Game Studies

The Child in Games: Representations of Children in Video Games (2009 - 2019)

Abstract

This paper examines representations of children in contemporary video games through content analysis. A selection of commercially successful and critically acclaimed video games published within the last ten years (n=506) was sampled in order to determine what proportion of these titles contained child characters. The games that contained child characters (including non-human and quasi-human child characters) were analyzed to ascertain the relative importance of these child characters. If they were found to play a significant role, the child characters were then coded for race, gender and age. The sample was categorized according to genre to discern whether child characters were more prevalent in some genres over others. Finally, the corpus was organized by year of publication to see if the proportion of games containing child characters varied over time. The results show that the majority of successful video games published between 2009 and 2019 did not contain any child characters at all. 19% of the total games sampled contained significant child characters, of which around half were playable characters. Most child characters were aged between 6 and 11 years old, and white, male children outnumbered non-white children of different genders. Child characters appeared more frequently in games categorised as “Action,” “Role-playing” and “Adventure” than they did in games from the “Sports,” “Strategy,” “Rhythm” and “Sandbox” genres, and the proportion of games containing child characters has remained fairly constant over the past ten years. The paper concludes by suggesting future directions for research conducted at the intersection of childhood studies and game studies.

Keywords: Childhood studies, representation, content analyses, children, age

Introduction

Ten years ago, Williams, Consalvo, Martins and Ivory (2009) published a large-scale content analysis of characters in video games entitled “The Virtual Census: representations of gender, race and age in video games.”’ Their aim was to “obtain a baseline measure of race, gender and age distribution across the current universe of videogame characters” (p.816) and to compare this distribution with that of the US population. Their sample consisted of 150 games mediated across 9 different platforms from a single year, and the results were weighted according to the sales rates of individual titles. From this sample, they were able to make generalized statements about representation in popular video games. The results showed an over-representation of male people, white people, and adults, and a systematic under-representation of non-male people, Black people, Hispanic people, Indigenous Americans, children and the elderly.

Williams et al. are far from the only researchers who have used content analysis to document and critique representation in video games, but their study is one of very few to consider “age” as an important identity marker. Williams et al. found that adults appeared at a rate in games 47.33 percent higher than their prevalence in the actual population of the United States, while children and the elderly were underrepresented. Children -- defined by Williams et al. as “people aged 13 and under” -- made up 21.41 percent of the population of the United States in 2009, but constituted only 3.58 percent of characters in the video games published that year. The absence of child characters in video games intersects with the systemic underrepresentation of non-white, non-male people in significant ways; however, the implications of children’s invisibility in gameworlds are different enough from other forms of erasure to warrant separate analysis.

This article uses content analysis to assemble and critique a corpus of video games, building on over two decades of video game research projects employing this method. Several of these studies look at games released in a single year (Dill et al. 2007; Downs & Smith 2009; Burgess et al. 2011), but others consider games released across several decades (Lynch et al. 2016). Some data sets include dozens of individual games (Dietz 1998; Heintz et al. 2001; Brand et al. 2003) whilst others consist of just 4 key texts (Waddell et al. 2014), and whereas some studies have selected and weighted games according to sales figures (Williams et al. 2009), others have used game genres (Passmore et al. 2017), age-rating categories (Haninger et al. 2001, 2004a, 2004b), review sites (Wilberg 2011) and industry awards (Perreault et al. 2018) to set parameters for their samples. Some studies have specifically considered self-creation as a form of representation by focussing on the possibilities and limitations of avatar customisation in Role Playing Games (RPGs) and Massive Multiplayer Online Games (MMOs) (Dietrich 2013). A small number of content analyses document video game paratexts rather than the games themselves (Burgess et al. 2011), and an even smaller number examines mobile games rather than PC and console games (Wohn 2011). In spite of the diversity of approaches to content analysis, the conclusions reached in these studies are markedly similar: straight, cisgender, white men are systematically over-represented in video games, women are more likely to be sexualised and positioned in secondary, supporting roles and non-white women are frequently exoticized. When Black, Indigenous, Asian, Middle Eastern and Hispanic characters appear in gameworlds they tend to be positioned in peripheral roles and are often heavily stereotyped. Queer characters are rare, and usually function either as comic relief or as symbols of depraved villainy. The exception to these general trends is found in Wohn’s study of representation in casual games, which affirms the under-representation of non-white characters, but finds that female characters appear more often than male characters, and that these female characters are not overtly sexualised. The dominant trends that emerge from these content analyses clearly demand consideration from feminist, post-colonial, queer and critical race studies perspectives. However, the lack of emphasis on “age” as an identity marker in the majority of these surveys means that perspectives from childhood studies are rarely employed to synthesize and critique data.⁠[1] This content analysis takes an intersectional approach that considers constructions of childhood in relation to constructions of other social identities. Building on the work of Thomas (2019), Toliver (2018) and Sharpe (2016) that examines the impossibility of the Black girlhood in the dominant cultural imagination, this article posits that documenting the intersections of race and gender with childhood is essential to understanding the factors that cohere to produce “the child,” both as a social role and as an ideological configuration.

Method

The games sampled in this study were chosen to represent both commercial success and critical acclaim. Since the aim of this content analysis is to identify the dominant patterns that characterize depictions of children, the reach, reception and prestige of individual games was used to determine whether a game’s influence should be considered “dominant.” In other words, this content analysis intends to document ideologies about childhood that are expressed in spaces that the majority of video game stakeholders would accept as “mainstream,” rather than those located in contexts considered extreme or unorthodox. Commercial success was determined using both sales data and consumer-generated Metacritic ratings. Firstly, annual statistics provided by the NDP group were used to determine the top ten best-selling PC and console video games for each year from 2009 to 2019 (n = 110). Secondly, the games that received the highest annual Metacritic ratings on each of the four most popular gaming platforms of a specific year were recorded for 2009 through to 2019 (n = 440). Critical success was determined through industry awards. Games that received awards from BAFTA, D.I.C.E. (Design Innovate Communicate Entertain), The Golden Joysticks and the Game Developers’ Choice between 2009 and 2019 were recorded (n = 674). Games that had been recognised by The Game Awards were also included from 2014 onwards, when this awarding body was established (n = 111). Duplicates were removed and games that were published before 2008 were excluded, unless they had been re-released, re-mastered, or re-made within the last ten years, leaving a sample of 506 individual titles. These 506 entries were recorded along with their publication year and age-rating.

The games were then categorised according to genre. Genre was determined using the tags specified in each game’s online promotional material. This generated 66 different categories, many of which were characterized by strings of descriptors such as “Stealth-Action-Adventure” or “Run and Gun-Puzzle-Platformer.” These hyphenated strings were broken down into their constituent parts (e.g., “Stealth-Action-Adventure” was recorded as “Stealth,” “Action and “Adventure”) in order to assemble a list of 33 discreet descriptors. This list was further condensed into 15 categories by incorporating subsidiary genres under umbrella terms -- for example, Metroidvania games were identified as “Platform” games, Third-Person Shooters and First-Person Shooters were consolidated under the heading “Shooter,” and Racing games were listed as “Sports” games. The final 15 genres were Shooter, Strategy, Role Playing, Online Multiplayer, Sports, Platform, Sandbox, Puzzle, Adventure, Simulation, Stealth, Action, Rhythm, Fighting, and Other. Although this process of simplification resulted in the loss of some granular detail, it was necessary in order to make meaningful comparisons between categories.

To answer the binary question “does this game contain any child characters?” it was necessary to set the parameters for what constitutes a child. For the purposes of this study, a “child” was defined as a character whose representation suggested that it was aged between 0 and 14 years. If a character’s age was not made explicit in the game itself and was not specified in the game’s paratexts, then its age was determined through a combination of anatomical markers in its audiovisual representation, its social relationships with other characters and its associated game mechanics. Unlike other content analyses (e.g., Williams et al. 2009), non-human and quasi-human characters were included this study, following Passmore et al.’s precedent in which all characters that are “positioned in a human-relatable context, e.g., behind the wheel of a car” (2017, p.143) were included. This inclusive approach extends to anthropomorphic animals, supernatural creatures, toys, automata, robots and cyborgs, because, as Jaques notes, there is a long media tradition “in which the child and the animal overlap, address and reflect one another” (2015, p.13) and children are frequently aligned with or symbolized by living toys.

Video reviews, Wikifan pages and gameplay walkthroughs shared on YouTube were used to determine whether a game contained any child characters. This process involved first watching a video review published on a commercial entertainment review site (IGN, Polygon) to get an overview of the game. This was followed by reading plot summaries and characters bios to look for mentions of child characters (Wikifan), and finally by doing targeted searches for gameplay videos hosted on YouTube based on this information. If a game contained a child character, it was recorded in one of the following categories: “Player-character” (e.g. Isaac in The Binding of Isaac), “Player-character (Flashback)” (e.g. young Aveline de Granpré in Assassin’s Creed III: Liberation), playable “Supporting Character” (e.g. Ellie in The Last of Us), playable ”Ensemble Character” (e.g. Baby Mario in Mario Kart 8 or Ezra in Kentucky Route Zero), “Ensemble Character (NPC)” (e.g. the Little Sisters in Bioshock), “Supporting Character (NPC)” (e.g. Atreus in God of War IV), “Antagonist (NPC)” (e.g. the Newts in Days Gone), “Protagonist (NPC)” (e.g. Klaus/Karin in My Child Lebensborn) or “Set Dressing” (e.g. the unnamed village children who populate certain locations in Dragon Age: Origins).

If a game contained child characters that belonged to more than one category, the most significant child character was recorded for the purpose of answering the binary question “are there any child characters in this game?” and then the remaining child characters were recorded separately for the purpose of determining the distribution of race, gender, and age. A child character was considered significant if it met several (but not necessarily all) of the following criteria: it was playable, it had a name, personality and backstory, its presence was integral to the plot, its presence demonstrably influenced the game’s atmosphere or aesthetic and it played a key role in navigating a game’s ludic challenges.

Race, gender, and age were determined through a combination of close readings of video walkthroughs hosted on YouTube, and an examination of a game’s paratexts including promotional material, official websites, and associated fansites. The category of “race” was not coded solely according to “the heuristic cue of avatar skin color” (Waddell et al. 2014, p.5) as it has been in other studies because a proportion of characters -- particularly those rendered in an anime style -- had “white” skin tones but were best described as mukokuseki or “culturally odourless” (Hutchinson 2019, p.106). Koichi Iwabuchi (2002) first suggested the term “culturally odourless” to describe the way in which Japanese creators sublimated the national and racial identity of their central characters in order to increase their global appeal. Instead, chromatic race was considered alongside a range of racialised characteristics including clothing, language, accent and associated cultural objects in order to categorise characters as “White,” “Black,” “Mixed Race,” “Latin American,” “Asian,” “Middle Eastern,” “Indigenous American” or “Mukokuseki.” “No Race” was used to describe non-racialised animals, toys, robots, and supernatural creatures, and “Customisable” was used to describe characters whose race was defined by players. Many races and ethnicities were not represented in the games sampled and so were not included as categories. Gender was coded as “Male,” “Female,” ” “Ungendered” or “Customisable.” No child characters in the games sampled were represented as being transgender, non-binary, or gender non-conforming, and so these were not included as categories. If child characters appeared as a group representing a range of genders -- which frequently happened when child characters were used as set dressing -- this was recorded as “All Genders.” Since rough estimations were sometimes used to determine a character’s age when this information was not explicit in the game or in its peritexts, this variable was coded using the categories “Infant,” “Child,” “Preteen,” “Early Teens” or “Undefined” rather than discreet ages in years. 

Findings

Of a total of 506 games included in the study, 331 (65%) did not contain a single child character.: their virtual worlds were entirely child-free zones. 176 games included child characters in their gameworlds, but of this subset 19 titles used child characters merely as set dressing. Of the remaining 157 games, 97 featured a significant child character, 45 of which were available as Player-Characters for the duration of the game. There seems to have been a slight increase in the proportion of games containing child characters over the past ten years, with 2009 marking a low of only 24% of games containing child characters, and 2018 marking a high of 56% of games containing child characters. 2019, however, fell just below the ten-year average of 38% of games, with only 36% of games released that year containing child characters. This pattern over time remained consistent when looking only at significant child characters.

When broken down by genre, Rhythm games and Sandbox games in this sample had no child characters at all. Only 3% of Sports games and 11% of Strategy games contained child characters. At the other end of the spectrum, 51% of Action games, 58% of Adventure games, and 78% of Role-playing games in this corpus contained child characters, albeit not necessarily in significant roles. In fact, when looking only at significant child characters, the number of Role-playing games that met these criteria fell drastically to 32%, showing that the majority of child characters in Role-playing games are in peripheral roles or are used as set dressing. 35% of Adventure games and 34% of Action games featured significant child characters, and Stealth games had the highest number of playable child-avatars relative to the total number of games from this genre. The likelihood of a game in this sample containing child characters did not increase when the game was deemed appropriate for child-players by an independent age-rating system (PEGI, ESRB). In fact, games that were rated “16+” and “18+” were more likely to contain child characters, while only 15% of games rated “3+” contained child characters. 36% of games rated “7+” contained child characters, 29% of games rated “12+” contained child characters, 41% of games rated “16+” contained child characters, and 54% of games rated “18+” contained child characters.

Discussion

In his writing on the spatial politics of childhood, Jenks notes that childhood

is that status of personhood which is by definition often in the wrong place, like the parental bedroom, Daddy’s chair, the public house or even crossing the busy road. All people in any society are subject to geographical and spatial prohibitions, whether delineated by discretion, private possession, or political embargo, but the child’s experience of such parameters is particularly paradoxical, often unprincipled and certainly erratic. In terms of social space children are sited, insulated and distanced, and their very gradual emergence into wider, adult space is by accident, by degrees, as an award or as part of a gradualist rite de passage. (2005, p. 73-4)

The fact that child characters simply do not exist in the majority of video games reflects children’s exclusion from social, public and professional spheres in contemporary Anglo-American society. Jenks’ association between modern society’s desire to “insulate” children and keep children at a “distance” holds true in virtual spaces, where children are seemingly “protected” through total erasure. The specific lack of playable child characters speaks to the idea of the child as object rather than subject, or as “symbol” rather than “agent.” Most child characters who appear in digital worlds are allocated secondary, supporting, or decorative roles. The decentring of child characters in video games affirms the presumed passivity and the peripherality of “the child” as a social position. The overrepresentation of whiteness and maleness documented in other content analyses is replicated within populations of child characters, which -- when combined with the frequent use of a repetitive set of child-stereotypes -- reinforces the notion of “the child” as a homogenous, monolithic social status, disconnected from class, race, gender and ability. The following discussion will explore some of the potential reasons behind the limited representation -- or total absence -- of child characters in game worlds, looking first at playable child characters and then at non-playable child characters.

Playable Child characters

In this sample, less than 9% of the total games surveyed had child-avatars, and within certain genres there wasn’t a single playable child character. Games that simulated adult professions did not contain playable child characters, reflecting modern, Western children’s legal and cultural exclusion from the workplace. Sports simulation games -- particularly those that feature living athletes such as the FIFA series or the Madden NFL series -- did not include any child characters, and this holds true for “realistic” racing games such as the F1 series, MotorStorm Arctic Edge, Race Driver: GRID, and the Need for Speed series. The Just Dance series, Dance Central, SingStar series, SongPop, Rock Band and the Guitar Hero series only depicted professional musicians and performers as adults, and games that simulated aspects of warfare represented all playable combatants as adults. Wii Fit Plus,Zumba Fitness and other similar titles represented the player on-screen using the outline of an adult body, and creative-training games such as Colors 3D and the Art Academy series were all facilitated by adult virtual instructors. One could argue that the absence of children in these games is concomitant with the way these games use levels of professionalism and “career-progression” systems of advancement to ascribe meaning to the playing experience and to assign value to the player’s interactions. In other words, since the child is a symbol for both domesticity and leisure time -- which are seen in contemporary Western culture as separate from the world of work -- the exclusion of the child from these games signals their connection to real-world industry counterparts. In contrast, the two racing games in this corpus that included child characters -- the Mario Kart series and Crash Team Racing Nitro Fuel -- were not only characterised by surrealism, absurdity, and fantasy, but were also designed to welcome “non-serious” playstyles that valued slapstick humour and lighthearted tomfoolery as much as they valued high levels of technical skill and ludic proficiency. The Mario Kart games, for example, include significant elements of chance that mitigate the importance of skill, and this accommodation seems to be directly communicated through the neotonised avatars.

Another possible explanation for the absence of playable child characters from certain game genres -- for instance, from Action games and Fighting games -- is that the kinds of playing experiences associated with these genres are often designed to fulfil implied players’ power fantasies. As a representative of a social group that is both physically and structurally vulnerable, “the child” does not connote strength, influence, or importance, and is therefore not an obvious choice to represent an overpowered heroic figure whose abilities greatly exceed those of an average person. In fact, the converse is true: in this corpus, one third of Stealth games -- a genre in which the avatar is deliberately weak and underpowered -- had playable child characters.

In her work on racialised video game characters, Anna Everett notes, “generic stereotypes are part and parcel of entertainment media’s shorthand narrative structures and communicative devices” (2009, p.115). Dominant stereotypes about children mean child characters can function as legible icons that communicate game mechanics to the player. Games with stealth mechanics such as INSIDE and Resident Evil 7 use playable child characters to express and explain the avatar’s debility and defencelessness, which in turn encourages an evasive, careful, strategic playstyle and facilitates a tense, unnerving playing experience. In short, games use audiovisual signifiers to express their rules, and a child-avatar is an effective shorthand for delineating a restricted set of combat options. While some video games in this sample did feature powerful child-heroes (for example, the player-character Aurora in Child of Light), they generally required an additional interpretive step in order to explain why this particular child is unlike other children. Without supernatural powers, royal heritage, or exceptional circumstances, the figure of the child is synonymous with the perennial underdog. Animal Crossing provides an example of the exceptional circumstances in which a childly avatar can drive action and be positioned in a role of responsibility. Since power is relational, the replacement of adult figures in the game world with animals -- who are reminiscent of cute, cuddly, stuffed toys -- installs the child-avatar at the top of the hierarchy, or at least flattens the hierarchy, granting the child importance as a creator, consumer, and community leader. Similarly, the anthropomorphised animals are also elevated through the inclusion of non-anthropomorphic animals in the gameworld that exist on a level with inanimate objects and comestibles.

A proportion of games had child characters that were playable only for a limited section of the game. 29% of games with a playable child character involved a temporary flashback to the adult avatar’s youth. Childhood as a critical site of identity-formation is a common psychoanalytic trope used across media (Byrnes 1995), and so in these video games, the brief sequences featuring playable child-avatars usually served the purpose of explaining and rationalising the adult avatar’s personality, motivations, and behaviour. That is to say, these flashbacks provided an “origin story” intended to deepen the avatar’s characterisation and enhance narrative cohesion. Additionally, flashbacks to avatars’ childhoods were designed to create a greater sense of intimacy between player and avatar, in part because childhood is often perceived as a universal and, therefore, deeply relatable experience. In titles such as the mobile game Florence, the flashback to the eponymous character’s childhood provides context for her current life as a bored, unfulfilled accountant: she was pushed towards studying maths rather than pursuing her interest in art by a controlling mother. The pattern of giving up on -- and then rediscovering -- her childhood passion provides structure for the game’s narrative arc, and the relatable childhood experience of victimisation at the hands of a well-meaning adult universalises what is ultimately a culturally and socially specific story.

Flashbacks were also used to generate “seamless” in-game tutorials. Horizon Zero Dawn, for example, begins when Aloy, the player-character, is a baby. The opening cutscene allows the adult male protector, Rost, to provide some basic exposition for players in the guise of talking in a direct and simple manner to the innocent infant Aloy. When players first assume control of Aloy she is around five-years-old. Players’ unfamiliarity with their environment, their inability to perform basic movements, and their lack of confidence is narrativized by placing them in the position of a young child who has not yet learned the impressive athletic abilities and hunting skills that come to distinguish Aloy as an avatar. Similarly, Uncharted 4 uses a flashback to the player-character’s childhood as a means of introducing controls sequentially. The game opens with a set piece aboard a speedboat to familiarise players with basic directional movement and camera controls, before jumping backwards in time to the protagonist’s childhood in order to teach players more complex movement controls such as crouch, run, climb and jump. The avatar’s youth in this segment provides a narrative explanation for the way the game deliberately delays the introduction of combat and shooting interactions until after players have mastered the controls for traversing the gameworld.

Other Action games that featured flashbacks to childhood include Batman: Arkham Asylum, Wolfenstein II: The New Colossus and Call of Duty: Modern Warfare 2019 -- in fact, 13% of Action games in this corpus featured achronological flashbacks to the player-character’s childhood. These flashbacks tended to centre on an incident of childhood trauma as a means of humanising the powerful, fearless, heroic avatars by depicting them at their most vulnerable, as well as working to justify the violent, vengeful behaviour of the adult avatars. Karen Lury writes that children are understood to be “perfect victims, since they are blameless” (2010, p.105), which permits overly-simplified “perceptions of right and wrong, despite the moral complexity inherent in any [social issue]” (p.107). Anneke Meyer expands on this, arguing “the discourse of innocence makes crimes against children ‘worse’ than crimes against adults by constructing the child as innately weak, vulnerable and defenceless. In this context, adult crimes against children become unequal and unfair, cowardly and ‘bad’” (2007, p.96). By depicting their overpowered avatars as vulnerable children, these games make their consequent annihilation of the forces responsible for their trauma seem not only morally “right” but also ludically “fair.” Players are permitted to enjoy the catharsis of unleashing excessive violence on their in-game foes because it is framed as just retribution for the unforgivable, indefensible crime of harming a child. In other words, depicting in-game enemies as child-killers sublimates some of the ludonarrative dissonance inherent in the idea of a hero-avatar who terrorises hundreds of henchmen during the course of the game. This process also works in the opposite direction. Rather than the figure of the child functioning to produce a simple moral binary of “good” and “bad,” flashbacks to childhood can serve to partially exonerate a game’s villain. Both Marvel’s Spiderman and Heavy Rain contain flashbacks to the antagonists’ childhoods, revealing their murderous, pathological personalities to be rooted in childhood traumas. Players are allowed to feel sympathy for the adult antagonists by way of their innately innocent child-selves, adding a degree of complexity to players’ perception of the villains and, consequently, to their own actions as “heroes.”

In other titles, a child-avatar functioned as an icon expressing a game’s non-violence. Essentially, the child-avatar itself was used to signify the game’s appropriateness for young players -- much like a visualisation of an age-rating certificate. This is the case for Splatoon, which features customisable childly-squidlings as avatars. The game belongs to the Shooter genre -- and the presence of guns usually rules out the possibility of a child-avatar -- however, in this instance, the virtual weapons are akin to water pistols or paintball guns, and the child-avatars underscore the fact that the game’s central mechanic should be understood as non-serious, non-fatal, sporting hijinks. Childly-avatars also seem to be used by games in this sample to encourage a pro-social playstyle. The cheery, neotonised protagonists of Scribblenauts make the game’s central mechanics of helping one’s community through acts of creative, playful service more robust and intuitive. Equally, the appealing, childlike toy-avatar in Unravel functions to generate an emotional attachment between the player and the character, eliciting a sentimental and nostalgic reaction to the gameworld and a sense of childly wonder at the majesty of nature. The “cuteness” of child-avatars, however, does not always (or does not only) give rise to a caring response: it can also permit licit acts of aggression towards avatars. The pliant naivety of the Pikmin in Pikmin 3, for example, or the tactile toyness of Sackboy in Little Big Planet make these player-characters undeniably “cute,” but also script a form of “rough play,” in which these avatars are endlessly dismembered, reassembled and replaced.

Of the playable child characters that were assigned a gender, 25 were male and 6 were female. The lack of playable non-male child characters reinforces the idea that while boys drive action, girls exist only as helpmeets, caretakers, damsels in distress, trophies, sidekicks and cheerleaders (Sarkeesian 2013). It also positions male children as the default gender, and non-male children as fundamentally “other.” Ageing down the “damsel in distress” does not fully uncouple her debility from her gender, and an age disparity between male primary characters and female secondary characters neither masks nor erases the gendered nature of the power imbalance.

In terms of race, 15 child player-characters had “no race” by virtue of being non-racialised animals, non-racialised fantasy creatures, or non-racialised automatons (e.g., the hero of Ori and the Blind Forest or the player-character in Astrobot). Similarly, 9 player-characters were categorised as “Mukokuseki” and 1 player-character had a customisable race. The “de-racing” of child characters parallels a trend in Western children’s literature, where studies have found that although animal and other non-human characters appear less frequently as protagonists than white child characters, they significantly outnumber child characters of all other races (Serroukh 2018, 2020). Presumably, removing race as a fixed identity marker is intended to render the avatar a cipher onto which racially diverse players can project themselves. However, considering the fact that 18 child player-characters were White, and only 3 child player-characters were non-white, this elision of children’s race works as another compounding layer of symbolic annihilation. Henry Jenkins comments on the racialised nature of childhood when he writes, “[i]n our culture, the most persistent image of the innocent child is that of a white, blond-haired, blue-eyed boy…and the markers of middle-classness, whiteness, and masculinity are read as standing for all children” (1998:13). The concept of a white, universal “Everychild” figure that can “stand for all children” disturbingly implies that there is something inherently alienating about non-white children that would impede player-avatar identification. Contemporary video games seem to affirm that whiteness is a property of childhood, and this has serious consequences for non-white children who may be denied the protections afforded to other children on the basis of their race. Lury writes that the coalescence of the descriptors “white,” “little” and “girl” in films produce and legitimate a range of sentimental responses, whereas “the little black girl, it seems, has been lost and nobody is looking for her” (2010, p. 54). When writing about white readers' inability to “see” the child character Rue from The Hunger Games series as Black, Toliver notes that the perceived incompatibility of childliness and Blackness “creates a form of age compression, in which the young girls are likened more to adults than to children, rendering Black girlhood interchangeable with Black womanhood” (2018, p.6). This results in the societal belief that Black girls “require less nurturing, protection, support, and comfort than White girls” (p.7). If a child character’s purpose is to elicit a protective or affectionate response from the (presumed white) player, then Blackness reduces the stereotype’s functionality. The friction between a character’s identity as a “child” and its racialised identity suggests the two identities are perceived by some as mutually exclusive: online debates in fan spaces about the racial identity of the child character Clementine from The Walking Dead series, for example, repeatedly query and re-write her Blackness. [2]

Many of the child-avatars whose race was categorised as Mukokuseki appeared in Role Playing Games (RPGs). Hutchinson has argued that, “[t]he hero of [RPGs] is the shōnen, the youth, on the verge of becoming an adult but still young enough to jump around, play and get into trouble” (2019, p.107). This suggests that the concept of childhood as a temporal event on dynamic journey towards adult stasis is used in RPGs to express the one of the key mechanics of this genre: character upgrading and customisation. The more time invested in “growing up,” the more powerful the avatar will be, even if the character does not age within the game narrative. Since the child is commonly defined “in terms of its vacuity and lack of form” (Balanzategui 2018, p.9), child-avatars are ideal signifiers of malleability and potential growth, and thus it seems game designers have intuitively reimagined Locke’s tabula rasa as an unlockable skill tree. Removing racial identifiers to create a “culturally-odorless” character is perhaps an expression of the “vacuity” of the childly avatar.

Non-Playable Child characters

One could argue that certain game spaces are “adults only” in order to shield virtual children from scenarios involving sex, violence, and other taboos activities. Since virtual children can only ever be in virtual danger -- and their suffering is only ever simulated -- concern for digital children is perhaps best understood as a need to defend a specific ideological construction of the child. Meyer notes the tendency to conflate concern for children with concern for “the child” when she comments, “adult indignation [about child abuse] is not only motivated by the harm inflicted on children but also by the infringement of adult ideals of childhood” (2007, p.102). Chris Jenks adds that the treatment of “the child” is a moral barometer for a society: he writes, “[W]hatever the general condition of childhood in society (treated violently, exploited, pornographized) it may be regarded as an index of the state of the wider social relation, the moral bond in society” (2005, p.38). Media that present the mistreatment of children without emphatically condemning it, threaten to undermine the moral authority of the society that produced those media (Reay, 2020a). Since what game developers make “possible” in gameworlds is often conflated with what they deem “permissible,” creators may choose to exclude child non-player characters (NPCs) from their digital worlds to avoid being accused of facilitating virtual child abuse. This accounts for the absence of child NPCs in violent, open world games such as those in the Grand Theft Auto (GTA) series, which encourage players to test the limits both of what is possible and what is permissible. Sjöblom reports that, “children and school busses were included in the beta version of GTA III but were scrapped before the final release of the game” (2015, p.72), demonstrating that the systematic exclusion of children from these otherwise detailed, comprehensive gameworlds was a conscious design decision. Considering the GTA series is infamous for being flagrantly immoral and offensive, it is notable that killing digital children is a taboo it refuses to embrace. Nonetheless, one could argue that excising children entirely from GTA’s simulation of society is -- in itself -- a form of violence: the developers eliminate children before players can.

When violent, open world games contain child characters, they tend to limit the extent to which players can interact with the virtual children, pre-emptively policing player behaviour towards this social group. Red Dead Redemption 2, for example, confines the single significant child character to the camp -- a designated safe space in which players are unable to use their weapons and violent interactions are prohibited. The only other game location in which players are subject to similar restrictions is the Wapiti reservation. In this way, the game presents players with an implicit moral code that equates violence against children with violence against Indigenous Americans. Considering the fact that the game features an “Honor” meter that provides quantitative feedback condoning or condemning players’ moral decisions, it is significant that the choice to commit violence against Indigenous American characters and child characters is withheld from the player. This seems to vindicate Holland’s observation that “[a]s the parameters of childhood are marked out and held firm -- children are dependent, vulnerable, in need of instruction and protection -- many other groups have been rhetorically bestowed with childish characteristics: women, people from ethnic minorities, and the whole of the previously colonised world have come to stand in childish relation to the exercise of power” (2006, p.148). Preventing the player from simulating genocide against virtual Indigenous Americans reflects colonisers’ general unwillingness to acknowledge and take ownership of historical and ongoing atrocities committed against Native populations globally, but it also places the Indigenous characters in childish relation both to the white, ethnically European player-character, Arthur, and to players themselves. As a result, both of these social communities seem out of place within this gameworld -- they sit “at odds with the dominating logic” of the virtual space (Sjöblom, 2015 p.79). Children and Indigenous Americans are rendered less like non-player characters and more like environmental objects, with their encoded immortality making them fundamentally “Other.”

The special moral status accorded to children gives their deaths additional weight, and so placing virtual children in mortal peril intensifies player’s investment in ludic decisions and narrative outcomes. It also encourages the player to balance strategic thinking with moral and emotional cognition. The Bioshock series, the Banner Saga series, TheWalking Dead series, and in dystopian city-sims like Frostpunk or This War of Mine use child characters to lend moral weight to in-game interactions. Decisions regarding basic resource management in these games, for example, gain complexity when players made to feel that the optimal ludic choice -- the choice that is mostly likely to bring about a win condition in the game -- is not a morally satisfying choice because of its effect on in-game children. In representing the characters that are impacted by player decisions as children, these games approve emotional, “irrational,” sub-optimal playstyles. Encouraging an emotional investment in dependent characters that runs counter to the player’s emotional investment in winning the game can make even the smallest ludic interactions feel weighty and challenging. Although the ludic consequences of killing or sparing the Little Sisters in Bioshock, for example, are not significantly different, this repeated choice works to distil the game’s central message about “ethical egoism,” agency and obedience.

Although players are generally prevented or strongly discouraged from harming child NPCs, violence against virtual children perpetrated by adult NPCs was not an uncommon occurrence in the video games sampled here. In fact, 21 child NPCs in this corpus are murder victims, whose deaths are non-optional plot points in the games’ narratives. Susan Tan writes, “[t]he vision of the dead child is one of the most horrific images in cultural imaginations. It is also one of the most pervasive” (2013, p.54). She traces the literary trope of the “sacrificial child” from Abraham’s near murder of Isaac in Judeo-Christian texts through to the dystopian pageant of child-on-child slaughter in The Hunger Games series. The same patterns can be found in video games. Child characters such as Petruccio in Assassin’s Creed II, Khemu in Assassin’s Creed: Origins and Phoibe in Assassin’s Creed: Odyssey are brutally murdered by in-game antagonists as a means of emphasising the difference between the cruel, inexcusable violence committed by the dastardly Templars and the justifiable, retributive, necessary violence committed by the heroic Assassins. Children’s deaths are also used to underline the brutality and hostility of an inhospitable, dystopian gameworld. The deaths of Riley, Sam and Sarah in The Last of Us and The Last of Us: Left Behind not only signal the dark, “edgy” and mature nature of the series, but also drive home the fact that the terrifying but beautiful apocalyptic gameworld -- which is lush and verdant with flora -- is inimical to human life.

Supporting child characters in this corpus frequently served to humanise the aggressive, surly, burly, hyper-masculine adult heroes. The violence of male protagonists such as Joel in The Last of Us or Kratos in God of War: Chains of Olympus and God of War IV is justified when it is in the service of protecting or avenging a child character, and both morally-questionable men are offered a path to redemption through their relationship with their (surrogate) children. This dynamic has been explored in a series of academic and journalistic articles on “the Daddening” of video games; however, less critical attention has been paid to how the presence of a supporting child character functions to humanise the player. The relationship that forms between the player and the secondary child character demarcates the player’s identity through a parasocial bond. The player is conditioned to behave in a more socially responsible and more emotionally vulnerable way as a result of characters like Ellie, Atreus, and Alice (from Detroit: Become Human), who communicate through dialogue, body language and impressively realistic motion capture facial expressions the extent to which the player’s interactions affect their wellbeing and their worldviews. In fact, Detroit: Become Human is explicitly about the humanising potential of adult/child relationships, and repeatedly requires the player to affirm their commitment to fulfilling the role of Alice’s guardian throughout the course of its interactive narrative (Reay 2020b).

The notion of the child as a moral barometer is made explicit in The Last of Us, wherein a non-player character called Henry is identified as an ally because he is accompanied by his little brother, Sam -- a boy the same age as Ellie. The Hunters, on the other hand, view children as burdens who cannot contribute to the group’s survival, and therefore choose to kill children rather than care for them. This makes the group irredeemable and justifies their total elimination by Joel and Ellie. A parallel exists in The Walking Dead series, wherein the Crawford community -- which kills its weakest members, including all those under the age of fourteen -- is shown to have collapsed internally due to its unnatural, evil ideology. The demise of all its members is presented as both inevitable and appropriate, considering the player-character’s central purpose is to protect the young Clementine. The flexibility of “the child” as a rhetorical sign is such that it can be used to emphasise the insupportable, indefensible, reprehensible nature of some murders, whilst highlighting the justice and appropriateness of other murders.

Limitations, Conclusions, and Future Directions

There are a number of limitations that should qualify any generalised conclusions drawn from this dataset. Firstly, games were selected for inclusion based on sales figures and industry awards, which meant that games designed predominantly for child-playerships rarely appear, as these games tend to sell in comparatively low numbers and rarely receive critical acclaim. Although there was a fairly even spread of games across each age-rating category (110 games in 3+, 87 games in 7+, 109 games in 12+, 90 games in 16+, and 108 games in 18+), this does not necessarily correspond to the age of the target audiences. The conceptual category “children’s video games” is unstable and, arguably, unhelpful (Reay 2018); however, it could have been used to generate a data set with significantly different representations of children and childhood. Furthermore, deliberately choosing games that meet conventional standards of success has resulted in my data set mainly consisting of AAA games, which further marginalises the boundary-pushing texts that exist at the periphery of the hegemonic games industry.

Secondly, and more significantly, this data set was coded by one person, which means decisions about liminal cases were not corroborated by another researcher. Although I took into consideration crowd-sourced opinions about specific characters, interpretations of this data set would have been more robust had there been multiple researchers participating in the coding process. I hope to publish my dataset as an open access, community-edited resource in the future so that others who wish to research this topic may act as intercoders and thereby strengthen the reliability of the content analysis.

With these limitations in mind, this article concludes that child characters were wholly absent from 331 of the 506 games surveyed here, meaning that less than 35% of games contained child characters. When child characters did appear in these virtual spaces, they were often positioned in secondary, supporting roles that relied on stereotypes about children in order to visually communicate game rules and efficiently elicit a specific set of affective responses in players. The roles assigned to child characters in video games are perhaps best expressed by connecting and consolidating the dominant “functions” of individual characters. Child characters had a range of functions in the video games sampled here, including the following:

• They served to explain restrictions to an avatar’s set of abilities (specifically combat abilities) and justify the limits of possible interactions within the gameworld
• They were used to create “seamless” tutorials
• They elicited feelings of protectiveness and intensified players’ emotional investment in ludic outcomes
• Their mistreatment exonerated extremely violent behaviour enacted by the player / avatar
• They were used as conventional signs to express vulnerability and their deaths were used to underscore the brutality of the gameworld
• They acted as iconic symbols for a game’s non-violence, and promoted pro-social playstyles
• They embodied the concept of the “blank slate,” thereby expressing character customisation and upgrading mechanics
• They were a stand-in for gendered stereotypes, specifically the “damsel-in-distress” trope and the “women in refrigerators” trope (Sarkeesian 2013)
• Their childly qualities were juxtaposed with grotesque, horrific qualities in order in elicit a disturbing, unheimlich player experience

I hope that this preliminary -- and predominantly descriptive -- overview will tempt other scholars into the interdisciplinary space between childhood studies and game studies. Research conducted at this intersection is as relevant for game designers looking to exploit ideologies about childhood in order to elicit intended player experiences as it is for childhood studies scholars interested in how particular player experiences shore up ideologies about “the child.” Furthermore, in-depth research into representations of child characters will not only complement and nuance the numerous ethnographic, anthropological and developmental studies of child players, but will also mark out an arena in which video games can contest and defend their relationship to the ideological figure of “the child.” After decades of video games being simultaneously infantilised as a children’s pastime and vilified as a threat to the sanctity of childhood (Reay 2018), embracing the digital kids that populate virtual worlds is an act of running “towards the trouble rather than away from it” (2020, p.3), as Amanda Phillips would put it. Since the figure of the child is a “lightning rod” for political and cultural controversy, making space for child characters in academic discussions of video games readies the field to engage in discussions surrounding the impact of video games on child players and on ideological constructions of “the child” in contemporary society.

Endnotes

⁠[1] It is also worth noting that content analyses that code for disability are equally rare, despite the fact video games researchers using other methodologies have used critical disabilities studies as a lens for documenting representation.

[2] One FANDOM user wrote, “Why is Clementine listed as African-American on her wiki page? I don’t remember any mention of it in the game and she looks more Asian or Caucasian to me.” (December 1, 2017). Another commented on the same thread, “She’s Blasian.” (August 12, 2018), prompting a heated discussion as to whether two Black parents could biologically produce an Asian child. Similar conversations were taking place years earlier. In 2014, for example, a commenter opined, “I think TellTale intended for her to be aracial (sic), like how the Coen brothers wanted Anton Chigurh in “No Country for Old Men” to be unidentifiable in terms of nationality. The fact people are claiming her race is unbacked and unproven” (July 15, 2014).

References

Balanzategui, J. (2018). The Uncanny Child in Transnational Cinema: Ghosts of Futurity at the Turn of the Twenty-First Century. Amsterdam: Amsterdam University Press.

Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavioral change. Psychological Review, 84, 191-215.

Behm-Morawitz, E. (2017). Examining the intersection of race and gender in video game advertising, Journal of Marketing Communications 23:3, 220-239.

Brand, J. E., Knight, S., & Majewski, J. (2003). The diverse worlds of computer games: A content analysis of spaces, populations, styles and narratives. Paper presented at Digital Games Research Conference 2003, The Netherlands.

Brock, A. (2011) “When Keeping it Real Goes Wrong”: Resident Evil 5, Racial Representation and Gamers, Games and Culture 6:5, 429-453.

Carr, D. (2019). Methodology, Representation, and Games. Games and Culture, 14:7-8, 707-723.

Cash, J. (2018). ‘We must be better’: Hegemonic Masculinity and Dadification in God of War, First Person Scholar, http://www.firstpersonscholar.com/we-must-be-better/

Consalvo, M. (2008). Crunched by passion: Women game developers and workplace challenges. Cambridge, MA: MIT Press.

Dahlen, S. (2019). Picture This: Diversity in Children’s Books 2018 Infographic, retrieved from https://readingspark.wordpress.com/2019/06/19/picture-this-diversity-in-childrens-books-2018-infographic/

Dietz, T. (1998). An Examination of Violence and Gender Role Portrayals in Video Games: Implications for Gender Socialization and Aggressive Behavior. Sex Roles, 38:5/6, 425-442.

Dill, K.E., & Thill, K.P. (2007). Video Game Characters and the Socialization of Gender

Roles: Young People’s Perceptions Mirror Sexist Media Depictions. Sex Roles, 57, 851-864.

Downs, E.P. & Smith, S. (2005). Keeping Abreast of Hypersexuality: A Videogame Character Content Analysis, paper presented at the International Communication Association, New York.

Everett, A. (2009). Digital Diaspora: A Race for Cyberspace, New York: New York Press.

Gerbner, G. (1997). Gender and age in prime-time television. In S. Kirschner & D. A. Kirschner (Eds.), Perspectives on psychology and the media, 69-94. Washington, DC: American Psychological Association.

Goetz, C. (2018). Coin of Another Realm, Game Studies,18:3 http://gamestudies.org/1803/articles/goetz

Gray, K.L., (2015). Cultural Production, Knowledge Validation, and Women’s Digital Resilience: Examining the Distinctness of Virtual Feminisms, in Fan Girls and the Media: Consuming Culture, ed. A. Trier-Bieniek. Lanham, MD: Rowman & Little-field. 

Haines, C.B., (2019) Race, Gender, and Sexuality Representation in Contemporary Triple-A Video Game Narratives, Thesis submitted to the faculty of the Virginia Polytechnic Institute and State University.

Hall, S. (1996). Stuart Hall: Critical Dialogues in Cultural Studies, ed. David Morley and Juan-Hsing Chen. London: Routledge.

Haninger, K. & Thompson. K.M. (2001). Violence in E-rated video games. JAMA, 286, 591-598.

---- (2004a). Contenting and Ratings of Teen-Rated Video Games, JAMA, 291, 856-865.

Haninger, K., Ryan, M.S., Thompson, K.M. (2004b). Violence in teen-rated video games, MedGenMed, 11, 6-10.

Heintz-Knowles, K., J. Henderson, C. Glaubke, P. Miller, M.A. Parker and E. Espejo (2001). Fair Play? Violence, Gender and Race in Videogames. Oakland, CA: Children Now.

Hutchinson, Rachael. (2019). Japanese Culture Through Videogames, Routledge: London & New York.

---- (2017). Representing Race and Disability: Grand Theft Auto San Andreas as a Whole Text in Gaming Representation: Race, Gender, and Sexuality in Video Games, Malkowski, J. & Russworm, T. M. (Eds.) Indiana: Indiana University Press.

Lawlor, S. (20180 “Your Daughter Is in Another Castle”: Essential Paternal Masculinity in Video Games. The Velvet Light Trap, 81, 29-42, doi.org/10.7560/VLT8104.

Leonard, D.J., (2003). “Live in Your World, Play in Ours”: Race, Video Games, and Consuming the Other. Studies in Media & Information Literacy Education, 3:4, 1-9.

Lesnik-Oberstein, K. (2011). Children in Culture, Revisited: Further Approaches to Childhood. Basingstoke, Palgrave Macmillan.

Lury, Karen. (2010). The Child in Film: Tears, Fears, and Fairy Tales. London: I.B. Tauris & Co.

Lynch, T., Tompkins, J.E., van Driel, I., Fritz, N. (2016). Sexy, Strong, and Secondary: A Content Analysis of Female Characters in Video Games across 31 Years, Journal of Communication 66, 564-584.

Malkowski, J. & Russworm, T.M., (Eds.) (2017). Gaming Representation: Race, Gender, and Sexuality in Video Games, Indiana: Indiana University Press.

Massanari, A. (2017). #Gamergate and The Fappening: How Reddit’s algorithm, governance, and culture support toxic technocultures, New Media and Society.

Mastro, D., E. Behm-Morawitz and M. Ortiz (2007). The Cultivation of Social Perceptions of Latinos: A Mental Models Approach, Media Psychology, 9:2, 347-65.

Meyer, Anneke. (2007). The Moral Rhetoric of Childhood. Childhood, 14, 85-104.

Nakamura, L. (2008). Digitizing Race: Visual Cultures of the Internet, Minnesota: University of Minnesota Press.

Parker, Felan, and Jessica Aldred, (eds.) (2018). Beyond the Sea: Navigating Bioshock. Montreal, McGill-Queen’s University Press.

Passmore, C., Yates, R., Birk, M.V., & Mandryk, R. (2017). Racial Diversity in Indie Games: Patterns, Challenges, and Opportunities, Extended Abstracts Publication of the Annual Symposium on Computer-Human Interaction in Play (CHI PLAY ’17 Extended Abstracts). Association for Computing Machinery, New York, NY, USA, 137-151.

Perreault, M., Perreault, G., Jenkins, J., Morrison, A. (2018). Depictions of Female Protagonists in Digital Games: A Narrative Analysis of 2013 DICE Award-Winning Digital Games, Games and Culture, 13:8, 843-860.

Phillips, A. (2020). Gamer Trouble : Feminist Confrontations in Digital Culture. New York: New York University Press.

Reay, E. (2018). Appraising the Poetic Power of Children’s Video Ga- mes. International Research in Children’s Literature, 11:1, 17-32, doi.org/10.3366/ircl.2018.0251.

--- (2019). Kideogames: Reimagining the Fringe of Literary Studies as the Forefront. Games and Culture, doi.org/10.1177/1555412019841476.

---- (2020a). Secrets, Stealth, and Survival: The Silent Child in Video Games, Barnboken, 43:1.

---- (2020b). Who Thinks Beating a Child is Entertainment: Ideological constructions of the figure of the child in Detroit: Become Human, DiGRA Conference Proceedings 2020.

Sarkeesian, A. (2013). Tropes vs. Women in Video Games, Feminist Frequency, retrieved from https://www.youtube.com/channel/UC7Edgk9RxP7Fm7vjQ1d-cDA

Serroukh, F. (2020). Reflecting Realities: Survey of Ethnic Representation within UK Children’s Literature 2019, CLPE.

Serroukh, F. (2018). Reflecting Realities: Survey of Ethnic Representation within UK Children’s Literature 2017, CLPE.

Sjöblom, B. (2015). Killing Digital Children: Design, Discourse, and Player Agency. The Dark Side of Game Play: Controversial Issues in Playful Environments, edited by Torill Elvira Mortensen et al., New York, Routledge, 67-81.

Stone, K. (2018). Time and Reparative Game Design: Queerness, Disability, and Affect, Game Studies, 18:3.

Sharpe, C. (2016). In the Wake: On Blackness and Being, Durham, NC: Duke University Press.

Thomas, E. E. (2019). The Dark Fantastic: Race and the imagination from Harry Potter to the Hunger Games, New York: New York University Press.

Toliver, S. R. (2018). Alterity and Innocence: The Hunger Games, Rue, and Black Girl Adultification, Journal of Children’s Literature, 44:2.

Waddell, T.F., Ivory, J., Conde, R., Long, C., & McDonnell, R. (2014). White Man’s Virtual World A Systematic Content Analysis of Gender and Race in Massively Multiplayer Online Games, Journal of Virtual Worlds Research, 7:2.

Williams, D., Martins, N.; Consalvo, M., Ivory, J.D., (2009). The Virtual Census: representations of gender, race and age in video games, New Media and Society, 11:5, 815-834.

Wilberg, H. (2011). What’s in a Game? Race, Gender, and LGB Representation in Video Games, A Thesis Submitted to the Faculty of the College of Arts and Sciences in Candidacy for the Degree of Master of Arts in Women’s and Gender Studies, Roosevelt University.

Wohn, D.V., (2011). Gender and Race Representation in Casual Games, Sex Roles, 65, 198-207.