This article reviews the history of the concept of the magic circle, its criticism and the numerous other metaphors that have been used to capture the zone of play or the border that surrounds it, such as world, frame, bubble, net, screen, reality, membrane, zone, environment, or attitude. The various conceptions of social, mental and cultural borders are reviewed and identified. Finally, a model is put forward where the psychological bubble of playfulness, the social contract of the magic circle and the cultural game forms are separated.

magic circle, psychological bubble, arena, play, game, boundary of play, safety, game space

Magic circle continues to be a hotly debated term in game studies. The term is intuitively accepted, useful in describing the difference between play and non-play; a handy metaphor that acts as intellectual shorthand for a more complex set of social relations. However, if it is not clear what ‘magic circle’ is shorthand for, then the term can muddle thinking. For example, when taken too literally, it can cloud our understanding of how play is bounded.

The metaphor of magic circle stands for a border that delimits an instance of playing. However, it is only one of many formulations of a border that surrounds and envelopes play, a border that has numerous ethical, legal and practical implications, a border that functions as a design aid, and a border that is relevant for an understanding of what play is. In order to come up with useful theoretical tools, the personal mindset of the participant and the socially negotiated and upheld contract that yields a site of play, need to be separated. Though these psychological and sociological objects are interwoven, it makes sense to separate them for purposes of analysis (for example Montola et al. 2009, 257-278; Stenros 2010). The psychological border set up by adopting a playful mindset and the border set up socially through negotiation often coincide, but they are two different things. In addition, residue of these two, as well as established forms of rule-structured playing create game spaces, sites and artefacts that are culturally recognized as arenas of play – even when they are empty and unused. When playing an existing game the social negotiation often means accepting a predefined set of rules. Confusing the three (mindset, social contract and game space) leads to muddled conceptions of playfulness, play and games.

In the following I shall review the history of the magic circle and its criticism; explore different formulations of the social, mental and cultural borders of play; and, finally, formulate a synthesis view of the boundedness of play.

The concept of magic circle is traced back to Johan Huizinga (1938), who wrote:

For Huizinga, magic circle is a space created for playing, a material or conceptual temporary world dedicated to the act apart. However, though the concept of the magic circle is usually attributed to Huizinga, and the above passage does capture the idea behind how the concept is used even today, he did not dwell on the subject for long. Indeed, the term ‘magic circle’ only appears six times in (the English translation of) Homo Ludens, and only three times in the chapter game scholars usually refer to: in the passage quoted above, once in the metaphoric meaning it has since acquired in game studies (Ibid, 11) and once as an example of a sacred space (as opposed to playful space, though Huizinga stresses that formally such a distinction cannot be made, Ibid, 20). However, in the context of the whole book it is not just an example or a metaphor among many, but a core feature of the examples given (as pointed out by Calleja 2012, see Huizinga 1938, 77, 210, 212).¹

In the book Rules of Play, Katie Salen and Eric Zimmerman (2004, 93-99) developed and defined the concept of the magic circle as it is understood in game studies today. Though the term and the central idea are adapted from Huizinga, the works of Apter (1991) and Sniderman (1999) also contributed to the framing. Zimmerman (2012) later clarified his view of the evolution of the term:

This is important, as there are critics who question Salen & Zimmerman based on what Huizinga wrote, and vice versa (e.g. Egenfeldt-Nielsen et al. 2008, 24-25). The two terms are connected, but the formulations – and especially the ontological contexts of the formulations – are different. Salen & Zimmerman’s (2004) magic circle of gameplay is entered voluntarily, it is self-sufficient, set apart from ordinary life in locality and duration, and it has rules that differ from ordinary life. These features it shares with Huizinga’s conception. However, for Salen & Zimmerman’s formulation the possible development of culture from play is not relevant, nor the alignment of play and ritual. They describe the magic circle as shorthand for “the idea of a special place in time and space created by a game,” (Ibid, 95) and go on to explain why they feel that it is a fitting metaphor; the circle is both finite and infinite, while the magic is in the transformation of meaning within the circle. In their formulation, the boundaries of play are fuzzy and permeable, but the borders of games are more formal (Ibid, 94-95). The magic circle is entered as play begins, or it is generated with that initiation (Ibid, 95). While within the magic circle, a temporary world is created where meaning is handled differently and rules of the game have authority (Ibid, 96). While in the circle, the players adopt a lusory attitude (Ibid, 97; see Suits 1978). The temporary world of the game is an open system and a closed system depending on whether games are framed as rules (closed system), play (open or closed system) or culture (open system) (Salen & Zimmerman, 2004, 96-97). Salen & Zimmerman also provide examples of the blurring of the boundaries from live action roleplay and what later would be called pervasive games, when actions are indexical or the activity is not formally recognized as play (Ibid, 574-579).

This conceptualization of magic circle has faced strong criticism, mostly because many scholars feel that the division between play and ordinary life is ultimately invalid. T.L. Taylor (2006, 151-155), writing in her book Play Between Worlds about exploring online game cultures, questions whether or not the division between game and life needs to be as strong as her reading of Salen & Zimmerman makes it out to be (see also Castronova 2005, 147-160). She sees the discussion about the division between game and life mirroring the discussion about the separateness of online and offline, and calls for non-dichotomous models. Her criticism is irrelevant if virtual worlds are not considered games but spaces where, among other things, play happens. Yet, even as she points this limitation out, she wonders whether or not this is actually a problem with the definition of games.

Thomas Malaby (2007; 2009, 79-106) is on the same track, but words his criticism much more harshly. He questions if a clear division between play and ordinary everyday life exists at all.² He also does not target just the magic circle, but play itself as long as it is understood inherently as separate from ordinary life, safe and free of consequences, and pleasurable (cf. Sutton-Smith & Kelly-Byrne 1984). This criticism is based on the work of virtual world ethnographers like Taylor and Malaby’s own ethnography in Greece amongst backgammon gamblers. “[A]ny game can have important consequences not only materially but also socially and culturally,” Malaby writes (2007, 107), and continues that this finding was very relevant also in non-gambling contexts when status and relationships are on the table in place of money. However, he does consider games “relatively separate,” pointing out that the degree of separation is highly dependent on cultural context. He underlines (Ibid, 111) that games are socially constructed to be “separable to some degree from everyday experience.”

Malaby’s criticism is also reminiscent of what Marinka Copier (2005) wrote a few years earlier, when she drew attention to magic circle as a metaphor; Huizinga’s circle was a sacred space whereas Salen & Zimmerman use a picture of a chalk circle to recontextualize the metaphor as child’s play:

Sybille Lammes (2006) has also criticised the magic circle metaphor; she sees it as a simplification of the relationship between the game and the world. Propelled by Bruno Latour’s work on actor-network theory, she has proposed moving to the term magic node. Jesper Juul (2008) has advocated the term puzzle piece to underline how games interface with the world around them, and Edward Castronova (2005, 147) has proposed that membrane would be a better term for the barrier that separates synthetic (that is, online) worlds from the Earth. Others who have voiced this kind of criticism include Mia Consalvo (2009, 411), who has argued that the magic circle upholds a structuralist conceptualization of games and that it emphasizes form at the cost of function. Especially the context of play is lost, and often context is key in deepening the understanding of instances of play.

Much of the criticism of Salen & Zimmerman’s magic circle seems to stem from their ill-worded explanation behind the choice of the term (2004, 95): “The fact that the magic circle is just that – a circle – is an important feature of this concept. As a closed circle, the space it circumscribes is enclosed and separate from the real world.” The usage of the words enclosed and separate here seems unfortunate, as their larger description of the concept is hardly that closed. It is almost as if this passage has turned the concept of magic circle into a straw man. For example, Malaby’s demand that the nature of games as socially constructed is not really in conflict with what Salen & Zimmerman wrote.

It is hard to find a scholar who has insisted on a strict border between play and non-play. Perhaps the most severe proponent of a clear division was Roger Caillois. He devoted a whole chapter in Man, Play and Games (1958, 43-55) to describing the corruption of games: if play ceases to be free, separate, uncertain, unproductive, regulated or fictive, he no longer sees it as play. A player who cannot properly stay aloof and separate play from non-play is corrupted. Upholding this normative border was extremely important to Caillois: cheaters and professionals destroy play, and in general the corruption of play leads, he felt, to alienation, superstition, violence, alcoholism and drugs(!). Yet even Caillois does not claim that there is a strict division between play and non-play; on the contrary; he claims that there should be a division.

Daniel Pargman and Peter Jakobsson (2008) also attack the “strong-boundary hypothesis” of Salen & Zimmerman’s magic circle, based on interviews and observations of hardcore digital gamers in their homes. Their findings make it difficult to conceive of a clear, closed-off border between play and everyday life. Their separationist reading of Salen & Zimmerman’s model is unfair, yet their criticism hits home when it moves from the ‘circle’ part to the ‘magic’ part. Their interviewees rarely experience any kind of magic or enchantment when playing digital games. Quite the opposite; it is common to play and watch television, or to play for a short time while waiting for the pasta to cook. Though Salen & Zimmerman mostly conceive of the ‘magic’ to refer to the transformation of meaning within play, the connection to sacred is very much present in Huizinga’s original work. Thus, it is more a criticism of the idealization of play than of the magic circle as a separate semiotic domain. Based on Pargman and Jakobsson’s work, it seems clear that there are many instances of digital game playing that are better characterized as mundane than as magical. This is actually in line with Huizinga’s (1938, 20) thinking that there is no formal difference between ritual and play, sacred and playful.

Michael Liebe (2008) and Gordon Calleja doubt the existence of magic circle specifically in digital games, since “the majority of actions possible are programmed into the game system and cannot be changed” (Calleja, 2012). There are three problems with this criticism: first, the interpretation of magic circle is quite narrow. For Liebe and Calleja only the agreement on constitutive rules (what is possible) is relevant, leaving out interpretation of rules, extra-ludic motivations or consequences, player-created goals, etc. Secondly, the difference to traditional games is not as severe as presented here. There are numerous “rules” in sports that an athlete cannot ignore (like the weight of the equipment, law of gravity) (see Montola 2012, 32-47), and – as Calleja points out – digital games also have social rules. Thirdly, this view casts digital games in a narrow light: only the events displayed on a screen (or even just the events within a field depicted on a screen) are within the magic circle (and thus the game). The player and the controller in front of the screen are not considered.

Zimmerman (2012) addresses some of these criticisms in a popular article, and rejects the strong boundary hypothesis attributed to him and Salen. Zimmerman asserts that the core idea of a magic circle is that “games are a context from which meaning can emerge.”³ Indeed, Jesper Juul (2008, 59) has pointed out that many of the critics of the magic circle (he cites Copier, Malaby, Pargman & Jakobsson) “claim to counter Huizinga, Salen and Zimmerman by stressing the exact social nature of the magic circle that Huizinga, Salen and Zim¬merman also stress.” Zimmerman (2012) also claims that the concept he and Salen put forward was mainly meant as a tool for design, and thus it should be evaluated based on its utility.

Despite the criticism, magic circle continues to be used.⁴ It seems to be a useful, powerful metaphor, though it has not been exactly clear what it is a metaphor for. Some of the problems seem to be connected to the idea of games as pre-existing artefacts that players enter into, others to the concept of the magic circle as necessarily having a material form. A particularly lucid new formulation for magic circle has been offered by Markus Montola, who conceives of

This formulation takes a further step away from addressing the relation between play and culture and instead underlines the social nature of the play contract. Conceptions of such a social barrier between play and non-play are quite common, and numerous interesting formulations have been proposed without the term magic circle attached. We shall next review some of these formulations of social boundaries.

The boundedness of play has been postulated not just in game studies, but also at least in philosophy, sociology, psychology, performance studies, library and information studies and legal studies. Metaphors that have been used to encapsulate play or the border around it include: world, frame, bubble, screen, membrane, reality, zone, environment and net. Note that some of these metaphors highlight the border, others the delimited space, though most refer to both.

Philosopher Kurt Riezler (1941) makes two distinctions: a social division between ordinary life and what he calls playworld, and a mental division between serious and playful attitudes.⁵ Playful and serious are opposites in his thinking, and there is a clear separation: “Though man’s mood can move things to and fro over the borderline between play and seriousness, he can not move the borderline itself, which demarcates attitudes, not things.” Riezler sees this playworld as separate from the ordinary, something you enter voluntarily:

In the real world everything is connected in chains of cause and effect, but in the playworld the chains of causes and effects have limits. However, the game can have goals that are connected to the real world, as in gambling or professional sports. Note especially that Riezler also considers the playworld to be a social construct, a “man-made agreement.” He introduces the playworld with these words:

This is the exact quote that sociologist Erving Goffman (1961, 27) cites before summing it up with the oft-quoted line: “Games, then, are world-building activities.” But before discussing Goffman in detail, let us consider Gregory Bateson’s conception of frame.

In the same essay where Bateson (1955) discusses metacommunication and the signal “this is play,” he also introduces psychological frames. These frames delimit what are meaningful actions and as such they are metacommunications. Bateson discussed the frames as psychological, but he also considers how they work in communication as messages – basically saying that they are also social. Though Bateson’s formulation of frame is ultimately a little unclear – perhaps due to its function as a tool in psychotherapy – it has been hugely influential. The idea that there could be a metacommunicative frame that declares “this is play” has been picked up by numerous scholars.

For example, in their discussion of the idealization of play, Sutton-Smith and Kelly-Byrne (1984, 317-318) come to the conclusion that one of the few things that can be said about play is that it is always a Batesonian framed event. They go on to point out that it is essential for the participants to keep in mind that they are playing, “otherwise the activity will break down into anxiety or violence as indeed it often does.” They also list numerous cues that can be used in keeping the playfulness of the frame at the forefront of everyone’s minds: certain types of actions (e.g. exaggeration, repetition), objects, physical scenes, vocalizations (e.g. registers for iconic sounds for cars or babies), characters, and attitudes.

The most well-known extrapolations of Bateson’s frame comes from Goffman (1961, 20): “games place a ‘frame’ around a spate of immediate events, determining the type of ‘sense’ that will be accorded everything within the frame.” Frame is thus social, shared and provides meaning in an encounter, a social situation (whereas Bateson’s frames are more akin to mental representations).⁶ An example of how sense is made in a game context is provided by the rules of irrelevance: during gaming, the participants forswear interest in the value of the game equipment. For example a chess set can be made of cardboard, wood or gold, yet during the game it is treated in the same way. Similarly the players’ background or social status should not influence the playing of the game (Ibid, 19-26). However, certain characteristics of the player do influence the game (Ibid, 29-31) (couples do not play bridge together, age might determine the order of turns, social status might be translated to a ‘better’ character in an RPG). This incorporation of external elements is guided by what Goffman (Ibid, 33) calls transformation rules: these rules tell us what modifications take place as external patterns of properties, which are given expression within the game.

Where the rules of irrelevance tell what is left out of the game frame, Goffman uses the Riezler quote to tease out what is kept in the frame. Games are world-building activities, as they set an “engine of meaning,” which makes it possible for events, roles and identities to emerge that would not be understandable or meaningful in any other frame (like grounding out to third, atari or lawful-good half-elf warrior). However, since Goffman’s interests do not just lie in games, he uses them as a stepping stone to say something about the social world. Unlike Riezler, who conceived of an ordinary life outside playworlds, Goffman sees everyday life as similar to games: only in the context of the street do terms like pedestrian or motorist become meaningful (Ibid, 26-29).

Goffman (Ibid, 65-66) introduces the metaphor of an interaction membrane as the boundary around an encounter. The border around play is permeable; as the wider world passes through more than just application of the transformation rules takes place. Yes, some elements are ignored and repressed, others are transformed, but it is also possible for the external elements to endanger the transformation rules and thus the encounter itself. It is possible for play to be collapsed by external events.

Peter L. Berger and Thomas Luckmann have tackled the delimiting of play from non-play in their foundational text The Social Construction of Reality (1966). They postulate a paramount ultimate reality that can enclose “other realities” which are “finite provinces of meaning” (also Schutz 1945, 551-560; James 1890):

Berger and Luckmann, like Riezler, Bateson and Goffman, postulate a structure constituted by meaning. These other realities are commuted to and from, and play and games are by no means the only finite provinces of meaning of this type; art, theatre, religion and even dreams offer similar enclaves.

Goffman further developed the theory on encounters; in Frame Analysis he again begins by discussing Bateson’s frames, and their application of it to games (1974, 7, 40-43). However, he discards most of the terms he introduced earlier, like encounter and membrane, and instead introduces terms such as framing and keying.

Instead of everyday life, Goffman discusses primary frame as the basis for our interpretation (Ibid, 21-39). Things that we say ‘really’ or ‘actually’ occur, occur in the primary frame (Ibid, 47); this activity is meaningful in its own right (Ibid, 560). The primary frame is itself also a construct, as culture, religion and cosmology influences it for social groups (Ibid, 27) – and Goffman makes a distinction between natural and social primary frameworks (Ibid, 21-22).⁷ The primary framework is not enough to make sense of what is going on; a transformation of meaning takes place. This is referred to as keying, and examples include make-believe, contests, ceremonials, technical redoings, or copies (48-78). Once an activity has been keyed, it can be further transformed by rekeying. Furthermore, keying is not the only way that activity can be transformed. The other possibility is though fabrication, which Goffman defines (Ibid, 83) as “the intentional effort of one or more individuals to manage activity so that a party of one or more others will be induced to have a false belief about what is going on.”⁸

During play numerous frames are present, and the participant has multiple roles he adopts. Goffman (1961, 51) uses the example of a bowler who takes a bad shot, and when he turns back to face his fellow players, he makes a facial expression that signals that the shot was not representative of his skills. An ideal player would not need to send such a signal, but the human participant does so. It is not part of the system of the game, but of the social encounter. Indeed, good sportsmanship is not about playing the game, but about navigating the social frame around it.⁹

The conceptualizations of delimited spaces within everyday life capable of transforming social reality are abundant. For example in the realm of psychodrama, Jacob L. Moreno (1965) has proposed the concept of surplus reality, based on Marx’s conception of surplus value. Surplus reality is a kind of alternative reality, a shared social space, where a group can act out or rehearse painful situations or relations from a participant’s life. Andrew Letcher (2001) has added the concept of temporary tribal zones into the pot. Writing in the context of religious studies, Letcher makes observations about a temporary spatial arrangement. Building on Mikhail Bakhtin’s (1965) carnivalesque and Hakim Bey’s (1985) temporary autonomous zones, Letcher proposes a temporary transformed space through the rules and conventions of a neo-tribe. The transformation is so strong that it creates an illusion of autonomy, though does not actually achieve it.

J. Tuomas Harviainen and Andreas Lieberoth (2012) have compared rituals and games in the context of library and information sciences.¹⁰ They build on the works of Huizinga, Letcher and Bateson, and discuss a local information environment, which both of these cultural forms share. It determines what parts of the real world are allowed to affect the participant’s behaviour. According to them, the separation of the real world and the game or ritual is made possible, from an informational angle, by three key features: “resignification of elements within the situation, increased attention to shared intentionality, and the fact that during such activities, access to information outside of the activity is limited.”

The resignification is very similar to Goffman’s rules of transformation. However, notice that for Harviainen and Lieberoth (as well as for Riezler and Bateson) the border between play and non-play is not just social, but has also a strong psychological element in attention to shared intentionality. Next, we shall move on to considering the border as personal and mental instead of social.

Similar to the sociologists cited above, Mihaly Csikszentmihalyi (1975) does not see games as exceptional. There is an important division, one that Riezler also made, of playfulness and play. Playfulness (a mindset), or flow, can take place not just in play and games (social setting), but also in work. However, it is interesting to note that Csikszentmihalyi’s characteristics of flow are not dissimilar to those of the magic circle; in fact one of them, centering of attention on limited stimulus field, seems familiar in this context (Ibid, 80-82).¹¹ In Csikszentmihalyi’s discussion of the experience of rock climbers, he writes that in contrast to

Csikszentmihalyi also discussed the limitation of the stimulus field in relation to doctors performing surgery, and points out (Ibid, 131) how it is important for the surgeon “to adopt a neutral attitude toward the future of the patient’s life.” The playful activity, flow, is circumscribed from the experience of the normative everyday life.

For Michael J. Apter (1991) play is a phenomenological state. He discussed two mindsets or metamotivational states: telic and paratelic. Telic is a serious mindset, an activity is engaged in for a purpose. Paratelic is a playful mindset, with the activity in itself as the goal (or a goal is adopted in the service of the activity). In constructing a structural-phenomenology of play, Apter also talks about a border. He sees it as a psychological bubble and terms it protective frame.¹² The paratelic state is characterized by freedom and it being voluntary: there is a feeling of being able to turn off the television, to be able to walk out of the game, or packing away gardening tools.

There is a private world, but it is not cut off from the real world. Like Goffman’s interaction membrane, when properties from non-play world enter, they are transformed. Another important feature is that when a person is in a paratelic mindset, when she is within this psychological bubble, she feels secure and unthreatened:

The major difference, then, between the psychological formulations, and the social formulations, is that in the former the border and its construction are conceived of mainly as phenomenological and personal – even if it can take physical and other culturally recognized forms. This helps in explaining why different people have differing interpretations of playful situations – or even as to what counts as playful – as the protective psychological bubble is not uniform and shared, but personal. Considering these in relation to danger is especially illuminating: it does not matter if a situation is objectively speaking dangerous or not, the personal experience and perception of it (and the person experiencing and perceiving) is what influences the presence or absence of the bubble. This helps in explaining and understanding deep play and dark play (Schechner 1988, 12-14; Geertz 1973, 432-433; Sutton-Smith & Kelly-Byrne 1984, 314-316; Csikszentmihalyi 1975, 74-101).

Compare this to Bernard Suits’ (1978) concept of lusory attitude. He sees it as one of the building blocks of games, even a requirement for the constitution of a game. A player with a lusory attitude accepts the rules of the game just because they make possible such activity as the game (Ibid, 41). Suits’ formulation is interesting as he marries the social and the personal: games require rules, which – though they can be personal – are usually socially shared. But in order for the game to be possible, all the participants must accept the rules. This is a personal choice. The player can have ulterior motives (like being near someone else who plays the game, making money as a player), but they still need to adapt an attitude where they take the rules seriously in order for the game to take place (Ibid, 142-146).

The attitude Suits discusses is not directly comparable to Apter’s bubble or Csikszentmihalyi’s screen; it seems more like something that can help in achieving that phenomenological state. It is an attitude one can choose to have, not something one strives for. Yet it does underline how the player is knowingly fostering a playful approach, even if she may not be able to switch from a telic to a paratelic or autotelic mindset on a whim. Humans not only play, but they are aware that they play. Suits also notes:

The borders as postulated in psychology tend to be phenomenological and personal. These mental metaphors – though they describe a border around play – are ultimately different from the social construct of the magic circle. Like between playfulness and the act of playing, there is a connection between the psychological bubble and the magic circle, and that relation is not as clear as one might hope.

Accepting a lusory attitude, accepting the rules of a game or playing is a social process, yet often the rules are not created on the spot, at least not from scratch. The next section discusses the culturally recognized games and their boundaries.

Huizinga’s formulation of the magic circle is something that is in place as play begins. Salen & Zimmerman see it as something that is either already in place or is generated as gameplay begins. While play may generate its own space, the playing of a game can be seen as entering a pre-existing space.

There is a long history of conceiving of games as their rules, going back to at least John von Neumann and Oskar Morgenstern (1944, 49), who write: “The rules of the game [...] are absolute commands. If they are ever infringed, then the whole transaction by definition ceases to be the game described by those rules.” A player can adopt different strategies, but the closed formal system of rules does not change (cf. Suits 1978, 41; Crawford 1982; Makedon 1984; Salen & Zimmerman 2004, 81; Juul 2005, 36; Myers 2009). Rules are constitutive. They not only regulate the activity of play, but enable it.¹³

Games establish a correspondence with an external referent system, such as the world around them; as they simulate and model something they are representational (Myers 2009; Gee 2008). Furthermore, they take cultural forms (such as images, artifacts and abstractions), and become semiotic domains (Gee 2007, 17-43; also Riezler 1941, 505), structures of interrelated sign systems. In relation to videogames Ian Bogost (2007, 241; also Gee 2007, 81-87) has argued that games are particular, embedded with specific cultural meaning: “The abstract processes that underlie a game may confer general lessons about strategy, mastery, and interconnectedness, but they also remain coupled to a specific topic.”

In a discussion of the syntax rules of games, operational gaming researchers Vadim Marshev and A.K. Popov (1983) define game space:

When a game exists as a cultural artefact, for example as a designed, ready-to-play game, possibly with a physical arena attached, then initiating play is a step into the game space. The formal boundaries of the game have been set by the designer or by tradition. The same concept has later been discussed in game studies as space of possibility (Salen & Zimmerman 2004, 67; see also Walz 2010, 92-119; Juul 2005, 164-167). Marshev and Popov further outline the role of the player:

Accepting the player position within the game (cf. a lusory attitude) the player produces and reproduces the game, its boundaries and its space. In The Magic Circle: Principles of Gaming & Simulation Jan H. G. Klabbers argues that games have a dual nature as social systems¹⁴ (2006, 38-46), and that a system is always in place as gaming begins.

It is worth noting that Klabbers’ conception of the magic circle is mostly based on Huizinga’s formulation, although he uses it primarily in the context of formal games and simulations, not spontaneous play. He conflates the magic circle as a social contract and the game as a formalized artifact of a social contract or cultural negotiation.

However, though games are recursive and reproduce their form through time, they do contain the possibility for emergent change (Malaby 2007, 104). The extent to which the rules are established before play commences is quite relevant. An established sport with official rules in a specific custom-built arena and a spontaneous bout of social play on a rainy street are both play, but only one of these is a pre-existing cultural artefact, and only one of these has a predefined game space.¹⁵ Yet, as Montola’s (2012, 40) discussion of the ball in basketball as a materially encoded representation of formal rules show, the material (or virtual) pieces or sites used in playing a game are not the game space, but manifestations of the rules.

Another interesting thing to note about the borders of play as cultural entities is that they are often recognized by other cultural systems, such as legal systems. As Greg Lastowka (2009), a scholar of law, has pointed out, violence is legally accepted in a boxing ring,¹⁶ and subjecting oneself to an “unreasonable risk of harm” is inherent to American football (Ibid, 386). Building on Huizinga, Lastowka notes that games are not just separate from the everyday, but they are ordered separate spheres that have their own jurisdictions and special rules; courts, for example, do not review the rulings of game referees (though they do uphold contract law) (Ibid, 385, 390-391). Games are not the only social sphere where there are special rules (compare dormitories, religious communities), but games can have rules that are in stark contrast with state rules. Often play where there is a very high psychological, physical or monetary risk involved (i.e. deep play) is legislated as in boxing, gambling and bungee jumping (Ibid, 388-389).

The fact that many state legal systems recognize games as happening in a different jurisdiction, of course, says very little about the border around play. Legal systems are social constructs just as games are, and there is no reason why one construct would not recognize another. However, what it does show is that on a cultural level that border is – at least up to a point – recognized and respected. Interestingly, Lastowka’s main argument addresses the legality of real-money trade in virtual worlds: he believes virtual worlds are games and thus courts should not interfere in the upkeep of the rules in these separate spheres of jurisdiction (Ibid, 392-394). If virtual worlds are games, then that assessment makes sense. But, again, if they are sites where play takes place, but where also non-play happens, then the situation is more complex. Yet the end user license agreements of MMOGs have explicit contractual statements that require the user to only use them for ‘play,’ whatever a legal duty to play might mean (Reynolds and de Zwart 2010).

The border that surrounds play is most visible when it is questioned, threatened or played with. Transgressive play draws attention to the border – and even questions if any border exists. Yet without limits, it is impossible to push oneself past them. According to Juul (2008, 64) “the magic circle is best understood as the boundary that players negotiate.” Bad, dangerous, transgressive and harmful play seemingly challenges the idea of play as separate. Especially gambling has been used as proof that play is inseparable from everyday life and that play can have grievous repercussions for ordinary life. However, psychologists have no problem incorporating “bad play” within a framework of separate play. Performance scholar Richard Schechner has some ideas about that as well. For him the idea that play is dangerous is absolutely central:

Schechner also recognizes that play sets itself apart in its own realm, behind a porous border – and he discusses both the mental and the social aspects of it. Later in the same article he describes playing as a creative destabilizing action that neither declares its existence nor intention:

As an example of play that is not socially shared, metacommunicated, he points out dark play. He even goes so far as to suggest that the person engaging in dark play may not even be sure that she is playing; it is possible that the action becomes (dark) play in hindsight, in the retelling, reframing and narrativization of the event (Schechner 1988, 14; see also Denzin 1982, 13-14).

However, Schechner also points out that the players need to feel secure in order to begin play (also, Weisler & McCall 1976). Trust is a key element. Indeed, the idea that play and games are safe is deeply ingrained in the discourse of game studies and especially game design. It ties into the idea that play is separate from everyday life and actions taken during play bear few consequences beyond the play session (see e.g. Caillois 1958; Rodriguez 2006). The typical way of framing that is by saying that games are a safe platform to practise. Game designer Chris Crawford has expressed this argument nicely:

Similarly another game designer Bernie DeKoven (2002, 12-13) considers trust among players and a feeling of safety (not risking more than we are willing to risk) as integral elements in establishing the intention of playing well together. This general argument has also received specific formulations. For example, in regards to role-playing games expert hobbyist Toni Sihvonen has written about what he calls the role-playing contract:

Basically the contract states that the participants should not make judgements about the player based on the character, or vice versa. There is a disconnect between the player and the character. Though people do make such assumptions, the social contract makes it possible for role-players to take on roles that are very dissimilar from themselves (cf. Goffman 1974, 194-195, 275-284). What is interesting about Sihvonen’s formulation is that it explicates the contractual nature of the border that is drawn between play and non-play; Crawford takes that contract as given.

What all these formulations have in common is, again, the idea that trust is built socially. Malaby (2007, 110) has called games artifactual to underline that they are not only manmade, but specifically socially constructed to be separate (to a varying degree) from everyday life.

It is also interesting to note that there are numerous games that take advantage of playing around with the borders of play. Either the borders are blurred and expanded, as in pervasive games (Montola et al. 2009), or the playing happens knowingly on the border, as what Cindy Poremba (2007) calls brink play. Though the metaphors are different, the phenomenon is the same: both use, as Poremba puts it, “the contested space at the boundary of games and life.” For many players the central draw of pervasive games is that they create an ambiguous zone between play and non-play – and inhabiting this not-knowing is quite pleasurable for some players (Montola et al. 2009; McGonigal 2006).

According to Poremba, (2007) brink games¹⁷ knowingly play with the metacommunicative statement “this is play.” Players who engage in brink play¹⁸ will be able to do things that in the normal social frame would not be acceptable, but are acceptable as they are “just playing” – but at the same time the possibility to be able to do those transgressive things for real is the reason they want to play. A game qualifies as brink play if a conflict between implicit social rules and implicit (or even explicit) game rules is integral to the playing (see also Consalvo 2005, 10).

Of course, it is also possible to use games to generate danger that spills outside its borders. Some sorts of dark and deep play do not so much use games as an alibi, but as an engine of strong emotional experiences. Bleed play (Montola 2010), where role-playing games are played in order for the player to experience some kind of shock, is an example of this.

The ambiguity produced by pervasive games and the winking at metacommunication in brink play presupposes a difference between play and non-play (or other play), or at least a friction between different frames or different sets of social rules. The rhetorics surrounding bleed play do this as well, though perhaps the social contract around such play is a little different to begin with.

Though it certainly can be questioned if the idea of play as relatively separate is a romantic notion in itself, at least this idea has been widely discussed and found useful in numerous fields of inquiry. Yet it is important to underline that play is not seen as exceptional in its delimited nature.

Based on the review a synthesis of the boundaries of play is now constructed. There are three different boundaries of play: the ‘protective frame’ that surrounds a person in a playful state of mind (psychological bubble), the social contract that constitutes the action of playing (magic circle of play), and the spatial, temporal or ideal, rule-based cultural game space where play is expected to happen (arena). The clear analytic differentiation and articulation of these three boundaries is essential as otherwise the usefulness of the terms as tools is diminished.

The Apterian psychological bubble is personal, a phenomenological experience of safety in a playful (paratelic/autotelic) mindset. If a person is playful alone, she need not negotiate or metacommunicate with others (though usually she does signal play unconsciously). There is a ‘border’ around her experience that guides her interpretation of the situation. A person needs to feel safe in order to be playful, though it is not necessary to actually be safe.

The magic circle of play is the social contract that is created through implicit or explicit social negotiation and metacommunication in the act of playing. This social contract can become societal as other social frameworks (law, economics) can recognize it. It is created when there is more than one person engaged in playful activity, though once established it is no longer necessary for everyone to constantly remain in a playful mindset. There is a connection between a playful mindset and play, but as a result of social negotiation and shared structuring of an encounter, it is possible to be in a telic mindset and still remain within the socially agreed borders. However, if enough participants slip into a telic mindset, then it can be questioned whether what is contained within the borders remains play even if it is still a game. The concept applies to the playing of single player games as well: though they can be played alone, they are socially recognized as domains of special meaning. However, the concept of a magic circle is more useful in relation to social play.

The arena of play is a temporal, spatial or conceptual site that is culturally recognized as a rule-governed structure for ludic action, or an inert game product. As the social negotiation of a magic circle becomes culturally established and the border physically represented, arenas emerge as residue of the playing (tennis courts, April Fool’s Day).Alternatively a rule structure can be culturally coded as a game product, one with a designed game space. These sites are recognized as structures that foster play even when empty (and they can be constructed in ways that seek to foster playfulness), but they require use to be activated as the border of the magic circle remains social. As socially recognized cultural sites they have severed the need to be engaged in with a playful mindset.

The difference between a socially constructed magic circle of play and a designed arena of game space can be unclear, as the cultural and social boundaries certainly imply and complement each other. The potential of the cultural is mediated by and actualized in the social. Yet keeping the two separate for analytic purposes is important. The magic circle does not travel with a game product, but is social, produced by the people present in the act of playing. It is not the line drawn on the ground, but the social contract attached to it. However, it does often align with the dormant possibility space provided by the rules of a ready-to-play game.

The players are rarely completely absorbed by the playing, which makes (meta)communication about play possible. It is possible to change the social contract during play – unless such changes are forbidden by the initial social contract (as in institutionalized games). Furthermore, the contract can be played with, which heightens its existence and its nature as a social construct.

The participants are supposed to treat the encounter within the borders of the social contract as disconnected from the external world and they are not supposed to bring external motivations or other carryovers from the non-play to the play, yet often they do. This can also be negotiated, or players can ignore it and pretend that they do not notice. As a contractual barrier is established, the events within the border are loaded with special significance. However, the border is porous and allows for traffic in and out, though passing through the border results in a re-signifying transformation – but it is also possible for the barrier to collapse due to pressure from the inside or out.

The events that take place while the contract is in effect are real, though their meanings may be altered. As the encounter is set up through social negotiation and special signification, it is possible to have numerous overlapping social contracts and frames of signification. The participant is able to view and interpret the events that take place through these various frames. Though this formulation is written with play and games in mind, it may be useful for deciphering other social encounters as well.

The concept of the magic circle has been widely debated in game studies. However, it makes sense to talk about magic circles, as the one described by Huizinga and the one introduced by Salen & Zimmerman are different in their meaning. This article has concentrated on the formulations of Salen & Zimmerman, and after tackling its criticism still finds the concept a useful metaphor. In order to determine what it is a metaphor for, various takes on the social, mental and cultural boundaries of playfulness, playing and games were reviewed and discussed. Finally, a synthesis of these works was offered.

The idea of a magic circle of play is that as playing begins, a special space with a porous boundary is created though social negotiation. The formulation presented at the end of this article of the magic circle is different from Salen & Zimmerman’s formulation mainly in its explicit basis as social contract and its focus on play (not just gameplay). It is also clearly separated from the mindset of the participant and the rules of a pre-existing arena.

This research has been supported by Finnish Cultural Fund and Academy of Finland. Special thanks to J. Tuomas Harviainen, Markus Montola and Frans Mäyrä.

1. Huizinga’s critics have mounted a convincing case that his conceptions of play and seriousness as well as play and everyday life are muddled: Ehrmann (1968, 32-33) questions the lack of definition for “reality” or “everyday life”, and how it is possible that reality exists prior to its component, play. Anchor (1978, 87) questions how Huizinga maintains both that play does not exclude seriousness, and that play and seriousness are two separate categories. Calleja (2012) points out that “[r]eality does not contain play; like any other socio-culture construction, play is an intractable manifestation of reality.” See also Rodriguez (2006) and Lammes (2006). This article, however, concentrates on the formulation of the magic circle forwarded by Salen & Zimmerman (2004) and thus sidesteps these ontological criticisms.

2. Malaby’s criticism is aimed at digital games and game studies in the form that it emerged around the turn of the millennium. The permeability of the border surrounding play has been discussed by other scholars of games for a long while.

3. Montola (2012, 30-31) has noted that the contexts discussed by Salen & Zimmerman are what Searle (1995) discusses as contexts if his formulation “X counts as Y in context C”. Montola (2012, 52-53) also points out that there is wide agreement on magic circle being about transformation, not isolation.

4. The concept of magic circle has been used in analysis of games and play especially when borders of play are discussed (e.g. Montola et al. 2009; Poremba 2007).

5. Note that Riezler (1941, 508) writes about adult play. As he sees it, in child’s play the real world and the play world have not yet separated.

6. Note that the essay addresses “the kind of games that are played around a table” (Goffman, 1961, 8) as Goffman was mostly interested in face-to-face interaction. It is important to keep this delimitation in mind though many of the concepts introduced there are applicable in a wider context of gaming.

7. A more useful way of conceptualizing the construction of everyday life and primary frameworks is offered by John Searle (1995).

8. In Bateson’s terms, the difference between keying and fabrication is that in keying the frame is correctly meta-communicated to all actors and in fabrication some of the actors are intentionally misled.

9. Though Goffman certainly does see play as taking place in a frame, the boundaries are far from clear. Frames within frames mean that there is a frame for the administration of a spectacle (such as rituals around a game) and then a frame for the game proper (Goffman 1974, 261-265; see also Fine 1983, 181-204). For example, he considers that all discussions about the rim of the frame between play and non-play lead to paradox; discussing the edge of the frame takes place in the framework (Goffman 1974, 249). The division between the social play in games and the sociability that surrounds them have been discussed also for example by Gee (2008, 24), Stenros et al. (2011) and Elias et al. (2012, 203-205).

10. For a review and discussion on the magic circle as re-signifying and sorting information barrier, see Harviainen (2012).

11. This has been discussed in psychology sometimes as “selective inattention.”

12. Though conceptually Apter’s protective frame is similar to and inspired by Bateson’s and Goffman’s frames (Kerr 1991, 34), it is important to note that it is personal and not necessarily social. In order to avoid confusion I’ll refer to Apter’s protective frame with his own metaphor, psychological bubble.

13. For an analysis of constitutive and regulative rules in games, see Montola (2012, 32-47). Note also that Salen & Zimmerman (2004, 96) wrote: “The magic circle of a game is the boundary of the game space and within this boundary the rules of the game play out and have authority.” While this implies that magic circles are reserved for rule-based play, something that Staffen P. Walz has criticized (2010, 110), Salen & Zimmerman note that the framing of games as either 'rules', 'play' or 'culture' is relevant.

14. This opens up a whole library of theory from social sciences, which is ignored in this article.

15. There have been different ways to form categories on the continuum of play and gaming – such as Caillois’ (1958, 27-33) paidia and ludus, Shubik’s (1983, 17-19) rigid-rule and free-form – but an exact typology is not relevant for the discussion at hand.

16. For an account of the development of modern boxing from prizefighting, and the role of courts and legislation in that, see Anderson (2001).

17. Salen & Zimmerman (2004, 478-481) discuss the same subjects under the header of forbidden play, but I prefer the term brink play.

18. Examples of brink play can be found in spin the bottle, Twister, Pillow Time, and the various kissing games (see Frasca 2007, 160-177; Sutton-Smith 1959).