Emerge Whence?
Defining, Designing and Categorizing Emergent Gameplay
This essay serves as both a review of the concept of emergence in games literature and an elaboration on previous writing in an attempt to categorize various forms of emergent gameplay. The practice of speedrunning — completing a videogame as fast as one possibly can — as a typified form of emergent gameplay serves to anchor this discussion. My aim is to provide a document that addresses the inherent contradictions of “designed emergence” so that it may be referenced in future writing and game-making practice.
EMERGENCE IN GAMES LITERATURE
Despite its frequent invocation both in game studies and games culture, the concept of emergence is seldom written about definitionally. In order to proceed there must be a clear understanding of three nearly-synonymous terms: emergence, emergent play and emergent gameplay. Instead of attempting to tackle one or the other, let us go about as chronologically as possible. While gesturing toward Brian Sutton-Smith’s (1997) The Ambiguity of Play, Cameron Kunzelman (2014) writes that “the process of collecting and collating the various ways that play has been written about in recent history is deeply revealing of how those thinking of play see it as an emergent property that comes out of creatures living life.” (14) Thus the term “emergent play” is at risk of redundancy even before videogames enter the picture.
Slightly outside of game studies we can look to the work of Ruth Aylett (1999) on “Narrative in Virtual Environments” for a canned definition of emergence: “the creation of complexity bottom up via interaction between essentially simple components.” (84) This idea of complexity from simplicity carries through to game designer Harvey Smith’s (2001) use and probable coinage of “emergent gameplay”, where he defines emergence “as an event that occurs, but that could not have simply been inferred from a system’s rules”. Smith goes on to describe an emergent strategy found by QA testers in Deus Ex (1999) where a unit which explodes when defeated can be lured towards locked doors and safes, thereby opening them “without spending any lock picking resources”. Smith comments “We did not plan this or even foresee it-it just worked.” Thus emergent play is implicitly tied to the uncertain, the unforeseen, the unexpected and the unpredictable. For our purposes moving forward we will say that “emergent gameplay” is a subset of emergent play that is done with and/or within videogames. Importantly this is not to take on a parallel definition of “gameplay” alone, whose exact meaning remains contested (see Janik 2020).
Once emergent gameplay entered the lexicon in the early aughts its definition began to distort. In the volume Game Design Perspectives, Joshua Mosquiera (2002) eschews Smith’s uncertainty for a language of consistency, writing that “[t]ogether, context, consistency, and story create a diagetic [sic] environment, from which emergent gameplay evolves — gameplay that is contextualized, consistent and mutually supportive with the story.” (71) In an issue of Xbox Nation, Evan Shamoon (2003) writes that “the emergent play [of Soul Calibur II] comes through enormous move palettes, which are blended to form an infinitely changing dance of death.” (5) In the first case I think emergence is simply being mischaracterized, however the second example of combos in a fighting game still adheres to the principle of complexity via simplicity. Yet still I bristle at this, the reason being that my personal view of emergent gameplay is caught up in the concept of intentionality — namely my belief that “intentional emergence” is an oxymoron.
To clear this up we can turn to Katie Salen Tekinbas and Eric Zimmerman (2003) who dedicate an entire chapter of their Rules of Play: Game Design Fundamentals to “Games as Emergent Systems”. They state that emergence “is the phenomenon of unplanned patterns appearing from within a system.” (152) With our fighting game example we can see that there is potential for unplanned combos, but not unintended combos, thus being an example of intentional emergence to my chagrin. By contrast something like the wavedash exploit in Super Smash Bros. Melee (2001) is definitely unintentional, yet like Smith’s exploding enemies it is not considered a glitch. It is my assertion then that intentionality is too narrow a concept to properly encompass or break down emergent gameplay. In the past I have preferred to use the idea of “presupposition”, which casts as wide a net as possible to include that which is unintended but does not fall into the realm of glitch, bug or “otherwise break the communicated logic of [the] game design” (Howard 2019).
Part of the reason for the heterogeneity of emergent gameplay as it is evoked stems from the fact that “[e]mergence can come about through complex programmed mechanisms that simulate adaptive agents and systems, but it can also happen on an experiential level, where extremely simple rules give rise to complex social or psychological relationships among players.” (Salen Tekinbas and Zimmerman 159) When we talk about emergent gameplay we are really talking about two or more different forms of emergence, what I will henceforth refer to as “social emergence” and “structural emergence”. I make this distinction not to pull apart the social from the structural or vice versa, but rather to highlight their interdependencies as accounted for in speedrunning practice. Bugs, glitches and exploits used in speedrunning are a form of structural emergence, but are discovered and disseminated through socially emergent means and channels.
Salen Tekinbas and Zimmerman also state “what makes a system emergent is that there is a special disconnect between the rules of the system and the ways those rules play out” (160). However in my opinion there is nothing special about emergence when you consider that it is always already baked into the possibility space of a given videogame, and the vast majority of videogames have potential for it, at least in its “weak” form. Furthermore the locus of this disconnect does not rest in the rules and their simulation but in the mind of the game designer. As Salen Tekinbas and Zimmerman later put it “[a]s a game designer, you are tackling a second-order design problem[…]you can never directly design play. You can only design the rules that give rise to it. Game designers create experience, but only indirectly.” (168) So while emergence may lay within the realm of the presupposed, it allegedly cannot be designed.
While others wrote about emergence in and around game studies in the early aughts (see Henry Jenkins [2004], Celia Pearce [2004] and Sandy Louchart and Ruth Aylett [2004] on emergent narrative), the next scholar of note is Jesper Juul (2005) and his Half-Real: Video Games between Real Rules and Fictional Worlds, where he defines emergence as “a number of simple rules combining to form interesting variations” (15). For Juul games of emergence are “the primordial game structure” (Ibid.) and contrast with games of progression, those in which “the player has to perform a predefined set of actions in order to complete the game.” (16) However because Juul only considers adventure games as “complete-once” (17) experiences he does not anticipate how the act of speedrunning slices through this dichotomy: games of progression are made into games of emergence through the discovery of tricks/strategies, bugs/glitches and exploits, while games of emergence such as Animal Crossing (2001) and Minecraft (2011) are made into games of progression by superimposing semi-arbitrary goals to race towards. Juul himself ultimately admits that “[m]any games can be found on a scale between emergence and progression” and that these “hybrids” contain “components” of the opposite structure (101).
Similar to my social/structural dichotomy, Juul asks “whether emergence is a feature of the game systems themselves or a feature of human cognition” (108) and concludes that it is “an interaction between the game system and human cognition” (117, emphasis added). On the relationship between emergence and game design Juul says “The practical argument is that it allows content to be created faster”, and while I think this has merit I also think Juul underestimates the labour of play-testing games with large possibility spaces (109, emphasis original). Still Juul’s games of emergence and progression would prove highly influential within games studies (see in particular Dormans [2011]), and the concept of emergent gameplay would be bolstered by Smith’s collaborator Warren Spector who gave talks espousing its values in the late aughts and early 2010’s (Spector [2007], Alexander [2013]).
I now turn to the aforementioned Celia Pearce (2009) and her Communities of Play: Emergent Cultures in Multiplayer Games and Virtual Worlds. As one might surmise, Pearce’s work largely focuses on social emergence rather than structural, however she does speak more directly to the interrelatedness of game design and emergent gameplay. Pearce says that “The more agency players are given, the larger the quantity and variety of emergent behaviours that are likely to occur.” For example, players of Uru (2003) would enact games of hide-and-seek but could not replicate this phenomenon within There.com (2003) or Second Life (2003) because there was no way to obfuscate the username above their avatars.
Pearce observes that “‘Emergence happens,’ regardless of whether the virtual world has affordances for it or not” and further that “emergence is the inevitable outcome of a large number of players within a network.” Though she stops short of claiming that emergent gameplay can be expressly designed, she implores that “[t]he more conscious we are of the patterns that emerge from specific design features and technical constraints, the more able we will be to work with emergence as a ‘material’ of game design” and “while we cannot entirely control [emergence], it may be possible to integrate its patterns into our design process.” Pearce also cautions that “we must be wary of forms of emergence that are undesirable, as these are almost impossible to reverse once they are under way.”
Where other scholars in the mid-late aughts such as Penelope Sweetster (2006) and Michael D. Kickmeier-Rust and Dietrich Albert (2009) seemed to take the ability to design emergence as a given in their research, the later work of Joan Soler-Adillon (2019) sought to critique and clarify the adoption of emergence in game studies, arguing that there are only two forms of emergence: self-organization and novelty. For Soler-Adillon the “key importance [is] that self-organization emergence is only such when the micro level interactions are not driven by the intent, or will, to affect the macro level.” Here intent does not refer to artistic intent by a designer but rather the “intralevel interaction” of player and game entities. Emergence as novelty on the other hand is negated through the course of its own discovery: “once incorporated into the model, [the structure or pattern] ceases to be emergent, as it is no longer new.” We can see this play out in speedrunning practice, where emergent strategies become canonized within a given speedrun.
Soler-Adillon also argues against the conflation of emergence and openness, stating that “anything emergent needs a large degree of openness, but not everything that results from openness is emergent.” He concludes echoing Salen Tekinbas and Zimmerman’s second-order problem, albeit more explicitly: “By definition, emergence cannot be designed. Only the conditions for its appearance can be. Thus, when designing for emergent systems, one sets up the system that allows for a big enough space of possibility and hopes that emergence will occur.”
CATEGORIES OF EMERGENT GAMEPLAY
In this section I would like to expand (no pun intended) on my model of possibility space that I introduced in “Edge Case Scenario: On Possibility Space, Speedrunning and Parallel Universes” where I argue that “the possibility space of a videogame can be divided into presupposed and expanded play areas” (Howard 2019). In my model possibility space is figured as a circle that expands from the edge of another circle representing the game-maker’s idea of a videogame. However if one were to represent this graphically as a Venn diagram, they would find it to be incomplete in its labelling, so let us first correct that.
The circle on the left side of the diagram is the “idea space” while the circle on the right side is the possibility space. The intersection of these two circles is the “presupposition space” and it represents the videogame as the collective game-maker knows it, what Boluk and LeMieux (2017) would call the “standard metagame” (36). The segment to the right of the presupposition space is the “expanded space”, which includes “any incidental phenomena such as bugs [or] glitches” (Howard 2019). Note that the Venn diagram is not meant to be proportional and the fact that the expanded space may appear smaller or larger than the presupposition space is an illustration rather than a measurement. To the left of the presupposition space is the “extracted space” which comprises “any cut content (concept art, unused assets, beta levels, placeholder text, etc.)” (Ibid.).
With the stage set let us now introduce the concept of emergence. As I have previously alluded to, emergent gameplay is able to cross over the boundary between the presupposition space and the expanded space. We can represent this with another circle at the edge of the presupposition space. This is the first way that I wish to break down emergent gameplay, by introducing the concept of “divergent gameplay”, which is emergent gameplay that takes place within the expanded space and therefore “diverges” from presupposition. Emergent gameplay that remains within the presupposition space will remain known as emergent gameplay for the sake of posterity. To show how the emergent/divergent dichotomy is useful we can turn to speedrunning and see how Scully-Blaker’s (2014) notion of the “finesse run” and the “deconstructive run” clearly maps onto it, i.e. finesse runs are emergent gameplay while deconstructive runs are divergent gameplay. While divergent play is close in spirit to Scully-Blaker’s (2016) “(re)curatorial play”, ultimately that concept is too bound up in intentionality to fit my model that is contingent on presupposition.
This model is adequate so far but I think it can be pushed further into territory unexplored in previous scholarship. Where others have turned to rules to flesh out their taxonomies (see Parker 2008; Scully-Blaker 2014, 2016) I turn instead to goals, specifically Juul’s (2007) dichotomy of enforced goals and non-enforced goals. Emergent gameplay that involves enforced goals will henceforth be referred to as “codified”, while emergent gameplay that involves non-enforced goals will be referred to as “improvised”. One may be tempted to apply Roger Caillois’ “Ludus” and “Paidia” to my use of codified and improvised, however this would be erroneous as improvised gameplay does not imply a lack of rules but a lack of enforcement via code. Rules are still present whether they be as strict as “the floor is lava” or as lenient as “yes, and…”.
Let us now lift these two dichotomies out of the Venn diagram and into a matrix. This yields four categories of emergent gameplay: codified emergent gameplay, codified divergent gameplay, improvised emergent gameplay and improvised divergent gameplay. Codified emergent gameplay occurs within the presupposition space in service of enforced goals. It includes activities such as pacifist runs, no hit/damageless runs and finesse speedruns. Codified divergent gameplay occurs within the expanded space also in service of enforced goals. It includes deconstructive speedruns and other challenges such as the “A Button Challenge” tool-assisted speedrun (TAS) in Super Mario 64 (1996) (see Howard 2019) and the “Minimum Captures” run within Super Mario Odyssey (2017) (see “How Speedrunners…”). Note how codified gameplay tends towards activities which James Newman (2008) dubs “superplay”.
Continuing on, improvised emergent gameplay occurs within the presupposition space in service of non-enforced goals. It includes “expansive gameplay” metagames such as “jeep tag” in Halo: Combat Evolved (2001) (see Parker 2008) or “the floor is lava” challenges in Mario Odyssey as well as improv comedy performances within videogames such as Polygon’s Monster Factory video series. It also includes some of Pearce’s examples of emergence, such as weddings and protests staged within MMOs. The final category of emergent gameplay is improvised divergent gameplay which occurs within the expanded space in service of non-enforced goals. It is perhaps the most diverse category as it can include “Action Replay/Game Genie, ROM hacks, mods, Machinima, glitch art practice, file-ripping or other user-generated tools and content” as well as glitch-hunting like that seen on the “Supper Mario Broth” blog and Twitter account (Howard 2019). Pearce’s account of “Avie Bowling” in Uru which uses a collision flaw where avatars get stuck in the ground then run around attempting to knock over traffic cones is also improvised divergent gameplay. An artistic example of improvised divergent gameplay is pannekoek2012’s “April Fool’s Day Video 2021”, a 44-minute tool-assisted playthrough of Super Mario 64’s Tall Tall Mountain which uses obscure movement, camera and “cloning” techniques to generate a cluster of enemies and particle effects that looks like Mario’s face when viewed from above.
While my matrix is not the most comprehensive analysis of emergence in videogames compared to Juul, Sweetster or Soler-Adillon, it does attempt to account for a wider array of gameplay practices than I have encountered before. However I can identify two major shortcomings of my four categories as they exist. Firstly, it is difficult for any given play session to be entirely divergent i.e. within the expanded space; for no matter how many parallel universes Mario traverses in pannenkoek2012’s “Watch For Rolling Rocks — 0.5x A-Presses” TAS, he must always return to his home universe to collect the star. Secondly, gameplay situations may include multiple goals simultaneously that cross the threshold between enforced and non-enforced. Super Mario Maker 2 (2019) glitch levels are an example of this, where the enforced goal is to complete the level and the non-enforced goal is to spectate a glitch or series of glitches (see Howard 2021). Beyond that however I believe keeping in mind these four categories of emergent gameplay gives me a robust vocabulary with which to tackle the question of emergence in my own work.
CITATIONS
Alexander, Leigh. “Spector: Go emergent — game design is not all about you.” Game Developer, 16 November 2013, www.gamedeveloper.com/design/spector-go-emergent---game-design-is-not-all-about-you.
“April Fool’s Day Video 2021.” YouTube, uploaded by UncommentatedPannen, https://www.youtube.com/watch?v=5B0bE_Mpptc.
Aylett, Ruth. “Narrative in Virtual Environments-Towards Emergent Narrative.” Proceedings of the AAAI fall symposium on narrative intelligence. 1999.
Boluk, Stephanie, and Patrick LeMieux. Metagaming: Playing, competing, spectating, cheating, trading, making, and breaking videogames. U of Minnesota Press, 2017.
Caillois, Roger. Man, Play, and Games. University of Illinois press, 2001.
Dormans, Joris. “Integrating Emergence and Progression.” DiGRA Conference. 2011.
Howard, David R. “Edge Case Scenario: On Possibility Space, Speedrunning and Parallel Universes.” 3 June 2019, rhombical.medium.com/edge-case-scenario-on-possibility-space-speedrunning-and-parallel-universes-795785d58210.
Howard, David R. “It’s a Beautiful City: Troll and Glitch Levels in Mario Maker as Visionary Environments” 16 March 2021, rhombical.medium.com/its-a-beautiful-city-troll-and-glitch-levels-in-mario-maker-as-visionary-environments-b3cb1eeecbe8.
“How Speedrunners beat Mario Odyssey without Cappy (almost).” YouTube, uploaded by Lowest Percent, 6 November 2020, www.youtube.com/watch?v=L5aQzCJW5-s.
Janik, Justyna. “Negotiating Textures of Digital Play: Gameplay and the Production of Space.” Game Studies 20.4 (2020).
Jenkins, Henry. “Game Design as Narrative Architecture.” Computer 44.3 (2004): 118–130.
Juul, Jesper. Half-Real: Video Games between Real Rules and Fictional Worlds. Kindle ed., MIT Press, 2005.
Juul, Jesper. “Without a Goal: On Open and Expressive Games.” Videogame/Player/Text, edited by Tanya Krzywinska and Barry Atkins. Manchester University Press, 2007. www.jesperjuul.net/text/withoutagoal/.
Kickmeier-Rust, Michael D., and Dietrich Albert. “Emergent Design: Serendipity in Digital Educational Games.” International Conference on Virtual and Mixed Reality. Springer, Berlin, Heidelberg, 2009.
Kunzelman, Cameron. “The Nonhuman Lives of Videogames.” Thesis, Georgia State University, 2014. www.scholarworks.gsu.edu/communication_theses/110
Louchart, Sandy, and Ruth Aylett. “The Emergent Narrative Theoretical Investigation.” the 2004 Conference on Narrative and Interactive Learning Environments. 2004.
Mosquiera, Joshua. “World Building: From Paper to Polygons.” Game Design Perspectives, edited by François Dominic Laramée. Charles River Media, Inc., 2002, pp. 67–77.
Newman, James. Playing with videogames. Routledge, 2008.
Parker, Felan. “The Significance of Jeep Tag: On Player-Imposed Rules in Video Games.” Loading… 2.3 (2008).
Pearce, Celia. “Towards a Game Theory of Game.” First person: New media as story, performance, and game 1 (2004): 143–153.
Pearce, Celia. Communities of Play: Emergent Cultures in Multiplayer Games and Virtual Worlds. Kindle ed., MIT press, 2009.
Salen Tekinbas, Katie, and Eric Zimmerman. Rules of Play: Game Design Fundamentals. MIT press, 2003.
Scully-Blaker, Rainforest. “A Practiced Practice: Speedrunning Through Space with de Certeau and Virilio.” Game Studies 14.1 (2014).
Scully-Blaker, Rainforest. “Re-curating the Accident: Speedrunning as Community and Practice.” Thesis, Concordia University, 2016.
Shamoon, Evan. “Here come the Xbots….” Xbox Nation, Nov. 2003, pp. 5.
Smith, Harvey. “The Future of Game Design: Moving Beyond Deus Ex and Other Dated Paradigms.” 2001, www.witchboy.net/articles/the-future-of-game-design-moving-beyond-deus-ex-and-other-dated-paradigms/.
Soler-Adillon, Joan. “The Open, the Closed and the Emergent: Theorizing Emergence for Videogame Studies.” Game Studies 19.2 (2019). http://gamestudies.org/1902/articles/soleradillon.
Spector, Warren. “Emergent Gameplay: Potential, Pitfalls and Progress.” Proceedings of the 2007 ACM SIGGRAPH symposium on Video games. 2007.
Sutton-Smith, Brian. The Ambiguity of Play. Harvard University Press, 1997.
Sweetser, Penelope. “An Emergent Approach to Game Design: Development and Play.” PhD diss.: University of Queensland, 2006.
