What is the Avatar?

Front matter

—

WHATIS THE AVATAR?

Rune Klevjer What is the Avatar?

Contents

Preface: to this editioheciscnieiedeinatomaina a 1 GoldenEye Q07 seau nans a ede nee een ee tbe ts 12 PICHON? 4. sAscieatdeacihe nhieu ahah bed dharani nb a a 16 AVATAR LNG OLY scrssssceccsvecseussestesstesaes tush a a Neat ava E Ea E a EE e E 17 Everyday virtuality. c.c:s.ccccsscsccedeeccsdelecticsponsetbededesceadsntsrdelestichectovigenedengesdarerdeteaaya deere 18 Acknowledgments to the original edition «0... esesecessecseseeseeseceeseeseeseeeeseceeseeeeenee 20 Chapter 1: Introduction 0.0.0... cc ccscsesesesescscsccescseseseseseseneseseseseseeeteeeseaees 21

Chapter 2: Simulations, games and make-believe The WOrldS:Of Play cscs. sieschisassionaitcecsiteestichaseschsneacentesveaeaavecntiaeses World as the diegetic 0.0... esssssesseesssseeesseseeesscssscsccesseeceeesees Fictional subjectivity: fiction as simulation ...0... eee eeeeeeeeeee Work World Snine eiia i avenachadomennaal aca: RECO MUNG t e ee a ea E E ET Eae aesae Theparadoxf play nienie e ea EEE Eataa

First-person EXPERIENCE’ sirieias aa ii 49 Cybertext isic e E E E E 53 LUGOIOGY APE EEA EE EEEE 56 Textplay oiia eira ar e EEEE Ee aas E iS i aad 58 Half-RO Ale iicn a deed unseat aed A E EEEN 60 The. CUPSOF theory. aininn a i aa ni 69

Chapter 1: Introduction

In this study I will suggest a theory of the avatar in singleplayer computer games’, and discuss how avatar-based games are different from other kinds of computer games as well as from other kinds of media. I believe that a closer study of the role of the avatar will cast light on some of the central aesthetic parameters that structure both how we play and why we play computer games. This knowledge will also have potential implications for our understanding of other formats and genres in the wider domain of digital media and culture.

The notion of the avatar that I am suggesting is not concerned with playable characters as a vehicle of communication and self-expression, but addresses how players engage with singleplayer gameworlds through fictional and vicarious embodiment. This approach connects computer game avatars to a broader category of avatars, from radio-controlled model planes to Lego men and paper dolls. In computer games, the vicarious body can take different forms; a character, a racing car, a rolling ball, a camera, a gun.

The emphasis on the role of the avatar also reflects a theoretical concern with the notion of embodiment and the notion of fictionality in computer games. My general claim is that the concepts of ‘fiction’ and ‘representation’, as these are typically being employed in game research and analysis, should be critically discussed and revised. Game studies needs a new concept of fiction, which can better account for the relationship between play and simulation, and which is more sensitive to the distinctive characteristics of computer-simulated and screen-projected gaming environments. This re-orientation needs to question the assumption that ‘fiction’ is synonymous with recounted (or diegetic) fiction, and it needs to be able to address the different mechanisms of embodied interaction

1 While it is common to differentiate between ‘video games’ (console games) and ‘computer games’ (PC or Mac), the two terms are also sometimes used interchangeably. Used as a general term covering games on all platforms, | have chosen in this thesis to use computer game’ rather than the more widespread term ‘video game’. As | will be arguing in chapters 3 and 5, the central element that distinguishes computer games from other types of games is the computer, not the (‘video’) screen.

What is the Avatar?

and corporealized pleasure that are involved in computer game play. My analysis of avatar-based play in computer games is meant as a step in this direction’.

The computer game avatar, as I will define it in this thesis, exploits the digital computer’s unique capacity for realistic simulation, and acts as a mediator of the player’s embodied interaction with the gameworld. The relationship between the player and the avatar is a prosthetic relationship; through a process of learning and habituation, the avatar becomes an extension of the player’s own body. Via the interface of screen, speakers and controllers, the player incorporates the computer game avatar as second nature, and the avatar disciplines the player’s body.

However, the computer game avatar is not to be understood as a tool or a mouse cursor; it gives the player a subject-position within a simulated environment, a vicarious body through which the player can act as an agent in a fictional world. This vicarious body is not merely a mediator of agency or ‘interactivity’ in a general sense, but belongs to and is exposed to its environment. In other words, an avatar is interesting and playable not just because of what it makes us able to do or perform, but because of what happens to us in the world that the avatar lets us inhabit. The avatar is the embodied manifestation of the player’ engagement with the gameworld; it is the player incarnated.

There are different kinds of avatars, and different forms of avatar-based play. A central concern in this thesis is with the difference between the 2D avatar and the 3D avatar, as well as with the various ways in which the relationship between the player and the avatar can be configured in different types of games. The 3D avatar is the more radical and ambitious variant of the computer game avatar. The central perceiving body of the 3D avatar, I will argue, is the navigable camera, which situates the player perceptually within a gameworld that is no longer flat, and no longer a miniature. This camera-avatar (or avatarial camera) brings avatar-based games closer to the aesthetics and discourses of cinema and Virtual Reality. However, in games, spatial continuity and visual realism has a different role to play. The goal of visual realism in avatar-based 3D is not to imitate cinema or to make cinema interactive, but to give the player realistic agency within the gameworld. On the other hand, the strict disciplining of avatarial embodiment, and the embracing of the simulated cinematic camera as a prosthetic perceptual apparatus, can only be seen as counterintuitive and inhibiting from the point of view of fully immersive VR.

2 Myapproach here differs from, and partly conflicts with, my earlier ‘In Defence of Cutscenes’ (Klevjer 2002), which was, broadly speaking, an attempt to analyse narration in games from the point of view of communication and rhetoric. While this effort was not entirely unproductive, it suffers from a lack of theoretical tools to describe the basic operations of fiction in simulation-based game environments. Also, my discussion pays little attention to the fact that different genres of games follow very different principles and mechanisms of fictional participation.

Chapter 1: Introduction

The generic category of ‘avatar-based’ singleplayer computer games does not refer to a clearly delimited group of games, which any given game would either belong to or not. It defines a particular kind of play, a game form, which is centred around avatarial embodiment as the primary mediator of interaction with the game space. In this sense we can talk about ‘weak’ or ‘strong’ avatars, depending on the relative importance of the interaction that is performed independently of the avatarial prosthesis. As a generic form, the avatar has a central position in the contemporary computer games market, especially on consoles, which have been more or less designed for this particular mode of play’.

I will also argue that avatar-based play is a form of make-believe, a fictional form, which defines a particular way of participating with simulated environments and fictional worlds. Considered as a fictional form, avatar-based interaction has grown out of and is embedded in computer games and computer game culture. On one hand, the avatar can be thought of as relatively independent from computer game play - as a principle for interacting with computer-simulated environments in general. On the other hand, the notion of avatarial embodiment that I am suggesting implies that as long as there is an avatar, fictional participation cannot be detached from play. Without a gameworld, or at least some kind of playworld, the central motivation for vicarious embodiment falls away.

My emphasis on the significance of avatar-based play in computer games is not meant to support any notion of ‘invisible’ (or ‘embodied’ for that matter) interface design. Avatarial embodiment is a particular way of fictionalising play and interaction. From the point of view of human-computer interaction, this fictionalisation is rooted in the principles of reification and concretisation. As I will be arguing in chapters 5 and 6, this process is antagonistic to other and more transparent forms of human-computer interaction, including many kinds of computer games’.

My aim is not to argue that avatar-based play is more valuable or more artistically interesting that other forms of play - and certainly not that it should be a general norm for human-computer interaction — but to describe what avatarial embodiment is, and how it structures our interaction with game spaces. The analysis does not aim to produce rules or guidelines for computer game design, but it may still carry some implications for how to think about the role of the

3 This tendency excludes, notably, the Nintendo DS’ touch-screen interface, as well as, possibly, the innovative hardware interface of the upcoming Nintendo Wii.

4 Theconcept of reification that I will be using in chapter 5 overlaps with but is far more limited that the Marxist and more socially oriented concept of reification. At the same time, it can be argued that reified forms of computer interaction, when applied as a norm for interface design, also have problematic ideological implications. See Kirkpatrick (2004:53ff) for a discussion of reification (or ‘double reification’) as a paradigm for human-computer interaction design.

What is the Avatar?

avatar within the overall player experience; what kind of play and what kind of make-believe are we designing for if we choose to build a design around the vicarious embodiment of an avatar?

The concept of the avatar that I am suggesting here is different from how the term is typically used. In computer game discourse, the most general and accepted meaning of ‘avatar’ seems to be synonymous with ‘playable character’, in all its diverse aspects. However there is still an important common ground between this general meaning and my own definition, namely the basic idea of embodiment or ‘incarnation’, which connects computerised avatarhood to the original religious meaning of the concept.

The dominant definition of the concept of the avatar in computer game discourses originates in the tradition of role-playing games’, but its typical use has been expanded to include also distinctly non-configurable and ready-made playable characters like Mario and Lara Croft. More narrowly even, and further removed from my own use of the concept, ‘avatar’ is also sometimes used to refer to the playable character as a mediator of communication and self-expression in multi-user virtual worlds. This use of the term - which refers to the ‘virtual persona’ of the user or player — was introduced by the massive multiplayer online game Habitat (1987:103), and popularised through Neal Stephenson’s influential 1992 novel Snow Crash.

The notion of the ‘avatar’ as prosthetic vicarious embodiment, in contrast, emerges from the tradition of action, racing and action adventure computer games, from Spacewar! (Russel/Graetz/Wiitanen 2006[1962]) and Donkey Kong (Nintendo 1981), via Super Mario Bros. (Nintendo 2004[1985]) and The Legend of Zelda (Nintendo 2004[1986]), to Tomb Raider (Core Design 1996), Halo (Bungie 2001), Gran Turismo (Polyphony Digital 1998) or Super Monkey Ball (Amusement Vision 2002). In all these games, the relationship between the player and the gameworld is rooted in the principle of prosthetic and vicarious embodiment.

The category of the ‘action adventure’, as I will be using it here, is broader than how the term is typically used in the gaming press and gamers’ communities. In those contexts, ‘action adventure’ (2D or 3D) is usually taken to be a very specific genre, which includes as part of its definition a significant proportion of puzzle-based challenges. While this definition of the genre includes Prince of Persia (Broderbund Software 1990), Tomb Raider and their followers, it does not include

— as my own broader category does - platform games, First Person Shooters or

5 Thenotion of the ‘avatar’ was introduced in the Ultima role-playing game series with Ultima IV: Quest of the Avatar (Origin 1985). In role-playing games, the ‘avatar’ is the player’s customisable on-screen character or persona in the game.

Chapter 1: Introduction action-oriented role-playing games, all of which do not necessarily put a lot of emphasis on puzzle-solving®.

Fighting games and action-oriented sport games are left partly on a sidetrack in this study, even if they definitely belong to the broad family of ‘avatar-based’ games. This relative neglect is mainly the result of the analysis’ central focus on avatar-based 3D as a specific form of avatar-based computer gaming. Fighting games like Ready 2 Rumble Boxing (Point of View 1999) or Dead or Alive (Temco 1998), and action-based sport games like FIFA 06 (EA Canada 2005a) or NBA Live 06 (EA Canada 2005b) do not adopt the navigable camera as part of the player-avatar relationship in the same way that the 3D action adventures and racing games do. This means that the distinction between two-dimensional game spaces and three-dimensional games spaces, which is one of the central concerns of this thesis, becomes less important. It also means that, from a historical point of view, sport games have not to the same extent gone through a marked transition between a 2D and a 3D ‘era’; with respect to avatar-based play, the difference between FIFA International Soccer (Extended Play 1993) and FIFA 06 is of less significance than the difference between The Legend of Zelda: A Link to the Past (Nintendo 2003[1991]) and The Legend of Zelda: Ocarina of Time (Nintendo 1998).

As noted above, the role playing Avatar - somewhat paradoxically — is not my object of study. Role-playing games have their own specific characteristics and their own history, a topic that would need a dedicated study beyond the scope of this thesis. Nor do I discuss in much detail the particular characteristics of action- RPG’s like Diablo (Blizzard 1996) or Fable (Lionhead Studios 2004) - games that marry the stats-oriented play of role-playing games with the prosthetic extension of the avatar. For the purposes of this study, in other words, the action-RPG genre is treated as a sub-genre of the broad action adventure genre rather than as a sub-genre of role playing. The distinctive role-playing elements of these games may serve to put avatar-based play and avatar-based fiction into sharper focus; one of the defining features of the prosthetic avatar is precisely that it does not depend on role playing or character customisation, and that it structures the relationship between the player and the gameworld in a different way than the role-playing avatar does.

6 Next to the action adventure, racing is the other major genre of avatar-based play among today’s computer games. Compared to the action adventure, racing in its various forms — and here | would include also games like SSX (EA Canada 2000) or Tony Hawk's Pro Skater (Neversoft 1999) — representa relatively ‘pure’ form of avatar-based play, placing considerably less emphasis on narrative, dramatic and cinematic elements. This makes racing games a good case for illustrating some of the basic principles of avatarial embodiment, but makes them less relevant to the study of hybridisations and tensions between avatar-based interaction and other forms of play and fiction.

What is the Avatar?

My analysis of avatar-based play is restricted to singleplayer games, and is motivated by a specific interest in singleplayer games as a particular form of play andasa unique type of gaming experience; the focus is on how the avatar mediates between the player and the game, not how it mediates between the player and other players. This implies addressing the relationship between the player and the game system, between the player and the simulated environment, and between the player and a fictional world. These concerns are of course relevant to any kind of game, but in multiplayer and online environments the social interaction of play nevertheless demands primary attention.

Methodologically, the study implies a relatively wide sweep of empirical observation and analysis. I cannot of course aim to get a complete first-hand knowledge of all games that could be relevant to the concerns of the analysis. In trying to capture the essentials of avatar-based play and its major variations, my strategy has been to play through a limited number of popular games that have been recognised as classics in some respect within their (sub-)genre, and then add to this by playing a larger number of games only briefly (1-4 hours of play). I have also been able to draw on fan-based knowledge from the numerous reviews, FAQs and walkthroughs that exist for practically every game that exists out there. Such written sources are particularly valuable in this kind of broad descriptive-analytical research.

The research process itself has evolved in a relatively disorderly hermeneutical fashion. I started out with a few games that were perceived to be central to my concerns, with no clear idea as to what would be the conceptual and generic boundaries of the study. Originally, the project started out as a study of a more specific genre within the action adventure umbrella: the First Person Shooter. I eventually discovered, however, that the aspects that this genre shares with a wider category of games needed a more dedicated focus, particularly because very little research has been done in this area. While the FPS definitely boasts a few interesting ‘special features’ that are specific this particular genre, those features would be difficult to describe without positioning them within a more general generic domain. Of particular importance to my change of emphasis from the FPS genre to avatar-based play as a game form - and to avatar-based 3D in particular — was the realisation that the ‘first-person’ camera, which is supposed to be the ultimate marker of the genre (hence its name), is not different in its basic functioning from the camera in Tomb Raider or Super Mario 64 (Nintendo 1996). In other words: the most distinctive feature of the FPS is not the perspective but the gun.

In this chapter I will discuss the differences and overlaps between games, simulations and diegetic fictions as distinct cultural forms. Drawing on the theories of Kendall L Walton, Marie-Laure Ryan and Gregory Bateson, I suggest a concept of fiction that is based on simulation and play, and which provides analytical tools to distinguish between different modalities and principles of participation in different media. In particular, the idea that fictional participation is a practice of re-positioning or ‘re-centring’ — whether in the worlds of paintings, books or children’s games of make-believe — is particularly useful for rethinking the relationship between players, avatars and fictional worlds in computer games.

Procedural representation

Any implementation of a model is a simulation. A model is, following the standard military use of the term, “A physical, mathematical or otherwise logical representation of a system, entity, phenomenon, or process”. This definition covers all models, from scientific climate models to Barbie dolls or model cars. A simulation represents the world not through description, narration or argument, but through the construction of a model that is meant to in some respect mirror the underlying regularities of selected phenomena, events and processes in the world.

Espen Aarseth argues that simulation is a highly distinctive type of discourse, quite unlike other forms of communication. He calls it a ‘virtual hermeneutic’, emphasising that the simulation represents an ever more influential alternative to the established story-mode of understanding the world. The simulation represents, Aarseth argues, a distinct way of interpreting and understanding the world (Aarseth 2000). From the point of view of semiotics, Umberto Eco, even if he is not addressing simulations directly, notes that certain forms of sign-relations are established on the basis of an ‘identity of function’; a broom handle stands for a horse not via an iconic relation, but because it can be straddled. A broom

7 See Online M&S Glossary, Defense Modelling and Simulation Office, at https://www.dmso.mil/ public/resources/glossary/ [accessed 25. July 2005].

What is the Avatar?

handle can be used as a substitute (‘Ersatz’) for an actual horse because it “serves an analogous function” (Eco 1976:209).

A simulation re-constructs (in some respects) a behaviour that we can recognise as familiar from the world outside the simulation. It does so by using models, which are functional or in some way ‘logical’ representations. Models can be abstract (a mathematical model) or concrete (a tin soldier). Traditionally, models can only be implemented by humans, who run simulations for the purpose of play, training and cultural expression (as with Eco’s example of the broom handle). However, algorithmic abstract models are built as a system of instructions and procedures that can be implemented by computers as well as by humans. Computer simulations are simulations that are run by computers (which are simulating machines or ‘simulators’), or by humans and computers in concert. Taking a cue from Janet Murray’s categories in Hamlet on the Holodeck (1997), we may say that a computer simulation is a ‘procedural representation’; the world interpreted in terms of a system of instructions or rules (Klevjer 2002:200)’.

The worlds of play

The majority of games are in fact also simulations (although not necessarily computerised simulations), because they are governed by sets of rules that model some phenomenon in the world. Chess, for example, models the conflict between two forces on a battlefield (the board). As a model it may leave a lot to be desired in terms of scope, detail and accuracy, but it still qualifies as an abstract, procedural representation. This does not mean that all games are simulations, or that all simulations are games. Simulations do not necessarily stage or prescribe a contest between the participants. They do not necessarily, as games do, model a conflict,

8 Aconcrete model is “A model in which at least one component represented is a tangible object; for example, a physical replica of a building.” For definitions of terms of simulation and modelling as they are used by the US military, see Online M&S Glossary by the Defense Modelling and Simulation Office (2005).

9 Hamlet on the Holodeck focuses on narrative structure and narrative agency in ‘cyberspace’ — that is in digital environments of all kinds, from sprawling databases to VR installations. All these computerised environments have, Murray suggests, four essential properties. They are ‘procedural’, ‘participatory’, ‘spatial’ and ‘encyclopaedic’ (1997:71). My own adaptation of the concept of ‘procedural representation’ is also a more generalised version of the term as it can be found in the field of computer graphics. See David D. Grossman's “Procedural Representation of Three-dimensional Objects” (1976). Salen and Zimmermann (2004) make use of the same concept, expanding considerably from the basic idea by discussing various implications for game design.

and they do not have to define a goal for the participants in terms of a winning or loosing condition”.

Nevertheless, both simulations and games establish a separate realm of activity that is governed by a set of formal procedures. The activity is motivated in the external reality in which they exist as sub-systems, and the activity may even have very serious implications beyond the boundaries of the system itself (think of for example Russian roulette, gladiator contests or military simulations). The important point is that this relationship will always be, as Jesper Juul (2005) points out, ‘negotiable’; the participants define the real-world consequences in advance. They can do so because games and simulations are autonomous systems of meaning with clear (although permeable) boundaries, and therefore are meaningful in themselves; Russian roulette may well be played with harmless blanks, and military simulations are often enjoyed on the couch, with cheese doodles.

With games as well as for simulations, their significance in relation to their contexts is premised on the fact that they possess a basic autonomy. This autonomy, according to Johan Huizinga (1955[1950]), is no less than the historical and aesthetic essence of all kinds of competitive play; all artificially staged contests. The key quality of play is irreducible, Huizinga argues; its meaning cannot be attributed to any purpose outside play. Play is meaningful in itself. Autonomy and non-instrumentality is at the heart of the ‘play function’ in culture, a principle which can be traced in all human activity through history. The essence of play, states Huizinga, with reference to the religious and ritualistic practices of pre-modern cultures, is the encapsulation of imagination and conflict within a magic circle. The magic circle signifies a separate realm of internally defined meaning, a ‘world’ of objective and shared truths within which the participants make serious intellectual effort and emotional investment. This is the familiar paradox of games - they mean nothing (because a game is just a game), yet seen from the inside of the magic circle they mean everything.

Play becomes serious, sometimes even deadly, not in spite of but because of its characteristic as a separate realm, according to Huizinga. The magic circle is a ‘sacred circle’, rooted in the rituals and contest of archaic cultures.

The arena, the card-table, the magic circle, the temple, the stage, the screen, the tennis court, the court of justice, etc., are all in form and function play-grounds,

10 In literature on games well as in the newer literature on computer games there are a variety of definitions as to how games are different from playful activity in general. All of them in one way or another emphasise the importance of conflict, as well as a winning condition or alternatively a ‘negotiable and quantifiable outcome’ (Juul 2005). See Salen and Zimmerman (Salen and Zimmerman 2004) fora review of definitions found in the literature.

What is the Avatar?

i.e. forbidden spots, isolated, hedged round, hallowed, within which special rules obtain. All are temporary worlds within the ordinary world, dedicated to the performance of an act apart. (Huizinga 1955[1950]:19)

It is precisely the meanings guaranteed by the principle of the magic circle that make the practices of ritualistic contest, warfare, law, poetry and philosophy possible. Modern computer-simulated systems tap into this same cultural realm of meaning. Through their self-contained detachment, they bring forward the autotelic ‘play function’ in (or of) culture. This cultural heritage partly explains why we will always encounter, with any participatory dynamic model, the latent temptation of non-serious play; the invitation to fool around. Computer games do not abandon the didactic and scientific rationale of abstract models, but re-contextualise this rationale within the cultural realm of play, turning latent temptations into shameless pleasure. We could say that computer games are simulations in reverse: they draw on our familiarity with the world to empower us within the simulation rather than using the simulation to empower us to handle the world. The primary function of modelling in computer games is to provide a playground, a material magic circle, a pointless system of meanings.

It is because games and simulations present themselves to us as autonomous meaning-making systems that we sometimes refer to them as worlds. In everyday language, when given no further qualifications or specifications, the term ‘world’ will usually refer to something like ‘the totality of our existence’. The degrees and modalities of metaphorical meaning vary from the presumably literal (‘world’ meaning our planet) to the more abstract (the ‘world’ of dating). All variants and shades of the term, however, resonate with the basic notion of totality; a presumed all-encompassing boundary, an outer rim that conditions meaningful practice. In the ‘world’ of dating, behaviours that would be ridiculous outside that game can still be perfectly meaningful within it. Equally, when we say that a person ‘lives in his own world’ it would typically mean that his behaviour (or a certain aspect of it) does not make sense to us. In order for something apparently meaningless to be able to make sense after all, it needs to bring its own world - its own sub-totality, its own magic circle — along with it. Given this premise, as Huizinga’s detailed historical accounts illustrate, pointless exercises can generate all sorts of serious meanings and consequences in their interfaces with the outside world, and typically will do so — even if they do not depend on such consequences to be meaningful and engaging.

Contest versus mimesis

The ‘world’ of the sacred circle captures a cultural modality that is common to games, simulated systems (whether serious or not) and to play in general. Still, imaginary worlds of literary or cinematic fiction do not easily fit into this picture. There seems to be, ona general level, a natural affinity between Huizinga’s ‘magic circle’ and Coleridge’s ‘suspension of disbelief’. Both refer to an experience of — somehow - stepping into an alternative reality, a separate and encapsulating realm of meaningful experience. However, the concept of the magic circle does not address the phenomenon of imaginary worlds — let alone the worlds of narrative.

While Huizinga’s magic circle allows for fiction and fantasy as a natural element of playful activity, the phenomenon of make-believe is not considered part of the core of play; it is not integral to the play-principle of culture. This principle, or cultural function, is an agonistic principle; its paradigmatic and originating form is to be found in the ritualistic contests of archaic culture. In Homo Ludens (1955[1950]), Huizinga never addresses the role of drama or storytelling as such, seen as separate from the functions of the contest. All play-derived civilising practices that he discusses throughout the book, including music and poetry, are described in terms of how they instantiate and develop the agonistic principle in culture. At heart, the ‘magic circle’ is a realm of artificial or ‘staged’ contest, not fiction.

In privileging the principle of the contest, Huizinga’s theoretical perspective mirrors Roger Caillois’ classifications of play in Man, Play and Games (2001 [1961]). Caillois focuses on play and games as concrete activities rather than any abstract ‘function’ or ‘principle’ in culture, and he is therefore interested in the mimetic as well as the agonistic elements of play. However, Caillois stresses the inherent conflict and incompatibility between agonistic and mimetic play”. If they mix, he argues, mimetic play will unavoidably ruin agonistic play, and vice versa. Huizinga, in contrast, never argues that the contest should be seen as incompatible with mimetic play, probably because he does not pay any attention to the ‘worlds’ of imagination and fiction at all. In any case, we can use neither Huizinga nor Caillois to argue that the ‘world’ of a game is similar to the ‘world’ of a novel ora film.