiction in much the same way as traditional fictive works, in depicting a world that has no actual existence but rather is merely imagined to exist. There may be alternative ways to describe these aspects of videogames, however. In a short conference paper on fiction and virtuality in videogames that has drawn a considerable amount of attention in games studies, games scholar Espen Aarseth claims that some elements of videogames are not fictional, but present virtual or simulated items (Aarseth 2005). The apparent fictive aspects of Oblivion referred to above may be virtual rather than fictional.

What is the status of the objects represented in videogames: are they virtual or fictional? Are videogames works of fiction? Indeed, exactly what are we committed to if we claim that videogames, or the items depicted therein, are fictions? In this paper I will attempt to assess the thesis that videogames are fictions, arguing there to be strong and modest versions of the claim. I will defend a modest sense of the thesis from arguments of the type forwarded by Aarseth, showing that these arguments depend on various confusions about the nature of fiction. The theory of fiction, developed particularly in the last 25 years within the analytic philosophy of the arts, proves to be very illuminating when focused on videogames. I also offer a conceptual reconciliation that characterises the virtual worlds found in videogames as a representationally and interactively rich species of fiction, explaining the intuitions that motivate referring to videogames as depicting virtual items, but not leading us to the perplexing conclusion that the goblins, ruins and treasures of Oblivion are not fictional.

2 Videogames and Fiction

There are a number of potential variations or confusions in the thesis that videogames are fictions. First, we need to distinguish this videogames as fictional thesis from the more specific claim that videogames are always instances of the genre of interactive fiction. It seems increasingly reasonable that the fictions seen in videogames do count as distinctly “interactive” (Tavinor 2009; Lopes 2001; Smuts 2009). The genre of interactive fiction, however, is a type of fiction in digital and non-digital media that sets out a branching narrative; the Choose Your Own Adventure books popular in the 1980s and the videogame Zork being prominent examples. Though it is true that some videogames or aspects of videogames are of this genre form of interactive fiction, it is also clear that most modern videogames that have ostensible fictional elements are not comprised of branching fictional narratives. In Microsoft Flight Simulator, one fictionally flies an aircraft; in Oblivion, one fictionally battles goblins and vampires. Though there are a number of narratives in Oblivion, they do not typically have the branching structure of the interactive fiction genre.

Similarly, to argue that a videogame is a fiction is not to be committed one way or another over its status as a narrative. Of course, the issue of whether or not videogames are narratives has been of particular interest to many games theorists (Murray 1998; Poole 2000). Narrative is a concept that recent writers have used to refer to any number of items, so much so that the term is now apt to strike many readers as being almost vacuous (Livingston 2001). But in a classical sense, narrative—or the near synonym “story”—seems to be a formal feature of certain representational artefacts, perhaps amounting to how they structure their content into a temporal arrangement providing a point of view—often, but not necessarily, that of a narrator—that motivates and guides an interpretation of that material. As I will argue later in this paper, the fiction/non-fiction distinction is a fact of the pragmatics of how depictive content is used, particularly with respect to what it is meant to refer: in the case of non-fiction, some aspect of the real world, and in the case of fiction, situations with an imagined existence only. Thus defined, narrative and fiction are conceptually distinct in that though there are clear examples of fictive narratives, many narratives are equally clearly not fictional—historical narratives, for example—and some (indeed many) fictions are not narrative in form, an example being a painting of a fictional landscape. Therefore, to claim that a videogame is a fiction is not necessarily to claim that it is also a narrative; the fictive thesis to be argued for here is thus largely independent of narrativism concerning videogames.

There are clearly a weak and a strong version of the thesis that videogames are fictions. Though there are certain complications with the following analysis, to be clarified later, here I will define a work of fiction as one in which the characters, places, events, objects, and actions referred to are fictional rather than real. A strong fictive thesis might claim that videogames are essentially fictions in that they necessarily depict fictional characters, places, objects, events, and actions. Potentially, one might claim that videogames can be defined in terms of their fictive qualities (Tavinor 2009: 23–25). Definition often comes in the form of a set of conditions that are claimed to be necessary and sufficient for an item to count as a member of the defined class. A strong fictive thesis might amount to a definition of videogaming in which it is claimed that fiction is a necessary condition. It need not be argued that this fictive condition is sufficient all by itself to make an item a videogame, and given that fiction is shared with a great many non-videogames this is not something we would ever want to claim. Hence, such a definition might also pick out other necessary features that are jointly sufficient to make an item a videogame. But claiming fiction to be a necessary condition of videogaming is still a particularly strong thesis.

Furthermore, it is quite obviously too strong. Phrasing the strong videogames as fictional thesis in terms of a necessary and sufficient condition definition gives us an obvious method to refute it: to find a counterexample of an artefact that is a videogame but is nevertheless not a work of fiction in not depicting fictional characters, places, objects, events, and actions. One of the very first videogames, OXO, provides such an example. OXO is a videogame version of the traditional pen and paper game tic-tac-toe and it seems no part of this game that it presents a fiction. Rather, OXO is a “transmedial” form of tic-tac-toe (Juul 2005: 48). Similarly, videogame chess, Sudoku, and solitaire do not seem to present a fiction that one is playing these games in the sense that Oblivion presents a fiction that one is fighting a goblin or exploring an ancient ruin. Again, these seem to be transmedial forms of games that originated in non-digital media, that is, real chess, Sudoku, and solitaire played in a computer setting. Thus the strong fictive thesis—that videogames essentially involve fictive elements—is immediately prone to refutation; indeed I am just not sure that anyone has been bold enough to assert such a thesis.

In fact, it is simply not necessary to hold the strong thesis in order to argue that videogames are often fictions, or are even usually so. A weakened form of the videogames as fictional thesis might state that while videogames often involve fictive elements, they do not necessarily do so. Videogames are sometimes works of fiction, and sometimes not. Moreover, this modest thesis might claim that if one is to define games, fiction need not count among the necessary conditions. Videogames are clearly not monolithic, but involve a range of structural and media qualities, and engage their players in a number of different ways. The modest fictive thesis allows us to retain the initial plausibility of videogames being fictions, noted in the opening of this essay, but to avoid the obvious counterexamples to the strong thesis.

The weaker thesis, of course, might also be appealing to those who simply doubt that videogames can be defined in terms of a necessary/sufficient condition definition. But equally, it means that offering an example of a videogame, or an aspect of a videogame, that is not fictional, does not refute the thesis that videogames are sometimes or even usually works of fiction. For this reason, the weak thesis seems very reasonable; indeed, almost unexceptionable. Games theorist Jesper Juul seems committed to some form of the modest thesis, given his use of fiction in explaining how games such as Simcity, while not sitting squarely within his “classic games model,” count as videogames nevertheless (2005). For Juul, videogames seem to be a game-fiction hybrid. Elsewhere I’ve formalized a similar theory, asserting a weak version of that claim that games are fictions by providing a disjunctive definition of gaming where fiction is not counted as essential, but is seen as characteristic of a partial range of videogames (Tavinor 2008; Tavinor, 2009: 15–33).

3 Fictional vs. Virtual

In his widely discussed paper, Aarseth seems to challenge some form of the fictive thesis by claiming that some of the apparent fictive elements in videogames—dragons, doors, mazes—are not fictional but instead virtual items. I should point out here that Aarseth does not acknowledge either the analysis of fictive works or the strong/modest fictive distinction I have described above, so I cannot make conclusions on his behalf. Aarseth’s claims are principally about the ontological status of game worlds and objects, and not videogames as works. Indeed, Aarseth seems happy to see some aspects of videogames as being fictional—though, as we will see, in a rather idiosyncratic sense of “fiction”—and at strongest his claim might be that “the category of fiction is problematic when applied to ‘game content’” (2005: 1).

What I want to do here is to develop Aarseth’s arguments and question what the arguments themselves really establish concerning the status of videogames as fictions. I think there is the basis in Aarseth’s arguments on which to develop a response to even the modest fictive thesis described above. I have claimed that a work of fiction is one in which the characters, places, events, objects, and actions referred to are fictional rather than real. Aarseth’s arguments show that there is an alternative way to characterize these ostensible fictional elements of videogames, in that we could refer to the goblins, ruins, and treasures found in Oblivion as virtual elements. If the aspects of videogames identified in the beginning of the paper and that led to the initial plausibility of the fictive thesis do turn out to be virtual rather than fictional, then this would constitute something of a challenge to even the modest thesis. Videogames in this case would not be works of fiction, but virtual works or simulations.

Aarseth argues that “game worlds and their objects are ontologically different from fictional worlds” (2005: 1) by which I take him to mean that the depicted elements in games have a different mode of being to those depicted in fictions like novels and television shows. His argument notes a number of key differences between some depictive elements found in videogames and traditional fictions and infers from these differences that certain videogame elements are not fictional. Referring to a difference between the dragon Smaug in Tolkien’s The Hobbit, and a dragon as represented in the videogame EverQuest, Aarseth notes that the former “is made solely of signs, the other of signs and a dynamic model” (2005: 2, emphasis in the original). It seems clear enough that the claimed difference between Smaug and the dragon in EverQuest is here principally in terms of their representational media: one is represented through propositions and pictures, and the other through these things and a dynamic 3D model.

Furthermore, because of the dynamic model, the EverQuest dragon makes possible a number of modes of engagement that Smaug does not: “Simulations allow us to test their limits, comprehend causalities, establish strategies, and effect changes, in ways clearly denied us by fictions, but quite like reality” (2005: 2). Virtual objects “can typically be acted upon in ways that fictional content is not acted upon” (2005: 1, emphasis in original). This claim seems for the most part true, because computer games do involve their players in forms of engagement that are quite different to those seen in fictions such as novels and television shows. With a literary work such as The Hobbit, our participation with the work is limited mostly to following or interpreting the set of fictive details that comprise a fixed narrative. Such works also call on appreciators to fill out the story with imaginative and imagistic detail, perhaps imagining the dragon to have very particular qualities that are not referred to by the fictional work. In a videogame like EverQuest, however, we play the game, and in doing so we seem to interact with the dragon depicted therein: we might battle it, run from it, and so on. Moreover, the dragon is more richly depicted that a literary dragon: wondering what the dragon looks like, a videogame player does not have to imagine these facts, rather he might just more carefully inspect the dragon, perhaps by moving around it to get a better view.

I think the argument here is most credibly understood in the following way. In fictions such as The Hobbit, the objects are depicted in such a way that they do not allow for the reader to have an effect on the fiction, because for one thing, the depiction is unresponsive. A goblin in The Hobbit is depicted by linguistic descriptions, and its qualities are largely fixed by the act of Tolkien’s authorship (though, again, readers will no doubt imagine the qualities of the goblin in different ways). Because of the depictive features of the literary goblin, the reader’s relationship to the goblin seems distanced and “one way”: a reader can read about its qualities, and she may be cognitively or emotionally affected by it—being curious about its qualities, or perhaps disgusted by it—but she cannot interact with the goblin. One reason for this is patently clear: the goblin does not really exist, because literary goblins are imaginary goblins. And so, as a number of philosophers have noted, readers and viewers of fictions are ontologically separated from the objects and events depicted in fictional worlds; and indeed, this gives rise to certain puzzles about why we react to them as we do, such as how we can have emotions for the characters and events depicted in fictional worlds when we know they do not exist (Radford 1975; Walton 1978).

But in videogames, the media through which objects such as goblins are depicted, including dynamic graphical models, allow players to perform actions on the depicted object, of a kind denied by regular fictions. The player of Oblivion can battle a goblin, and formulate strategies about how to defeat them; indeed, doing so is a prerequisite for performing well in the game given how goblin-filled this particular gameworld is. Moreover, where the non-existence of objects in literary fictions is abundantly clear, in the case of videogames there does seem to be something there with which I am interacting: in Oblivion, my eyes track the movement of the goblin, and I manipulate my character toward it so that I can strike it with my sword. The interactions which are crucial to the playing of videogames quite clearly demand that there exists something with which to interact: and on the evidence of how we describe our interactions with videogames, it is tempting to say that we interact with goblins. If we take this analysis of the situation seriously, then the “ontological gap” that seems so clear in traditional fictions may not exist with videogames and other virtual worlds. That no such gap exists would certainly fit with Aarseth’s conclusion that “game worlds and their objects are ontologically different from fictional worlds” (2005: 1).

To emphasize this claim, Aarseth describes a number of virtual artefacts that seem to engage participants in modes of interaction that are denied by their fictional counterparts. In a key example he argues that the labyrinths or mazes found in computer games often have a real existence that those depicted in fictions do not. The maze depicted in the final part of Kubrick’s The Shining is clearly fictional, whereas the labyrinth depicted in a game like Pac Man is not. This is because the latter depicts a maze that you can actually trace, and your success at playing Pac Man demands that you do so while avoiding the ghosts who are also navigating the maze. The Shining, however, only presents cinematic glimpses of the labyrinth as a location for the action of the film’s narrative. In fictional depictions of mazes, the claim is, the viewer or reader is never presented with a labyrinth she might actually trace, and of course doing so is never a precondition of experiencing the fiction. But if “If a 2D drawing or a painted or tiled floor can be a proper labyrinth […] then a 3D virtual labyrinth in a computer-simulated world is a real labyrinth since it can be navigated by the by the same rules as the one at Hampton Court” (2005: 3). Again, the interactive potential of the videogame artefact is claimed to set it apart from a fictional counterpart.

The final example I will look at here is of virtual doors, and indeed later we will find that it is an ideal example to really grasp the issues here. Aarseth contends that there is a difference between virtual and merely fictional doors. Discussing the first-person shooter videogame Return to Castle Wolfenstein, Aarseth notes that “only some of the doors in the game work as doors should. Most of the doors are merely textures on the walls that look like doors, but whose function is purely decorative. Other doors actually do behave in a door-like manner; they can be opened, closed, seen through, walked through and fired through. Clearly, these two types of door are very different….” (2005: 3). The purely decorative doors, claims Aarseth, are fictional, while the doors that can be opened and walked through are not fictional doors, but are virtual or simulated doors. Again, the claim is that an interactive difference—the virtual doors function as doors, allowing for the egress to other virtual spaces—constitutes a difference that sets these items apart from fictional doors. Fictional doors, for example “the hatch” that constitutes an important part of the action in the first two seasons of television show Lost, cannot be used by viewers of the fiction to access new areas of world that is depicted: the viewer of Lost simply looks on passively as the trapdoor is used by the characters within the fiction. Viewers of the TV show are ontologically distanced from the hatch and the fictional world in which it exists.

I think it is clear enough that there is a genuine distinction that has been located in these examples: but Aarseth intends to draw a rather strong ontological conclusion from it, in that the nature of the distinction concerns the “mode of existence” of virtual objects, which he claims is different to the mode of existence seen in both fictional and real items (2005: 4). He concludes that “there are at least three different ontological layers to game content: the real, the virtual and the fictional” (2005: 4). This seems to me a very adventurous conclusion to make on behalf of the representational media of videogames, and it would be easy to dismiss this ontological claim if Aarseth was alone in these strong ontological intuitions. In fact, a significant number of new media theorists have been similarly tempted to claim that virtual worlds have implications for our understanding of ontology, and it is not uncommon to discover claims that virtual items have a unique mode of existence that might alter our conception of what is real (Wertheim 1999; Heim 1993).

4 Fiction and Prop-Based Make-Believe

Do the depictive and participative features evident in videogames and other virtual items establish that the items depicted therein are not fictional, and hence, that videogames involving such depictions are not works of fiction? Obviously we need some clear idea of what fiction really is. Unfortunately, Aarseth does not supply a clear explanation of what he takes the concept to signify, relying quite oddly on a rather poor definition drawn from Microsoft Encarta that takes fiction to be comprised of:

  1. 1.

    novels and stories that describe imaginary people and events; and

  2. 2.

    something that is untrue and has been made up to deceive people (2005: 2).

It is obvious why this definition is tempting for Aarseth. The former clause defines fictions partly in terms of their media (novels and stories) and partly in terms of their imaginary nature. The second defines fictions as lies. Neither seems apt to describing videogames (which are clearly not novels or stories, or lies) and hence videogames are not fictions.

But this, simply put, is an awful analysis of fiction. Aarseth actually begins his paper by criticizing previous theories of games as fiction for using the term fiction without qualification, but he then notes that he will “not engage” with fiction theories such as those from Thomas Pavel and Kendall Walton (Aarseth 2005: 1). But this is exactly what Aarseth and others need to do if they are to make a credible claim that videogames or their depicted objects are not fictions. Aarseth is not the only games theorist to have an under-developed theory of fiction: Miguel Sicart’s study of computer game ethics lacks preciseness on the concept, and seems to equate the fiction of a videogame with its “visual” or “narrative” elements (2009: 21, 24–25). Even though his theory has a good deal of detail on the role of fiction in videogames, Jesper Juul’s work also lacks a rigorous theory of fiction, equating it, without much detail, with the philosophical notion of possible worlds (2005: 122). Arguably, this lack of a clearly articulated theory of fiction is problematic for the theories developed in both of these works.

The arguments presented in the previous section rest on the assumption that the representational and participative differences seen in videogames motivate distinguishing them from uncontested fictions, implying that fiction is properly characterized by its media and the modes of interaction those media representations support. However, there is clearly a good sense in which fiction refers not to works of fiction or their media existence as novels or stories, but to the imagined scenarios that are presented by such media artefacts. This indeed seems to be the more fundamental sense of the concept of fiction given that such imagined scenarios are both historically and creatively prior to media instantiations of fictions: simple imaginings and oral stories predate films and novels by many thousands of years, and creative imaginings are ultimately the source of the fictions that find their way into fictive works. Hence, we might give a basic analysis of fiction as referring to imagined states of affairs, a sense of the concept that abstracts fiction away from any particular depictive medium. And this analysis would seem to apply to videogames: the nuclear holocaust that is depicted in the post-apocalyptic role-playing game Fallout 3 does not represent actuality, but has been invented through an act of imaginative creation; likewise the goblins and ogres in Oblivion or the characters and city suburbs in Grand Theft Auto IV.

Fiction, under this analysis, turns out to be a fact concerning the pragmatics of representation (Tavinor 2009: 38–44). The typical way to show this is to reflect on the fact that a pair of formally identical representation tokens—portrait paintings say—can differ in their status vis-à-vis fiction. Imagine two portraits, one painted to represent a mythical figure such as Odysseus, but based on the sitting of a model, and the second painted to represent the model himself. Perhaps the sitter for the painting of the mythical figure liked the original painting so much that he requested that the painter repeat the effort to depict him dressed in the mythical garb. In this case, there might exist two formally indiscernible paintings, one which depicts a fictional person, and the other depicting a real person in fancy dress. It is not the media or representational form that makes one fiction and the other non-fiction, because they share the representational form of portraiture; rather, it is fact about their intended function that distinguishes them, a fact which surely has to do with what the painter had in mind when producing the artefact: the fictional work is painted to depict a person with an imagined existence only, the non-fictional portrait to depict a person who actually exists. Thus, a pair of formally and perceptually matched items may be fictional and non-fictional depending on their intended functions.

It is this basic analysis of the concept of fiction that is developed in the philosophical theories offered by Walton (1990), Greg Currie (1990), Peter Lamarque (1996), Lamarque and Olsen (1994), and many others working within the analytic philosophy of the arts. Under these theories, fiction is a classification that depends on the intention with which a depictive artefact is produced and used for the purposes of imagination. Walton argues that fictions engage us in “games of make-believe” that often involve linguistic props such as in novels and short stories, but also involve works of visual art and even sculpture (1990: 63). Likewise, Lamarque points out that there is nothing about the semantic or syntactic structures of fictional representations that make them fictional; rather it is the “fictional stance” that is fostered toward them that determines their fictive status (Lamarque 1996). Many other philosophers have characterized the distinctive nature of our cognitive, perceptual, and emotional attitudes vis-à-vis fictions, though there is naturally a great deal of detail and subtle variation to the accounts (Carroll 1990, 1998; Feagin 1996; Robinson 2007; Scruton 1974).

I have argued elsewhere that Walton’s theory of prop-based make-believe is particularly apt for explaining the fictive nature of videogames (Tavinor 2009). Walton argues that the imaginative games that we play, obvious from childhood onward, are often augmented by fictive props that lend the games of make-believe a richness and seeming objectivity they would not otherwise have. He notes that, “The role of props in generating fictional truths is enormously important. They give fictional worlds and their contents a kind of objectivity, an independence from cognizers and their experiences which contributes much to the excitement of our adventures within them” (1990: 42). To take Walton’s key example, in a childhood game of make-believe where stumps are meant to represent or stand proxy for bears, a large stump might represent a large (and probably ferocious) bear (1990: 37–39). The stump contributes to the game of make-believe by objectifying various facts of the fictional world that the children imagine; and note that in this example, given the physical nature of the prop, it might even allow of the imaginative game that the bear can be fictionally wrestled if the children decide to grapple with the stump.

In the rather more sophisticated games of make-believe that constitute our grown-up imaginary adventures, props take the form of the linguistic inscriptions, pictures, verbalizations, physical gestures, and sculpted forms that comprise the media of the representational arts. And note that there are already participative variations in the traditional representational arts given the variations in their depictive forms. Novels and plays differ to representational paintings in the extent of their temporal duration. Though it takes some time to view a painting such as The Rape of the Sabine Women by Poussin, and to understand its meaning, a novel like David Foster Wallace’s Infinite Jest unfolds over a significantly greater period of time, and so leads to a quite distinctive mode of participation as a reader incorporates more and more information into her reconstruction of the novel’s many details, information that can alter her interpretation over time. Linguistic fictions also demand that appreciators “fill out” the fiction with imagistic detail, imagining for example, the precise appearance of Smaug from the descriptions of the dragon given in The Hobbit. Illustrated versions of the book may more strongly guide these imaginings. Peter Jackson’s forthcoming movie adaptation will no doubt provide a particularly vivid take on the dragon by employing the modern representational means of CGI, and will require less in the way of imaginative involvement from the audience. Representational artworks such as sculptures may even demand physical movement from appreciators, as they move in relation to the object to see the full extent of its form. Thus, fictive props come in any number of different media, and it is commonplace that the differences between their media can lead to different modes of engagement in their respective fictions. But in all these cases the events depicted are clearly fictional and the objects are designed with in intention of grounding the imaginative engagement of an audience. This understanding opens the way for concluding that videogames are fictions, but which similarly have a distinctive media that alters their characteristic modes of participation.

In fact, Aarseth warns against the conclusion that these differences merely make videogames a different kind of fiction: “Of course, it can be argued that the fictionality of Tolkien’s dragon lies in the fact that it simply has no counterpart in reality, and not in the material way it happens to be presented to us in games and stories. In other words, the argument would go, both dragons are equally fictitious, they just happen to be presented in different media” (2005: 2). In response to this, Aarseth notes that simulations can also represent non-fictive things, and that our intuitions about such cases make it hard to sustain the fictive/non-fictive distinction for simulations or virtual items generally. Aarseth notes that many of the events in the first-person shooter Brothers in Arms: Road to Hill 30 are made up of “documentary” (real) events, but that this game is “ontologically similar, and practically identical” with the videogame Call of Duty which is not as closely based on reality, and subsequently to “classify one as fictional and the other as documentary would make little sense” (2005: 2).

Unfortunately, Aarseth’s intuitions about this case arise only because he has failed to connect his examples to the relevant cases from fiction and non-fiction. The “documentary” aspects of Brothers in Arms: Road to Hill 30 that Aarseth thinks distinguish videogames from traditional fictions can also be clearly seen in traditional fictions, and so cannot be a motivation for counting videogames as ontologically different from fictions. The documentary facets in Brothers in Arms are comparable to those in traditional fictions such as the James Bond movies which depict real events such as the Cold War, as the material setting of the fiction. Nevertheless, the exact circumstances represented in a movie about Bond are fictional, and hence are part of a work of fiction, because the story is intended not as a retelling of real events but as a telling of events with no real existence. Similarly, the exact events depicted in Brothers in Arms are no less fictional for the fact they are set in the context of places and events with a real existence. This is the necessary proviso on my earlier claim that works of fiction are those in which the characters, places, events, objects, and actions referred to are fictional rather than real: it is an unexceptionable fact that fictions also often depict elements that are conceptually derived from things with a real existence, both of a general and particular kind: in the first instance they refer to properties or kinds of things with a real existence (war, countries, people), and secondly they refer to individual things with a real existence (the Second World War, Germany, Hitler).

5 Virtual Fictive Props

The claim here then, is that videogames are fictions, and that their distinctive participative features—which might tempt us to conclude that they or the objects depicted therein are different to fictions—derive from the nature of their computational props. To really explain the participative variations that might have led some to distinguish between videogames and other more familiar fictions, we need to carefully examine the nature of their props.

The fictive props seen in videogames are often virtual depictions (Tavinor 2009: 61–85). To see what this means, the concept of virtuality itself needs some analysis. If we look at how the notion of virtuality first entered computer science, we see that the concept calls attention to a functional correspondence between items. Virtual computers, which were common in the early days of computing, exist where a computational program is carried out in a non-electronic medium, typically through pen and paper calculations. Because algorithms are substrate independent, programs can be carried out in any medium where the functional nature of the program is preserved. Thus, pen and paper operations can instantiate the same computational process that is run on an electronic computer. I have argued elsewhere that this sense of virtuality refers to the fact that one object can serve as an interactive proxy for another kind of object because it replicates the functional structure of the target object (Tavinor 2009: 48–51; Tavinor 2011). Indeed, this constitutes one of the core meanings of the concept of virtuality: a virtual war is an event that is functionally equivalent to a war, though perhaps not meeting precisely with some material condition of genuine wars (perhaps by being undeclared).

In this sense a virtual depiction is a depiction that preserves some functional aspect of its target, and so allows for an interaction of the kind one might have with the target object. The most obvious and illustrative instance of this is the virtual camera, a depictive artefact involved in 3D graphics and hence used in many modern videogames. Videogames do not involve actual cameras, rather virtual camera is an idiom employed by game designers to describe a key functional aspect of three-dimensional representation. In particular, the virtual camera is crucial in opening up the possibility of three-dimensional spaces, and allowing virtual movement through those spaces. Alongside polygonal 3D objects, the virtual camera is one of the key developments in virtual representation, and illustrates the definition of virtuality given here in that the structures it employs are the algorithmic transformations of various vector functions of a 3D model. This software function is apt to be treated as a camera, because these algorithmic geometrical manipulations, and their subsequent display on a 2D screen, can be made to match quite closely the changes that would occur if an actual camera was used to film an actual scene. Subsequently the virtual camera finds a host of first-person, third-person and cinematic uses in videogame depiction.

But, thus defined, the concepts of fiction and virtuality are overlapping rather than conceptually opposed. Because I have explained the virtuality of depictions in terms of the interactive structure of their media, and fictionality in terms of the pragmatics of representations, it is clear that virtual depictions can represent real and fictional items. Take the example of Google Street View, an internet application that depicts the topography of real places through the means of photos taken from a camera mounted on a car that is driven through the actual locations, and then stitched together and arranged as a graphical hypertext document with a number of individual 3D scenes. Street View is a case of virtual representation because it allows the user to explore a depictive structure in a way that corresponds to the actual exploration of the place represented, because the depictive structure maps onto the topography of the real place. Clicking on an arrow modifies the “point of view” of the depiction in a way that corresponds to movement through the depicted city. But the cities that the user can thus virtually explore are real cities. Indeed, one could imagine a Street View version of a fictional place such as Liberty City from Grand Theft Auto IV, showing how this virtual depictive application would work equally well with fictional places. The difference, of course, is that the scenes themselves would not be derived from photographs of an actual place, but from video captures of a designed 3D environment.

In videogames, virtuality most often manifests itself where a depiction allows for a kind of interactive involvement that corresponds to an interaction one might have with a target item were it actual, and most often the virtually depicted items are also fictional. It is in these terms that we can address the arguments about virtual goblins, labyrinths, and doors. A goblin in Oblivion is a fictional goblin: no such goblin exists. But the media of its depiction are structured in a way that it responds to the interactions of the player. Technically, it is a 3D polygonal model appended with a collection of fictive affordances (Tavinor 2009: 61–85). Hence, virtuality clearly relates to the notion of “affordances” a term that has sometimes been used to describe how videogames allow for player action (Juul 2005; Cogburn and Silcox 2009). The virtual nature of certain videogame depictions derives from the fact that they afford various modes of interaction. When one fictionally approaches a goblin in the game, the depictions of the game allow for an interaction that corresponds to an interaction that one might have with an actual goblin (complicated in this case by the fact that goblins as a kind are fictional things). Note also that these virtual affordances are often tagged with graphical artefacts so as to make their potential for interaction obvious the player: in Oblivion, as one gets close to various objects, an icon, such as a cross-hair, appears, signifying the potential for interaction with the object.

Similarly, most labyrinths in videogames are fictional but also virtual. Take the maze-like structures of Wolfenstein 3D. The rooms, corridors and Nazis depicted in this game are fictional: no such rooms, corridors or Nazis exist. Nevertheless, the labyrinth in this case is depicted by a 3D virtual model, and a player is able to fictionally and virtually explore the labyrinth because his character’s position is depicted by a virtual camera of which he is in control through his input into the controls of the game. Hence, the virtual space can be virtually navigated, and should the 3D maze structure be sufficiently complicated, the player might become virtually (and fictionally) lost.

This example also calls attention to the fact that the props used to depict fictions often do so in virtue of replicating—in a real or virtual way—the properties they make fictional, because many of the mazes found in videogames are represented by depictions that might themselves quite properly be referred to as mazes. The maze that appears on the screen during the playing of Pac Man is, apart from its virtual medium, more or less identical to a maze one might find in a puzzle book: here the functional correspondence derives from the fact that both kinds of maze are comprised of complicated geometrical configurations. But the maze in Pac Man is also a fictional maze because it is depicted that there are ghosts floating around the maze, and these ghosts are clearly imaginary. Explaining the difference here is that a depiction of a maze might itself count as a maze because some mazes are simply depictions, but a depiction of a ghost is never itself a ghost, because ghosts are quite different from depictions: they are spirits of the deceased!

A similar thing occurs in non-virtual fictive media when a real sentence is used to represent a fictional sentence uttered by a movie character, in virtue of its being a real sentence. But this does not mean that the utterance this depicted is not a fictional utterance: it would be truly bizarre if the fictional sentences uttered by Luke Skywalker were not to be counted as fictional because the actor Mark Hamill used real sentences to represent them! The difference between these two examples is located solely in their depictive media: the maze in Pac Man is a 2D virtual representation, and as such can be virtually navigated, the sentence in Star Wars is linguistic token fixed at the time of the production of the movie, and depicted in such a way that is does not support the interactive functions which actual utterances support. Note however, that in some videogames there are what under my theory would count as virtual utterances, in the form of the dialogue mini-games in Mass Effect, Fallout 3, or Dragon Age: Origins. In these cases the depictive media of the fictional utterances are functionally defined so that the player can have virtual fictional conversations with the characters of those gameworlds. I’ve never had a real conversation with Moira in the game Fallout 3, though my player-character has had a numerous fictional conversations with her (usually about topics such as mole-rats, mines, and radiation sickness).

Finally, and providing another good illustration of the nature of virtuality and its relationship to fiction, are the virtual doors that play an important role in Aarseth’s argument. The genuine difference that Aarseth refers to between merely decorative doors and usable doors does not amount to a difference between fictional and non-fictional doors as he contends, but to a difference between fictional doors depicted in a non-virtual way, and fictional doors depicted in a virtual way. There are no real doors whatsoever involved in Grand Theft Auto IV, but some of the doors in the game, meeting my analysis of virtuality as those cases where an item might stand as a functional or interactive proxy of its target, allow for virtual use because the depictions are structured in such a way to cue an affordance of entering a new virtual space. Incidentally, this example further illustrates the fact that fictionality and virtuality are distinct categories in that it shows that a single fictional item can be depicted in virtual and non-virtual ways. In the cut-scenes in Grand Theft Auto IV, a particular door may be depicted in a non-virtual way in not allowing use. But during the subsequent gameplay, the item may be represented in a virtual way in that it can now be used to exit from the virtual space.

Hence videogames present fictions, but fictions that differ in their media by involving virtual depictions that allow for a kind of participation not seen in most traditional forms of fiction. Though I will not discuss this further here, the virtuality seen in certain videogame elements, because it is defined in terms of a propensity to support the interaction of the player, may be a species of “interactivity,” a concept that has come in for increasing recent philosophical discussion (Smuts 2009; Lopes 2001, 2009; Tavinor 2009). Videogames are thus often virtual fictional works. Characteristic of such virtual fictions are their rich representational media, their responsive nature, and their consequent interactive opportunities. Aarseth’s own account of the relationship is that virtuality is ontologically distinct from fiction. In a way he is correct, but he gets the nature of the distinction wrong. The two classifications are not opposed, but somewhat overlapping. This reformulation of the concepts of fiction and virtuality, as deriving from different considerations—media and pragmatics—but being somewhat overlapping in that there are virtual fictions, accounts for the media differences that Aarseth notes, but does not lead us to reject to very strong intuitions that games—with their goblins, dragons, Russian civil wars, and ghosts—are fictional. Videogames, modestly, are sometimes works of fiction, though of a different kind to many traditional fictions.

Ultimately, of course, it is the interactivity and virtuality discussed above that allows the fictions found in videogames to function as games. Indeed, virtual fictions are ideal for situating games because they allow for the depiction of activities that lack the costs of their real counterparts, meeting one of the familiar criteria of games as being separate or isolated from reality (Caillois 1961; Huizinga 1950). Elsewhere I have developed a theory of how the fictional aspects of videogames—in a game like The Elder Scrolls IV: Oblivion, the environments, characters, monsters, weapons and so on—provide the content that is structured into the rules of the game (2009: 92–102). In Oblivion playing the game is comprised of exploring the environments, trading and conversing with characters, and battling goblins. All of these things are fictional, but they can provide the formal aspects of a game because of their interactive and virtual structure.