Ritual Discourse in RPG's

Christopher I. Lehrich 

Introduction

Theoretical analysis of RPG's remains largely cut off from other theoretical discourses, a situation that tends of itself toward sterility.  Two reasons for this isolation predominate.  First, RPG theorists come from a wide range of educational backgrounds, and as such have no shared body of theoretical models or discourse on which to draw.  Second, RPG theory hopes to serve a constructive function, rather than a purely analytical one: where the anthropologist for example traditionally understands herself as necessarily exterior to the people and situations she analyzes, the RPG theorist wishes to employ the results of his analysis to improve his own gaming.
The former difficulty need not concern us unduly.  So long as theoretical models from outside current RPG discourse receive adequate formulation and explication in RPG terms, only an a priori hostility to other theoretical constructs would dismiss them out of hand.  It is worth considering that such hostility does appear mutual -- that is, much RPG discourse formulates itself in opposition to academic theoretical discourse, while many academics continue to express disdain and scorn if not outright hostility for role-playing games as an activity -- but resolution of this can only come about in a historical situation as yet hard to imagine.  Thus I shall set the issue aside, stating only that I intend to explain fully whatever theoretical constructs I deploy.
The second problem, however, inheres in the nature of RPG's themselves.  A purely theoretical analytical model of RPG's, i.e. one without any practical application whatever, will generally be received poorly, if at all, within RPG communities.  Indeed, even RPG theorists who go to considerable lengths to formulate the practical implications of their models are sometimes derided as airy pseudo-intellectuals.  Fortunately, some recent RPG publications by members of the theoretical community have received accolades,[1] and this will presumably have the long-term salutary effect of legitimizing theoretical work within the hobby at large.
At the same time, analyses of RPG's have come to formulate practical, essential divisions and categories, and argued that these may be unbridgeable.  For example, Ron Edwards's tripartite GNS model rests upon the notion that the three categories must remain discrete in order to avoid paradigmatic clash and attendant misunderstandings among players, leading in turn to poor play.  That is, a group of players with strongly Narrativist tendencies should be wary of playing a strongly Gamist-structured game, or introducing into the group a player with such an approach.  While "hybrids" -- games that effectively serve more than one of the three major play-types -- are conceived as possible, a central point for Edwards is that Narrativist-oriented play is not well-suited to Gamist-oriented games, and that groups who attempt such may need to revise the game extensively to fit their needs.  Similarly, a single player who cannot conform to the paradigmatic norms of the group in which she plays will probably find herself continually at odds with other players, leading to social conflict; this player would be best advised to find another game.[2]
In his recent article "Story and Narrative Paradigms in Role-Playing Games,"[3] John Kim argues that underlying such categories we find two approaches: "Collaborative Storytelling" and "Virtual Experience."  These tend, like Edwards's categories, to remain divided.  In what Kim calls "Paradigm Clash," we find a naturally-occurring conflict between perspectives: 
To the storytelling point of view, the experiential view seems to result in an unnecessarily limited set of techniques. . . . Experiential play may also seem passive, letting events happen rather than actively controlling them. . . . [Conversely,] To the experiential point of view, storytelling play seems to be creating a product for a nonexistent reader. . . . Experiential players faced with storytelling play may complain about breaking suspension of disbelief, or lack of depth.
Conflict arising from disjuncture, narrative or otherwise, is not only theoretical.  Most gamers have experienced it, and one great strength of Edwards's model (derived from the earlier Threefold Model developed in the Advocacy newsgroup[4]) is to emphasize recognition and classification as means to avoiding the problem.  In both his and Kim's models, players and groups who recognize their preferences in a categorical sense can select games to fit their desires, or revise them so, leading to enjoyable play with a minimum of fuss and trouble.
While I support this general constructive point, and do not presently wish to challenge the classification itself (a much-contested issue), I suggest that a hard-line division within analysis leads toward weaknesses in a general understanding and formulation of how RPG's really function.  By drawing on some theoretical models outside of RPG's, I would like to propose a more unified model of RPG narrativity.
A word about practicality: I do not, in the present article, formulate the practical implications of this model for game design or play.  I do not see this as a weakness in itself: if the model serves analytically, it can have synthetic value.  But the two operations have at least a notional distinction, and can operate well in isolation.  If theory must face a practical proof-critique, then all analysis is already crypto-synthesis; logically speaking, there is thus insufficient distance postulated to ensure the validity of the analysis.  In short, without the ability to distinguish at least heuristically between theory and practice, theoretical work can never have real logical force, lending weight to the criticisms mentioned at the outset.
A further point: I intend to propose a ritual model for RPG play, based upon recent understandings of ritual within the academic discourses of anthropology, sociology, and history of religions.  This model would appear to fall squarely into the common discourse of analogy as theory, of proposing that RPG's are "like" something else in order to help emphasize a point otherwise unclear.  Such analogical reasoning is founded upon an essential methodological principle: the analogy is not identity.  Thus response to the proposal is constrained to two related moves.  On the one hand, one may move to expand the analogy, picking up additional aspects of the metaphorized object or activity and further relating them to RPG's; on the other, one may move to limit the analogy, demanding that the metaphor not be taken to the point of absurdity.[5]
Some find this mode of analysis useful, primarily in a creative sense.  If one "gets" the analogy, in its logical extension and intension, one thinks about the hobby in a somewhat new way, perhaps leading to new creative engagement with design or play.  But if one does not "get" the analogy, the tendency, naturally, is to dismiss it as unhelpful, or to reformulate it endlessly until one does "get it."  Either way, the reason to analyze such a metaphor is generally synthetic, to create new ways of engaging with the hobby.  In other words, the proposal of yet another analogy serves no analytic function.
In proposing a ritual model of RPG's, I do not wish to add another analogy to the lists.  I do not mean that RPG play is like ritual at all; I mean that it is ritual.  Therefore classical and recent tools of ritual analysis apply fully to RPG's, for analytical purposes, for making sense of RPG's as something other than an entirely isolated hobby, indeed for seeing RPG's as a human cultural product not particularly distinctive to modern society.  If to some this seems a claim that RPG's are not special and extraordinary, I suggest on the contrary that this grants to RPG's a legitimacy and "specialness" attendant upon their roots in wider humanity and culture.[6]


Ritual
An obvious first step in proposing this model is the formulation of a definition of ritual.  Unfortunately, perhaps, such definitions have been the focus of extensive debate for more than a century now, with no clear end in sight.  More models have been proposed of what ritual "is" than many readers might believe.  I have no intention of summarizing this whole history; I will instead simply propose a starting-point.
The above-mentioned disjuncture between "Collaborative Storytelling" and "Virtual Experience" parallels, in a number of respects, two recent emphases in ritual theory.
Virtual Experience correlates well with Ronald Grimes's and Victor Turner's focus on "performance," which ultimately amounts to a notion of total involvement in ritual activity.[7]  In ritual, according to this perspective, humans engage the totality of hearts, minds, and bodies, setting them to work creatively and dynamically to produce effects within the social and mental worlds of the participants.  Thus in zazen (Sitting Zen), one does nothing but sit, generally in an approved posture; one's mind and heart should be similarly focused on nothing but sitting, not in the sense that one should think continuously, "I'm sitting," but rather that one's mind should be in a state parallel to the body's state, thinking nothing, resting, yet remaining alert and awake, receptive to outside contact.  In the Catholic Eucharist (Mass), to take a quite different sort of example, liturgical tradition emphasizes that the communicant should be fully involved in the process, such that when the miraculous transformation of the substance of wafer and wine (Transubstantiation) occurs, and when in fact the communicant receives these into the mouth, it is not only one's body that receives the body and blood of Christ, but the totality of body, mind, and soul.  Thus this understanding of ritual emphasizes what in RPG terms is called "immersion," a total involvement in the activity.  Failure on this score would be seen as ineffective (zazen), impious (Eucharist), or shallow (RPG).
The Collaborative Storytelling model is less obviously commensurate with a ritual model.  Two directions, however, support this formulation.  First, there is Claude Lévi-Strauss's structuralist interpretation of mythic and ritual thought as bricolage, and second, there is the movement largely associated with Pierre Bourdieu, Sherry Ortner, and Catherine Bell toward understanding ritual as "practice" (or "praxis" in the more overtly Marxist formulations).[8]
Lévi-Strauss's idea, in simple terms, is that cultures think like oddly artistic hobbyists. [9] Imagine you have a basement full of stuff from which to build whatever you like.  You have bits of old machines, things your neighbors threw out, scraps of wood, and tail-ends of old projects, as well as the taken-apart bits of all your old projects.  Now you decide to build something, and you have some ideas -- aesthetic and practical -- about how that should be done; you are very skilled and talented, and can see possibilities in all sorts of things.  But you do not have a Home Depot available, or you consider it "cheating" to go buy things.  At any rate, you have to build the thing you're going to build from what you already have in your basement.
A nice example is a Rube Goldberg cartoon, though those are deliberately silly.  You fly a kite, and the kite string pulls a lever, and this pushes an old boot, and that turns on your iron, and the iron burns some old pants, and smoke goes into a tree, and....  A brilliant example is the recent Honda advertisement called "the cog," which can readily be found on the Internet.[10]  The point is that one constructs an elaborate machine out of bits and pieces already owned.
Lévi-Strauss's point is that each object used contains its own history; that is, the iron has already been used for something and the bricoleur then gives it a new use.  The iron, to focus on the single example, is a local source of heat; it can burn pants, or make a grilled-cheese sandwich, and of course can press a shirt.  But it cannot be a refrigerator.  And if, clever person that you are, you pull the heating coil out of the iron for some project that requires a heating coil, your iron now contains the history of its usage: it is now a heating coil and a heavy weight.
Every sign in myth and ritual, says Lévi-Strauss, is like this iron, and every living mythic culture is like this bricoleur.  When faced with a (social) situation, an intellectual problem of whatever kind, the bricoleur begins by running through his memory (the basement) to see what he already has that can be used to solve the problem.  He then builds the machine that solves the problem, in the process incorporating the entire history of every object in question, and furthermore altering (however slightly) each object so used; when he goes to build something else, later on, the current project will be part of the history of each object.
Technically speaking, every sign is thus constrained and yet free.  On the one hand, it is not constrained to the degree of a percept, a particular contingent mental encounter with an actual object; this percept is what is called a "perception" in the formalist model to which Kim refers.  A percept is entirely constrained, because when a person looks at a given object on two successive occasions, his or her mental equipment has altered -- to use a cliché, one cannot enter the same river twice.  At the same time, a sign is not fully liberated, as is a concept, an idea arising in reaction to a particular person's connections to a percept: when I look at the lamp on the table, I may think of my grandmother (who perhaps owned a similar lamp), and thus "grandmother" is a legitimate conceptual link, but no such connection may arise for you, and even if it did, it would be a different grandmother.  So a sign (Lévi-Strauss means the Saussurean version of the sign) is both constrained (the iron cannot be a refrigerator) and free (it can do a whole range of things involving local intense heat).  In Lévi-Strauss's linguistic analogy, this iron is a sign in the same way as a word is: the word "iron" can mean a range of things (the metal, the instrument) but it cannot mean anything at all.  Furthermore, this word only acquires meaning by its relations to other words: if I say "iron," you do not know until I go on with "a pair of pants" what sort of meaning I intend, even whether it is a verb or a noun.
The other approach I want to bring up, "practice" theory, arises from a number of rather technical difficulties with structuralism, and amounts to an attempt to understand manipulation of signs and symbols in strategic yet controlled ways.  With respect to ritual, practice theory argues for a continuity among behaviors, as against the disjuncture of ritual from other modes of action.  The signs used in ritual, that is, acquire meaning from their extra-ritual contexts, and furthermore the special meanings accorded to them in ritual carry over into other modes of life.
From a practice perspective, every ritual contains within itself a number of structures, just as in structuralism; these structures are in essence the Rube Goldberg machines constructed by the bricoleur.  As we know from Lévi-Strauss, the iron can be replaced by any other source of local heat, since its only function in the machine in question was to create smoke by burning a pair of pants.  Thus the machine has a structure, requiring a number of elements, but the specifics of which objects or signs are used to fill those element-slots are open.  What interests practice theorists is strategic choice: how do people decide whether to use an iron or a space heater?
Broadly, the question in practice theory is how people choose, from a limited range of culturally-available options, which techniques to apply at a given moment.  This depends on strategy: we want to maximize rewards in a specific situation.  But in order for strategy to work, we have to play the game; that is, one cannot go outside the structure of the system to manipulate signs as one likes, because to do so annuls the power of the strategy in the first place.  Thus every strategic use of signs is at once a free, liberated exercise of power by a situated person, and at the same time a contribution to keeping the system stable and intact without significant change.  The possibility of real change is thus undermined by the very strategies which seek to change the system, because they depend for their efficacy upon the structures in question.
If the dichotomy between virtual experience and collaborative storytelling parallels that between performativity and what we might call the practice of bricolage, as yet this parallel serves no analytical or synthetic function; it is once more an over-theorized and over-determined metaphor.  In addition, it is as yet under-explained, in that the theories may be formulated but their application to the specific situation of RPG's is not yet clear.  In short, while we can see a parallel division within both the two discourses and the two modes of behavior, this does not answer the question: why are RPG's ritual?

Semiotic Modeling of Ritual and RPG
I have noted that Kim's use of the formalist perception-discourse-conception model parallels the semiotic or structural percept-sign-concept model.  The difficulty with the formalist model for this purpose, however, is that it is focused primarily on an interpretive perspective, in which the analyst stands in a perceptive relationship to a given discourse; like the circular model in hermeneutics,[11] the central issue is how an interpreter can make sense of a discourse already present, how we approach meaning through interpretation of texts and signs already distant from their producers (authors).  Thus a central preoccupation of both formalist analysis and of hermeneutics has been the analysis of ways in which the reading situation is not conversational, in which reading a text is not having a conversation with the author.  But in RPG's, the situation is normally conversational in an obvious sense, and thus this mode of analysis focuses on problems seemingly distant from those in RPG's.
The structural model of signification, from which the practice theory also arose, is by contrast primarily concerned with the use of signs by a current producer, a situation more obviously commensurable with RPG play.  The question, in short, is not how players read a text produced for them by a game-master, but rather how the whole group in combination produces signs and texts that they themselves read.  The structural model of signification fits well here, as the primary issue is to understand ritual or mythic activity as a mode of discourse production.
In ritual, participants manipulate a range of signs within a constrained structure.  That structure can change through such manipulations, but only within narrow limits.  Every Catholic Eucharist differs significantly, in that the place, people, and physical environment of the ritual vary, but this variation is officially read by participants as within a fixed structure.  The post-Vatican II use of the vernacular in the Mass, for example, was at once a major transformation of the structure of the ritual, and at the same time theorized as not radically transformative: even in the vernacular, according to the Vatican II council, the Eucharist retains its sacramental efficacy.  From a semiotic perspective, the linguistic alteration represents a new negotiation of liturgical language as a discrete sign, where Vatican II agreed that the differences between Latin and the vernacular should not be understood as an essential structure of the ritual, but rather a relatively arbitrary sign amenable to conversion without undermining ritual structure itself.
At this same level of semiotic manipulation, we can see in RPG reconstruction and revision a parallel analytical discourse.  Taking to its extreme the Edwards et al. formulation that "system matters,"[12] the claim is a clearly structuralist one: transformation of system elements in RPG's effects concomitant transformation of gameplay and orientation.  For example, a combat system dominated by so-called "realism", usually meaning a high prioritization of real-world simulation in modes of action and effects of violence, is not a discrete sign that may be removed from a given game and replaced with an entirely stylized, anti-"realist" combat system.  Because such a system element is structural, it links to all other parts of the total game structure and its transformation thus strongly affects the whole.  Mike Holmes has made this point well, arguing that a "realist" combat system colors the whole game, such that all activity occurs with reference to such a preoccupation with violence;[13] as Kim puts it,
[E]ven if a gun is never fired during the game session, the mechanics for that [weapon] may influence the story  --  because they shape how the player conceives of guns within the fictional world. If the mechanics make all guns exceptionally deadly, it increases the tension in a scene where a gun appears even if the gun is never fired. 
Thus the "system does matter" principle argues that system elements are motivated signs, and thus contain structure; their transformation affects the totality of the structure.
Between the Vatican II approach to language and the Forge approach to system, however, we must recognize that the difference is not absolute; furthermore, the distinction drawn is ideological, not "factual."  There can be no question, for example, that the use of the vernacular in Catholic Mass has significantly changed the ways in which Catholics experience the ritual; indeed, were this not so, there would have been no reason to make the change in the first place.  Vatican II asserted a matter of aesthetic and theological priority: however far-reaching the effects of this transformation, they argued, the essential core of the ritual (transubstantiation in a broad sense) would not be affected, and whatever aesthetic loss of force might be entailed by the loss of the affective qualities of Latin (as traditional, foreign, ancient, powerful) would be more than made up for by gains in broader spiritual involvement (through understanding the liturgy intellectually, thus affectively through content rather than through an aura of ritualism).  Indeed, Martin Luther's move to the vernacular was intended partly to combat the affective dimension of Latin as itself powerful, arguing that this amounted to a kind of fetishism or idolatry: the focus should be, he thought, on the content of the words spoken, rather than on their linguistic medium.
In Forge RPG theory, conversely, there is an implicit distinction between system elements and other elements.  It is certainly plausible that the radical transformation of the combat system of Dungeons and Dragons from the AD&D system to the recent d20 system considerably changes all elements of gameplay, even those not overtly connected with combat; to replace the combat system with a more freeform model akin to The Pool would presumably effect further changes.  But first of all, it seems clear that transforming other elements of the game (setting, background, character generation) would also entail drastic concomitant changes in gameplay; for example, d20 games not based on Dungeons and Dragons genre and story conventions exist in considerable numbers, and certainly do not play exactly the same way as does Dungeons and Dragons.  In short, it is unclear how one is to classify elements into arbitrary and motivated, into those which can be shifted without large-scale structural effects and those which cannot.[14]  
More interestingly, RPG theorists (taken in the broadest sense) generally make a series of divisions among elements in their games, and implicitly argue for relative arbitrariness.  That is, the notion that a "combat system" is in any sense a discrete element, a discrete structure, should not be accepted uncritically.  If the Forge "system matters" principle argues that even apparently discrete structures like this are motivated and not arbitrary, we must recognize that this presumes a tendency to see such systems as arbitrary, that they are apparently discrete.  By emphasizing that "system" is motivated and structural, the Forge theorists further suggest a prioritization of elements, where motivation is taken as superior to arbitrariness, so that theoretical analysis and synthesis should focus on structure rather than sign.  To put this differently, it is implicit that RPG's consist of a vast group of interrelated elements, falling into a natural hierarchical order; those nearest the trunk of the tree, as it were, are relatively motivated and theoretically important, while those nearest the branch-tips are more arbitrary and of lesser theoretical weight.
At the same time, few would argue that the arbitrary, non-structural signs are trivial or unimportant.  Such arbitrary elements as Color (essentially affective set-dressing in imagined space) or snack choices by players are not irrelevant, and may in particular instances be elevated to structural elements: the game-concept Long Pig The Role-Playing Game made snack choice and usage into a system element, while Ars Magica troupes interested in medieval history may make set-dressing a primary focus for play.[15]  But the claim is that it is by shifting such elements from arbitrary to motivated, from incidental to system, that they become analytically important; in general, the analyst does not focus classification on such elements, but rather begins with system.
The important point here is that whether the issue is the relative weight of meaningful dimensions of liturgical language or the classification of structural elements in RPG's, the understanding is in both cases ideological, intended not only to classify and analyze the ritual in question but also to emphasize and push for improvement in the activity, thus making normative claims about what the ritual should be about.  Precisely at this point, predictably, the ideological weapon of "practicality" often comes into play in RPG discourse: because a more purely analytic classificatory model (e.g. the polythetic comparative model proposed for the humanities by Jonathan Z. Smith[16]) eschews normative claims in the form of practical suggestions for game design or ritual construction, the RPG theorist codes such classification as impractical, thus valueless.  This is equivalent to a Catholic liturgist saying of an academic theorist's analysis that it is irrelevant because it does not help formulate new dimensions in Mass.  For the academic, however, this is precisely the point: she may be interested to see the results of her analyses serving a constructive use to the liturgist, she does not wish to impose her perspective upon those she studies.  Ronald Grimes, for example, believes deeply that ritual theory can be of constructive value for people seeking to formulate or reformulate their rituals, but as a rule he does not tell them how to go about it.[17]  A ritualist who denounces Grimes for not proposing a "how-to" makes an entirely ideological -- and ultimately incoherent -- claim: if Grimes does not propose a "how-to," his work is useless; if on the other hand he does tell ritualists how to "fix" their rituals, he will (and should!) be denounced for telling others what they ought to believe.
I have come a long way around, but the notion of RPG's as ritual can now be asserted directly.  Between RPG theory and RPG practice there exists a dynamic relationship structurally identical to that between the theory and practice of ritual within lived ritual communities.  RPG theory, by this logic, is only commensurable to academic theory and analytical method through a deeper and more complex formulation; a relatively direct correlation links RPG's to rituals in their actuality.[18]  In order to recognize this link, we must accept the duality of theory and practice as integral to ritual performance itself; in other words, rituals are not actions or activities performed in isolation from their cultural worlds, but rather performances related to theoretical concerns in the same way as game-play relates to the theory and system-construction that surrounds it.
To put this differently, and more specifically, RPG play enacts theory, in the sense that standing behind and prior to play is a series of theoretical constructs: system design, GM notes, pre-play agreements and social contract, genre expectations, and other theoretical tools.  From this perspective, RPG play acts out this prior structure; this is equivalent to the old reading of ritual as acting out a liturgical text.  At the same time, the prior structure is to a degree open to challenge within game play, and furthermore does not fully constrain particular game actions, determining a range and a set of priorities rather than laying out a script.  As has been recognized for some decades now, the same can be said of the most formal ritual: within apparent constraint there is scope for contestation, not only of the various issues and questions related to a particular ritual's situation within the social context, but also of the ritual itself with all its symbols.  
Nevertheless, these two views are always in dynamic, creative tension: the available range of manipulations of ritual signs stands within a structural context only slightly accessible to interior challenge.  For example, radical transformation of Catholic liturgy cannot proceed from within ritual performance itself, while small-scale local transformation and contestation are fully expected.  Radical transformation of liturgy, as we have seen with Vatican II, must come from a theoretical discourse exterior to performance.  Conversely, such discourse acquires its ability to challenge ritual structurally by sacrificing its analytical and normative force at the local level; that is, while Vatican II could change liturgical language, a structural change not available to a given congregation at the moment of performance, the congregation can manipulate particular performances to effect social meanings inaccessible to the Vatican.  For example, a particular wedding ritual may be used, at a given moment and in a particular contingent historical situation, to enable deep consideration within the congregation about the traditions of marriage, divorce, and childbirth; these same issues can be discussed by the College of Cardinals, as indeed they are, but not at the level of particular people in particular time, since they can only formulate principles and cannot apply them individually.
Precisely the same dynamic obtains in RPG discourse.  While a given structural situation of notes, game system, theoretical models, and so forth formulates a contextual model within which play occurs, such structures do not extend to the level of individual particularity that is central to play experience; that is, no game structure can be so logically intensive as to dictate every action and speech by every participant at all times, because to do so (even were it possible) would annul the entire nature of the game as game.  In fact, this limitation of theoretical efficacy is granted the status of a virtue in Forge theory, through the double formulation of "practicality" as a rational anchor and the hierarchization of the relative motivation of system structures as relative theoretical importance.  Not surprisingly, we find that the usual model of RPG discourse has it that performance (play) is the "real" anchor of RPG's, and that theory is understood by its proponents as a potentially liberating source of creativity and energy for "real" play.

Liminality in Ritual and RPG: Preliminary Classification
If we recognize in RPG's a dynamic interaction of theoretical and practical reason, between structure and event, it is not clear how within the practical sphere the active, strategic manipulation of signs actually works.  That is, we have seen that in religious ritual, situated people deploy signs and structures within the context of larger, only partly flexible structures, and that RPG play stands within a similar context; we need now to understand how RPG players manipulate signs and structures for strategic reasons, and how such strategies are both free and subject to constraint.
For this purpose, I would like to propose a specific analogy, that of RPG play to a particular mode of ritual behavior.  At the outset, however, I should note that this is analogy and not identity; that is, while RPG is (and is not merely like) ritual, it is nevertheless a distinct and specific kind of ritual, one with no exact equivalent in other ritual spheres.  Thus this analysis must be effected within a deliberately constrained comparative model, in order to evade the methodological problems attendant upon the loose metaphoricities described in the introduction.
Every modern scholar of ritual is familiar with the liminal model of rites de passage (passage-rites), originally proposed by Arnold van Gennep in the eponymous book, and elevated to a critical analytical model in especially the earlier work of Victor Turner.[19]  In its classic formulation by van Gennep, such passage-rites as initiations consist of three stages.  First, the neophyte is separated from the symbolic and social structures which normally surround him; second, the neophyte passes through a liminal phase, in which a series of new and powerful symbols known as sacra are presented to the neophyte for consideration and reflection; and finally, the neophyte is aggregated back into the social structure, now in a new status.  
For example, in boys' puberty initiations, the boy is removed from boyhood and society in general, perhaps secluded in a special initiation hut or otherwise physically removed; in addition, he is visibly marked as unclassified, e.g. having his head shaved, being painted black or white, stripped of clothing, and so forth.  Once separation from boyhood has been effected, the neophyte is in a condition of liminality, "betwixt and between," neither this nor that; neither boy nor man, he is unclassifiable, a condition generally expressed through symbols marking status as not participating in even a larger range of classes: he may be dressed as an androgyne, marking him as neither male nor female (and both); he may be forced to lie on the ground in a posture normal for corpses, marking him as neither dead nor alive (and both); and so forth.
In this liminal phase, various sacred symbols (sacra) are presented to the boy and his co-initiates (such initiations usually involve several boys at once), in the form of monstrous and bizarre masks, objects, or behaviors, presented to the neophytes by already-initiated men.  All these signs serve as objects of thought, and are commonly distorted to emphasize reflection on particular issues; for example, a figurine or dancing costume might be shrunken and blurred in all its parts, but bear a wildly exaggerated phallus, encouraging reflection on sexuality and male sexual identity.  
In an example discussed by Turner,[20] Bemba girls are presented with an earthenware figurine of an exaggeratedly pregnant woman who carries four infants, two at her equally exaggerated breasts and two on her back; other features of this figure (arms and legs, for example) are shrunken to stubs.  The figurine in this case is accompanied by a riddling song about a mythical midwife, and initiated women say the riddle's point is straightforward: Bemba tradition demands that after giving birth women abstain from sexual intercourse for a year.  But a woman's husband may object to this, and one's mother or mother-in-law may also demand that the young woman get pregnant again, as the older woman wants grandchildren and the husband wants sexual satisfaction.  The point of the sacrum, then, is that a wife who does not respect the tradition of abstention will become like the figurine, dominated to destruction by babies and their care.  However much a woman may wish to give in to her husband or mother -- or her own desires -- she must abstain.  Thus the use of exaggerated symbols in the liminal phase focuses attention on traditional culture, its reasons and purposes, and ultimately promotes conformity.
Once this instructional phase has concluded, aggregation usually begins with more or less permanent markers of the new status, followed by social presentation of the neophyte to the relevant communities (initiates, then society at large).  For example, a boy may be circumcised, marking him permanently as an initiate (thus fully male), then dressed in men's clothing (not unlike the old British practice of a boy's changing permanently from short to long pants); the initiates are then presented to the men, who welcome them into the men's longhouse or equivalent male structure from which they were previously forbidden, and they depart this house to be greeted by the women of the community as men rather than boys.
The emphasis in the current analysis is, as for Turner, the liminal.  There is no difficulty spotting separation and aggregation in RPG's.  Depending on a particular group's habitual practices and preferences, separation may begin at the front door of the host's house or apartment; this is particularly apparent in more LARP-oriented play, where entry into the broadly-defined play space is marked by a transformation of manner and affect, even of clothing.  But the most limited table-top play generally marks a separation between game-play and out-of-game behavior.  This is perhaps most obvious negatively, in objections to players who do not focus on the game and continually introduce "irrelevant" topics (television shows, video games, current events, etc.) into play.  
I have marked the term "irrelevant" with quotes for a reason: these topics are only irrelevant if and to the degree that a given group marks them so, a point generally negotiated through piecemeal social contract means.  The LARP example, as an extreme of the Virtual Experience model, may tend to object to any introduction of topics or behaviors not previously formulated as "in-game."  A smaller-scale variant of this general dynamic is the issue of "in-character" as distinct from "out-of-character": in some groups, speech should be performed in-character, in that anything said by a given player should be taken as the speech of that player's current character; sometimes this takes the form of linguistic constraint, notably the demand that players speak of their characters in the first person rather than the third.  
At a more strategic level, groups may make a sharp distinction between in-character and out-of-character knowledge, raising as a problem whether a player may act in-character upon knowledge presumably not available to his character.  That is, if Alan (playing Thror the Barbarian) knows that Marler the Wizard (played by Barbara) has been captured by an evil sorcerer and is held in a deep dungeon below the castle in which Thror now stands, and Alan knows this because as a player he was present when Marler/Barbara was captured, but Thror was not on the scene and thus has no particular way to know what has occurred, a group must consider whether Alan may have Thror head for the deep dungeon to rescue Marler.  
The question is complex, and may be handled strategically at any number of levels.  For example, some groups feel that, so long as Thror's rescue of Marler would make an exciting story, the fact that Thror "knows" nothing about the capture is irrelevant.  Even within this perspective, however, we might note a distinction between Alan having Thror "happen accidentally" to head downwards, postulating an in-game coincidence to cover the out-of-game implausibility, as against Alan having Thror declaim in ringing tones that somehow he knows what has occurred, postulating a backwards revision of plot and thus annulling disjuncture.  Another strategic choice, of course, would have Alan simply ignore what has happened to Marler, since Thror is "actually" ignorant of it; Alan and Barbara may hope that events will transpire such that Thror can rescue Marler, but the interior logic of the game-world in this case does not permit Alan's use of out-of-character knowledge to alter events in this fashion.
At a theoretical level, the same issues obtain, particularly in the aesthetics of game design.  Some groups prefer to keep rules and systems as far in the background as possible, because they see such structures as irrelevant to the game-world; that is, since Thror himself cannot be imagined thinking that he has a +7 to hit but a -2 to damage if he swings his fist, while he has a +3 to hit and a +6 to damage if he swings his sword, the strategic choices made by Alan in selecting the appropriate attack for the situation can be read as interfering with the interior game-logic.  Other groups see such activity on Alan's part as an essential aspect of gaming as an activity.  For example, one can treat a Dungeons & Dragons "dungeon-crawl" as a competition by the players, as strategic manipulators of an intricate mechanical system, against the Dungeon Master who has similarly manipulated the system to construct a difficult challenge; in this case, Barbara's choice to cast Magic Missile rather than Fireball because she makes a trade-off between damage inflicted upon a chosen target and the collateral damage which comes from the fireball spell, not to mention the specifics of range, casting-time, and material components, is anything but irrelevant: indeed, at one extreme, this may constitute much of the fun of play.
In any event, the problem of negotiating the bridge between in-character and out-of-character is founded upon the structural separation effected at the outset of ritual.  The social aggregation at the close of play thus amounts to an undoing of this separation: players step back from the in-character world (to whatever extent they postulated themselves as in it) in order to receive rewards or accolades, rehash enjoyable events, and generally begin shifting from a relatively discontinuous and separated game-time to an ordinary social event, itself marked eventually by the dispersal of the participants to their everyday lives.
We have already seen that within the liminal phase, the "game itself," classification, and identity are sites of considerable contestation and difficulty.  But it is when we take into account the question of sacra and response that the parallel to initiation becomes particularly valuable.  In particular, when we consider the interrelation of freedom and conformity, i.e. the political nature of liminality, we can begin to dig under the surface of gaming to discern the social relations and contracts which make play possible.

Liminality in RPG's: The Social Rituals of Play
One of Turner's great achievements in the study of ritual was his explication of the socio-political implications of ritual activity; while he was hardly alone in formulating this general perspective, Turner has the advantage for present purposes of having a relatively clear model that does not depend on extensive prior reading in the literature of anthropology or sociology.
As liminality theory shaded into the origins of "practice" theory, it gave rise to a stock type of analysis.  The symbols of a given ritual, particularly its liminal phase, would be explicated for purposes of situation, giving sufficient data for the reader to make sense of the further argument.  The analyst would then attempt to demonstrate the following dynamic at work: within the liminal phase, neophytes -- and by extension, the society as a whole -- employ symbols and structures to challenge, test, and even undermine the structures and norms of authority; through the ritual process, however, particularly as the liminal phase moves towards conclusion in aggregation, all this "testing" ends up serving the purposes of established authority.  Thus the ritual gives the illusion of freedom and choice, but actually enforces conformity; ritual is thus read as a technique of mystification by which cultural authority can be produced and reproduced by deceiving participants in all walks of society into accepting these authority structures as natural, given, and ideal.
There is certainly truth in this reading.  For example, numerous carnivalesque rituals (Mardi Gras in New Orleans, Carnàval, Saturnalia, etc.) do indeed construct a special space and time in which to express discontent, disorder, radicalism, and challenge, all of which is then often deployed in a larger cultural context to emphasize the "rightness" of hegemonic discourses of authority.  But more recently scholars have begun to grant that this reading is simplistic: Mardi Gras has on numerous occasions been used precisely to foment revolt, for example.  Thus recent practice theory, when it has focused on ritual and liminality, has tended to admit that ritual does produce conformity through the illusion of free choice, but at the same time to grant that particular agents in particular historical situations have the ability to manipulate symbols to their own advantage, despite the apparent constraints (and apparent freedoms) of ritual structures.
At present, I will not push the socio-political reading of RPG's beyond the narrow, local community.  It would be interesting to consider how RPG's as ritual necessarily participate in and reconstitute the structures of society at large, but the data-set required to do such analysis meaningfully is prohibitively large.  In addition, ethnography of game-sessions has barely begun, if indeed it can be said to have begun at all, and thus we have only the most dubious sort of anecdotal data.  My concern, then, is with the socio-political workings within a gaming group, which amounts to an analytic perspective on the social contract of such a group as it intersects with other structures of gaming.
It is worth noting here that the dominant Forge theory generally takes social contract to be a maximally distanced structure, standing at the upper extreme of the hierarchy of RPG structure.  While there has been discussion of social contract and means by which it can be negotiated in order to avoid paradigmatic or personal conflict, the emphasis fits squarely within Edwards's overall approach.  That is, because social contract is seen as at a considerable remove from in-game play issues, the most efficient way to deal with contractual problems is to discuss them outside of play, e.g. by confronting a problem player outside of game time, by formulating explicit social expectations before play, and so forth.  But the fact remains that these problems generally arise within game play, and prior constraint cannot fully predict or forestall such difficulties.  I suggest, in fact, that precisely because RPG's are ritual behaviors, social conflict is inherent in the form.  At the same time, from a practical perspective, it is worth recognizing that because structural and sign-manipulation achieve their maximal expressions within liminality, with extra-ritual commentary discourse primarily functioning to protect ritual tradition against challenge, acting disjunctively to separate possible challenges from the fragile yet powerful matrix of ritual performance, play itself will necessary be the central locus of social contestation, and importantly it is only within its structures that conjunctive solutions are possible.  In other words, while extra-gameplay discourse may try to protect a game against social contract problems arising within gameplay, such strategies cannot of themselves achieve consensus; the means by which a group can resolve such questions must be sought within play.
Extending from this point, we may note a common tensive relationship between extra-ritual assertions of hegemony over performance on the one hand, and on the other a concomitant counter-balancing of the manipulation of ritual as a site for resistance.  Simply put, it is often the case that as authoritative discourse tries to increase control over what happens within ritual performance externally, resistant elements become increasing empowered within performance and have greater efficacy without.  In an RPG context specifically, it seems not unlikely that increasingly emphatic assertions of hegemonic control of appropriate play and in-game discourse will tend to evoke increasing resistance within play, which is to say that players within the game will tend to challenge strong norms asserted by the game-master (or the game text, the received tradition of appropriate play, etc.) the more forcefully they are expressed.  One classic example returns us to Advanced Dungeons and Dragons: the more Gary Gygax asserted his authority and authenticity in laying down constraints about "the right way to play," the more particular groups and players were drawn either to revise the game, to play other games, or to challenge Gygax's principles from within play.  With respect to more ordinary assertions of authority, e.g. "railroading,"[21] the more overt the railroading the greater the tendency to resist; that is, if GM railroading involves providing genuine incentives to follow the predetermined plot structure, resistance may be minimal, while if a GM simply blocks all choices but the "correct" one through ad hoc and increasingly ridiculous means (deus ex machina maneuvers, etc.), players may find themselves led to beat their heads against the imposed limitations rather than find creative and enjoyable means by which to "play along."[22]
My point is not simply that strong formulations of norms in play style and social interaction may produce the reverse of the desired effect, though this is worth consideration.  Rather, I wish to emphasize that semiotic manipulation within play reacts to functions in the given structural context, such that assertions of social or technical norms naturally constitute important objects of gameplay contestation.  As in initiation ritual, the imposition of social structures through such means as sacra or rules systems demands challenge and consideration within ritual; attempts to eliminate such semiotic manipulation within ritual liminality, including gameplay, can only provoke two kinds of response: resistance to the norms or elimination of ritual effectiveness.  Thus the nature of gameplay as ritual activity necessarily determines its focus on manipulation and challenge of given structures.
If RPG play can be read as reactive, it is neither mechanical nor passive, and a great strength of both structural and practice theories is the emphasis on dynamism in the relationship.  If on the one hand ritual imposes upon its participants a series of interlinked structures and motivated signs, to which participants are then forced to react by the normative view of ritual activity and thought, at the same time those participants actually have considerable flexibility in doing so.  This is where some of the earlier Marxist approaches overestimated the hegemony of authority-structures: they assumed that the imposition not only of signs but of structures through which to think them fully constrained initiates (for example) to conform to a rigid status quo; ritual could thus be read as a means of combating in advance nonconformity, resistance, and the potential for revolution, because it mystified the arbitrary, cultural nature of authority structures by transposing them into tradition, and then constructing a notion of tradition as natural and "given" in nature or meta-nature (the gods, the spirits, etc.).  But as numerous critics of such ritual theories noted, this implies a special division in society: there are those who create authority-structures, who to some degree know that these structures are merely inventions, and then there are those who are simply slates inscribed upon by such authority structures through ritual; the only flexible part of this formulation would be the first part, in that it is possible that authorities too are entirely subject to what they take to be given structures and traditions, such that everyone is enslaved by ignorance of the functions and methods of their own society.  Good Marxism this may be, but it does presume that people are entirely controlled and dominated by what they are told, and never think flexibly.[23]  In fact, the approach deconstructs itself: if this is all true, how can the academic analyst spot the problem at all?  Presumably, academia would constitute a constrained discourse that recognizes itself as an object of critical analysis, in which case how did it become so?  The logical conclusion essentially would assert that the members of critical academic discursive circles are a different sort of people than those constrained by discourse, such that radical elitism becomes a naturalized and normative structure -- precisely that which the analysis desired to challenge in the first place.
In RPG's, flexibility is relatively obvious: few if any players or observers would assert that gameplay is so constrained as to prevent flexibility in semiotic manipulation of any kind.  At the same time, this creativity is still generally taken as a marker of the distinctive or even unique character of RPG's.  Quite apart from the fact that this entails RPG theorists' participation in the reproduction of authoritarian notions of ritual behavior, a complex logical circle inserts itself in this understanding, common it seems from the inception of RPG's as a discrete ritual form.  With the explication of this circularity, it will become clear why I emphasize an analogical parallel to liminality in religious ritual.

Creativity as Circularity
Overt acceptance of creativity and flexibility within RPG play is indeed unusual in ritual.  Importantly, however, it is not the existence of such dynamism that marks a distinctive ritual mode, but the fact that participants of all levels recognize and accept this.  By contrast, the modern Catholic Eucharist permits considerable scope for flexibility and creativity in each and every performance, by every participant at every level, but this is not commonly accepted as either present or desirable; we might note that the common disdain for Neopagan ritual invention among relatively knowledgeable mainstream religious Americans includes (but is not limited to) a distinction between "real" or "traditional" ritual as opposed to those which Neopagans "make up."[24]  In this context, we can read the ideological split as a claim against creativity within the special context of ritual, importantly different from how RPG discourse consciously constructs itself as creative and dynamic.[25]
To put this in terms of initiation, we find that the liminal phase involves flexibility and invention on the parts of not only the neophytes but also the entire society; at the same time, such flexibility is commonly denied by the hegemonic discourse, as already indicated by the tendency to conceive of neophyte interaction with sacra as "instruction" rather than creative engagement.  Similarly, we find numerous discourses about carnivalesque ritual formulated in terms of what has been called a "hydraulic" theory: carnivals act as valves, allowing participants to "blow off steam" rather than harness it to antisocial ends.  By permitting marginal elements of society to "act out" their frustrations, authorities retain control of real power and maintain the stability of those they dominate.  Real challenge or engagement with social rules is annulled, because it "doesn't count" in ritual space.
Thus the demarcation of ritual space and time -- that formal construction of division between ritual and everything else central to what Catherine Bell calls "ritualization" -- lends itself to protection of social norms.  In RPG's, with their discourse of invention and creativity, such protection seems non-present or at least marginal.  But this accords with expectations: by asserting that RPG gameplay constitutes a protected space in which to deal with the limited range of issues at stake in a given game, RPG's naturally tend to assert not only that gameplay permits flexible engagement with social norms but also that the effects of exterior norms on players do not play a significant role in the game.  For example, the protection of RPG's allows a male player to play a female character, a heterosexual player to play a homosexual character, without its being read as relevant to the player's out-of-game identity; we do not, that is, assume that a male player who chooses a female character is actually conflicted about his sexual identity.  At the same time, this entails that the female character in question, if she appears as a chauvinist stereotype, cannot "officially" be read to imply chauvinism on the part of the player.
While for majority players -- white, male, middle-class -- this freedom may not appear problematic, it entails real difficulties when (especially) female players enter the game situation, most especially if such players have a romantic and/or sexual affiliation with another player.  Indeed, female players often find themselves read as "not serious," "just the GM's girlfriend," and so forth.  When such players experience events in game-time, whether plot events effected by other players or overtly structural elements constructed within the game rules, their responses may be read as problematic for in-game discourse.  To take an extreme example, if a female player reacts (in-character or out, in-game or out) negatively to a rape scene perpetrated upon her (or any) character, some groups will interpret this as a failure by the player to recognize the lines separating gameplay from ordinary discourse; more insidiously, perhaps, the player may feel that she should not overtly respond negatively, precisely because she accepts that other players grant this absolute division of discursive spaces, de-legitimizing her own emotional response as confirmation that she is not a "serious" player.
The common RPG theoretical response to such a situation, at least in recent times, is to grant the legitimacy of the player's response.  But this is formulated as a special case: certain types of in-game discourse "cross the lines" or "go overboard."  By implication, normative in-game activity does not require such responses, and thus this theoretically symptomatic treatment of the situation continues to emphasize that gameplay constitutes a protected space by constructing new social-contract rules to prevent specific problems.  That is, theoretical criticism of the rape situation proposed above amounts to this: RPG groups and games ought to have rules that say that players' characters cannot be raped.  But this misses the point.  On the one hand, it constrains RPG discourse to a limited range of social issues, making commentary and criticism of rape (for example) simply a prohibited discourse, undermining the very dynamic freedom which is supposed to permit a player to deal with situations that he or she would or could not encounter in real life; on the other, it retains and protects the hegemony of RPG discourse as something within which players may not respond personally or emotionally by making those situations in which such responses are legitimate into abnormal cases.
Continuing the comparison to initiatory ritual in particular, we have here an extra-ritual response to contingent historical circumstance through limitation.  In the case of the Bemba girls' initiation mentioned above, let us suppose that a girl responds to the figurine by saying, "If I become like the figurine, the white organizations that provide support and health services will give extra assistance even outside of infant care; therefore for my family in the current situation the appropriate answer to the riddle is that I should throw over tradition and use pregnancy to create a cargo-cult reciprocity with whites."[26]  Here we see a creative, dynamic response to the symbolic structures proposed, but with an ultimate response at odds with the hegemonic intent.  An obvious counter-response would add additional symbols and instructions to prevent this response by future neophytes, and perhaps provide extra-ritual instruction of this particular neophyte so as to annul the validity of her solution.
In RPG ritual discourse, the same structure of constraint through piecemeal placation consistently obtains.  To the extent that RPG players understand themselves as creative and dynamic, not controlled by encultured norms, they are enabled to reproduce challenged norms within gameplay as protected space.  That is, the liberation and protection afforded players with respect to uneasy social issues tends only to enable players who (often unconsciously) represent majority discourses to reenact the violence of those social categories in a hegemonically protected fashion, defended by the structure of the RPG as separated and distinct.  If the white, male player's black, female character enacts stereotypes, the notional freedom explored merely reproduces dubious social norms, an effect seen overtly in fantasy and science fiction book cover images (e.g. the work of Boris Vallejo), with their manly men with weapons and voluptuous women in revealing clothing.
To shift the modalities of play from reproductive to transformational may be desirable, but it is unclear how this might be effected.  While RPG ritual liminality permits exploration, its structured and constrained nature acts to defend stereotype reproduction as "freedom" while blocking challenges thereto as failures of player technique or understanding.  Logically, practical game-construction cannot merely strive to forestall deployment of stereotypes, but must work actively to undermine their function within gameplay; it is here that critical formation of counter-hegemonic moves (e.g. feminist game design) must focus effort, at the same time recognizing that simply formulating a game that pre-determines the boundaries of appropriate and inappropriate structure challenges cannot achieve anything.

Disjuncture and Continuity
As we have seen, the liminal phase of passage ritual, or more broadly the "sacred space" effected by social disjunctures outlining any ritual practice, affords a privileged site for examination and contestation of extra-ritual concerns; this sacred space in RPG's is found in gameplay, often understood as a "safe" place for exploration, and distinguished from other active spaces by a number of explicit and more subtle formations.  So far, I have focused on how such privilege and safety becomes a double-edged sword, permitting some forms of experimentation while denying others legitimacy, and also undercutting the radicalism of experiment to render it harmless.  But as with any ritual, the protective structures that reproduce hegemonic discourse formations are themselves genuinely threatened by in-ritual challenges.  It is worth considering how such challenge may be formulated through semiotic manipulation in gameplay.
In The Savage Mind, Claude Lévi-Strauss suggested that ritual tends to be conjunctive, as opposed to the disjunctive, classifying emphasis of myth.  His meaning is best expressed, perhaps, in a discussion of the difference between game and rite:
 All games are defined by a set of rules which in practice allow the playing of any number of matches.  Ritual, which is also 'played', is on the other hand, like a favoured instance of a game, remembered from among the possible ones because it is the only one which results in a particular type of equilibrium between the two sides.  The transposition is readily seen in the case of the Gahuku-Gama of New Guinea who have learnt football but who will play, several days running, as many matches as are necessary for both sides to reach the same score.  This is treating a game as a ritual....  Games thus appear to have a disjunctive effect: they end in the establishment of a difference between individual players or teams where originally there was no indication of inequality.  And at the end of the game they are distinguished into winners and losers.  Ritual, on the other hand, is the exact inverse: it conjoins, for it brings about a union ... or in any case an organic relation between two initially separate groups....[27]
The point is that a game like soccer or Monopoly takes a group of people not initially distinct in game terms and divides them into at least two classes (winners and losers).  By contrast, the ritual performance of soccer described here does not conclude until all players have been made equivalent; latent in Lévi-Strauss's formulation is that the natives project their preexisting social divisions upon the game by picking teams upon non-arbitrary given grounds.  For example, they might decide that each team will be made up exclusively of initiated men of a given moiety, so that the teams represent moieties; through the ritual process, they then construct a situation in which this difference is asserted as non-absolute.  This is arguably the point of the modern Olympic Games: national participation through representative athletes is supposed to assert that all men are brothers, that superiority is individual and not national, and so forth.
Setting aside the numerous quite serious problems with Lévi-Strauss's theory with respect to ritual as a broad range of behaviors -- indeed, I doubt he intended that it be taken as a general principle in the first place -- we can see this dynamic at work in a major RPG discourse, particularly that which emphasizes the collaborative nature of play.  As we have already seen, in Kim's Collaborative Storytelling model "play is understood as multiple authors producing a single discourse and a single story."  The same model discourages secrets among participants, and judges success partly by whether "all of the participants significantly contributed to that discourse."  Following up Lévi-Strauss's notion, we can see here a striving toward conjunction and unity, as against disjuncture in the form of "winning" or limited player dominance of the discourse.  In other words, one of the distinctive characteristics of RPG's as opposed to more traditional games is precisely that they fit a ritual rather than a game model.
At the same time, a more serious deployment of structural and practice perspectives on the semiotic elements of both religious and RPG ritual must recognize the oversimplification inherent in this conjunction/division split.  First, that there are no winners or losers cannot be accepted uncritically.  Precisely because a dominant RPG discourse denies such divisions, we must consider the possibility that play imposes upon players a notional unity by denying the option to seek or even accept division.  After all, if we extend this rhetoric of unity, it can be taken as a claim that in-game, all players are equal and in fact equivalent, which may be deployed strategically by situationally- or socially-dominant players to assert that complaints are anti-group and thus mark bad players.  In this context, the discourse of collaboration and unity can support the problematic use of hegemonic authoritarian or oppressive discourse, as discussed previously in the context of chauvinism.
But not all such challenge necessarily supports authority or serves as an instrument of oppression.  To take a simple example, the rhetoric of unity and conjunction may be deployed to block favoritism or to identify problem players as those who either try to dominate play or refuse to participate at all.  Especially in the latter case, the unifying effect of ritual process may enable a group to draw out a timid player, emphasizing further the liminal "safety" of game space.
More interestingly, however, the conjunctive nature of ritual process may act together with the aggregation of ritual closure to effect genuine social alteration.  A play group is often formed on an ad hoc basis, where some players do not know each other well outside of the game context, and indeed may not have met.  Through successful ritual collaboration in a shared space understood as distinct from other social spaces, a new social group forms, enabling friendship and other forms of collaboration that refer to the constructed game-space rather than to other social structures.  That is, precisely because gameplay is at once divided from other social spaces and nominally focused upon a limited set of predetermined issues, and because such rituals do act conjunctively by taking given divisions and annulling "winner and loser" categorizations, gameplay tends naturally to formulate an alternative social framework.  Particularly for those who find mainstream, dominant social frameworks problematic or dangerous, gameplay can constitute a controlled social space in which to succeed and seek liberation.
However psychologically supportive and validating such an alternative framework may be --  and it is worth noting that some psychologists have pointed to RPG's as valuable for self-exploration and validation among (especially) teenagers -- from a broader social perspective we should recognize that this essentially entails a continuation of the initiation discourse.  Turner notes that it is common that the neophytes, whatever their extra-ritual socio-economic status, are as part of the liminal leveling considered equivalent.  While friendships among those simultaneously initiated often extend beyond the ritual situation, social status, factored out within liminality, is not particularly affected by such friendships.  That is, it could be argued that the shared space of ritual, although it permits and even demands reflection upon social inequalities, ultimately acts not only to affirm these inequalities as natural and given, but also deludes those in inferior positions into thinking that they achieve a measure of equality that is in fact nonexistent.  From this perspective, we can see that RPG's may act simultaneously to affirm and assist players psychologically, and at the same time discourage them from acting upon or challenging the inequities of modern social dynamics.  Anecdotally, at least, we seem to see this in stereotypes of RPG players as "geeks" or "nerds" who, by participating in gaming, in conventions, and generally in a subculture, are thereby diverted or distracted from real social action or mobilization.  To formulate a rather overstated Marxist reading, the recognition of RPG's as ritual is confirmed by its ability to serve as an opiate for the oppressed.

Conclusions: Toward an RPG of Practical Reason
At present, RPG theory primarily acts as an exterior, supporting discourse referred toward the "real thing" -- gameplay.  Ironically, criticism of some RPG theory as irrelevant or trivial, on the ground that it is not practical for play goals, actually serves to grant power and hegemony to theoretical discourse: the very fact that gameplay so strongly formulates the barriers between in-game and out-of-game, play and system, in-character and out-of-character, reproduces the mystification of theory's active role in discourse construction.  As a way of concluding this somewhat dispersed series of analyses, then, I should like to propose some new directions in theory, directions which I think contain the possibility for real practical change.
First, theory must recognize a distinction between analysis and synthesis.  While it is important that such a distinction not become the object of fetishism, as it in a sense already has, the mystification of the aspect of RPG's traditionally associated with hierarchy and power can only lead to abuse on the one hand, analytic sterility on the other.  As Kim points out for Collaborative Storytelling, "It considers the rules system to be outside of the meaningful product.  Rules are judged on their results for shared play, not on how the participants view the process."  This perspective sets aside the impact of system and theory upon gameplay, asserting player freedom and collaboration instead.  While such a view may seem liberating, and indeed may be so as against old-fashioned GM authoritarianism, it implicitly claims that RPG performance occurs outside of structure, not in reaction to it.  But since social structures and presumptive traditions of play at the least are necessarily at work in RPG performance, there can be no doubt that gameplay has a structured context; were this somehow not the case, and gameplay fully liberated from exterior structures, there could be no possibility of conflict or its resolution, as no player would have a context within which to react conflictually.  Thus while a particular group or style may wish to formulate a liberated play modality as ideal, this has an ideological function and serves to replace one authoritarian structure (GM authority, game-system authority, etc.) with yet another.  In order for theory to advance the improvement of gameplay, then, it must work to distinguish between analytical activities and constructive or synthetic ones, and furthermore strive to bring this to consciousness within actual play.
Second, RPG theory needs to take seriously the contributions and insights of other disciplines.  Eventually this should be a reciprocal engagement, but this will require acceptance by academic and other mainstream intellectual theorists; insofar as RPG theory can support such a move, it must do so by engaging actively and constructively with such theorists, in language acceptable to their traditions.  In the meantime, RPG theory must set aside its tendency to see its analytical object as unique and thus special.  William James reminds us forcefully,
The first thing the intellect does with an object is to class it along with something else.  But any object that is infinitely important to us and awakens our devotion feels to us also as if it must be sui generis and unique.  Probably a crab would be filled with a sense of personal outrage if it could hear us class it without ado or apology as a crustacean, and thus dispose of it.  "I am no such thing," it would say; "I am myself, myself alone."[28]
James's point is clear: while we are willing to make all sorts of classifications within RPG's, we tend to think of RPG's as unique and thus special.  But "unique" is simply a logical category that can be applied to any object of analysis supporting formulation as a categorical object.  If RPG's are unique, that does not mean they are not ritual, or social behavior; it only means that they can, from a particular perspective, be formulated as having some distinctive characteristics.  So long as RPG theory continues to formulate itself otherwise, as unique in an illogical, strong sense with respect to other behaviors, such theory will continue to be marked by two unfortunate properties: first, it will be perpetually in the position of many religious discourses of having continually to defend its boundaries against the incursions of other discourses and analytical methods; and second, it will be incapable of real analytical force because it has built into its very self-definition essentialist biases that again require constant and vigilant defense.  Arguably, the tendency of much RPG theory toward rigid hierarchization and toward discourse-circle hegemony would thus constitute a parallel to more obviously religious dogmatisms.
Third, RPG theory requires models founded upon a productive and reproductive, as opposed to interpretive and receptive, situation of narrativity.  Two obvious examples, Kim's already-cited article and Liz Henry's "Power, Information, and Play in Role Playing Games,"[29] are admirable moves toward intelligent application of exterior models, but find themselves at odds with the purposes of those models.  Kim's awareness of this problem is clear:
There are many differences between RPGs and books [upon which the formalist model is built], but some are more subtle than others. It is clear that RPGs have no division between author and reader. Each participant both expresses and interprets. Further, this calls into question what the story is. The answer depends in part on what we define as the discourse or "text" of RPG play.
These questions are essential, and require answers; indeed, even cursory examination of recent RPG theory reveals a constant concern to formulate authorship, textuality, and so forth with respect to RPG's.  But these debates mostly run around in circles, die out, and get revived with new energy but no really new formulations, with endless repetitions of the cycle.  The problem, in short, is that formalist and hermeutical models are founded on confronting the genuinely difficult problem that interpreting a text is not comparable to a conversational situation; intricate and elegant strategies are deployed to make sense of how we make sense of text, if you will, given that it is not conversation.  But RPG's are conversational; the problem does not arise directly.  By attempting to read RPG's through such lenses, we are caught in circularity: conversations are like books (except that they are not face-to-face), and books are like RPG's (except that the latter are face-to-face).  Why not drop out the sidetrack and recognize RPG's as active, dynamic, conversational forms of symbolic manipulation?  I have attempted a beginning here, but a great deal more needs to be done.[30]
Fourth, stemming from the last point, RPG theory must take into account the social issues at stake and at work within the smallest, most apparently arbitrary activities of play.  That so much discussion of "problem games" focuses on social difficulties -- problem players or GM's, paradigmatic clashes, etc. -- reveals that the central issues in play are social.  To the extent that RPG theory tends to work hierarchically, from top-down (broad categorical strokes before specific game issues), it mistakes the actual dynamics by incorporating its analytic framework into problems needing resolution; this is another means by which theoretical discourse mystifies itself and its contributions, and it can most effectively be challenged from within theory itself.
Fifth, RPG theory must, through engagement with broader social theory -- particularly the mode of anthropological theory labeled "practice" -- become aware of symbolic and structural manipulation as a strategic part of everyday life, a set of techniques also employed (and refined) within the specifically RPG context.  This occurs at every level of play; there can be no absolute divisions between in-game and out-of-game, for the same reasons that the only absolute division between a Catholic Eucharist and a Catholic's everyday life is an ideological one.
Finally, RPG theory must move beyond hierarchical classification as a technique.  There is no question that classification is a valid, even necessary goal for serious analytical work.  But as in so many disciplines, most notably the study of religion, the tendency is to use the scientific character of classification to construct an aura of objectivity; we see this in discourses that stress "correctness".  The natural upshot of such an endeavor is to reify the categories as ontologically legitimate, mystify their constructed character, and thus naturalize the authority-claims latent within such structures.  Classification must recognize that the object does not exist outside of the construction of taxa; "religion" or "ritual" do not exist, but are means by which historically situated and motivated people classify certain behaviors.  Similarly, "RPG" is not a thing, a singular object, unique and discrete from others, and Narrativist orientations do not differ from Simulationist or Gamist ones except insofar as we construct them so.  Classification is the basis of comparison, not of truth or certainty.  Until RPG theory takes on board serious recognition of its comparative nature, it will remain an ideology and not a science.[31]


Notes

1. E.g. Ron Edwards' game Sorcerer (Chicago: Adept Press, 2001; see www.sorcerer-rpg.com).
2. Edward's views have been formulated in several articles, all of which may be found at The Forge (http://www.indie-rpgs.com).  Apart from the library articles, a useful recent discussion started by Edwards is "The whole model - this is it" (http://www.indie-rpgs.com/viewtopic.php?t=8655).
3. Stable URL: http://www.darkshire.net/~jhkim/rpg/theory/narrative/paradigms.html
4. Stable URL: rec.games.frp.advocacy.
5. The Forge has hosted lengthy discussions of how RPG play is like playing in a band (with the gamemaster playing bass), how RPG play is like playing a pinball machine, and so on. Examination of the range of such discussions will show the two discursive thrusts: the drive for clarification and precision in the metaphor, and the extension of the analogical range. As a rule, such discussions end when those who find the analogy helpful have formulated a version that is clear to them personally, when those who do not find it so grow tired of trying, and when most become frustrated with those who try to extend the analogy to ludicrous, literalist extremes.  These discussions are not worthless -- I intend no such general criticism.  But in order to work effectively as analytical models, such metaphors must be formulated rigorously, with their boundaries precisely set.  For more casual discussion, on the other hand, one of the best qualities of a forum like the Forge is that it permits this sort of open speculation and play; indeed, a close analysis of the ludic dimension in such RPG discourse would be valuable for understanding the interrelations of RPG play and theory.

6. On the issue of the "unique" as special, and its problematic applications to serious analysis within classificatory discourse, see Jonathan Z. Smith, "Fences and Neighbors." Imagining Religion (Chicago: University of Chicago Press, 1982), 1-18.
7. See Ronald L. Grimes, Beginnings in Ritual Studies (Washington, D.C.: University Press of America, 1982); Victor W. Turner, Dramas, Fields and Metaphors (Ithaca, N.Y.: Cornell UP, 1974); Turner, From Ritual to Theater: The Human Seriousness of Play (New York: Performing Arts Journal Publications, 1982). Essentially all of Grimes' work work since the late 1970's fits the model am describing here, as part of what he has dubbed "ritual studies".  Turner's work, however, took a strictly performative and dramatic turn; his earliest works, while excellent, do not directly fit this model, and can only be made to accord with the performative perspective with considerable hindsight and, I think, distortion.
8. See Claude L´vi-Strauss, The Savage Mind (Chicago: University of Chicago Press, 1966); Lévi-Strauss, The Naked Man, trans. John and Doreen Weightman (Chicago: University of Chicago Press, 1990); Pierre Bourdieu, The Logic of Practice, trans. Richard Nice (Stanford: Stanford UP, 1990); Sherry Ortner, by Since the Sixties,biety and History 26.1 (Jan. 1984), 126-66; Catherine Bell, Ritual Theory, Ritual Practice (Oxford: Oxford UP, 1992).
9. The French idea of bricolage is not directly translatable into English; we simply have no category quite like it. The bricoleur is a hobbyist of a sort, but elevated to a high artistic level. For the Lévi-Strauss formulation, see The Savage Mind, chapter 1, "The Science of the Concrete"; the translation is execrable, and those with a good command of French would be well advised to read La pensée sauvage, chapter 1, "La science du concret."
10. Stable URL: http://194.29.64.17/thecog/movie.html
11. I shall not go into detail on hermeneutics, as it is founded primarily on philosophical negotiation of the problems of interpretive reception, problems relevant but not central to the analysis of RPG's.  On this model, see Paul Ricoeur, Hermeneutics and the Human Sciences (Cambridge: Cambridge UP, 1981). See also Umberto Eco, Interpretation and Overinterpretation (Cambridge: Cambridge UP, 1992); and Hans Georg Gadamer, Philosophical Hermeneutics (Berkeley and Los Angeles: University of California Press, 1977).  Also useful, though less approachable, are Eco's The Limits of Interpretation (Bloomington, IN: Indiana UP, 1994) and A Theory of Semiotics (Bloomington, IN: Indiana UP, 1979).
12. A central tenet of hegemonic Forge theory.
13. See Mike Holmes, "Mike's Standard Rant #3: Combat System" 
    (http://www.indie-rpgs.com/viewtopic.php?t=2024). Holmes' essential point is this: "If you don't want combat to be the focus of a game, do not include special rules for it. Especially if you don't include special rules about anything else."  This "standard rant" has been discussed periodically on the Forge.
14. It should be pointed out that the Forge "system matters" principle does not claim that other elements do not matter; the question is one of emphasis, and is here an analytical distinction rather than a polemical one.
15. See iago [Fred Hicks], "Long Pig the RPG: Would You Play It?" 
    (http://www.indie-rpgs.com/viewtopic.php?t=6091).
16. Jonathan Z. Smith, "Fences and Neighbors," Imagining Religion: From Babylon to Jonestown (Chicago: University of Chicago Press, 1988), 1-18. The polythetic system is hardly perfectly objective, but as Smith argues persuasively, it is less inherently inclined toward normative claims and slippages than the monothetic, taxonomic sorts of systems founded on hierarchy.
17. Although see his Deeply Into the Bone: Reinventing Rites of Passage (Berkeley and Los Angeles: University of California Press, 2002), the purpose of which is explicitly to formulate ritual theory as a constructive discourse for people wishing to invent or reinvent their own rites of passage.
18. The commensuration of ritual discourses and discourses about ritual, between ritual in fact as analytical discourse and academic analysis as in fact ritual, is outside the scope of the present paper. The argument, founded upon a grammatological engagement with practice, performance, and structural analysis, juxtaposed to early modern magical practice and the theoretical dramaturgy of Zeami's Nö, will be part of the core of my book Magic in Theory and Practice, where I do not connect it with RPG's per se.

19. Arnold van Gennep, The Rites of Passage, trans. Monika B. Vizedon and Gabrielle L. Caffee (Chicago:
University of Chicago Press, 1961); Victor Turner, "Betwixt and Between: THe Liminal Phase in Rites de Passage," Proceedings of the American Ethnological Society, Symposium on New Approaches to the Study of Religion, 1964:4-20; Turner, The Ritual Process: Structure and Anti-Structure (Aldine de Gruyter, 1969); Turner, The Forest of Symbols: Aspects of Ndembu Ritual (Ithaca, NY: Cornell UP, 1970).
20. "Betwist and Between," 13, citing Audrey I. Richards, Chisungu (London: Faber and Faber, 1956), 209-10; the new edition is Richards, Chisungu: A Girl's Initiation Ceremony Among the Bemba of Zambia (London: Routledge, 1982).
21. "Railroading," for which there are numerous more or less equivalent terms, is the practice of a GM essentially scripting the majority of plot events and structures within a given play session or series of such.  For example, the GM may decide, prior to play, that he wants the PC characters, all cowboys, to engage in an OK Corral-style gunfight as the climax of play; when the PC's choose (via their players, of course) to ride out of town to investigate a lost silver mine, the GM uses various strategies to prevent them from doing this, because he needs them in town in order for the gunfight to take place.  Such strategies range from subtle hints to overt assertions of authority; a possible example would be to inform the players that several of their horses are lame and cannot be ridden, then to have no horses available at the town stable, then to ensure that nobody in town will sell his or her own horse. By the time the players have negotiated this many options, it is generally clear to everyone (though very often not stated) that no matter what they do, the PC's will be prevented from riding out of town.
22. This point has been emphasized in various RPG discussions. One common suggestion is that if, for some reason, the GM actually needs her players to follow a set of railroad tracks, the GM should react to repeated attempts to jump the rails out-of-game, by saying something like, "Okay, guys.  I'm really not that prepared, actually, and I kind of need you to go and do X. Is that okay?" While this may act practically to achieve the desired effect, it depends upon the rigidity of in-game/out-of-game divisions to acquire efficacy, and cannot in itself be deemed a resolution of a more fundamental difficulty.
23. I would agree with these thinkers that people never think truly independently, that is unconstrained in any manner by encultured structures; the point here is that even constrained thought and action has tremendous flexibility and ranges of possibility, and is not simply scripted or railroaded in the RPG sense.
24. This division is reproduced in strictly academic contexts not only with reference to ritual but also to myth: myths are not "really" myths if they are invented for that purpose (whatever such a purpose might be), just as rituals as not "really" rituals if they are consciously invented so. The intrusion of dubious ideas of consciousness, ontology, and category only deflect from the central point: academics by formulating critique in this fashion reproduce the ideology of authenticity that authorizes and legitimates certain religious behaviors as stable and non-inventive, as against the "wannabe" inventions of recent "flakes" and "crazies".  In a sense, we might see the division here as between those who are creative within an authorized framework and those who create their own framework. The critique thus becomes reflexive, as indeed we should have suspected it always was: the academic is really saying that she herself, by being creative (doing new analytical work) within an authorized or traditional framework (academic and disciplinary traditional discourse) is legitimate and critical, while "crazies" (those proposing unexpected critiques) fall outside the authorized framework (do not have Ph.D.s, for example) and thus need not be taken seriously.
25. It would be interesting to consider whether the apparent (though entirely anecdotal) overlap between RPG communities and Neopagan ones might be at least partly rooted here. In the absence of serious sociological data, I suspect that an effective technique here would be close analysis of White Wolf's various Neopagan-oriented games (especially Werewolf and several of the Ars Magica supplements) with respect to ritual/magical creativity, criticism of religion, and criticism of what the authors refer to as "traditional" games in their explanations of how their games are special and different.
26. This is a purely hypothetical construct; I know of no such actual response among Bemba, and the example is deliberately over-simplified for heuristic reasons.
27. Lévi-Strauss, The Savage Mind, 30-32; the reference on the Gahuku-Gama is to K. E. Read, "Leadership and Consensus in a New Guinea Society." American Anthropologist 61.3 (1959): 429.
28. William James, The Varieties of Religious Experience (New York: Longmans, Green, and Co., 1902), 9. See also Jonathan Z. Smith, "Fences and Neighbors" for a penetrating discussion of the "unique" in theoretical discourses.
29. http://www.darkshire.net/~jhkim/rpg/theory/liz-paper-2003/
30. The same point might be made about Edwards's dependence upon Lajos Egri's constructive models for creative writing, models poorly suited to analytical purposes. In essence, Edwards asserts that Egri's models fit RPG's, except that the product is entirely different, authorship is shared, and really the Threefold Model is analytic rather than constructive.
31. Here I take science to be a reflexive and self-critical attempt to differentiate and understand its analytical objects.  There can be no question that modern science, in the usual sense, does not always fulfill these criteria, in particular because it tends to claim objectivity instead of constructed reflexivity.  But given the need for such reflexive awareness, the goals and ideals of science remain worthy of theoretical discourse; see the introduction and first chapters of Bourdieu's The Logic of Practice for a brilliant (if dense) formulation of scientific analysis that recognizes and takes seriously its own constructed nature.  For comparison as a discourse and a method, Jonathan Z. Smith's Imagining Religion should be the starting-point of any attempt at theoretical construction.