The Ontology of Game Spatiality
by Frederik BakkerudAbstract
In the last fifteen years especially, a paradigmatic shift has been observed in game studies, a paradigm in which games are consistently conceptualised as spatial objects. This calls for methodological reflections, one of which is this article’s focus: the ontology of game spatiality. Scholars frequently depend on metaphorical concepts -- for example, “labyrinths,” “gardens” or “virtual cities” -- to describe something as intricate as game space, the ontological levels of which they frequently mistake. This article seeks to remedy these tendencies through a formal framework, achieved by taking a step back from particulars and approaching games from the general perspective of ontology. More specifically, it takes the cybermedia model as a point of departure in order to formalise a four-layered framework of spatiality comprising the representational, the mechanical, the material and the player level. This ontology allows for analysis and comparison of vastly different game systems, even those upheld entirely by the players themselves, such as physical sports like association football.
Keywords: game spatiality, game space, formal analysis, spatial analysis, game ontology, virtual environments
Introduction
In 2008, Stephan Günzel observed the beginning of a paradigmatic shift in game studies, a paradigm in which games are consistently conceptualised as spatial objects (see also Aarseth, 2001a; Aarseth and Günzel, 2019; Murray, 1997). This has called for methodological reflections -- it follows that one may take up this question from a variety of approaches, and while Günzel does so from a phenomenological perspective, the present study adopts an ontological view, asking more specifically: how may researchers practice formal analysis on the spatiality of games? In other words, how may we formalise the ontology of spatiality in game studies? Suggestions as to these questions can be found throughout game studies (most recently Debus, 2019), but no single framework has provided a satisfactory point of departure. Some frameworks are too general, that is, not specialised towards spatiality; some are unduly essentialist and some simply conflate the intricate layers of the game object. This article pursues an ontology that is not essentialist, which is to say, does not aim to define what games -- or their essential components -- are. On the contrary, it formulates four separate levels of game space available to humanistic analysis, and with which scholars may more rigorously study the object at hand.
This study is meta-methodological, in so far as it is concerned with theories of games more than empirical games in themselves. As with game ontologies generally, the questions are approached deductively -- and by necessity [1]. Game spatiality is evidently a rather indefinite concept, not least when it comes to the highly diverse phenomena people generally consider to be games in the first place. In fact, the entirety of the game object, and not merely the observable (more precisely virtual) environment, may be analysed in terms of spatiality. Interfaces, too, are topological structures -- as are hypertexts [2]. What may be considered spatial in games, in other words, exceeds that which is merely “moved around.” Indeed, one may humorously object that structure (as abstracted representation) is already in itself “spatial,” so let me emphasise -- in the ontology posited in this article, spatiality is considered to encompass not merely a “structural” (or ludological) level but several levels of the game object, including but not limited to the signifiers “on the surface.”
The four-layered framework, as will be explained, takes Espen Aarseth and Gordon Calleja’s cybermedia model (2015) as its point of departure. That model comprises four un-hierarchised elements of the game object as a whole: surface signs, mechanical systems, material mediums and players, respectively. These elements -- henceforth, levels -- are presented as a layered (meta-)structure in the present article and specialised towards a spatial framework, comprising four corresponding levels: the representational, the mechanical, the material and the player level. Most importantly, this framework is built on the assumption that the mechanical level is not considered to denote the software code of the game; in fact, I am going to argue that the software code is not a sustainable ground on which to build ontologies for the humanistic study of games in the first place. This view entails that the game is experienced by a player, acting as a researcher, and that the game world is treated as real or virtual and as such navigable -- regardless of what goes on in the binary code “underneath.” How else would the theory be general, that is, applicable to more than merely digital games?
The goal of such an ontology pursued in this article is to build clearly defined terminology and methodology, with which game scholars would not mistake, for example, the representational for the mechanical level as frequently as is currently observed. In terms of spatial structures, scholars need better formal terminology than, say, “open landscapes” (Aarseth, 2005), “one-room games” (Aarseth, 2012) or “virtual cities” (Nitsche, 2008), all of which are metaphors and amalgams of representative and cultural features. Likewise, the meaning of spatial concepts such as “paths” should never be taken for granted in game studies, as these concepts are so highly dependent on their disciplinary context (for an elaborate criticism of game spatial metaphors, see Bakkerud, 2022). I believe an ontology such as the present one may provide a step in the right direction -- towards fewer ambiguities and mistakes in studies of the intricacies of game space.
This ontology is presented -- and indeed, illustrated -- by the end of this article's second section. The first section outlines the assumptions on which this ontology is grounded. An analysis of the cybermedia model follows in the second section, after which the four levels are clarified separately, including an account of the concept of virtual environments that has proven necessary to consolidate the framework. Only then is the ontology presented with various examples from academic discourse. Many perspectives should benefit from these findings, seeing as a variety of research questions are dependent on concepts of game spatiality. As Aarseth and Sebastian Möring argue with regard to game hermeneutics, “every game interpretation rests in a particular game ontology, whether implicit or explicit. The quality of this ontology then, determines a vital aspect of the quality of the analysis” (2020).
Preliminary Assumptions
Accounting for the ontology of game spatiality necessitates a few principles about how this structure is observed in the first place. If one looks briefly to the narrative structure of “traditional” literature, one finds that it is commonly located “beyond” the surface of textual signifiers in a metonymical sense [3]. Gérard Genette, most notably, describes narrative as a “pseudo-time” transposed from the space of the linguistic text (1980, p. 34; p. 86) -- that is, as a sequence of events transposed from a syntagmatic order of the words. In games, though, time and space are not merely signified. Structure is not merely metonymical. Games are virtual structures of what Aarseth has called ergodic textuality (1997) -- inherently different from the literary artefacts characterised by Genette. The structure of game spatiality is at once navigable and signified. Structure not merely constrains the space available -- it is at the same time communicated to players through a layer of spatial signifiers.
Many studies approach game spatiality from the disciplines of either ontology or phenomenology -- to put it briefly, on the objects in themselves or the objects as phenomena of experience, respectively. The present article builds on the former, but that is not to say that this is a solid dichotomy, as research can hardly disregard one or the other. Analysis is always mediated by game experience, as opposed to the binary code running “underneath” the game software. Consider the opening sequence of Half-Life (Valve, 1998). The player finds oneself on a train that stops, during which there is a brief loading screen, but space is virtually unchanged. Due to technical constraints, the developers had to load another -- as far as the game software is concerned -- separate space with no visible changes, for the train to be openable (Yang, 2018). To the game engine, the door was otherwise another “wall.” This example raises the question, should these spaces be differentiated on an ontological basis? Arguably not. Players have no practical access to binary code while playing a game, and humanistic researchers are generally not interested in an ontology of the binary code itself. The ontology pursued here is of “the game” as a playable structure, as an object that allows for an almost (or occasionally) endless variety of different sequences to be played out. This structure, or architecture, is separate from any particular game session, and posited on the basis of content gathered from experience.
To this end, one may adopt Michael S. Debus’ notion of “phenomenological commitments” with regard to the development of game ontologies (2019). In essence, these commitments characterise how such ontologies are ultimately built on the game’s experienced content as opposed to its binary code. More precisely, the spatiality of digital games is inherently discrete, in so far as it is executed from binary code -- in Debus’ words, even motions that “appear to be fluid or continuous to the player… are ultimately occurring between miniscule discrete locations within the computation” (2019, p. 154). This commitment, then, allows us to ignore the aforementioned loading screen of Half-Life -- to consider it a segmentation only on the (omitted) level of binary code. Let me emphasise that this preliminary assumption is concerned mostly with making the ontology compatible with digital and analogue objects alike. The aim is, in other words, to reach a degree of generality that applies beyond particulars -- to formulate a general theory, whose applicability remains consistent even as different objects from the same subset emerge. The alternative of, for example, adopting Lev Manovich’s cultural layer and computer layer (see 2001, pp. 45-46) would forego generality in the research context of games, the consequences of which are entirely unnecessary. The board game of Agricola, for example, in its analogue edition (Rosenberg, 2016) would be incompatible with this perspective, while its digital adaptation (Playdek, Inc., 2020) would be compatible despite their differences being merely material (and to a certain degree representational). Notice how this example also manages to showcase the general applicability of this ontology, specifically how the levels of either editions of Agricola are effortlessly identified and separated.
But how is structure observed in games? Daniel Golding is critical of what he calls “the totalizing, ‘from above’ perspective” in game studies (2013, p. 29). The conception of games as “configurative” objects implies a holistic view, the author argues, one divorced from “lived experience” (Golding, 2013, p. 30). The present article is arguably guilty in that respect, though it does not claim to have “full and unhindered vision” (Golding, 2013, p. 30) of the object at hand. Formal analysis does not necessitate unbounded knowledge of any particular artefacts. But Golding is right, in so far as games must always be played, even if characterised as a holistic object. As long as the analysis is concerned with abstracted forms, that one identifies fairly effortlessly, this is not actually a challenge. To suggest that games are characterised by mechanical systems among several other levels is not totalising by any means, and it is a reasonable condition that such levels are extracted from the game object and “applicable across different individuations of games” (Debus, 2019, p. 88). I am, to be sure, not advocating for the sort of “hard-line” ludology (van Vught, 2016, p. 6) commonly attributed to scholars such as Markku Eskelinen, who (in)famously claimed that “If I throw a ball at you I don't expect you to drop it and wait until it starts telling stories” (2001).
I must also emphasise that it is not my intention to suggest that phenomenology has less application in studies of game spatiality, or, for that matter, that mechanical systems are the focus of ontologies whereas sign surfaces are the focus of phenomenologies. More accurately, I argue that ontology is better equipped for classifications of the intrinsic qualities of these artefacts, as phenomenologies arguably dissolve the levels of mechanical systems and surface signs. Like Debus, I believe game spatiality should be conceptualised in terms of that in which is moved, as opposed to that which is created by movement (2019, p. 196). Do note that Debus contrasts this view with that of post-structuralists Augé and Lefebvre, but I consider it reasonable to contrast it with the discipline of phenomenology too -- seeing as space, in that view, is arguably inseparable from experience.
Either discipline is incredibly useful, and the above summary is understandably reductive, but it is not feasible to explore their relationship much further within the scope of this article. That is a topic for further study, although this article fully acknowledges that any discussion of these phenomena’s “intrinsic qualities” naturally entails an experienced object, a structure observed and analysed -- indeed re-constructed (for analysis) -- by a player(-scholar). In any case, humanistic scholarship seems to largely agree that the binary code “underneath” the played game is not the (well, primary) object of study. In the following sections, this article explains the model in more depth before comparing various ontologies and practical examples that could benefit from this perspective.
The Cybermedia Model
This article builds on Aarseth and Calleja’s cybermedia model -- “a matrix constituting four elements: the representational (or surface) sign, the mechanical system [sic] the material medium and the player” (2015). Do note that cybermedia comprise more phenomena than merely games. In principle, the proposed ontology should characterise cybermedia too -- but for practical reasons, “games” will remain the analytical focus. Researchers seem to always find exceptions to their scholarly distinctions with time and should as such limit their scope carefully. The following section accounts for each of the elements and explains the process of adoption towards this article's ontology. Henceforth, elements are considered levels and labelled accordingly -- the representational, the mechanical and the material level, with the player added “outside” this layered structure, which will be explained in due time.
Following “general semiotic theory,” surface signs refer to the interpretable surface of the game object, “the representational elements” including “alphanumeric text, imagery or sound or other expression types” (Aarseth and Calleja, 2015). This is the representational level, which naturally applies beyond computerised signifiers and extends to any representational phenomena of games. Mechanical systems, then, refer to the “mechanical operations which structure the process.” “Any mechanical changes that can potentially be recognized as such by an observer with sufficient access to the system’s informational sphere,” the authors argue, “will have been effected by the mechanical system” (Aarseth and Calleja, 2015).
The main challenge in terms of spatiality is to disentangle the surface signs from the underlying mechanical systems. What appears to be a door in the audiovisual gameworld may have no mechanics attached to it, nor any non-representational purpose, which effectively renders it a wall in terms of game space topology. Thus, the mechanical level is concerned with space not as it “appears” but as it is organised topologically. A given player needs not know, see or explore a spatial connection in order for it to be formally “there” in the game world. This view entails that space exists before it is populated by objects and agents, and that it may be extracted and analysed as such, as space “in itself” (see Debus, 2019, p. 196). The mechanical level of space may denote the game’s actual room-for-movement, but it needs not resemble real-life physics. In fact, even non-Euclidean game spaces whose disorientation seemingly defies humans’ experience of real-world space are organised by spatial, topological principles -- a network of paths, that is.
As with surface signs, the mechanical level is found in analogue and digital phenomena alike [4], even if mechanical activities are not computerised, seeing as they still have “processual consequence” (Aarseth and Calleja, 2015). This also applies to processes that are entirely upheld by the players themselves, for example, by verbal communication of actions and changes. In addition, as Jonne Arjoranta argues, what specifically pertains to computerised games mostly applies, at least in principle, to non-computerised games as well (2011, p. 3). Various physical books ranging from postmodernist works to the Choose Your Own Adventure series are topologically (or mechanically) organised as well, as Aarseth argues prior to the cybermedia model (1997). One consequence of this perspective is -- as previously mentioned -- that no formal differences are observed between the space topologies of analogue and digital sessions of Agricola (though of course they are different in terms of the representational and material levels).
The mechanical level arguably makes for the most distinctive (and controversial) level of games, especially when one considers the ergodics that Aarseth locates in “the mechanical organization of the text” (1997, p. 1). In non-ergodic, purely audiovisual media -- for example, cinema -- spatiality is merely visual as opposed to traversable (Murray, 1997, p. 79). In the words of Günzel, it is the “space-image” as opposed to the “image-space,” that characterises the computer game (2008, p. 171). Günzel may limit his arguments to digital phenomena, but the observation is relevant in any case, seeing as it elegantly captures the ergodic character of these virtual objects.
Representation, then, is not an exclusive (or essential) feature of games, and yet its structural level remains part of the ontology. The reason for this -- in part -- is that we can hardly define one and not the other. As Aarseth writes, “[t]he relationship between semiotics and mechanics is one of duality” (2014, pp. 488-489). It is precisely on account of this layered structure, that we must observe the surface signs in order to interpret “the level below.” As Aarseth argues in another article, “we usually have no problem seeing through the representation and discern the ludic reality” of games (2019, p. 128) -- which is to say, surface signs are “feeding” us information of the mechanical systems. In some cases, such as in text-based games, where formal space is signified by virtue of linguistic signifiers, there is arguably also a space signified by “the surface,” even if rather like an image of the mind. This world is invoked in the interpreter’s thoughts, slightly more similar to that of the classical novel (in terms of signifiers anyway), and simultaneously an inherent structure in the artefact itself.
To describe game spatiality in terms of a vertical structure is widespread in game studies, and for entirely valid reasons. It usually involves the two “upper” levels only, that of the representational and mechanical (which arguably stems from a time where game ontologies were more fashionable). One example is Debus’ formal ontology of game elements, in which he applies the dichotomy of gameworld and game space, the latter of which refers to “the navigable, limiting, and organizing formal structure” (2019, p. 213). Here, spatiality is conceptualised as a layered structure, as the “gameworld is considered to be added on top of the underlying spatial and navigable structure” (Debus, 2019, p. 214). With regard to game spatiality more specifically, Aarseth posits a conceptual dichotomy that is rather similar, although it is seemingly more concerned with visual spatiality in a geographical sense (and therefore less holistic) -- it comprises topology (mechanics) and topography (representation; 2019):
Ludic landscapes consist of two layers that are conceptually superimposed, but independent: The topographical, which is the sign-stream presented by the game engine to the player, and the topological, which is the actual room-for-movement through which the player’s tokens navigate. (Aarseth, 2019, p. 130)
According to Aarseth, “[t]opology rules the ludic world. Gameplay is topological, and the fidelity of the topography therefore yields to the ludic topology” (2019, p. 131). This conception, too, is clearly structured in layers, with signifiers superimposed on mechanical systems, like a surface.
Now, allow me to elaborate on the relationship between the mechanical and either of the additional three levels, especially seeing as the mechanical level can be said to “share” the same ontic spatiality. That is to say that the material level, should it -- as is the case in association football -- constitute a real lawn with stripes, is not easily distinguishable from the mechanical level “placed” on top of this. The same applies to a digital instance of the same mechanical space, in which the grass surface is situated on the representational level -- such as in the popular sports game series of FIFA. One may even argue that in game space where rules are entirely upheld by the players themselves, there is a shared ontic spatiality between the mechanical and the player level. In all instances, the mechanical level constitutes a set of spatial rules that co-exist with the other three levels (or surfaces). To solve conundrums like these, one may adopt the concept of the virtual environment, which Aarseth, too, has applied in many previous arguments. With this concept, I am -- like Aarseth -- referring to any actualisations of game objects, and not something in opposition to that which is “real” or “physical,” and certainly not something synonymous with “digital” or “computerised” -- in other words, a concept that is not limited to the description of video games only.
Aarseth in particular (2001b) has been critical of virtuality when the term is applied interchangeably with “digitality.” He defines “virtual worlds” -- generally interchangeable with “environments” -- as “dynamic representations of an artificial world” (Aarseth, 2001b, p. 228). “The distinguishing quality of the virtual world,” the author continues, “is that the system lets the participant observer play an active role, where he or she can test the system and discover the rules and structural qualities in the process” (Aarseth, 2001b, p. 229). Games are widely conceived of as simulations, but Aarseth posits virtuality as the more accurate concept, seeing as simulations depend on the existence of a “real counterpart” (2001b, p. 228). The present article will take this definition to include any system upheld by the players themselves too, such as before with association football [5].
The virtual, then, refers to the rule-bound structure that the game object itself invokes, with no regard as to whether that is played out in real life, in our minds or mediated through computer screens. As Debus writes, “from a structural/ludological perspective, a classification of game space can be universally applicable, despite the ontological differences of ‘real’ and ‘digital’ spaces” (2019, p. 193). Notice how Debus juxtaposes “structural” and “ludological,” so as to emphasise mechanical systems as the formal or structural element of games. A match of association football would constitute mechanical systems like any other game object -- even if these systems are material in the “real” sense (for many sports are indeed full of rules), and even if those systems would pertain only to the evaluation of winners and losers. In fact, Aarseth himself has argued that “football” is an example of ergodics (1997, p. 95). Sports simply uphold these mechanical systems on a conceptual level.
We may also approach this concept through Veli-Matti Karhulahti’s notion of “behavioural properties,” according to which virtual objects constitute “a mode of simulation” (2012, p. 3). According to Karhulahti, neither the sky nor the mountains of Half-Life 2 (Valve, 2004) hold any behavioural functionality, that is, neither may be explored nor manipulated by the player avatar. In essence, they are “ontologically mere textures of purely fictional nature” (Karhulahti, 2012, p. 4). What is here labelled “fictional” corresponds to that which is purely representational in the framework of the present article. The skies and mountains correspond to what Aarseth (1997), with reference to Genette (1982), characterises as mere “description” -- as opposed to “narration” or “ergodics” -- which is the equivalent of surface signs with no relation to mechanical systems.
In a text-based game like Dracula (CRL Group, 1986), there may be no visual spatial dimensions, but a virtual (and indeed traversable) environment is readily recognised (Karhulahti, 2012, p. 4). Space may at first seem “fictional,” that is, the mountains and forests are signified linguistically, but the player may explore these -- for example, by typing “examine mountains,” in which case the game responds: “High towering mountains hide the setting sun from a cold, still sky, far away in the distance.” Had the player examined the sky in a similar manner, the game would merely respond: “I can’t see any such thing” (either citation from Karhulahti, 2012). The sky, then, constitutes a representational level only, unless the player somehow comes across a typed command, not familiar to the researcher otherwise -- and as such, this also exemplifies very well the epistemological shortcomings of any ontologies of phenomena such as these.
In the aforementioned example, we further observe a mechanical level that is only “spatial” in so far as it is “structural.” This is the case in (analogue) card games like The Mind (Warsch, 2018), where “players draw cards which display numbers from one to 99 and keep their drawn numbers hidden from the other players. The players then, without communication, place the cards face up, one after the other, in the correct order of numbers” (Debus, 2019, p. 224). This space is arguably no more spatial than the syntagmatic order of linguistic texts, because we merely consider the order of cards. In terms of definitions, however, it exceeds the scope of this article to differentiate this from other “kinds” of spatiality. In an intuitive sense, one could employ the concept of virtual environments only to characterise game spatiality more comparable with movement in “real life,” as opposed to any level of space only accessible in a conceptual manner. That indeed seems more natural with regard to the notion of “an environment.”
Some pen-and-paper games such as Dungeons & Dragons may employ spatial combat environments based entirely on the mind of the dungeon master, who then explains to the players their approximate positions in an imagined space -- as opposed to the common variant of employing a visual map with a clearly defined grid representing players’ exact positions in the environments (the former known as “theatre of the mind”). As such, these approaches clearly differ on all levels of the model presented in this article despite the fact that they, simply put, may be telling the same stories. They are represented differently, the latter verbally and visually and the former entirely verbally. The latter has a material representation of the environment that is furthermore mechanically organised in a grid with players’ positions clearly defined -- whereas the former relies on the less defined mechanical organisation of the dungeon master’s mind. Put more practically: Should a player want to know the range of their fireballs, they may refer to the grid in the latter and to the dungeon master’s mind in the former.
To conclude on virtual environments, this concept should be considered to encompass the mechanical level -- and, only in so far as they may share the same ontic space, either of the three remaining levels -- of the object. In Figure 1, one observes an abstracted virtual environment of the field on which one plays association football -- thus none of the surrounding environment is considered, nor is the material nature of the ground.
_-_Association_football.png)
Figure 1. The abstracted virtual environment of association football (Wikipedia, 2022). Click image to enlarge.
Having established these relationships, I move on to the level of the player -- arguably the trickier level of this ontology. If the other levels are considered “extractable” from the game itself, and if they “exist” independently of players recognising it, then where does the spatial level of the player fit in the ontology? The player, most often, implies that phenomena are experienced and characterised from said experience. As such, our ontology is always mediated, and the player arguably essential in the collective structure. This level stands, so to say, “outside” the three-layered structure. The player may observe visual cues provided in the game world -- such as gravel paths between never-ending plains, which are likely there (among other reasons) to suggest a direction towards progress. But these cues do not constitute topological difference, which is to say, that all of that hypothetical space maps onto the exact same mechanical level.
In the vertical structure of this ontology -- as depicted in Figure 2 -- signifiers are spread over the mechanical level like a “surface,” that the player interprets in order to engage with the object. The mechanical level as such is not immediately present to the player, seeing as it is always mediated by the representational surface. As already mentioned, merely “seeing” a door in the digital game space does not make it “openable” in a mechanical sense. In order to grasp its mechanical behaviour, then, the player would have to, say, press the buttons associated with “activation,” or look for changes in their cursor’s behaviour, as it hovers over the door’s visual surface. The material level, to continue, “carries” these upper levels -- it is in many ways their condition, along with the player level. The player, as explained, experiences and observes these levels, while the virtual environment encompasses the mechanical level and occasionally the representational or material level.
_-_Layered_structure.jpg)
Figure 2. The layered structure of the ontology of game spatiality. Click image to enlarge.
Comparing Ontologies
In the following, various well-cited example studies of game spatiality are compared in order to stress the applicability of this ontology. A few are specifically spatial, and some are more general ontologies of games.
Meta-Ontology
One might begin with Aarseth and Pawel Grabarczyk’s ontological meta-model (2018). This model allows for mapping of vastly different game ontologies, comprising the four layers with three sub-layers each, that need not be discussed in their entirety in this article. Interestingly, Aarseth and Grabarczyk themselves map the cybermedia model, so that the two models correspond neatly: material medium to physical layer, mechanical system to structural layer, sign surface to communicational layer and player to mental layer, respectively (2018, p. 10).
There is not a lot to say with regard to this article's four levels; as with the cybermedia model, these fit their respective counterparts nicely. However, some fundamental problems are observed within the sub-layers of the communicational (representational) layer, which readily emphasise the reason why sub-levels were not added to the present article's ontology: these are the presentational, the semantic and the interface sub-layer. According to the authors, semantics “[refers] to any communicated semantic information, from a simple command to a whole narrative,” which would suggest that this sub-layer accounts for linguistic signifiers only, even if those are already in themselves “presentational,” such as in text-based games -- or novels. One may argue that this sub-layer is merely one of many presentational sub-layers “which [refer] to the aesthetic aspects of the game” (Aarseth and Grabarczyk, 2018, p. 7), unless the authors consider presentation necessarily visually manifested in games.
In either case, the distinction appears arbitrary, especially when we consider a space that is conveyed only by means of a linguistic -- as opposed to an (audio)visual -- semiotic surface (in our times, most game spaces “appear” by virtue of visual signifiers). Text-based games often feature, and mediate, a mechanical level -- spatially manifested, that is -- by virtue of alphanumeric signifiers, which is why the distinction between presentation and semantics is not really applicable. Lunar Landing Game (Storer, 1969) may feature no visual-spatial signifiers, but due to the ergodics of the landing mechanic (as seen in Figure 3), it constitutes a level of mechanical spatiality regardless of its representational modes [6]. It matters not whether the representational level is considered semantic or presentational -- or even a mere interface -- so long as it is decidedly communicational, that is, representational.
_-_Lunar_Landing_Game.jpg)
Figure 3. Landing the plane in Lunar Landing Game (Storer, 1969). Click image to enlarge.
One observes in all cases that the cardinality of game spatiality is highly dependent on our preliminary concepts of spatiality. It decidedly matters how we conceptualise spatiality, when we make arguments as to how particular game spaces are “navigated” -- whatever we understand by this term -- or how particular game spaces are “played” differently -- questions that often rely on the cardinality of game space, and as follows, what level of spatiality said cardinality pertains to.
Mistaken Phenomenology
Michael Nitsche posits five conceptual planes for the analysis of the player’s experience of digital game spaces (2008, pp. 15-17). The fact that this framework is derived from phenomenology, and yet remains conceptually similar to the arguments pursued in this article -- that is, if one observes only two of the five planes -- is testament to the wide applicability of the present article's ontology, despite disciplinary differences. It is furthermore testament to the very nature of an ontology in research; as Aarseth writes, “ontology is the study of the nature of games: their mode of being or existence, and of variation within their domain … So the preliminary steps of any game ontology must be to first establish a meta-ontology” (2014, p. 484).
The first plane is rule-based space, which is the “mathematical” level comprising the “architecture” of game spatiality. The second is mediated space, which is the screen presenting the space of the image, seeing as Nitsche’s object of interest is video games, specifically. The third is fictional space, which pertains to the mind and imagination as to what lies “beyond” the currently observed space. The fourth is play space, comprising players, hardware and other physical elements outside the screen itself. The fifth is social space, in which players interact with others, often via voice communication software. Nitsche does not position this framework as a hierarchy, but he does consider rule-based space “the basis” for mediated space, which clearly corresponds to the layered structure of space already discussed -- and pursued in this article.
Considering disciplines, an ontology should arguably merge the second and third plane -- the mediated and fictional spaces -- as space is still “there,” even if we do not see “all space” simultaneously (such as with games in first-person perspectives). Moreover, Nitsche’s mediated space appears to suggest that space is visually represented by necessity, in which case text-based spatiality would have to be considered “fictional” according to the author’s framework. To a formal ontology, this does not make for a productive distinction, as it is not sufficiently abstract and thus excludes a range of games. The same applies to the fourth plane, the play space, as it only applies to games played on digital screens. As to the fifth plane, the social space, this is not entirely translatable to the player level either. The player’s experience, so to say, is spread throughout Nitsche’s planes, such as the fictional space. This is also the reason that the player level of my framework in Figure 2 is seen to “hover” in a suspended state over the three remaining levels of game space.
One may object that the above comparison is unproductive, seeing as the present study and that of Nitsche’s explicitly build on two different disciplines: ontology and phenomenology, respectively. But as I argue in a previous paper concerned with game space topology (Bakkerud, 2022), Nitsche’s classification is seemingly and largely concerned with games as ontological entities. To take an example, “Nitsche is critical of metaphors such as playgrounds, sandboxes, and gardens, seeing as” (Bakkerud, 2022, p. 5) “they do not refer to the structures […] but focus on their use” (Nitsche, 2008, pp. 171-172). In this example (which does not stand alone), Nitsche seems to conflate the disciplines of ontology and phenomenology, respectively. This should further explain why the five planes are discussed in the present article [7].
Phenomenological Critique
Another phenomenological study with relevance to this article, despite it arguably being more carefully grounded in phenomenology, is Günzel’s theory of the “pictorial medium,” with which he posits a new type of images derived from computerised games: “simulation pictures, the perception or reception of which includes interaction” (2008, p. 171). Similar to Nitsche, his scope is that of (screen-mediated) video games. Whereas prior pictorial mediums were merely “image-spaces,” video games present “space-images,” in which movement is “induced by the viewer.” This space, then, is “rarely contained within a single frame and is indeed often transgressed, that is, outside frame” (Günzel, 2008, pp. 171-172). Günzel goes on to observe an important gap between the “navigability” and the “presentation” of video games (2008, p. 180), suggesting further similarities with the concepts of the present study; a level that is structural and mostly “navigable,” and a level that signifies or “presents.” If anything, then, the concepts of spatiality pursued in this article are at least partially compatible with studies derived from phenomenology.
Interestingly, Günzel is critical of Mark J.P. Wolf’s conceptual pair of “off- and on-screen” (Wolf, 2001), seeing as it could never account for the (formal) differences in game spatiality, but only “the visual result of interaction” (Günzel, 2008, p. 173). This critique is compatible with mine with regard to Nitsche’s fictional plane, and it may even be extended to another aspect of Wolf’s classification (2001). Indeed, Wolf appears to confuse several levels of spatiality when he suggests that split-screen configurations are structures of distinct spaces (2001, p. 65). Most would agree, though, that two players participating in, say, Gears of War (Epic Games, 2006) -- on the same console -- are navigating the same, continuous instance of one game space. In other words, we need merely acknowledge the singular and shared ontological status of both the representational and mechanical levels of this game space, even if one or more players are participating simultaneously.
It is probable that Wolf’s mistake may be attributed to the conflation of “space” and “screens” (that is, perspectives in computerised games), but even if that game session were played on two different consoles -- thus, not with shared screens -- the representational, the mechanical and arguably the player level all remain the same [8], while only the material level would deviate. The implications of Wolf’s classification are that any changes in perspective, even a mere “zoom-out,” would constitute another distinct space in the ontological sense. This ought to make a sufficient example as to how Wolf’s classification of game spatiality needs refining, an approach to build upon -- which the present article hopes to deliver.
Avatar Bias
In the ontological literature on game spatiality, there is a reoccurring type of argument reliant on what might be called an “avatar bias.” That would be the case in Clara Fernández-Vara, José P. Zagal and Michael Mateas’ ontological distinction between the cardinality -- that is, the number of axes (or dimensions) one “can use to move entities around” -- of gameplay, gameworld and representation, respectively (2005). Gameplay refers specifically to the player’s capacities within the game, whereas gameworld refers to the mechanical “world” of the game space, that is, the level of space in which there is mechanical activity, while not necessarily within reach of the player herself.
As such, this article's mechanical level comprises the authors’ levels of both gameplay and gameworld, so that the difference between the two is rather about player affordances than qualities of the space itself. That distinction is applicable in many specific cases, but it is troublesome in a game ontological sense: how is “shooting” not an act of navigation, that is, how is this action formally different from “the way the player moves within the gameworld” (Fernández-Vara et al., 2005)? The projectile is fired through the game space, some of which the player avatar may not traverse. This is a case of avatar bias, where the argument seems to describe not all “games” -- as is otherwise the purpose of most ontologies -- but only an implicit subset based on player avatars, that can be said to navigate independently of the rest of the game entities.
This bias is often present in research emphasising the “navigation” of game space as part of the space’s formal structure. I argue, though, that the mechanical level -- in terms of space -- is not concerned with navigable space per se. Indeed, from this perspective, there may be virtual objects positioned in space that is not necessarily navigable, that is, “reachable,” by any player-controlled entity (mostly, the avatar) -- that is, if we take “virtual” to denote anything behaving dynamically in games. As such, I am arguing that the “gameworld” and the “gameplay” of Fernández-Vara et al.’s study (2005) should be consolidated. In a similar manner, Wolf defines space as something “in which way-finding is necessary, a space made of interconnected spatial cells through which the player’s avatar moves, a network often organized like a maze. All of the space may be present onscreen at once” (2009). To necessitate “way-finding,” though, is to presuppose constant movement, which is not an inherent aspect of the game. Similarly, the game may feature a space so easily accessible that this space can hardly be characterised by way-finding. Indeed, that something is “often” the case is simply not sufficiently rigorous if we are to develop game ontologies. And finally, the idea that space must necessarily be presented on screens is problematic for reasons already covered, most notably because the distinction between on-screen space and off-screen space is not formalisable if one’s object of interest is games more broadly.
Calleja (2011) adopts Wolf’s understanding and considers navigable space to be “necessarily simulated space, as opposed to a space that exists only as a representation” (2011, p. 78). However, this view only accounts for game space that is traversable by players. Calleja does include some avatar-less games in his study, such as “real-time strategy games,” referred to as “miniatures” in which players hold “disembodied powers” (2011, p. 91) -- and in this case, he is right in arguing that these games are navigated, though not necessarily “explored,” as “the player’s point of view… [wanders] freely above the landscape” (Calleja, 2011, p. 90).
Nonetheless, Calleja’s concept would not account for the parts of space where the player’s avatar -- or “miniature entities” -- could never set foot (figuratively speaking). You may have agents or objects shooting at your avatar from an unreachable place, which nonetheless is a place on the mechanical level. This matters to your success from said encounter, as you would have to fire back in order to progress. As such, Calleja’s concept is reasonably applicable, but I argue that the very notion of navigation implies a certain subset of elements primarily found in audiovisual games, such as player-controlled avatars, that are not inherent features of what is more widely referred to as games. It would be misleading to describe all of the gameplay of the board game Agricola as “acts of navigation,” even if it is a highly strategic task to ensure the correct spatial positioning of one’s crops and houses.
As per Karhulahti (2012), the virtual objects of games may function independently of player capabilities. As such, it might be, in the context of the present study, more accurate to speak of “positioning” -- or “re-positioning” -- so as to not imply the active traversal of the player. Virtual guns will usually fire their projectiles along traceable trajectories, but would we describe these trajectories as navigation or positioning? In any case, the act implies a player or agent actively “steering.” The concept of spatiality pursued in this article, however, describes games in a sense “beyond” observations of player movement, that is, beyond space as it appears and affords actions to the actively engaging player.
Lack of Hierarchies
A final example, then: in Aarseth’s typology of quest game landscapes, he differentiates one-room games from open worlds (among other topologies; 2012). However, as Debus argues (2019, p. 222), there is no formal measure in order to differentiate the scale of either, the consequence of which is that the one-room’s proportions are only different on the representational level. As suggested in their metaphorical names, both types of quest game landscapes are separated based on the visual environment connoted by the author: one appears cramped and limited, while the other appears vast and lively. Of course, there is often correlation between these spaces’ appearances (representation) and proportions (volume of space players may explore) but that is never a given fact. Indeed, if players are controlling, say, a cockroach roaming around a human living room, that cockroach’s size relative to a human might compare to that of a human relative to a huge dragon’s lair. The same flaw is evident when Nitsche characterises one game space structure as a virtual city (2008, p. 171), as that term does not denote a concrete structure at all. Indeed, the aspects of said metaphor are mostly representational (for a more substantial critique of these metaphors, see Bakkerud, 2022).
One might object that hierarchies add little to the cybermedia model; why would the original model not suffice? This may be clarified with a brief example in the form of another question: which layer would a player entering any room (say, a laboratory) in a three-dimensional game space first see? The mechanical level -- the actual structure delineating the player’s spatial capabilities -- or the representational level? This article argues in favour of the latter, as the player would not yet know which (represented) doors are actually openable or which walls the player may actually climb (just to name a few examples). By applying the hierarchal framework of this article, then, researchers may more carefully understand how these levels of space function and intertwine.
Conclusion
Günzel (2008) has provided an excellent sketch for the future of studies into this topic -- one derived from the perspective of phenomenology. I found that a similar sketch was necessary from the perspective of ontology, as this topic’s success is indisputably contingent on rigorous terminology. A brief conclusion to this article, then, is the four-layered ontology adopted from the cybermedia model, comprising the levels of the representational, the mechanical, the material and the player.
The relationship between them is like that of a vertical structure: the representational level is considered to be the surface of the mechanical level, which structures the game object “from below,” even if this structure is entirely upheld by the players themselves. This level constitutes the virtual environment of the given game space, a concept that is introduced in order to differentiate the mechanical level, whenever this level shares the same ontic space as either of the other three.
Endnotes
[1] As Roland Barthes writes on the study of narratives: “faced as it is with millions of narratives” -- or games, naturally -- “it is condemned to a deductive procedure” by necessity (1990, p. 81).
[2] Indeed, the very dichotomy of space (or architecture) of or in games is indicative of the object’s roots in hypertextual phenomena.
[3] A metonymy is a trope based on contiguity between the signifier and the signified.
[4] In fact, this applies even in “biological” phenomena, according to Aarseth and Calleja (2015). For the sake of simplicity, however, this article does not differentiate between analog and biological phenomena.
[5] Do note that Aarseth’s later concept of “games in virtual environments” (2004) is not compatible with the one applied in this article. This article does not exclude digitised versions of “traditional games” from the concept of virtual environments -- comprising anything from “card games, board games, dice games” to “mechanical arcade games such as Pinball” (Aarseth, 2004, p. 364). The term as applied in this article is entirely motivated by its disregard of material mediums and representational modes.
[6] To clarify: in text-based games, communication is entirely verbal, as opposed to games in which written characters (ASCII) are employed as graphic elements -- for example, in Rogue (A.I. Design, 1980).
[7] In said study (Bakkerud, 2022), I demonstrate a widespread tendency in scholarship on game spatiality to conflate metaphors with structured terminology. For example, I argue that the meaning of metaphors such as Nitsche’s “[are] contingent on cultural and representational aspects” (2022, p. 6).
[8] It is -- in itself -- misguided to measure spatiality in terms of “screens.”
References
Aarseth, E. (1997). Cybertext: Perspectives on Ergodic Literature. Johns Hopkins University Press.
Aarseth, E. (2001a). Allegories of Space: The Question of Spatiality in Computer Games. In M. Eskelinen & R. Koskimaa (Eds.), Cybertext Yearbook 2000 (pp. 152-171). University of Jyväskylä.
Aarseth, E. (2001b). Virtual worlds, real knowledge: towards a hermeneutics of virtuality. European Review, 9(2).
Aarseth, E. (2004). Quest Games as Post-Narrative Discourse. In M-L. Ryan (Ed.), Narrative across Media: The Languages of Storytelling (pp. 357-372). University of Nebraska Press.
Aarseth, E. (2005). From Hunt the Wumpus to EverQuest: Introduction to Quest Theory. In F. Kishino, Y. Kitamura, H. Kato, & N. Nagata (Eds.), Entertainment Computing. ICEC 2005. Lecture Notes in Computer Science, vol. 3711 (pp. 496-506). Springer.
Aarseth, E. (2012). A Narrative Theory of Games [Conference paper]. In Proceedings of the International Conference on the Foundations of Digital Games. Raleigh: ACM.
Aarseth, E. (2014). Ontology. In B. Perron & M.J.P. Wolf (Eds.), The Routledge Companion to Video Game Studies (pp. 484-492). Routledge.
Aarseth, E. (2019). Ludoforming: Changing Actual, Historical or Fictional Topographies into Ludic Topologies. In E. Aarseth & S. Günzel (Eds.), Ludotopia: Spaces, Places and Territories in Computer Games (pp. 127-140). Transcript Verlag.
Aarseth, E., and Calleja, G. (2015). The Word Game: The ontology of an indefinable object [Conference paper]. Proceedings of the 10th International Conference on the Foundations of Digital Games (FDG 2015). Pacific Grove: FDG.
Aarseth, E., and Grabarczyk, P. (2018). An Ontological Meta-Model for Game Research [Conference paper]. Proceedings of DiGRA 2018.
Aarseth, E., and Günzel, S. (2019). Introduction: Space -- The Theoretical Frontier. In E. Aarseth & S. Günzel (Eds.), Ludotopia: Spaces, Places and Territories in Computer Games (pp. 7-9). Transcript Verlag.
Aarseth, E., and Möring, S. (2020). The game itself? Towards a Hermeneutics of Computer Games [Conference paper]. International Conference on the Foundations of Digital Games (FDG ’20).
A.I. Design. (1980). Rogue. Digital game published by Epyx.
Arjoranta, J. (2011). Do We Need Real-Time Hermeneutics? Structures of Meaning in Games [Conference paper]. Proceedings of DiGRA 2011 Conference: Think Design Play. Hilversum: DiGRA.
Bakkerud, F. (2022). Game space topology revisited: A review of labyrinthine terminology [Conference paper]. Proceedings of DiGRA 2022.
Barthes, R. (1990). Introduction to the Structural Analysis of Narratives. In Image--Music--Text (pp. 79-124). Fontana Press.
Bethesda Game Studios. (2002). The Elder Scrolls III: Morrowind. Digital game published by Bethesda Softworks.
Calleja, G. (2011). Spatial Involvement. In In-Game: From Immersion to Incorporation (pp. 73-92). MIT Press.
CRL Group. (1986). Dracula. Digital game published by CRL Group.
Debus, M.S. (2019). Unifying Game Ontology: A Faceted Classification of Game Elements [PhD dissertation]. IT University of Copenhagen.
Epic Games. (2006). Gears of War. Digital game published by Microsoft Game Studios.
Eskelinen, M. (2001). The Gaming Situation. Game Studies, 1(1). https://www.gamestudies.org/0101/eskelinen/
Fernández-Vara, C., Zagal, J.P., and Mateas, M. (2005). Evolution of Spatial Configurations In Videogames. Proceedings of DiGRA 2005 Conference: Changing Views -- Worlds in Play.
Genette, G. (1980). Narrative Discourse. Cornell University.
Genette, G. (1982). Figures of Literary Discourse. Columbia University Press.
Golding, D. (2013). To Configure or to Navigate? On Textual Frames. In Terms of Play: Essays on Words That Matter in Videogame Theory (pp. 28-46). McFarland & Company, Inc.
Günzel, S. (2008). The Space-Image: Interactivity and Spatiality of Computer Games. In S. Günzel, M. Liebe, & D. Mersch (Eds.), Conference Proceedings of the Philosophy of Computer Games 2008 (pp. 170-189). Potsdam University Press.
Karhulahti, V-M. (2012). Suspending Virtual Disbelief: A Perspective on Narrative Coherence. In D. Oyarzun, F. Peinado, R.M. Young, A. Elizalde, & G. Méndez (Eds.), Interactive Storytelling. ICIDS 2012. Lecture Notes in Computer Science, 7648 (pp. 1-17).
Manovich, L. (2001). The Language of New Media. The MIT Press.
Murray, J.H. (1997). Hamlet on the Holodeck: The Future of Narrative in Cyberspace. The MIT Press.
New World Computing. (1999). Heroes of Might and Magic III: Armageddon’s Blade. Digital game published by the 3DO Company.
Nitsche, M. (2008). Video game spaces: image, play, and structure in 3D game worlds. The MIT Press.
Playdek, Inc. (2020). Agricola Revised Edition. Digital game published by Asmodee Digital.
Rosenberg, U. (2016). Agricola Revised Edition. Board game published by Lookout Games.
Storer, J. (1969). Lunar Landing Game. Digital game with an unknown publisher.
Valve. (1998). Half-Life. Digital game published by Sierra Studios.
Valve. (2004). Half-Life 2. Digital game published by Valve.
van Vught, J. (2016). Neoformalist Game Analysis: A methodological exploration of single-player game violence. University of Waikato.
Warsch, W. (2018). The Mind. Board game published by Nürnberger-Spielkarten-Verlag.
Wikipedia: The Free Encyclopedia. (2022). Association football. Accessed on February 16, 2022. https://en.wikipedia.org/wiki/Association_football
Wolf, M.J.P. (2001). The Medium of the Video Game. University of Texas Press.
Wolf, M.J.P. (2009). Theorizing Navigable Spaces in Video Game Worlds in Structure, Sign, and Play. In DIGAREC. University of Potsdam.
Yang, R. (2018, November 19). Everything you liked about Half-Life was a secret train. PC Gamer. https://www.pcgamer.com/half-life-is-20-why-everything-you-liked-about-valves-classic-was-a-secret-train/