[In this abridgement of the first chapter of new book Imaginary Games, available via this Amazon link, game designer, philosopher, and writer Chris Bateman, best known for the game Discworld Noir and the book Game Writing: Narrative Skills for Videogames, examines the game-as-art debate from an interesting new angle.]
What is a Game?
In April 2010, esteemed film critic Roger Ebert walked unknowingly into the teeth of a rancorous beast when he posted on his blog that not only were video games not art, they could never be art. Given that the internet is packed to its virtual rafters with belligerent gamers who will argue to the death over the insignificant minutiae of their preferred forms of play, this inevitably unleashed a storm of criticism.
In many respects, it was a boon for the games industry that Ebert had chosen to wade in on this topic, since there were enumerable critics in various media who would simply have treated the entire subject with disdain. Whatever one makes of Ebert's claims, he at least had the respect for the medium of digital games to consider this topic seriously.
But what is art, and what is a game? There is a temptation, as Ebert observed, to think that this is simply a matter of semantics and thus not a big deal -- an attitude embodying a rather wide prejudice against philosophy which Ebert, thankfully, does not share.
He quotes from the Greek philosophers in saying that art "improves or alters nature through a passage through what we might call the artist's soul, or vision," and constructs an argument based on the premise that, as goal-oriented activities, games are precluded from being considered art or, to put it another way, the possibility of winning in a game is anathema to artistry.
Yet not all things we call a 'game' include the notion of winning. A child's game of make-believe need not, and neither do most tabletop role-playing games, which are, at heart, a more sophisticated form of exactly the same thing as children's make-believe.
A rhyming game like 'ring a ring a roses' doesn't involve winning either, and certain computer games are equally divorced from an overarching goal -- Will Wright has called his game SimCity (Maxis, 1989) a "software toy", and there are many other games with ambiguity in this regard, such as the classic 8-bit title Deus Ex Machina (Croucher, 1985).
Before we can do justice to Ebert's argument, we must first establish with some confidence what we mean by the term 'game', and this is no easy matter. In fact, this has been recurring theme in the literature of game studies, which from the outset has involved nearly endless discussions concerning the boundary conditions of games. For the most part, we are no closer to an answer than we were when we began, but it is interesting to note that a great deal of the debate presumes that there is a definitive answer to be reached. The fact that people seem confident the term can be unraveled gestures at an underlying unity to the concept of a game, and thus suggests that the problem is not wholly insoluble.
In his 2009 keynote for the Digital Games Research Association (DiGRA), Ian Bogost admits that so much of the discussion within game studies has been dominated by this very question, "what is a game?" In an insightful summary of the 'moves' offered thus far, Bogost covers the history of this crucial investigation.
First, there was the ludology vs. narratology debate, which hinged upon whether games were best understood as a system of rules, or as fictions. But as Bogost notes, there is a false dichotomy in this approach. The question being asked is akin to "is a game a system of rules, like a story is a system of narration?" -- and worded this way, the sense of disjunction is removed and the answer is simply returned in the affirmative.
Jesper Juul provided the next major move in this debate, by suggesting in his seminal book Half-Real (2005) that:
...video games are two rather different things at the same time: video games are real in that they are made of real rules that players actually interact with; that winning or losing a game is a real event. However, when winning a game by slaying a dragon, the dragon is not a real dragon, but a fictional one. To play a video game is therefore to interact with real rules while imagining a fictional world and a video game is a set of rules as well a fictional world.
Games are thus suggested to be both systems of rules and fictions. Bogost criticizes not this claim, but an underlying assumption that even if this is the case, the rules somehow have a kind of precedence -- some part of a game is more real than the other part (usually the rules). It is here that we reach the philosophical domain of ontology, where questions of being and reality are discussed. If someone makes a claim concerning what is real, they are asking an ontological question. This point will become important shortly.
Juul once again provides Bogost's "third move" -- the question of what is the appropriate area of study in respect of games: is it the games in themselves, or the players?
It leads to an idea that games 'exist' when players occupy them, which Bogost compares to Kant's breakthrough realization that whatever things may exist, we as humans only have access to them via our thought and senses. Once again, seen in this way we are no longer addressing the question "what is a game?" so much as we are dealing with the ontological implications of games.
Bogost has his sights in this keynote on introducing his own ontological move, based on the platform studies he has conducted with Nick Montfort. Here, a number of different component levels of digital games are systematically uncovered -- looking as games as just rules misses out on many key aspects.
In the case of the Atari VCS that Montfort and Bogost study in Racing the Beam (2009), the hardware and software constraints had distinct effects on the games that were (or could be) made. There are hidden elements in the nature of digital games to be teased out.
Drawing on the work of Levi Bryant and Graham Harman in ontology, and in particular the notion of a "flat ontology", Bogost boldly suggests that we entirely abandon attempts to claim a hierarchy of some kind in understanding games.
A game, he offers, is better understood from the perspective of such a flat ontology, one in which no one kind of entity has precedence over another (as in the case of the rules taking precedence over the fiction in Juul's half-real paradigm). Bogost goes further, suggesting we can look beyond the ontological elements that involve humans and throw the remit far wider such that:
...game studies means not just studies about games-for-players, or as rules-for-games, but also as computers-for-rules, or as operational logics-for computers, or as silicon wafer-for-cartridge casing, or as register-for-instruction, or as radio frequencies-for-electron gun. And game is game not just for humans but also for processor, for plastic cartridge casing, for cartridge bus, for consumer... and so on.
It's a fascinating discussion that Bogost develops, one that takes a great deal of contemporary philosophy in its stride, and offers a refreshingly wide stance of its subject matter. But while his application of Harman and Bryant's object oriented ontology reveals some interesting questions, it's not clear that it answers the question we set out to explore in this chapter.
The matter at hand, you may recall, is "what is a game?" and it's far from clear that this is best dealt with as an ontological question. Ontology is principally concerned with what exists, the nature of being, or, in its wider scope, the grouping and relationships between entities. There is an ontological aspect associated with games, as we've already seen with Juul's concept of half-real, but to get to this kind of discussion requires a prior conception of what we mean by "game".
Bogost could not reach the conclusion that game studies should include such esoteric areas of exploration as the relationship between registers and instructions, or radio frequencies and electron guns, had he not already established that registers, instructions, radio frequencies and electron guns were all involved with games in some way. His conclusion presupposes a certain concept of a game. It is only by deploying this concept (whatever it is) that he is able to recognize the many things involved in digital games.
Treating "what is a game?" as an ontological question will not settle it once and for all, although that is not to say that ontology doesn't have an important role in a philosophical investigation of games. There are in fact some rather crucial questions in the intersection between games and reality -- and particular that nebulous concept "virtual reality" -- that warrant addressing.
For the time being, though, we must set this domain of philosophy to one side in order to undertake a philosophical investigation as to what the unifying concept behind "game" might be given that we can so easily and confidently act as if we know what a game is, despite not actually agreeing on any particular answer to the question "what is a game?"
Games and Play
The whole of philosophy can be understood as conceptual investigation -- as attempts to explore how the concepts of language (and those behind language) are deployed, as enquiries as to the relationship between our concepts and what we term reality, and as rigorous examination of the consequences that concepts and systems of concepts produce. The British moral philosopher Mary Midgley has suggested that one can appreciate the purpose of philosophy by a comparison with plumbing (2005).
Most of the time, we just accept that our conceptual plumbing is doing its job, but every now and then we detect weird smells from underneath the floor boards and must take them up and examine what's going on behind the scenes. It is time to take a crowbar to the floor of game studies and find out what lies underneath.
In her 1974 philosophy paper "The Game Game", Midgley became only the second philosopher to tackle the question of "what is a game?" This paper, I'm sad to report, is largely unknown in both philosophical and game studies circles, despite its relevance to the foundational question in the latter domain's area of exploration.
The first philosopher to explore this space, Bernard Suits, initially approached the subject in a 1967 paper actually entitled "What is a Game?" which he later revised and expanded into his 1978 book The Grasshopper: Games, Life and Utopia.
Sadly, Sarah Hoffman (2009) has suggested that among the philosophical community Suits' work remains largely unknown, and Midgley's paper is similarly quite obscure.
This is unfortunate, since Midgley and Suits between them have much to offer that is useful in decoding the game concept, and interestingly both of their approaches involve something of a swipe at another philosopher who is far more well-known -- Ludwig Wittgenstein. Indeed, Thomas Hurka in his 2005 introduction to Suits' The Grasshopper has suggested that Suit's book is "a precisely placed boot in Wittgenstein's balls."
Working towards a deeper understanding of language in his magnum opus, Philosophical Investigations (1953), Wittgenstein specifically singles out games as an example of what he calls family resemblance. He observes that the vast variety of games -- board games, card games, ball games and so forth -- have nothing specific common between them, but instead are tied together by a series of commonalities and relationships.
He relates this to the way in which members of a family display common traits -- a similar nose, or build, or hair color, for instance. It is precisely Wittgenstein's claim that "you will not see something that is common to all [games]" that Midgley and Suits take task with.
Suit's complaint is that Wittgenstein asks us to "look and see" if there is anything common in all things we call games, but then doesn't do so himself. Suits thus objects that Wittgenstein had "decided beforehand that games are indefinable", and indeed accuses Wittgenstein of believing in the "futility of attempting to define anything whatever".
Alas, Suits seems to have thoroughly misunderstood Wittgenstein's purposes, for despite the explicit reference to games it is a point about language that the Austrian philosopher was trying to make, namely that the way words come to be used does not originate in definitions; definitions are post-hoc justifications for the way words are used, and it is this usage that Wittgenstein insists is the genuine meaning of the word, not any definition we might propose.
Midgley accepts Wittgenstein's main point, but disagrees with his use of family resemblance to characterize the underlying concept. As she noted to me earlier this year (2010), words such as 'game':
...have neither a single, fixed meaning (which was what Wittgenstein pointed out) nor merely a vague string of resembling meanings (as his idea of family resemblance suggested) but a definite shape, an underlying organic unity which is often mysterious but must be present in the background to account for e.g. their being usable as metaphors.
She observes, indeed, that Wittgenstein is quite dependent upon understanding the word "game" in a particularly subtle way, for without this he cannot make use of his idea of a "language game" which is a central concept in his later philosophy. This is only possible because we do have a general grasp of the concept of a "game" and can thus understand appeals to this concept in a wider context, such as in the case of Wittgenstein's language games.
In "The Game Game", Midgley (1974) draws from the work of Julius Kovesi to develop her argument. Kovesi had very similar issues with the apparent nebulosity of Wittgenstein's concept of "family resemblance", and his argument can be felt resonating in Midgley's paper. In his book Moral Notions (1967), he had been pursuing a rigorous argument counter to the attacks on moral philosophy by A.J. Ayer and G.E. Moore and others that had attained considerable notoriety in the first half of the twentieth century.
Kovesi demonstrated the relationship between needs and concepts by the example of particular kinds of furniture, claiming that provided you understand the need that (say) a chair embodies (i.e. to support a person while sitting), you know what characteristics are relevant in distinguishing a chair from other kinds of objects. This example generalizes to other cases. As Midgley observes (directly following Kovesi) "in general, provided you understand the need, you know what characteristics to look for. To know what a chair is just is to understand that need."
Thus -- despite disagreements over the details concerning games -- we are all perfectly able to deploy the concept of a game precisely because there is an underlying unity to it. It is because games meet human needs (and, for that matter, animal needs), and because human nature has its own structure, that we can identify what constitutes a game. Those needs that a game meets are precisely what is involved in understanding what the concept 'game' must mean.
Grip and Grind
One way of exploring our need to play is to dig deep into the biology of the gaming experience. There are a number of key chemicals that can be identified, such as the neurotransmitter epinephrine (or adrenaline), which is the underpinning of excitement -- the most primal of the emotions of play.
Games of chance and competition add to this raw excitement a winning state that produces feelings of elation that cause the victor to punch the air or raise their arms, what Suits describes as "the truly magnificent exhilaration that can be produced only by a supreme triumph". There is no word in English for this emotion, but the researcher Paul Ekman (2003) notes that in Italian it is called fiero, the personal triumph over adversity. Since this word is unfamiliar to most English speakers I will use the term triumph as a synonym.
The experience of triumph is intimately involved with some of the most popular forms of play and this emotion, and its watered-down version satisfaction, can be correlated with the neurotransmitter dopamine. This reward chemical is principally released by the reward center of the brain's limbic system, the nucleus accumbens, and is involved in emotional experiences of satisfaction and triumph.
This neural apparatus is vital to the formation of behavior, and can also generate compulsion via reward schedules (or schedules of reinforcement) of the kind identified by B.F. Skinner (1938) and Charles Ferster (1957).
Game mechanics based upon these systems are endemic to digital games, as John Hopson (2004) and others have observed, especially in the case of computer role-playing games and MMOs. While the player maintains interest in what the game is asking them to do, they will merrily jump through whatever hoops they are pointed towards, provided there is some reward to be paid out.
However, when they begin to lose interest they will become aware that they are being asked to perform a series of highly repetitive tasks in order to achieve some measure of progress. Players call this the grind and the associated activity grinding, and compare it to a metaphorical treadmill. The comparison is apposite -- but it is important to remember that the hamster often enjoys running on their wheel.
When a player is grinding, they are expected to keep accumulating an in-game resource (usually either money or experience points) in order to reach progressively higher targets, each target affording the player a reward in terms of increased power or capability. Economist Edward Castronova has gainfully compared this to Camus' account of the myth of Sisyphus, the mythological Greek king who was condemned to roll a boulder up a hill repeatedly, only to have it roll back down again.
But Castronova notes that it is not quite this futile in a game with grinding: when the grinding player reaches the top of the mountain, the stone goes over the top and rolls into the next valley, and as Sisyphus goes to roll it up the next mountain he discovers he has become more powerful. Again and again, the grinding Sisyphus conquers one mountain to find another behind it -- but all the while, they have the sense of achievement from having gained a little bit of power in the process.
Grind is often singled out for criticism, but this is because when players notice they are grinding, they have already lost interest in the activity they are being asked to perform. Pragmatically, those games which rely on reward schedules to structure their play can maintain interest for radically greater lengths of time than those that do not -- and provided what the player is asked to do does retain some interest, the grind is precisely what maintains the player's interest. While some game designers try to develop mechanics which avoid the grind, games that make use of grinding are becoming increasingly significant in the commercial market for digital games.
There are sound reasons for certain players to prefer games which include grinding, since the mechanics behind grinding usually afford progressive advantages to players for continuing to play, and this means that players with a lack of skill in a particular relevant area can compensate by simply grinding to increase power ("level up").
The sense of triumph may ultimately be less because of this self-adjusting element of difficulty, but this also means that success is not restricted to the players with the necessary skills (or tenacity) to overcome. Furthermore, players open to grinding can still achieve triumph by aiming for thoroughness -- completing collections, for instance, or doing everything possible that the game presents as a goal.
Grinding is an important part of the appeal of certain digital games, namely those which utilize reward schedules to structure the play. Although these mechanics began in fantasy role-playing games, they are now found in a great many other games. The appeal of the car simulator series Gran Turismo (Polyphony Digital, 1997 onwards) lies in part in the capacity to grind races in order to earn money and thus buy bigger cars.
The Modern Warfare franchise (Infinity Ward, 2007 onwards) has applied reward schedules to the already successful first person shooter (FPS) genre of digital games and as a consequence set new benchmarks for sales, selling 20 million units in a market that previously topped out at about 8 million. A great many games now directly target their players' reward centers via reward structures and grind.
The reward center of the limbic system also has incredibly close structural ties to another part of the brain, the orbito-frontal cortex which lies in the brain at a position just above the eyes.
This is essentially the rational decision center of the brain, and there is no other part of our neural architecture more closely tied to the nucleus accumbens, where most dopamine is produced. Making a good decision is pleasurable, and indeed this explains why solving puzzles is enjoyable: when we find the solution, it triggers a release of dopamine.
There are many games which rely upon this neurobiological mechanism for their enjoyment, including chess, checkers and so forth, where direct competition further enhances the enjoyment of making a good decision. The fun of a game of chess lies not just in winning the game but in solving the challenging problem of how to beat the other player given the current state of the board, which is why chess puzzles in newspapers and the like enjoy an audience: solving puzzles is inherently enjoyable, if you're sufficiently interested in the kind of puzzle to want to solve it.
Chess is in effect a game which generates puzzles to be solved, and the same is true of a great many hobby games, the relatively complex, decision-focused board games that enjoy a cult following among geeks of all stripes, although these often have other dimensions to their appeal as well.
But there is more to this connection between the decision center and the reward center than solving puzzles. Researchers at Cambridge University (Clark et al, 2009) have shown that even if we fail at an activity, the decision center will release dopamine if it assesses that we nearly won.
In other words, when we come close to triumph, the limbic system spurs us into another attempt. I have called this mechanism grip, and it explains why gambling and certain digital games can be addictive even in the absence of reward schedules.
When a particular game gets the player into a state of wanting "just one more go", it is because of grip: the feeling that one might succeed (or do better) on another attempt fosters the desire to persevere. Slot machines depend upon this for their appeal (if you didn't win with that coin, surely you will have a better chance of winning next time!).
Although related to grind at the neurobiological level, grip is quite distinct in character. Recall that a player who is grinding is repeating the same activities, accumulating an in-game resource. A player caught in grip is also pursuing future reward, but its attainment is uncertain.
A grinding player knows they will eventually collect enough of their resource to win their reward; a player experiencing grip only believes they will eventually win -- which is partly why slot machines are so effective at making money for casinos.
As it happens, grind also generates grip. The decision center assesses the future reward -- and this reward will certainly be achieved, the only uncertainty is when. As a result, players caught in grinding often have great difficulty stopping, and if they do manage to stop they remain under the game's spell, and anxiously desire to return to it at the earliest possible moment. Players who tend towards goal-orientation are particularly at risk to both grip and grind, and as a result are more likely to become addicted to a game than a process-oriented player.
I myself tend heavily towards goal-orientation, and as my wife will testify I become terribly addicted to computer role-playing games -- to the extent that nowadays I'm not allowed to play them except under special circumstances (such as I am working on the development of one). While living in Knoxville in 2000, I had terrible insomnia because I was playing Pokémon (Game Freak, 1996), a computer role-playing game about training fantasy creatures.
I was so caught up in the grind of the game's reward structures (i.e. improving the abilities of my pet creatures) that I was in a state of perpetual grip, always wanting to get back so that I could evolve my Bulbasaur or get my Pikachu to just the right state. I spent several hundred hours playing those games that year, and I was by no means the only player to be sucked into the fictional world of Pokémon: the original Game Boy games sold an astonishing 45 million units between them.
In the absence of grind, grip occurs when the player is in pursuit of triumph, which corresponds to a large release of dopamine: the player wants to win, and while they believe they can win, grip will motivate them to continue trying. But in order to produce this potent experience of triumph it is necessary for there to be adversity to overcome.
In games of chance, this can be attained by raising the stakes -- the bigger the risk, the greater the reward -- but in games of competition this is attained by direct conflict. This brings in frustrations (i.e. anger) which are associated with the neurotransmitter norepinephrine. Nicole Lazzaro (2004) has also correctly recognized that difficult puzzles also produce triumph; this appears to be because of the close link between the decision and reward centers of the limbic system we have already seen in connection with grip.
It is also worth briefly noting the role of testosterone in sustaining competitive play. Since the 1970s, testosterone has been connected with persistence and tenacity, and testosterone levels spike when player's triumph over adversity. It is not that high testosterone is a requirement to experience Ekman's fiero, it is that people with high levels of this androgen (male or female -- its effect on behavior is not gender-specific) are more likely to persist against a challenge and thus attain victory (Andrews et al, 1972; Oades 1978; Booth et al, 1989; Mazur et al, 1997). To put this another way: players with high testosterone levels are more susceptible to the grip of an undefeated challenge.
Recall that excitement relates to epinephrine, and that competition can be related to norepinephrine. This makes two of the basic patterns of play correspond with the two sides of the fight-or-flight response, first observed by Walter Cannon in 1929. Anger is the underpinning of the desire to fight, and corresponds to competition, while fear lies beneath the urge to flee, and relates to experiences of vertigo and excitement.
It must be noted, however, that a game of competition need not be angry, and an exciting play experience need not be fearful. A low level of frustration (anger) will not even reach the conscious awareness of a competitive player, and although their neurochemistry is identical, excitement and fear are distinct experiences, with only the latter involving the amygdala, the brain's fear center.
What's more, there is a connection between the chemical correlates of the fight-or-flight response (epinephrine and norepinephrine) and the reward chemical dopamine involved in grip, grind and triumph, since the former two chemicals are synthesized from the latter. In fact, essentially every multi-cellular animal that has evolved since the Cambrian makes use of these three catecholamine chemicals -- from ants to zebras -- although the actual chemistry involved is slightly different in certain cases. The chemicals that support play at the biological level are the same as those involved in the most basic behavioral mechanisms of reward and survival.
The anthropologist Thomas Malaby has taken a particular interest in play and games, and has published a number of fascinating papers on the subject (2007, 2009), with particular reference to time he spent studying the role of play in contemporary Greek culture. One of Malaby's key observations concerning games are that they are processes, sustained by human practice. But what kind of process?
He notes that "Games are, at root, about disorder", recognizing a central role for contingency in games, and suggesting that the incredible unpredictability of our everyday experience bridges the gap between games and life in general: games contrive unpredictability, but life is by its very nature always already unpredictable. Contriving contingency is one of the things that games excel at, since games which are readily predictable rapidly become boring.
The element of uncertainty, while crucial, is not the whole of the matter. Malaby (2007) observes that a second crucial aspect of games is their capacity to generate meaning. The many kinds of situation that can occur within a game (including but in no way restricted to the goal states and final outcomes, such as winning) happen in never wholly predictable ways and are "subject to interpretations by which more or less stable culturally shared meanings are generated."
This generation of meaning is a critical aspect of the game experience, and it is thoroughly open-ended. Not only can the way games are played alter within any particular social group, but the meanings that a game can generate can also change.
This appreciation of the meaning of the internal states of games is crucial to understanding the play of a great many games, and particular of digital games. The more complex games are not always undertaken for the sake of winning, even if this forms part of the framework of motivation.
No, what is rewarding in a game is the interpretability of the states the rules of the game throw into the player's awareness. Nowhere is this more clear than with a game like The Sims (Maxis, 2000) or the game I designed with Gregg Barnett, Ghost Master (Sick Puppies, 2003), where a great deal of the player's enjoyment is in the stories they tell about the little people running around on screen.
There is an important connection at this point with stories. Stories too are processes, and like Malaby's games they aim to be compelling or engaging, and possess a characteristic capacity to generate meaning by their internal states. There is a temptation to say that, unlike games, the content of stories is fixed, static -- but this reaction is premature.
Perhaps the most important states generated by a story are the emotion