1 Introduction

The Gamer’s Dilemma refers to the philosophical challenge of justifying the intuitive difference people seem to see between the moral permissibility of enacting virtual murder and the moral impermissibility of enacting virtual child molestation in video games (Luck, 2009). This Dilemma has generated a large philosophical literature (e.g. (Ali, 2015; Davnall, 2021; Kjeldgaard-Christiansen, 2020; Luck & Ellerby, 2013; Montefiore & Formosa, 2022; Ramirez, 2020)) and a recent empirical exploration (Formosa et al., 2023). Recently, Luck (2022) has argued that the Gamer’s Dilemma is actually an instance of a more general “paradox”, which he calls the “paradox of treating wrongdoing lightly”. For example, while watching a light-hearted or comedic movie or TV show that includes murder may seem morally permissible, watching a similarly light-hearted or comedic movie or TV show that features child molestation may seem morally impermissible. Recasting the Gamer’s Dilemma in this way could justify an important shift in the literature in two respects. First, it points both to the irrelevance of the interactive element in video games for the Dilemma; that is, the fact that you enact the virtual action in question through pushing buttons (and so on) rather than merely passively consuming it as with a film or TV show. This could undermine attempts to solve the dilemma by focusing on these interactive elements (see Ali, 2015; Ramirez, 2020). Secondly, it suggests that further engagement with philosophical discussions of other forms of media are essential for properly contextualising the Gamer’s Dilemma.

In this paper, we explore these issues as follows. First, we set up the debate by exploring the relevant background literature. We then look at whether the Gamer’s Dilemma really is an instance of a more general dilemma or paradox, and we argue (with some qualifications) that it is. Next, we critically discuss Luck’s (2022) proposed graveness resolution. This is an attempt to distinguish between the moral permissibility of light, playful or comedic treatment in various forms of media, including video games, of fair-game virtual wrongdoings, such as virtual murder, and the impermissibility of the similar treatment of off-limits virtual wrongdoings, such as virtual child molestation. Finally, we consider various upshots of our argument. In particular, we argue for the importance both of looking to the treatments of related problems in other literatures and the limitations of the implicit moralism in the Gamer’s Dilemma debate that engagement with these other literatures helps to expose.

2 Background Literature

The Gamer’s Dilemma (Luck, 2009) presents the problem that while some wrongdoings enacted in video games (which we refer to as ‘virtual wrongdoings’Footnote 1), such as virtual murder, seem to be “fair-game” (or morally permissible), and others, such as virtual child molestation, seem to be “off-limits” (or morally impermissible), a morally relevant distinction is unavailable to distinguish the former from the latter in terms of moral permissibility. As a result, one is left having to either accept that off-limits virtual wrongdoings, such as virtual child molestation, are counterintuitively fair-game, or that fair-game virtual wrongdoings, such as virtual murder, are counterintuitively off-limits.

The most popular approach to addressing the Gamer’s Dilemma is to resolve it by identifying relevant moral differences between off-limits virtual wrongdoings and fair-game virtual wrongdoings that would justify why the former is off-limits and the latter is not. With virtual child molestation being the paradigm case in the literature, some noteworthy attempts to distinguish it as off-limits include: virtual child molestation amounting to child pornography, which is itself wrong (Bartel, 2012); contributing to the harm experienced by women by its endorsing incorrigible social norms which eroticise inequality (Levy, 2002; Patridge, 2011, 2013); or virtue ethical approaches which consider virtual child molestation sufficiently more harmful to a gamer’s moral character than virtual murder (Bartel, 2020; McCormick, 2001). Some resolving attempts have promise and can (arguably) resolve the Gamer’s Dilemma for a limited set of cases. For example, the virtual rape in the infamous Atari game “Custers Revenge” may be morally objectionable on the grounds that it unambiguously advances harmful and existing real-world social norms via its eroticising the inequality of women (as well as for other reasons) (Levy, 2002; Patridge, 2011, 2013), and this kind of moral objection may not be applicable to a virtual murder in the same context, as social norms that promote murder do not exist, at least to the same extent. Therefore, a morally relevant distinction is available, and the Gamer’s Dilemma is resolved for these cases. However, no resolution has generated consensus in the literature as resolving the dilemma in all cases (Montefiore & Formosa, 2022).

The alternative strategy in the literature has been to dissolve the Gamer’s Dilemma by showing that because the dilemma’s intuitions regarding virtual murder and virtual child molestation aren’t really held, or at least aren’t morally justified in being held in all cases, then it is not the type of virtual wrongdoing that is morally relevant but rather the video game features that contextualise it. For example, the type of game being played (Ali, 2015), the context realism and perspective fidelity of the game (Ramirez, 2020), or some other non-moral feature, such as norms regarding what is in poor taste (see Young, 2017) or unconventional in video games (Montefiore & Formosa, 2022), may be the relevant factor(s) in determining the moral permissibility of virtual wrongdoings. For example, following Ramirez (2020), the conflicting intuitions of the Gamer’s Dilemma do not plausibly get off the ground in highly realistic video game settings which produce “virtually real experiences”,Footnote 2 as both murder and child molestation may be justifiably intuited as morally unacceptable in such contexts. A recent empirical study lends some support to these views as it shows that, while there is a significant difference between how people respond to the moral acceptability of virtual murder and virtual molestation, there is also a significant difference between how people respond to the moral acceptability of high and low realism virtual actions (Formosa et al., 2023). However, as with resolving strategies, no consensus has emerged about any dissolving approach being considered to have solved the dilemma in all cases, but rather only for a limited set of cases (see Luck, 2018).

To account for dissolving and resolving approaches, the literature has moved toward focusing on a narrowed version of the Gamer’s Dilemma limited to cases that, when contextual features are held constant, such as genre, degree of contextual realism, perspective fidelity and so on, may justifiably trigger the competing intuitions that ground the dilemma. Once narrowed, the Gamer’s Dilemma appears to remain, albeit limited in scope (see Luck, 2018; Luck, 2022; Montefiore & Formosa, 2022; Nader, 2020). Luck’s (2022) recent claim is that while narrowing the Gamer’s Dilemma is a needed move, to maximise explanatory strength a solution to the narrowed Gamer’s Dilemma ought to be expanded to account for not only all instances of the narrowed Gamer’s Dilemma, but also other moral dilemmas that share the same formal structure and which are all captured under the umbrella of “the paradox of treating wrongdoing lightly” (Luck, 2022). Luck thereby claims both that the Gamer’s Dilemma is an instance of a broader set of moral dilemmas that involve fictional, rather than merely virtual, wrongdoings, and that a resolution to this broader paradox will also resolve the Gamer’s Dilemma.Footnote 3 We turn to these claims in the next section.

3 Is the Gamer’s Dilemma an Instance of the Paradox of Treating Wrongdoing Lightly?

3.1 Luck’s Paradox of Treating Wrongdoings Lightly

The broader paradox Luck identifies is based on the intuitive moral impermissibility of engaging with a representation of a wrongdoing in various modes of entertainment that treat those fictional wrongdoings “too lightly” (Luck, 2022, p. 12). Engaging with a form of media too lightly includes either: 1) “playful engagement” with a wrongdoing, such as playfully engaging with a video game about a wrongdoing, as opposed to being a professional e-sport gamer or a game critic who may have a more “serious” attitude toward that engagement; or 2) the treatment of a wrongdoing being in “poor taste”, such as a crude joke that trivialises a wrongdoing such as the Holocaust. Note that the first condition refers to the players’ mode of engagement with the game, whereas the second refers to the game’s treatment of a wrongdoing. Presumably these two are linked, since the way that players engage with a game is partly dependent on how the game presents itself, and the latter is likely to encourage a certain mode of engagement with it by players. In any case, Luck claims that there seems to be an intuitive moral difference such that there are a range of cases in which treating fictional murder lightly is permissible (a “fair-game” wrongdoing) while treating fictional child molestation lightly is impermissible (an “off-limits” wrongdoing), and there does not seem to be an obvious property that fictional murder lacks and child molestation possesses that would morally justify this intuitive difference. Since the Gamer’s Dilemma has the same structure as this paradox, and this paradox is broader than the Dilemma as it covers a range of media (including video games), it seems to follow that the Gamer’s Dilemma is a specific instance of this more general “paradox”.

Is this argument plausible? To explore this point, we consider two objections to this argument. The first objection is that the interactive element of video games is essential to the Gamer’s Dilemma and this level of interactivity is absent in other forms of media, and thus the Gamer’s Dilemma cannot be subsumed under this broader paradox since it does not take interactivity to be an essential feature of cases that fall under it. The second objection is that there are instances of the Gamer’s Dilemma that do not involve treating a fictional wrongdoing “lightly”, and thus (once again) the Gamer’s Dilemma cannot be subsumed under a broader paradox that makes treating a wrongdoing lightly an essential feature of cases that fall under it. We shall consider each objection in turn.

3.2 Is Interactivity Essential to the Gamer’s Dilemma?

A common way of differentiating video games from other forms of media is in terms of the interactivity of the medium, which leads to greater levels of immersion and agency (see Brey, 1999, 2003; Ramirez, 2020; Cogburn & Silcox, 2014; Murray, 2017; Ali, 2015; Davnall, 2021; Polito & Hitchens, 2021). Interactivity has consequently played its part in some attempts to deal with the Gamer’s Dilemma (see Ali, 2015; Ramirez, 2020; Tillson, 2018). For example, Ali (2015) argues that virtual wrongdoings in video games, be it murder or child molestation, which are simulated by the player through interactivity, rather than merely witnessed as part of a game’s unfolding narrative (i.e. lacks interactivity), aligns a game’s in-game context with the context the gamer is desiring to realise, and this feature of virtual wrongdoings is what, according to Ali, makes them morally salient.

While interactivity is certainly an important element of video games, is it essential to the Gamer’s Dilemma? That is, is the fact that one does not merely observe a depiction of a virtual wrongdoing but in some way causes the virtual wrongdoing through pushing buttons (or moving a mouse or joystick and so on) essential to grounding the intuitions that form the basis of the Gamer’s Dilemma? For Luck’s move to recast the Gamer’s Dilemma as an instance of a broader “paradox” that goes beyond interactive media to work, it must be the case that interactivity is not essential to the Dilemma.

Interactivity in video games raises several contentious issues, including: what the nature of video game interactivity amounts to given the broad diversity of virtual “environments” and virtual entities in those environments (Davnall, 2021); how interactive video games can really be said to be (Cogburn & Silcox, 2014); and whether a clear binary distinction can be drawn between video games and other forms of media along the lines of interactivity given the cognitive processes involved in one’s engagement with ‘non-interactive’ media are interactive (see Walton, 1990; Davnall, 2021; Rorty, 2006). Nonetheless, as a recent empirical study (Polito & Hitchens, 2021) has shown, there is a clear psychological difference to be noted between those media that have direct controls that give the perception of influence, such as video games (labelled as “interactive media”), and those that do not, such as film, television, and literature (labelled as “non-interactive media”). In this study, some of the noted psychological differences for participants using interactive, rather than non-interactive, media included “reduced agentive involuntariness, but increased flow, presence, character involvement, and time perception” (Polito & Hitchens, 2021).

What is unclear, however, is if these immersive features of interactivity can underwrite the moral difference people seem to intuit between virtual murder and virtual child molestation in video games. The property of flow (or automaticity) and distorted time perception (e.g. losing track of time) seem less morally relevant here (see Rau et al., 2006), whereas increased presence (i.e. conceiving of oneself as actually present in the virtual world) (see Sanchez-Vives & Slater, 2005; Sheridan, 1992) and character identification (i.e. identifying with the character in the game) (see Murray, 2017) do seem morally relevant. When a player controls an avatar in an interactive medium, such as video games, as compared to watching an identical visual representation in a non-interactive film or tv show, studies such as this show us that they will tend to feel more “present” and feel greater “identification” with what that avatar does in the game. Greater presence and identification will make it feel more like the player themself is doing the wrongful actions depicted in the video game, rather than watching someone else perform those identical wrongful actions in non-interactive media. This may be even more of a concern in mediums such as VR (Ramirez & LaBarge, 2018) which can potentially produce even higher degrees of presence and/or identification (see Sanchez-Vives & Slater, 2005; Slater et al., 2006). Insofar as this makes it more likely that an interactive medium, such as video games, will produce a more “virtually real experience” (Ramirez & LaBarge, 2018), that is, where we experience a simulation as if it were real (i.e. it feels like we are really murdering or molesting someone), than in non-interactive media, and insofar as these seem like morally relevant features, it follows that the interactivity of video games is an essential feature of (at least some cases of) the Gamer’s Dilemma. If so, then the Gamer’s Dilemma and the paradox of treating wrongdoing lightly (assuming there is such a thing), would seem to be about different things (at least in some cases), as the former relies (in at least some cases) on a moral concern with high degrees of presence and identification caused by interactive mediums, such as video games, whereas the latter does not.

We think that this provides us with a compelling reason to believe that the interactivity of video games may, in at least some cases, be an important moral feature that may explain why engaging with a virtual wrongdoing could be morally impermissible, since it might be wrong to have the virtually real experience of murdering or molesting others (for more on this line of argument, see Ramirez, 2020). However, this consideration may not be enough to justify the stronger claim that interactivity is an essential feature of the Gamer’s Dilemma for the following three reasons.

First, video games that don’t produce, or produce to a high degree, a virtually real experience with high degrees of presence and identification, may still be intuited as morally off-limits or not fair-game. For example, imagine a highly offensive game that glorifies the sexual abuse or murder of children, but does so with very low-quality graphics and with a story and mechanics designed to alienate players, such as being played from a god’s eye perspective, which together do not encourage the player’s identifying with the perspective of the in-game avatar (see Ramirez (2020) for discussion of the former and Davnall (2021) for discussion of the latter factor). Nonetheless, such a game may still generate Gamer’s Dilemma cases as people may intuit that murder, but not sexual molestation of children, is permissible in this game, even though it produces low levels of presence and identification in the player.

Second, virtual wrongdoings that generate a high level of presence and identification via their being interactive may be in some cases morally unobjectionable regardless of those features being generated, due to the kind of agency manifested by the player (or morally objectionable in virtue of the kind of agency manifested). For example, Mildenberger (2017) distinguishes between three kinds of virtual agency in regard to virtual killing in multiplayer games, which we shall adapt here for our purposes to take place in high presence, single-player video games and be applied to other virtual wrongdoings, such as virtual child molestation. These three kinds of virtual agency are role-playing, virtual acting, and acting. “Role-play” involves a player virtually acting as if they shared the beliefs and desires of their virtual identity (the character of the avatar they are controlling); for example, one desires to run over an innocent pedestrian NPC in GTA insofar as they are role-playing that this is what they qua their avatar desires. “Virtual acting” involves a player virtually acting in the interest of their own beliefs and desires to achieve in-game or virtual ends; for example, the desire to run over an innocent pedestrian NPC in GTA insofar as this is what they as a gamer desire, where their desires have in-game ends (i.e. to progress through the game). Finally, “acting” involves a player virtually acting in the interest of their own non-virtual beliefs and desires by using the virtual medium to realise non-virtual ends; for example, a player desires to run over an innocent pedestrian NPC in GTA to upset their friend who is watching and asked them not to. Mildenberger argues that, in multiplayer video games, it is only the last kind of virtual agency—acting—which could be morally objectionable, and not because of the content of the virtual wrongdoing (i.e. it being an instance of virtual murder or virtual child molestation) but rather because of it “deliberately and non-consensually evoking disagreeable emotions in others” (Mildenberger, 2017, 185). All that is necessary to accept for our purposes is that moral assessments of virtual wrongdoings involve more than just a consideration of how interactive an action is qua immersive features or qua agency, which the above agential distinctions from Mildenberger highlights. Thus, it may be the case that ‘role-play’ experiences of virtual killing are morally unobjectionable even though they qualify as virtually real and agential via their being highly interactive.

Third, there is something importantly correct about Luck’s move, following Davnall’s (2021) “deflationary approach”, to identify the Gamer’s Dilemma as an instance of a broader phenomenon concerning the depiction of fictional acts qua fictional representations, rather than virtual acts per se. This is because there is a consistency to be noted between the far different standards that are commonly applied to the presence of (non-sexual) violence and the presence of sexual violence, especially against children, in various forms of media. For example, the Australian Guidelines for the Classification of Computer Games, 2012 (Cth) state that the weaker condition of a game merely containing “descriptions or depictions of child sexual abuse” are sufficient to refuse classification, whereas the much stronger condition of “violence with a very high degree of impact which are excessively frequent, prolonged, detailed or repetitive” is needed to refuse classification. Unsurprisingly, these varying standards track onto similar cultural standards and legal frameworks being applied to various forms of media, including both interactive and non-interactive mediums (e.g. the Australian Guidelines for the Classification of Computer Games, 2012 is based on the Commonwealth Classification (Publications, Films and Computer Games) Act 1995). It therefore makes sense to see the Gamer’s Dilemma as an instance of this broader phenomenon, even if the interactivity of the medium complicates that story.

However, there may nonetheless be important differences between different forms of media. For example, literature, as a non-audio-visual medium, may have more scope to deal with some of these issues, such as child sexual abuse, than audio-visual mediums, such as films and video games. The fact that books don’t face the same sort of classificatory regime that audio-visual mediums such as films, television and video games are subject to provides some evidence in support of this suggestion. Further, games which promote very high levels of presence and identification may be more morally problematic when they contain depictions of wrongdoings as they make it feel more like the gamer is actually perpetrating those wrongs themselves (which could harm their moral character and so on). For example, Ramirez (2020, 144) notes that the sexual assault which fictionally occurs in Shakespeare’s play Titus Andronicus fails “to generate the kind of controversy that games involving murder, sexual assault, or pedophilia generate”. While for Ramirez this may indicate an important dimension of the Gamer’s Dilemma specific to video games, it may instead indicate that the medium via which a fictional wrongdoing is expressed adds a contextually relevant factor for determining its intuitive moral acceptability.Footnote 4

3.3 Does a Broader Gamer’s Dilemma Need Wrongdoings to be Treated Lightly?

This leads to Luck’s second key claim, that what is going on in such cases is that we are treating fictional wrongdoings lightly, and some wrongs are “off limits” to treat in this way (such as child molestation) and others are “fair game” to treat in this way (such as murder), across a range of media. However, an important objection to this claim is that our intuitive concern with fictional representations of child molestation can occur even when these cases are not treated lightly. Recall that treating lightly can occur either because one engages with it in playful manner (see Seddon, 2013) or because the treatment is in poor taste (see Young, 2019). Of course, it is a mistake to think that all video games and all video game play is “playful”, insofar as this contrasts with “serious”. Luck does not make this mistake, referring to e-sport gamers and film (or we could add game) critics as examples not fitting the category of playful engagement. More broadly, lots of other modes of engagement with games, such as “superplay” which involves play orientated around game mastery, such as high score pursuit, speedrunning, sequence breaking and pacifist play, would also seem to fit this category of ‘non-playful’ play (see Newman, 2008). While nobody may be taking Candy Crush Saga too seriously, other morally intense games, such as The Walking Dead or The Last of Us, are games that many gamers play seriously (Nguyen, 2020), demanding the same kind of serious aesthetic engagement with which one watches a morally challenging film or play. That is, these games do not treat their content in a light, comedic or playful manner and appropriate engagement (see Ali, 2015; Nader, 2020) with that content by gamers is also not playful, even though they are still only “playing” the game and not, for example, “reviewing” it as a professional critic.

Building on this point, games involving objectionable depictions of child molestation might be intuitively seen as wrong even when the games encourage players to take the depicted act seriously (rather than playfully), and some gamers do in fact take it seriously. For example, a game designed to allow paedophiles to take seriously virtual child molestation might be seen as morally unacceptable even though the game, and gamers, do not engage with the depicted wrongdoings “playfully” (in any sense of the term). However, perhaps such cases are instead matters of “poor taste”. But that also seems not to be the case, at least in the sense Luck (following Young, 2019) uses the term to mean trivialising an objectively wrong act. The above-mentioned game, far from trivialising the objectionable depictions of child molestation, makes this wrongdoing its core and serious focus. Such a game seeks to glorify or intensify, rather than trivialise, the virtual wrongdoing in question, and again, this could plausibly make us see it as more morally unacceptable because of its lack of playfulness.Footnote 5 In contrast, our intuitive concern with fictional off-limits wrongdoings may be in some cases diffused in virtue of their fictional representation being treated lightly.Footnote 6 Take for example Lubitsch’s WWII comedy film To be or Not to Be [Film] (Lubitsch, 1942) and its use of satire in exposing the ridiculousness of the Nazi’s ideology. By treating the Nazis as a comedic object, the film (arguably) successfully shows how in some cases making light of serious wrongdoings may be an appropriate fictional approach (or at least not clearly an inappropriate approach) to fictionally representing grave events.

Drawing these points together, there does seem to be an intuitive moral difference, or at least different cultural standards, regarding how murder (and other forms of nonsexual violence), compared to how child molestation (and other forms of sexual violence), are treated in a range of interactive and non-interactive mediums, with much more permissive standards in the case of murder rather than in the case of child molestation. However, whether or not the depicted wrongdoings in question are treated lightly, or engaged with lightly, does not seem essential to this difference, since it also holds in some cases when those wrongdoings are treated, or engaged with, seriously. The remaining philosophical challenge then becomes to show whether some property or feature can be pointed to that would justify the different moral standards applied to these two cases. In the next section, we look at whether Luck’s (2022) suggestion that the graveness of the depicted wrongdoing can successfully justify this difference.

4 Does Graveness Resolve the Paradox?

The previous section suggests that there is a more general problem here, but it does not seem to be the problem of treating wrongdoings lightly. Rather, the problem seems to be how to justify that, intuitively, a much wider range of cases of murder and other forms of non-sexual violence are seen as morally acceptable to depict and engage with in a range of different media (including video games), compared to the comparatively narrower range of cases of sexual violence and child molestation that are seen as morally acceptable to depict and engage with. One way to account for this difference might be, as Luck suggests, in terms of graveness.

For Luck (Luck, 2022, 14), “graveness refers here to something like seriousness or solemnity”. Graveness is an “objective property” that is “sensitive to both extrinsic and intrinsic factors” (Luck, 2022, 15). The latter refers to inherent moral features of the act, and the former to things such as “the social context of the wrongdoing; how long ago the wrongdoing occurred; how distanced from reality a fictional wrongdoing may be; and who committed the wrongdoing” (Luck, 2022, 15). Off-limit wrongs, such as “rape, molestation and homophobic insults” are, Luck argues, also instances of “oppression”, or the long-term systematic denial of justice, which are very grave, whereas fair-game wrongs, such as murder, are not typically instances of oppression. The fact that they are also instances of oppression makes off-limit wrongdoings, for Luck, sufficiently graver than fair-game wrongdoings to the extent that it is permissible to treat the latter but not the former lightly. The other ways that acts can be more grave is if the wrongdoings being depicted happened very recently (the “too soon” phenomenon in comedy), or if the depicted situations or acts are too close to reality (the “too close” phenomenon), or if it involves the use of a misrepresentation of a group (e.g. black people) by another group (e.g. white people) that has a long history of systematically benefiting from such misrepresentation (Luck, 2022, 17). While Luck does not consider this list of ways that acts can be very grave to be exhaustive, he clearly takes it to be illustrative of at least the key cases.

Luck’s list includes a seemingly ad hoc range of issues lumped together under the broader rubric of graveness, and while some of these issues clearly are morally relevant, it is not clear in all cases that they really make a wrongdoing (or its depiction) graver to a sufficient extent. We shall start with the “too soon” and “too close” phenomena. These are indeed morally relevant considerations. But the moral issue is not so much temporal or fictional distance per se. For example, many people may get very upset if certain religious or national figures or symbols are not treated with sufficient reverence. But some of these depicted figures can be very temporally distant, such as Jesus or Muhammad. Time is relevant only because people tend to care more deeply about more recent events and persons (e.g. people will likely be more upset about a crude joke made about a recently departed loved one compared to the same joke made 40 years later), although not exclusively as instances of depictions of historically distant religious figures make clear. To give an inverse example, a depiction of the humiliation of a homeless person that nobody seems to care about might involve a matter that happened very recently (i.e. too soon) and be very accurate in its depictions (i.e. too close), but the fact that it is temporally and fictionally nearby doesn’t seem to make its depictions more morally grave than it otherwise would be. In this case, the temporal and fictional nearness has little extra impact because no one (let’s assume) cares deeply about this person.

What such cases reveal is that what matters, morally, is that some people will be very upset if something they care about deeply, or is central to their self-image, is treated in ways they see as being disrespectful. This is a morally relevant consideration since actions that greatly upset others give us defeasible reasons not to perform such actions. Of course, these considerations need not be over-riding or absolute, since, all things considered, it may be permissible or appropriate to, for example, mock a religious figure even if it will upset a lot of people. Similarly, these considerations are also not infallible, since the fact that most people don’t seem to care that a homeless person was humiliated (as in our above example) doesn’t make that act morally better. However, the fact that a certain depiction will deeply upset a lot of people is a morally relevant consideration which may require looking for more sensitive and respectful depictions unless there are good overriding reasons not to do so, such as because one is making an important political point. We could extend this observation to the core cases of the Gamer’s Dilemma by noting that many people will tend to care deeply about, and thus become upset by, certain depictions of child molestation in video games, and they tend not to care nearly as deeply, and thus not get as upset, about similar depictions of murder in video games. But that is just another way of repeating the grounding moral intuitions at play in the Gamer’s Dilemma, and it doesn’t really help us to see if those intuitions have robust moral foundations such that we can point to a relevant feature that justifies the moral difference in question without being viciously circular.

Further, we don’t think it is correct that all intuitively off-limit wrongdoings are in fact graver, or even graver at all, than some intuitively fair-game wrongdoings. While Luck is right to note that the broader category of moral wrongdoing can be broken up into other sub-categories, such as oppression, he does not also engage with the literature and category of moral evil which is deeply concerned with the issues of moral graveness and degrees of moral wrongness (e.g. Bernstein, 2002; Calder, 2013; Card, 2002; Formosa, 2008, 2019; Garrard, 1998; Morton, 2004; Russell, 2014). Indeed, a central focus of that literature is the nature of moral graveness and the extent to which “evil” is reducible (or not) to an action being very, very wrong. Features typically discussed in that literature that make an act morally graver, and thus potentially an evil action, include the type and degree of harm suffered by victims, such as whether a harm is sufficiently “life-wrecking” (Formosa, 2008), the types of motives that perpetrators of those acts have, such as whether they were motivated by sadism or acted intentionally (Kramer, 2014), and responses of incomprehension when trying to understand how others could act so badly (Russell, 2012).

In the literature on (real-world) evil, acts such as murder, torture, and chopping people’s limbs off when inflicted for the sake of sadistic enjoyment or competitive advantage are typically considered to be very morally grave and thus candidate evil actions, given the extreme harm they inflict from reprehensible motives. However, virtual versions of these actions are routinely considered fair-game wrongdoings in video games and other forms of media. In contrast, actions which might not be candidate evil actions in the real-world, may be seen as off-limits in video games even though they are less morally grave than some evil actions. For example, committing mild but repeated and intentional mockery of a disabled child from a misguided desire to “impress” some friends, rather than from taking any pleasure in the act, would probably be considered by many to be off-limits in a video game, whereas the sadistic torture, murder, or chopping off of limbs of other characters for competitive advantage would be considered by many to be fair-game in a video game. However, on most (if not all) theories of evil about the real-world equivalents of these actions, the latter actions would be the morally graver, given the life-wrecking or life-ending harms perpetrated for sadistic purposes, in contrast to the comparatively minor harm involved in the (still morally reprehensible) mild mockery of a disabled child. Or at the very least, it seems that no plausible theory of evil would hold that the mild mockery of a disabled child is so significantly morally graver than sadistic murder, torture, or limb chopping-off that this could ground a substantive moral difference in the permissibility of depicting the latter but not the former. Even so, many people would find the latter (i.e. murder, torture, etc.) more morally acceptable than the former (i.e. mild but repeated mockery of a disabled child) to perpetrate and depict in a video game, indicating that whatever may be underwriting or fuelling these intuitions, it does not consistently track onto the degree of the moral graveness of the depicted act.

To further strengthen this last point, consider as another indication of moral graveness the severity of punishment for different crimes. While the relationship here is far from perfect, for most penal codes and retributivist lines of thinking (e.g. Corlett, 2001), the moral graveness or seriousness of the crime or wrong in question typically bears at least some degree of proportionality to the punishment it warrants (even if there are other relevant considerations, such as deterrence or remorse, that are considered). Murder is typically considered in penal codes to be a very serious wrong deserving of some of the harshest punishment. While various forms of sexual violence, such as rape or child molestation, are also considered very serious crimes, they are typically not punished more severely than murder (see, for example, section 19(a)(1) & section 66(a) to 66(d) of the Crimes Act 1900 (NSW); Holmes, 2013; Taussig, 2012). Again, while in no way conclusive on its own, it adds some prima facie weight against the claim that child molestation and other off-limit wrongdoings are significantly morally graver than murder and other fair-game wrongdoings. Indeed, our sentencing regimes seem to suggest, in line with much of the literature on evil, the opposite view, namely that murder is more morally grave, or at least not significantly less morally grave, than molestation (all else being equal). This suggests once again that, whatever is underwriting the differing intuitions people have about the moral acceptability of depictions of murder and molestation in video games and other forms of media, it is not, contra Luck, the significantly different degrees of moral graveness of the depicted actions that is doing that work.

5 From the Gamer’s Dilemma to the Paradox of Fictionally Going Too Far

Thus far, we have agreed with Luck’s (2022) move to expand the Gamer’s Dilemma into other fictional settings to bolster the Dilemma’s explanatory scope. This endorsement, however, comes with the caveat, argued for above, that those fictional instances need not be treated or engaged with lightly to trigger the intuitions grounding the Gamer’s Dilemma. This shifts the problem from one of being unable to rationally explain why virtual murder is justifiably intuited to be morally acceptable and virtual child molestation morally unacceptable in video games (i.e. the Gamer’s Dilemma), to the problem of being unable to rationally explain why some fictional wrongdoings are justifiably intuited as fair-game, such as many fictional instances of murder, and others are justifiably intuited as off-limits, such as many fictional depictions of sexual assault or the molestation of children. To make it distinct, we shall call this latter problem the “Paradox of Fictionally Going Too Far” (which is a slight modification of Luck’s (2022) “paradox of treating wrongdoing lightly” that removes the focus, argued for above, on treating wrongdoings lightly). This move allows us to see the Gamer’s Dilemma as an instance of this broader paradox applied to specific examples in the context of video games.

One of the main benefits of recontextualising the Gamer’s Dilemma as an instance of a broader paradox is that we can both reuse the resources of the Gamer’s Dilemma literature and apply them to this broader paradox, and we can reuse the resources of a range of other literatures that are relevant to the broader paradox and apply them to the Gamer’s Dilemma. We hope to demonstrate these benefits by attempting to expose a largely implicit moralism in the Gamer’s Dilemma literature (but, for a recent exception, see Luck, 2023) that a focus on the broader paradox may help to disabuse us of and thereby open us up to new alternatives. We argued above that Luck’s attempt to resolve the Paradox of Fictionally Going Too Far by appealing to a moral property, that of the graveness of actions, fails to account for the (purported) intuitive difference between fair-game and off-limit fictional wrongs in all relevant cases. One response to this argument might be to search around for some other moral property (or properties) that can account for this intuitive difference. However, to assume that we need to look for a moral property to explain this intuitive difference is the implicit moralism that we wish to contest here. We can see a similar move in recent discussions of the Gamer’s Dilemma where it has been argued that the lack of consensus around any successful resolution or dissolution to the Gamer’s Dilemma that appeals to moral properties might suggest that we need to look to something else besides moral properties to explain what is going on with the Dilemma and its grounding intuitions (Montefiore & Formosa, 2022; Young, 2016). This leads to an analogous view that a moral explanation may not be able to account for the intuitive difference between fair-game and off-limits fictional wrongdoings that grounds the Paradox of Fictionally Going Too Far. While by rejecting the graveness resolution to this paradox we have not ruled out the possible success of other moral approaches to resolving this paradox, we do claim that its failure and the current state of the Gamer’s Dilemma at least motivates us to explore the alternatives, which we shall now do.

If differing moral properties cannot explain the intuitive difference between off-limits and fair-game wrongdoings in fictions, what else can? Here we claim that a focus on aesthetic, cultural, and conventional norms that function both in and across different media, and which are a common feature in other literatures, can help us to provide plausible answers to this question. This appeal harks back to Luck’s initial, and quickly dismissed, proposed resolution to the Gamer’s Dilemma concerning social convention. Luck was right to reject this approach to the extent that an appeal to conventional normativity won’t necessarily entail a moral norm (see Kass, 2002; Luck, 2009; Luck, 2019; Young, 2016). However, this is only a problem if we seek to resolve the dilemma by providing a morally relevant distinction between fair-game and off-limits fictional wrongdoings. Instead, one might, in the same spirit as dissolving approaches in the literature (e.g., Ali, 2015; Öhman, 2020; Ramirez, 2020), appeal to the normativity of various cultural and aesthetic conventions to further contextualise our intuitions to the extent that their seeming moral force is (arguably) explained away on other non-moral grounds.

We turn first to cultural norms. It is easy to treat our intuitions as if they are infallible trackers of morality (see Singer, 2005) and ignore the extent to which non-moral normative concerns, such as our cultural context, tend to shape our intuitive appraisals. For example, via this approach we can note that conventional permissibility around fictional representations will likely differ from culture to culture, and as a result the intuitions that ground the Paradox of Fictionally Going Too Far may only get off the ground in certain cultural contexts. For example, Young (2016, 22) argues that if “we take Japan as a contemporary example of a society with different social conventions regarding sexualized imagery,” then it is not hard to see how those in other cultural contexts either might not intuit virtual child molestation as morally impermissible or might not be so ready to intuit virtual murder as morally permissible (for empirical support for the claim that virtual murder may not be seen as morally acceptable, see Formosa et al., 2023). Of course, we may just bang the table and insist that our intuitions are morally correct and that everyone else’s are mistaken, but at the very least we should concede that it is at least possible that our so-called moral intuitions about such cases may either be mistaken or may in fact not really be moral intuitions but rather may be misinterpreted cultural intuitions (or something else).

We can add further weight to this suggestion by considering individual differences. A recent empirical study (Formosa et al., 2023) has shown that intuitions about the moral acceptability of virtual murder and virtual molestation in video games vary by individual gaming experience. In particular, participants with high levels of gaming experience found virtual wrongdoings, including both murder and molestation cases, to be more morally acceptable than those with lower levels of gaming experience (Formosa et al., 2023). Intuitions about the Gamer’s Dilemma therefore seem to vary according to the amount of gaming experience an individual has, with more experienced gamers taking a less moralising view of virtual actions. Whether this is because they are desensitised to the morality of virtual violence or because they are better able to see that virtual actions are “just a game” and therefore morally acceptable is, however, unclear (Formosa et al., 2023). What this suggests is that a further non-moral feature, namely gaming experience, impacts the intuitive acceptability of fictional wrongdoings for individuals, and this raises the broader puzzle of whether those with high or low amounts of relevant experience are in a better or worse position to be making moral judgements regarding the moral acceptability of fictional wrongdoings.

Further, what may be intuited as fair-game or off-limits fictional wrongdoings in one medium, or within one genre of a medium, may differ when placed in a different media context. Consider, for example, the varying, while all controversial, receptions of Nabokov’s (1955) novel Lolita and the subsequent Kubrick (1962) and Lyne (1997) film adaptations. Kubrick’s adaptation was edited by censors to make it “appropriate” for film audiences by lessoning the erotic aspect of Humbert’s relationship with Lolita (see Green, 2020), which is informative as to the different kinds of presumed limits afforded to representations of immorality across media. One way to account for these differing limits may be due to intrinsic features of the medium. Perhaps, for example, fictional depictions of child molestation are less confronting in non-audio-visual media such as literature (as compared to video games or films), and this may allow for more nuance and scope for tackling morally challenging content in literature (as opposed to video games or films), thereby justifying a more morally permissive attitude toward fictional wrongdoings in that media context.

Alternatively, there may be nothing fundamentally remarkable about literature that makes morally challenging content fair-game in that media context and not in other media contexts. It might instead be historically contingent associations with literature, and now cinema, as “artistic” mediums able to tackle morally challenging ideas, while other “non-artistic” mediums, such as video games, that are not conventionally seen in this way, that guides conventional norms concerning the acceptability of fictional wrongdoings across different mediums. For example, the video game Heavy Rain, while arguably not comparable as a work of art with the novel Lolita, attempts to achieve a similar kind of sympathetic attitude towards fictional wrongdoers in a way that video games seem primed to do, and in principle capable of, through their capacity to produce heightened levels of perspective fidelity. Of course, these historically contingent associations are malleable, and over time, video gaming may take on the associations as a medium for artistic expression (at least more so than it currently does) (see Nguyen, 2020). Therefore, it is unclear why it is morally acceptable, if done in similarly tasteful and mature ways, for fictional wrongdoings to be explored in one medium but not another. In any case, seeing the Gamer’s Dilemma as a sub-problem of a broader Paradox of Fictionally Going Too Far should encourage us to look more broadly at the treatment of, and changing moral, aesthetic, or cultural judgments about both violence and sexual violence against adults and children in a range of mediums.

An important objection to our suggestion here that there may be a range of non-moral features, such as cultural norms, individual differences in familiarity with media, and various medium-specific norms, that may account for our intuitions about what constitutes fair-game and off-limits fictional wrongdoings is: why then do these intuitions appear to many people to be at least prima facie moral? Why, to at least some people, do off-limits fictional wrongdoings appear morally off-limits and not merely conventionally off-limits? Luck (2023), in response to Montefiore and Formosa (2022), answers by appealing to what appears to be the simplest explanation, namely that our moral intuitions are indeed morally grounded. Without clear evidence it is hard to answer this question, but drawing on the only empirical study so far of this question suggests four explanations. There may be others.

The qualitative component of the study by Formosa and colleagues (Formosa et al., 2023) found that the three most common negative themes participants used to explain why the relevant virtual action was morally unacceptable was that “it was just wrong”, that it is “disgusting or sickening”, and that it “makes us morally worse”. The most common positive theme in support of the virtual action being morally acceptable, and indeed the only positive theme widely mentioned in the study, was that it was “just a game”. The first theme mentioned above points, according to the study, to the potential problem of “moral dumbfounding”, that is, of “maintaining a moral judgement, without supporting reasons” (McHugh et al., 2017), since participants did not clearly explain why the virtual action in question was wrong. One potential reason why people might not be able to come up with moral reasons in support of their judgment is that they are confusing what might be a conventional norm, such as the prohibition against incest (see Haidt et al., 2000), for a moral one. The second theme points to a related issue, which is the (contested) role that disgust responses might have in colouring moral judgments (see Haidt et al., 1993). Some people may be confusing a disgust response to, for example, virtual molestation, for a moral assessment (See Kjeldgaard-Christiansen, 2020). The third negative theme points to the problem that people might be basing their moral assessment of fictional wrongdoings on either an unproven or false belief, in this case about the causal role that playing video games with objectionable content will have in making us morally worse. As discussed in much detail in the Gamer’s Dilemma literature, we lack good evidence for this claim, which is not to say that the claim is false, merely that it is unproven (see Young, 2016). Finally, the most used positive theme that “it is just a game” points back to a range of amoralist views in the literature that argue that virtual actions take place inside a morally insulating magic circle and should not accordingly be appraised through a (real-world) moral lens (see Huizinga, 1944; Juul, 2008; Nguyen, 2017; Salen & Zimmerman, 2003; Stenros, 2014a, 2014b). This leaves open the view that off-limits fictional wrongdoings might be what Patridge (2011, 306) calls “juvenile, or in bad taste, or even boring, but not morally objectionable”. In other words, some people may be confusing violations of conventional or taste norms in different fictional contexts, such as their being in bad taste, with their being a violation of moral norms because the off-limits fictional wrongdoing appears to be morally charged.

It is of course still possible that we may yet find moral grounds for the prima facie intuitive immorality of off-limits fictional wrongdoings. Further, individuals may simply have been unable to expand upon, or articulate in an experimental condition, the moral reasons that they may have had that provide the justificatory grounds for their moral intuitions (see Luck, 2023). However, what we have shown here is that by considering the available empirical evidence and the paradox across a range of media, a stronger candidate for what is causing the prima facie intuitive immorality of off-limits fictional wrongdoings in many, if not most, cases of the Paradox of Fictionally Going Too Far seems to be non-morally relevant features of the fictional wrongdoing (such as its violation of medium-specific conventions) and the observer of that wrongdoing (such as their cultural context or experience with a medium).

Further, a refocusing of the Gamer’s Dilemma in the context of the Paradox of Fictionally Going Too Far may also illuminate interesting points of comparison as to how fictional wrongdoings are treated across a range of media. In particular, there seems to be an asymmetry whereby the moral status of video game wrongdoings is being subject to moralist assumptions in the Gamer’s Dilemma literature much more readily than other fictional wrongdoings, such as those in film or theatre, are in other literatures. Consider the following three nearby examples alongside the Gamer’s Dilemma. In the film studies literature concerning the Paradox of HorrorFootnote 7, an amoralist approach is commonly assumed (e.g., Carroll, 1990; but see Woodcock, 2013 & DiMuzio, 2006 for exceptions), followed by a consideration of what it is about fictional wrongdoings (which involve horrific imagery), that make them appealing. In the philosophy of fiction literature on the puzzle of imaginative resistanceFootnote 8 (see Gendler, 2000; Gendler & Liao, 2015; Levy, 2005) an amoralist approach is once again commonly assumed, followed by considerations about why it is that some fictional wrongdoings provoke imaginative resistance more than others. Finally, in the analytic aesthetics literature on the Paradox of Fiction, an amoralist approach is once again commonly assumed, followed by a consideration of whether it is rational to be emotionally moved by fictional events at all (e.g. see Friend, 2008; Radford & Weston, 1975).Footnote 9

In contrast, in the Gamer’s Dilemma literature it is not only commonly implied that it is rational to be emotionally (or at least morally) moved by off-limit video game wrongdoings, but it is also rational to be so moved on justified moral grounds. The treatment of fictional wrongdoings in these other literatures therefore make their central foci weaker claims concerning whether it is rational to have a particular emotional response to, find appealing, or imaginatively resist (off-limits) fictional wrongdoings, whereas the stronger claim of whether engaging with an off-limits fictional wrongdoing is immoral over and above its being fictionally immoral is largely not entertained (see e.g. Brey, 1999; but see Young, 2015 for an exception). This suggests that if the Gamer’s Dilemma is an instance of the broader Paradox of Fictionally Going Too Far, and we ought to therefore treat off-limit video game wrongdoings as a subset of off-limits fictional wrongdoings, then we are either overestimating the importance of moral intuitions in video game contexts or underestimating their importance in these other literatures. However, given that we have argued above that we seem to have stronger grounds for suspecting that there are non-moral rather than moral bases that can best explain the intuitions that ground the Gamer’s Dilemma, we therefore have little reason to hold that, for example, imaginative resistance to an off-limits fictional wrongdoing entails the immorality of engaging with that fictional wrongdoing.

6 Conclusion

In this paper, we have argued for four key moves. First, we have accepted Luck’s expansion of the Gamer’s Dilemma to be applicable to a wider set of media, but given a novel recasting of this in terms of the Paradox of Fictionally Going Too Far. This move was motivated by a rejection of Luck’s requirement that fictional wrongdoings need to be treated lightly to be subject to the paradox. Second, we developed a novel criticism of Luck’s (Luck, 2022) graveness resolution to this broader paradox. Third, we argued that the Paradox of Fictionally Going Too Far helps to expose an implicit moralism in the Gamer’s Dilemma literature when compared to relevant nearby literatures about other forms of media. Finally, we considered a range of non-moral, cultural and media conventions that plausibly help to dissolve the intuitive moral gap between non-sexual and sexual violence that is central to the broadened paradox. However, further research is needed to support our argument about the role different forms of media and degrees of expertise in these forms of media may play in influencing intuitions concerning the moral and non-moral acceptability of fictional off-limits wrongdoings, and to explore the practical implications (if any) of our claims for classificatory guidelines and the social acceptance of morally confronting video games.