Video Games and Emotion : Early Models and Research

The research on emotions in video games has largely been confined to experimental attempts to prove the aggressive effects of gameplay upon players (Anderson & Morrow, 1995; Ballard & Wiest, 1996; Cooper & Mackie, 1986; Dominick, 1984; Fling, Smith, Rodriguez, Thornton, Atkins, & Nixon, 1992; Funk & Buchman, 1996; Griffiths & Dancaster, 1995; Irwin & Gross, 1995; Kubey & Larson, 1990; McClure & Mears, 1984; Nelson & Carlson, 1985; Schutte, Malouf, Post-gorden & Rodasta, 1988; Scott, 1994; Silvern, 1987). The methodology of these experiments may be seen, with some variation (and qualification), as taking the following general form. First, a group of players are measured for an aggressive state. Second, some players are given a so-called “violent” game, while other players are given a so-called “non-violent” game. Third, after a measured period of play, the aggressive states of both groups are re-measured. Fourth, the subsequent difference in the level of aggression is found to be either short-lived, minimal, or non-existent.

There are many problems with this experimental methodology. Aside from Jenkins' (1990) observation that early research tended to abstract “violence” from its context of reception, the experimental conditions of these types of research produce an artificial context for producing and expressing aggressive tendencies, and tend to confuse correlation with causation (Durkin, 1995, 1999). Certainly, there are broad, historically specific correlations between the male demographic and the aesthetics associated with challenge, competition, and aggressive mastery. However, aggressive personality types often choose to engage with media that reflect and reinforce their existing tendencies, making the measurements correlative, not causal.

A more obvious criticism is that such research does not adequately account for the cathartic experience of games, that is, the extent to which emotions are aroused and purged and thereby nullified, reduced, or redirected into a socially acceptable form. Catharsis is an aesthetic evident in Malone's (1981), Myers' (1990c) and Turkle's (1984) analysis of players' preferences, but assumptions about catharsis in gameplay need qualification. First, like aggression, the catharsis players expect from games is less likely to be measurable in an unfamiliar experimental context: the everyday routines one seeks relief from are attenuated or disrupted by the experimental conditions. Second, it is simplistic to see the reduction of aggression as evidence of beneficial catharsis. To presume that aggression felt in relation to a video game is necessarily negative and anti-social, and therefore undesirable, is the same as evaluating a person's anger at their football team losing without reference to the psychological, sociological, and ideological functions of fandom. Some players may find it useful to induce manageable anger at a game rather than deal with anger, or other emotion, resulting from some unmanageable social situation.

Third, measuring a purged emotion after its purging ignores or misrepresents the desired aesthetic experience of playing a video game. D etermining whether or not some players end up with a (temporary) higher level of stress or aggression after play may be beside the point if what players seek from a game is what they experience while playing it, not the state they expect to be in afterwards. In this respect, a “black box” experimental approach draws attention away from the actual processes of gameplay, especially the real-time regulation of emotions. Video games may produce stress relief that is measurable by comparing the “before” and “after” levels of physiological arousal. However, not only may this purging be episodic, significant emotional moments of play may occur early on in, or before the end of, a gaming session, and the actual closure may be poorly defined or arbitrary. The underlying problem here is that the catharsis hypothesis was developed for traditional dramatic forms—stage plays, print novels and feature films—which have an aesthetic formula organised around discrete emotion-episodes polarised around central concerns, in which elements are drawn together and resolved in something that subsequently resembles a total, organic text. Video games have been played for the last three decades not as ineffective dramatic forms but as a different kind of dramatic form, or, more precisely, as a different kind of emotion-regulator.

Within the bulk of the research addressed in this chapter there has, it should be noted, been a tendency to consider a range of emotional states in video games. However, those emotions tends to be addressed indirectly and/or are seen as debased, compared with print and film. The “altered states” of “meditation,” “kinaesthesia,” “immersion,” “agency,” and “flow” may be positive, but they are generalised from certain genres or aesthetics, principally first person shooters, and refer to general states. They therefore provide a limited basis for analysing the dynamic regulation of emotions in any particular game, and it remains necessary to provide a more systematic theoretical account of the regulation of meaning and emotion in gameplay.

Frijda's Model of Emotion

While Bordwell's (1986) account does not address the emotional experience of film viewing, researchers such as Smith (1995), Tan (1997), and Grodal (1997) see emotional responses as of equal or primary significance. For Tan (1997), while there may be many motives for watching films, “what all natural viewers of the traditional feature film have in common is their desire to experience emotion as intensely and abundantly as possible, within the safe margins of guided fantasy and a closed episode” (1997, p. 39). In an analysis of video games, in which the player's sensorimotor system is engaged and emotional responses are not merely virtual, it becomes even more necessary to appreciate the motivating role of emotion.

Following Tan, we can turn to Frijda's definition (1986, 1988, 1989) of emotions as changes in action readiness through the appraisal of a situation. “Action readiness” refers to how emotions activate the nervous system, giving way to rapid physiological changes, such as heart rate, bronchial dilation, and inhibition of the digestive system. These not only prepare an individual for physical action in a general sense, they are tied to specific “action tendencies” (Frijda, 1986, p. 70): to run, attack, seek intimacy, kiss, and so on. Some theorists, such as Izard (1984), argue that each emotion has its own neurophysiological substrate, such that a basic set of emotions can be identified. Most of these “basic” or “elementary” emotions—notably interest, happiness, surprise, anger, fear, sadness, and disgust—are generally accepted. However, cognitive (Kagan, 1984), structural-developmental (Sroufe, 1996), functionalist (Campos & Barret, 1984; Frijda, 1986), and socio-cultural (Abu-Loghod & Lutz, 1992) traditions in the study of emotion argue that the quantity and/or quality of emotions is seen as depending upon the stage of development, cognitive appraisal, and/or the cultural context (see Mascolo & Griffin, 1998, pp. 4-14).

The functionalist position is privileged here, following Frijda (1986), Lazarus (1991), and Ortony, Clore and Collins (1988), who see emotion as having a control precedence. That is, “emotion is functional in that it takes control, precluding extended reasoning: a decision to take action—any action—is better than failing to act in time” (Tan, 1997, p. 43). Emotion, then, influences one's attentiveness and motivation towards a particular problem, and, by extension, one's ability to deal with this problem. However, a key difference between theorists is the extent to which emphasis is placed upon associated physiological and cognitive processes. Without minimising the extent to which the different traditions offer conflicting interpretations, it is possible to take the moderate position that each tradition principally emphasises different aspects of processes that cross the biological and psycho-social.

Upon this basis we can appreciate Tan's (1997) distinction between “primary” and “secondary” appraisal. For Tan, following Frijda (1986), “primary appraisal” is an advanced, automated, and immediate process directly related to the stimulus, premised upon such basic elements as intensity, match-mismatch, hedonic quality, agency and causality. This evaluation is limited to “immediately observable affordances” (Tan, 1997, p. 47), that is, those forms of action which are obviously associated with the object or situation, such as fight or flight (Gibson, 1979; see also Baron, 1980). Indeed, the term “affect” can be usefully reserved for the biological component of arousal consequent upon primary appraisal (Basch, 1976). Tomkins' (1962, 1963, 1991) classic analyses of the complex assemblies of nine innate affects is paradigmatic of this type of usage. While the feeling-states that Grodal (1997) refers to are more general than Tomkins' (1962) affects, they may also be seen as affective in the limited sense that they are premised upon primary appraisal. “Secondary appraisal,” by contrast, is a continued cognitive elaboration accompanied by “emotional significance that is not directly evident from the situation itself” (Tan, 1997, p. 47) (see Frijda, 1986; Lazarus, 1966). In Frijda's (1986) terms it is only when secondary appraisal occurs that we can speak of true emotion, and this study reserves the term “emotion” for this usage. Indeed, for Frijda, each emotion has a distinctive situational meaning structure, or cognitive evaluation. For example, fear is not merely a physiological response, it is a state of arousal combined with a cognitive evaluation of an imminent threat.

A central question for the study of emotion in film is if the viewer experiences true emotions. Tan (1997), following Frijda, is aligned with the cognitive branch of film theory (Allen, 1997; Carrol, 1990; Grodal, 1998), which critiques the assumption that viewers mistake the screen for reality. Tan (1997) differentiates between an involuntary “illusion of motion,” based upon the physiological phenomena of “flicker fusion” and “apparent motion,” as well as a series of voluntary illusions: the “illusion of the diegetic effect,” “the illusion of the controlled witness,” and “the illusion of the observational attitude.” These “illusions” comprise a sequence which most viewers voluntarily and progressively accept and invest in. The viewer thereby takes up a voluntary observational position as an invisible witness and enters into the diegetic effect. Being in the position of witness the viewer knows that events do not affect his/her welfare directly, but is still able to form an emotional relationship with the film and its characters. This relationship is not one of habitation or straightforward identification, but of empathetic concern from the position of spectator. This finds its analogue in other situations in which we find ourselves spectators, such as watching sport. Even if the events being watched are not real, our relationship to situations and characters as a spectator is genuine, such that we may feel genuine, albeit inhibited, emotions. As Frijda (1989) argues: “film does not elicit emotions by products known to be imaginary, but allows inhibition of emotions because the products are known to be imaginary” (¶ 4).

Tan (1997) elaborates upon this by distinguishing between “Artefact-emotions,” or “A-emotions,” which are responses to the film's non-diegetic elements, from “Fiction-emotions,” or “F-emotions,” which are responses to situations within the film's diegesis. In the classical feature film, A-emotions are minimal (or minimised). Tan consequently focuses on F-emotions, arguing that interest is the key emotion produced while watching a feature film, and that this is oriented around the production and confirmation of expectations about events in the diegesis. These expectations are governed by “thematic structures,” such as viewers' expectations about the course of a story about “Revenge,” and “character structures,” that is, how characters' goals and motives lead viewers to privilege and expect certain outcomes on behalf of those characters (p. 179). The action tendency of interest is a desire to keep watching, but more complex emotions, such as admiration, pity, and sadness, may be reinforced by, and reinforce, interest. These more complex emotional responses are likely to be inhibited because they pertain to a fiction, and the action tendencies associated with them are virtual, or imagined. Someone watching a film may be cued to dislike a particular character on the screen, but s/he is not obliged to act upon that emotion by abandoning their seat and assaulting the screen; s/he merely imagines and fears or hopes for certain outcomes.

Frijda's (1986) model of emotion, and Tan's (1997) application of it to film, provides a basis for some valuable questions and departure points. In video games, the cognitive response to the aesthetic object gives way to some form of sensorimotor-ergodic action, such that video games produce both virtual and actual action tendencies. This being the case, how does the player's position as both invisible witness (to narrative sequences) and active participant affect his/her emotional responses? Should we see participation as giving rise to a different sort of interest, and does the confluence of virtual and action tendencies affect the course of this interest? Should we expect video game characters to elicit the same kind of empathy felt towards filmic characters? And does a player's concern with his/her own interests as a player interfere with his/her ability to empathise with characters?