The gamer’s dilemma: an expressivist response

Abstract

In this paper, I support a hybrid form of expressivism called constructive ecumenical expressivism (CEE) which I have previously used (to attempt) to resolve the gamer’s dilemma. (Young, 2016. Resolving the gamer’s dilemma. London: Palgrave Macmillan.) In support of CEE, I argue that the various other attempts at either resolving, dissolving or resisting the dilemma are consistent with CEE’s moral framework. That is, with its way of explaining what a claim to morality is, with how moral norms are established, with the role intuition is able to play in establishing these norms, and therefore with how the gamer’s dilemma can be resolved. I also demonstrate, more broadly, how CEE advocates robust relativism as a means of justifying the constructed moral norm’s normative credentials, and therefore how one society’s norm can be judged morally superior to another’s.

Introduction

The gamer’s dilemma (Luck, 2009) can be presented as follows: If it is intuitively the case that virtual murder is permissible and intuitively the case that virtual paedophilia is impermissible, and if there is no moral difference between these two activities, then one of the aforementioned intuitions is incorrect and needs to be amended to avoid inconsistency. We should intuit either that virtual paedophilia ought to be permitted (like virtual murder), or that virtual murder ought to be proscribed (like virtual paedophilia). Each amendment is unpalatable, however, precisely because it runs contrary to our intuitions: hence, the dilemma.

A number of attempts have been made either to resolve the dilemma or dissolve it. In the former case, by arguing for the existence of various moral differences between virtual murder and virtual paedophilia that these intuitions align with, thereby validating them (Bartel, 2012, Coghlan & Cox, 2023; Kjeldgaard-Christiansen, 2020; Luck, 2022; Ostritsch, 2017, Patridge, 2013; Ulbricht, 2023; Young, 2016). Or, in the latter case, by arguing that claims about the intuitions on which the dilemma is grounded are simplistic, owing to their lack of context, and therefore inaccurate when applied broadly (Ali, 2015; Nader, 2020; Ramirez, 2020). Various context-specific, and therefore refined, versions of the dilemma—indicating narrower intuitions—continue to be discussed, however (Ali, 2022; Luck, 2018, 2022). A point I will return to.

More recently, Montefiore and Formosa (2022, 2023) have changed tack, and sought to resist the dilemma. They argue that while claims about the intuitions relating to the gamer’s dilemma have a normative foundation, such that there are reasons why, intuitively, one ought not to permit virtual paedophilia while allowing virtual murder, it just so happens that these reasons are non-moral (e.g., they stem from some psychological attitude or taste-based aesthetic).¹ It is therefore possible, they claim, for virtual murder and virtual paedophilia to be morally equivalent and for our intuitions not to signify this, but also for these same intuitions to have normative force, albeit in virtue of the non-moral properties they purportedly reflect.

I will not spend time explaining directly or in detail how these various responses differ from each other, and therefore what specifically is involved in resolving, dissolving or resisting the gamer’s dilemma, although relevant aspects of these previous responses will form part of a metacommentary² that I intend to engage in when supporting the expressivist approach to morality I originally presented in Young (2014, 2015) known as constructive ecumenical expressivism (CEE), which I then applied to the gamer’s dilemma (Young, 2016).

Support for CEE will involve both a general defence of its relativist credentials (that will include addressing published objections) and a specific defence of its response to the gamer’s dilemma. The latter defence will include an explanation of how CEE can in fact accommodate various other responses to the dilemma within its general approach, and indeed might arguably be said to anticipate these responses (in terms of their variations). The mapping of these other responses onto CEE will constitute the aforementioned metacommentary.

Before proceeding, however, it is worth noting that the argument in support of CEE will be limited to (a) what it (CEE) has to say about how it can resolve the gamer’ dilemma, and (b) its capacity to account for—owing to its compatibility with—other responses to the dilemma. Importantly, then, while part of this specific discussion requires a broader defence of CEE’s relativist credentials (which I will provide; see below), I will not engage in more general discussion on CEE, as this is beyond the score of this paper.

In particular, I will:

• 1. Outline CEE (Sects. 3 and 4) for the benefit of those less familiar with the approach, but also to make clear its anti-realist, expressivist and relativist credentials.

• 2. Demonstrate how various attempts to resolve the dilemma (Sect. 5), or dissolve or resist it (Sect. 8) are, collectively (despite their respective differences), consistent with the means by which moral norms are established within a particular community, according to CEE, and therefore compatible with CEE’s approach to morality, generally, and resolving the gamer’s dilemma, specifically.

• 3. Defend the claim that, despite its relativist credentials, which have been of concern to critics (Dennison, 2016; Luck, 2019; Ostritsch & Ulbricht, 2021; see Sect. 6 for details), CEE is nevertheless robust enough not only to withstand capricious changes to norms, but provide the means to justify the assertion that one society’s (non-capricious) norm is superior to another’s. To do this, I will draw a distinction between local and global normativity (Sect. 7). I will also demonstrate how CEE provides the means to differentiate between virtual paedophilia and virtual murder that, pace Montefiore and Formosa, is both normative and moral in nature, rather than normative and non-moral (Sect. 8).

Before discussing any of this, however, I will begin by unpacking the gamer’s dilemma.

Unpacking the dilemma in its original and broadest form

As noted, the original version of the gamer’s dilemma (Luck, 2009) is grounded on two claims about what we intuit when it comes to engaging (via one’s avatar) in virtual paedophilia and virtual murder (typically, either action is directed towards a non-player character): that, intuitively, the latter is permissible, whereas the former is not. Luck’s challenge in his original 2009 paper was to justify these contrasting intuitions given that arguments in favour of the permissibility of virtual murder appear to apply equally to virtual paedophilia and, conversely, arguments favouring the impermissibility of virtual paedophilia appear to apply to virtual murder (indicating no morally relevant difference between them).

In a later work, however, Luck (2022) expresses a preference for the dilemma to be presented as a paradox:

• 1. Virtual murder is permissible.

• 2. Virtual paedophilia is impermissible.

• 3. There is no relevant difference between virtual murder and virtual paedophilia with respect to permissibility.³

Statements 1–3 are paradoxical because, while each purports to describe a particular truth about the state of nature—that is, while each makes an ontological claim about the instantiation of moral properties by actions—when taken together, they contain an inconsistency. Notably, if 1 and 2 are true (i.e., if they describe a state of nature), then 3 (and therefore what it describes) cannot be true because the truth of 1 and 2 contradicts the truth of 3. If, however, 3 is true, then either 1 or 2 is false. Therefore, if one holds that 1 and 2 are true (i.e., if one holds that 1 and 2 are describing a state of nature), then one cannot (on pain of contradiction) hold that 3 is also true (i.e., hold that the state of nature statement 3 describes is also true), and vice-versa if one were to hold that 3 is true.

Of course, the gamer’s dilemma, as originally formulated, does not consist of a series of ontological claims (as presented in 1–3). Instead, it comprises two epistemic claims and an ontological one. To illustrate:

• 4. People intuitively believe that virtual murder is permissible [epistemic claim].

• 5. People intuitively believe that virtual paedophilia is impermissible [epistemic claim].

• 6. There is no relevant difference between virtual murder and virtual paedophilia with respect to permissibility [ontological claim].

In the case of statements 4–6, each can be true and non-contradictory when taken together. It may well be that the state of nature is such that there is no moral difference (regarding permissibility) between virtual murder and virtual paedophilia. Likewise, it may well be that people intuit differently when it comes to the permissibility of virtual murder and virtual paedophilia (in keeping with 4 and 5). However, as statements 4 and 5 are not making claims about the state of nature per se, but are instead describing people’s epistemic relations to this alleged independent reality—in terms of what they intuit about it when it comes to the permissibility of virtual murder and virtual paedophilia (namely, that there is a lack of equivalence)—then there is no contradiction within statements 4–6 when taken as a whole, even if at least one of the intuitions stated is necessarily wrong.

Before continuing, I acknowledge that reference to ‘people’ in statements 4 and 5 is somewhat vague when it comes to the question of numbers (i.e., how many people intuit this: some, most, all? ). Let us therefore tighten things up a little:

• 7. More people intuit that virtual murder is permissible than do not.

• 8. More people intuit that virtual paedophilia is impermissible than do not.

• 9. There is no relevant difference between virtual murder and virtual paedophilia with respect to permissibility.

The claims made in 7 and 8 are empirically testable (in principle),⁴ although it is likely (as Ali, 2015, noted) that they would need to be given more context to be informative. While conceding this point (although I will return to it briefly it in Sect. 8), I am nevertheless confident of the respective truths of each of these statements (i.e., 7 and 8), even when construed broadly, as they are here. Certainly both statements are compatible with current gaming conventions (i.e., video games incorporating virtual murder are readily available, commercially, whereas video games incorporating virtual paedophilia are not).⁵ In the absence of empirical confirmation of statements 7 and 8, however, let us proceed on the assumption they are true, for the sake of argument.

Accepting, then, for the sake of argument, the truth of statements 7 and 8, we must next consider the implications of their relationship to the state of nature. Does the fact that more people intuit the permissibility of virtual murder than do not reflect the fact that the state of nature is such that virtual murder is permissible (i.e., the activity does not instantiate immoral properties)? Mutatis mutandis, the same can be asked of the intuition about the impermissibility of virtual paedophilia. If so, then do our (i.e., the majority’s) respective intuitions about virtual murder and virtual paedophilia signify a form of moral wisdom precisely because they reflect the state of nature?⁶ Moreover, if both intuitions reflect the state of nature, then statement 9, which purports to describe the equivalent permissibility of these two virtual actions, must be mistaken (recall the paradox within statements 1–3). Alternatively, if either intuition fails to reflect the state of nature (either because there is no moral difference between these two virtual actions, or because virtual murder is in fact impermissible and virtual paedophilia permissible), then their lack of alignment (for either of these reasons) means that at least one of our (the majority’s) intuitions, possibly both, is leading us astray, morally speaking.

Whether our intuitions about the permissibility of virtual murder and the impermissibility of virtual paedophilia reflect the truth of the state of nature, or fail to do so, depends of course on the state of nature. Therefore, what is needed is a way to justify claims about the state of nature in the form of a convincing argument either for moral equivalence, which would count against our intuitions, or a lack of equivalence, which would support them. (In relation to this last point, please note that, for the sake of brevity, when discussing the possibility of a state of nature in which there is a lack of moral equivalence between these two virtual activities, I will only consider one in which virtual murder is permissible and virtual paedophilia is impermissible, rather than vice-versa.)

As noted in Sect. 1, various attempts have been made to resolve the gamer’s dilemma by arguing against moral equivalence, thereby supporting our intuitions (I will discuss attempts to dissolve and resist the dilemma in Sect. 8). Such “attempted resolutions” (as I will hereafter refer to them) are suggestive of a state of nature in which virtual murder is permissible and virtual paedophilia is impermissible (although, often, they fail to make explicit any moral-realist sympathies they may hold and, instead, differ—explicitly, at least—only in terms of the reason(s) for a lack of moral equivalence (I will return to this point, shortly).

Bartel (2012), for example, argues that virtual murder differs from virtual paedophilia insofar as the latter eroticizes inequality. More recently, Coghlan and Cox (2023) have argued that acts of virtual paedophilia, unlike virtual murder, warrant (by way of a fitting emotional response) moral repugnance directed towards the self (as the agent of the virtual act): something they refer to as self-repugnance. Without going into the finer details of each of these attempted resolutions, my point is this: these authors (by way of examples) have each proffered an argument that is suggestive of a particular, albeit different, state of nature.

Of course, when it comes to what we intuit about virtual paedophilia (or indeed intuit about anything), our intuition may be expressed simply as “It feels wrong” or “It’s just not right”; it may lack articulation beyond this. The job of the moral theorist is to justify in a more articulate way why it feels wrong or is not right, and therefore why our intuition about virtual paedophilia is a form of moral wisdom, rather than a case of moral dumbfounding (i.e., when an individual remains committed to their intuition-based judgement despite being unable to articulate reasons for it—other than to say that “it’s wrong”—even in the face of evidence one might otherwise use to support the claim that it is not wrong, or at least not harmful; see Haidt, 2001). When articulating different reasons for a lack of moral equivalence (thereby adding a more articulate voice to our intuition), it may be that the state of nature is such that independent moral properties do exist but, at best, only one of the attempted resolutions picks them out accurately. Or it may be that many, if not all of them, do (i.e., they may all pick out different, independently existing, moral properties of virtual paedophilia). Whatever the reality, as intimated above, what is not clear from the way many of these attempted resolutions are presented is whether they are actually claiming that our intuitions reflect objectively existing moral properties of the world, (even though they are suggestive of such a view).

An exception is Luck (2022) who, as part of his grave solution, states that “graveness is… intended to be an objective property. That is, it is not determined (at least not completely) by what people think is serious or solemn” (p. 1300; emphasis added). If Luck is correct, then our intuition about the impermissibility of paedophilia reflects an independently existing—objective—property of nature which, in the context of virtual paedophilia, he associates with not taking wrongdoing seriously. Thus, while our intuition about the impermissibility of virtual paedophilia does not, and indeed need not, have anything to say about ‘graveness’, for Luck at least, this is what it aligns with.

Likewise, whatever the actual position of many of the remaining attempted resolutions, our intuitions may reflect the immorality of virtual paedophilia without articulating any of the reasons proposed for its wrongness by these different accounts. Further to Luck’s unambiguous approach, however, is constructive ecumenical expressivism (CEE) which adopts an entirely different—anti-realist—perspective that I will now present.

Detailing CEE’s expressivist and hybrid credentials

CEE is committed to a foundational principle of expressivism: namely, that when a subject declares “X is immoral” they are not picking out, and therefore describing, some independently existing moral feature of the world or state of nature (or what Chrisman, 2014, refers to as value properties) but, rather, expressing their disapproval of X. Unlike more traditional versions of expressivism, however (e.g., Blackburn, 1984; Gibbard, 1990; Hare, 1952)—and its forerunner, emotivism (see Ayer, 1946; MacIntyre, 2007; Stevenson, 1937)—CEE cannot be accused of reducing moral claims exclusively and exhaustively to an attitude or emotional outburst akin to (in the case of immorality) “I disapprove of X” or “Boo X!” Instead, in accordance with the hybrid nature of its closest predecessor—ecumenical expressivism (Ridge, 2006, 2007, 2009)—CEE holds that moral utterances have a cognitive component as well as a better known conative one.

According to CEE (and its predecessor, ecumenical expressivism), when uttered, the proposition “X is immoral” reveals two interrelated facts about the mental states of the subject (Carr, 2015). The first is that S disapproves of X. This conative component is characteristic of emotivism and the more traditional versions of expressivism that followed from it. In the case of virtual paedophilia, S’s assertion that the act is immoral indicates that she has a negative attitude towards it. The manner in which CEE differs from these other accounts, however, is what gives it its hybrid status. S’s negative attitude towards virtual paedophilia, it is more accurate to say, indicates not so much S’s disapproval of the virtual activity as it does a property that S believes is realized by the act of virtual paedophilia. The belief about P amounts to a second—cognitive—component that interrelates with the conative component in order to construct S’s negative attitude towards (in this case) virtual paedophilia. Thus, in declaring that virtual paedophilia is impermissible:

(CEE) S disapproves of P and believes that virtual paedophilia realizes P.

According to CEE, property P amounts to different things to different people. S₁ may view P in terms of negative utility (e.g., the realizing of more displeasure than pleasure; say, in the form of increased harm). S₂, on the other hand, may hold P to be a violation of God’s law, or constitutive of a failure in one’s duty to others. S₃, in turn, may characterize P as a vice rather than a virtue, and so on. Two people may therefore express their negative attitude towards X for different reasons. In the case of virtual paedophilia, it may be that:

(CEE_a) S₁ disapproves of p (where p equates to increasing the suffering of others) and believes that X realizes p.

S₁ has a negative attitude towards X because she believes that it realizes a property (p, in this case) of which she disapproves. Her negative attitude towards X is therefore the result of her disapproval of p and her belief that X realizes p. Mutatis mutandis, the same can be said of S₂, below:

(CEE_b) S₂ disapproves of q (where q equates to violating God’s law) and believes that X realizes q.

Before continuing, a point of clarification. When referring to an unspecified property believed to be realized by some generic subject (S), I have and will continue to use upper case P. When referring to a specific property that a particular subject (S₁, S₂, and so on) believes is realized by a particular action-type (e.g., X), I have and will continue to use a lower case letter (p, q, r, and so on). A different lower case letter is used to denote the (potentially) different properties different subjects believe X realizes.

Constructing a social norm

If CEE (along with all other species of expressivism) rejects the existence of independent moral properties, then what are our intuitions to align with if they are to act as a form of moral wisdom? According to CEE, they must align with socially constructed moral norms which are established through the force of consensus (hereafter ‘norms’). There is, however, more to this relationship than simple alignment, which I will come to in a moment. First, the construction of a norm is supported by the following a posteriori truth: Where a shared moral attitude occurs with regard to some action, as a society, we are able to create or construct a norm that then acquires its own objectified moral standard. This foundational claim finds support from Prinz (2007) when he states: “Things that we construct or build come from us, but, once there, they are real entities that we perceive” (p. 168).

Consensus does not require complete agreement, however. To explain: We have seen how both S₁ and S₂ have a shared negative attitude towards virtual paedophilia, expressed through the utterance “virtual paedophilia is immoral”. But we have also discussed how each subject has a different reason for their shared negative attitude. In Young (2015), I explain how consensus can be achieved through a certain kind of agreement, even when disagreement of a different kind (or at least the potential for disagreement) remains. To do this, I distinguish between the group’s shared de re attitude, and their different de dicto attitudes.

To state that A and B have a shared attitude towards x, such that they both hold that x is immoral, is to declare that they have the same de re attitude. When considering the act that A and B’s attitude is directed towards (the intentional object), their attitude towards that act (the thing in itself) is the same. But this shared de re attitude exits in virtue of the belief that x realizes some property (P) which they both disapprove of, but which can be (and is) different for A and B: A believes that x realizes p and B believes it realizes q. Their differing belief about which property is realized by x means that they have different reasons for their shared de re attitude. One could say that they have different de dicto attitudes regarding x… (namely, different beliefs about why it is immoral). (Young, 2015, pp. 317 − 18)

For S₁ (or A in the quotation above), X (e.g., virtual paedophilia) is immoral because she believes it realizes p (which, for S₁, represents the increased suffering of others) of which she disapproves. S₂ (or B), on the other hand, holds that X is immoral for a different reason. She believes that it realizes q (qua a violation of God’s law), of which she happens to disapprove. If we now include the numerous other society members who express the same negative attitude towards virtual paedophilia, we will likely discover that they hold this view for a variety of reasons. Some will share the same de dicto attitude espoused by either S₁ or S₂, but others will disapprove of X for some other reason (i.e., have an altogether different de dicto attitude): different from both S₁ and S₂, and even from many of the other members of their society.

What I am describing, here, is, I contend, a typical society (or community). In a typical society, a posteriori, the construction of a norm is the result of a shared (de re) attitude directed towards a particular type of action (in this case, virtual paedophilia). But this (typical) society, with its shared de re attitude, will in all likelihood comprise members with a variety of de dicto attitudes that scaffold the norm to which these attitudes relate, thereby making the norm more robust (I point I will return to in Sect. 6). Moreover, where a consensus is produced in a manner that conforms to CEE (which I am claiming will typically be the case), then the force of this consensus, which has created the norm in this way, justifies its objectified or intersubjective (hereafter, just ‘objectified’) standard. As a consequence, it can be claimed, with justification, that a particular (agreed) de re attitude is the one we (qua our society) ought to have, at least with regard to this object of moral inquiry (namely, the act of virtual paedophilia). Such a possibility can be likened to Copp’s (2011) realist expressivism (see also Copp, 2001) because it allows that where S shares the same attitude as the norm, she should be commended (by her society) for doing so, even if only tacitly. Equally, where she does not, it is appropriate—in accordance with the constructed norm’s objectified status (about which I will have more to say, shortly)—to rebuke her for her alternate (some within her society might even say deviant) moral attitude. This is because, given the norm’s status as the objectified standard by which this society operates, both the rebuke and a subsequent change of attitude on the part of S are said to be warranted (Nichols, 2008).

Applying CEE to attempted resolutions of the gamer’s dilemma

It is my contention that the various attempts at resolving the gamer’s dilemma, which typically involve explaining a shared position—namely, the moral status of virtual paedophilia⁷ and why this differs from virtual murder—are consistent with the means by which CEE accounts for the construction of norms (recall, I will discuss attempts to dissolve and resist the dilemma in Sect. 8). Applying CEE to the earlier presented examples of Bartel (2012), Coghlan and Cox (2023), and Luck (2022), we get:

• Bartel disapproves of p (the eroticization of inequality) and believes that X realizes p.

• Coghlan and Cox disapprove of q (that which elicits self-repugnance) and believe that X realizes q.

• Luck disapproves of r (not taking wrongdoing associated with a grave act seriously) and believes that X realizes r.⁸

According to CEE, each theorist is expressing the same negative de re attitude towards virtual paedophilia because they believe that the virtual activity realizes a property, albeit a different property—p or q or r, and so on—of which they disapprove. They have arrived at a consensus when it comes to their attitude towards virtual paedophilia; but they have done so for different reasons (i.e., they each have the same de re attitude but different de dicto attitudes).

The theorists, of course, might not consider that what they are doing is simply expressing a negative attitude towards X on account of some property they believe it realizes (of which they disapprove) but, rather, and minimally, describing through their utterance (that “X is immoral”) an independently existing property of X, and therefore describing some objective state of nature.⁹ CEE accepts that this may well be what they believe but, importantly, the reason given by an individual for their moral judgement—their de dicto attitude (as CEE calls it)—does not have to be true. Instead, CEE allows that the reason given by a particular individual for their shared de re attitude could be false. In fact, CEE allows and even anticipates that a number of the reasons contributing to the objectified norm may turn out to be false or problematic: that is, be based on beliefs that are eventually shown to be false and/or internally inconsistent. It even allows (and again anticipates) that different de dicto attitudes will be inconsistent with each other (e.g., a consequentialist and deontological reason for the immorality of an action).

Given CEE’s fundamental opposition to moral realism, where a reason is allegedly referring to an objective property (as is the case with Luck’s ‘graveness’), then, according to CEE, this claim is necessarily grounded on a false belief because, for CEE, objective moral properties do not exist. However, if what CEE takes to be a false belief is otherwise internally consistent, and has not yet been shown to be false (insofar as there is no evidence available to challenge the moral-realist assumptions on which it is grounded), or is unfalsifiable (e.g., if the reason is based on a belief in the existence and moral authority of God), and the holder of the belief is open to the possibility of belief revision (e.g., is not dogmatic), then it can contribute, along with other de dicto attitudes that satisfy these same requirements (hereafter the “Requirements”) to establishing the objectified norm. Moreover, such reasons can contribute even if they are inconsistent with each other. CEE therefore allows Luck’s ‘graveness explanation’ to form part of the scaffolding that supports the objectified norm regarding the immorality of virtual paedophilia because, even if some of these different attempted resolutions—these reasons for claiming that virtual paedophilia is immoral—are based on what CEE takes to be false assumptions, as long as they satisfy the aforementioned Requirements, they can still contribute to the construction of this particular objectified, rather than objective, component of reality.

To be clear, CEE does not require that the de dicto attitudes that contribute to the norm (the de re attitude) are consistent with each other (as noted), but they have to be internally consistent, and not contrary to evidence, if they are to make a legitimate (i.e., justifiable) contribution to the societal norm. Of course, consistency (internal or otherwise) and not being empirically refuted are not moral properties; but they are properties that justify reasons, and these (reasons) must be satisfied in order for one’s expression of approval or disapproval to be justified as an expression of one’s attitude towards X (i.e., in terms of the reason for the attitude), given the role that each de dicto attitude is playing, if justified (by satisfying the Requirements), in the construction of the norm of that society.

The way we should think about our intuitions (and, as always, I am talking about moral intuitions) is also different under CEE. If independent moral properties were to exist, then, as I have argued, our intuitions about virtual murder and virtual paedophilia would have to reflect these—and therefore the objective state of nature to which these properties belong—if they (our intuitions) are to signify moral wisdom. As a consequence, our intuitions would always be set apart from moral truths. The best they could do is ‘happen to reflect them’, rather than be a part of what makes them the truths they are. With CEE, however, our intuitions can be, and certainly will always have the potential to be, constitutive of that which establishes the de re attitude, and therefore the objectified norm. This is because, for those members of a society who share the same de re attitude as other members—about (in this case) the immorality of virtual paedophilia—but who do so solely in virtue of their intuition which (let us allow) they are unable to articulate beyond saying “It’s just wrong”, what they intuit amounts to their reason for their negative attitude, which is something they consider to be reason enough.¹⁰ Importantly, though, as far as CEE is concerned, their particular reason (de dicto attitude)—even if it takes the form of a laconically articulated intuition and nothing more—sits alongside all the other de dicto attitudes held by all the other members of the community who, together, by force of consensus, construct the objectified norm that virtual paedophilia is immoral (in virtue of their shared negative de re attitude). In accordance with CEE, then, intuitions should not be thought of as something always (and necessarily) set apart from the objectified norm but, potentially, one of an indefinite number of means to its construction.¹¹¹²

Given my last point, we can still say that an intuition (even a laconic one), like any other reason for one’s attitude towards X will either align or misalign with the objectified norm of a particular community. What CEE is claiming in addition to this is that, where it does align, its alignment contributes to the robustness of the norm, rather than merely coincides with it. If independent moral truths existed, then they would be true irrespective of whether our intuitions align with them; but they would not become more robust truths because of the alignment of our intuitions. In the case of objectified norms, however, while these can be established even if our intuitions fail to align with them, the robustness of the norm is nevertheless strengthened by the alignment because it adds yet another reason (albeit a laconically articulated one) for why the norm expresses what ought to be the case; in the same way that it is strengthened by any additional reason (i.e., any additional de dicto attitude) that satisfies the Requirements noted previously for the shared de re attitude that constitutes the norm’s content.

Given that we are assuming that there are no objective moral properties, the question of relativism must be addressed and, alongside it, the matter of how the robustness of the objectified norm, and therefore its claim to morality, is able to stave off any anticipated criticism of CEE’s relativist stance.

CEE as a form of robust relativism

The first specific challenge to CEE’s relativism that I wish to address comes from Ostritsch and Ulbricht (2021) who claim that “we should not take moral disagreement to mean that there are no objective moral facts independently of the attitudes that people have towards them” (p. 203). I agree. Thus, in the same way that two people disagreeing about whether the Earth is flat and free floating, or spherical and resting on the top of a giant turtle, does not prove that there are no independent facts about the Earth, so two people disagreeing about the moral status of X does not prove that independent moral properties do not exist.

CEE has never claimed that moral disagreement is proof of moral relativism; rather CEE simply assumes (without defence) the absence of independent moral properties and the truth of expressivism, and builds its case on these assumptions. In short, it has always taken the following position and sought to address the following issue: If one assumes that independent moral properties do not exist, and that moral utterances about X express the utterer’s attitude towards X, then (a) how are moral norms established, and (b), in light of CEE’s approach to (a), how can one moral norm be considered superior to another, given the relativism inherent within CEE?

It is my intention to support CEE without having to defend the anti-realist/expressivist assumption on which it is grounded, and therefore my intention to seek a resolution to the gamer’s dilemma with this assumption in place. To do this, I must defend CEE against a potentially problematic consequence identified by Luck (2019) when he states: “if you are unwilling to accept that slavery is permissible merely because the society it occurs within does not have the appropriate belief/attitude toward it (i.e. there is insufficient disapproval), CEE is not for you” (p. 37). In response to what might be for many a hefty price to pay for endorsing CEE, I will explain how the possible consequence Luck articulates, which is indeed a moral position CEE would allow, needs to be understood within the context of local, rather than global, normative claims: a distinction I will clarify in Sect. 7.

Dennison (2016), for his part, is willing to concede the possibility that CEE is a descriptively accurate account of how societies form their moral norms, and even that its robustness makes capricious changes to these norm less likely. However, under CEE, a norm’s resistance to whimsical change does not alter the fact that its constructed nature is based on, and therefore relative to, the attitudes of that society. Therefore, to avoid the pitfalls of what we might call traditional moral relativism (which does not permit one to prioritise one society’s norm over another’s), Dennison asserts that CEE must show either that it does not promote moral relativism, or explain why doing so is not necessarily a criticism of CEE.

That CEE promotes a form of moral relativism is not in doubt. However, CEE provides the means by which one society’s constructed norm can justifiably be judged as morally superior to another’s without having to fall back on the existence of independent moral properties. To understand how, consider the following scenario: Suppose communities C₁ and C₂ share a negative de re attitude towards virtual paedophilia. According to CEE, this means that respective members of each community disapprove of some property (P) that they believe virtual paedophilia realizes. Closer inspection of their various de dicto attitudes reveals, however, that those in C₁ have a greater diversity of reasons (de dicto attitudes) for their negative de re attitude than members of C₂. If we accept that there are no objective moral properties, and therefore no independently existing moral reality for the community’s norm or indeed any of the supporting de dicto attitudes to accurately reflect, then the difference between these two hypothetical communities supports my contention that a de re attitude shared by the majority of people within a given community, such that it becomes their constructed norm, is more robust if it is the product of a greater number of different de dicto attitudes (i.e., if it is based on a number of different reasons). And it is this robustness that not only reduces the likelihood of a whimsical change of attitude (and hence norm) but, in response to Dennison’s objection, justifies the normative authority of the objectified standard of that community with regard to the moral status of virtual paedophilia. A claim I will now defend.

All de dicto attitudes should be subject to the same level of scrutiny. How one views them after that, however, will depend on the outcome of that scrutiny. Thus, even if a de dicto attitude supports a negative de re attitude towards virtual paedophilia (and therefore an established norm to that effect), as noted already, its mere alignment with this norm does not necessitate that it (the de dicto attitude) is true, nor can it. This is because, if there are no objective moral truths—as CEE claims—and one’s reason stems from a belief that is said to be grounded on one of these alleged truths, then the belief and hence the reason cannot be true. However, recall that CEE does not demand that a reason be true, only that it stems from a belief or beliefs that are justified insofar as they currently satisfy the previously discussed Requirements, and are therefore open to revision.

To give an example of where this would fail to be the case, suppose some members of community C₁ have a negative de re attitude towards virtual paedophilia only because they hold that it represents an act that, if done for real, would violate the oath of fidelity in marriage and celibacy outside of it. If further inquiry revealed that these members had no problem with child brides, then this suggests that their issue is not with paedophilia per se and, by association, virtual paedophilia, but sex outside of marriage (which is how virtual paedophilia is always represented, as far as I am aware).¹³ Such a de dicto attitude against virtual paedophilia is easily shown to be internally inconsistent (i.e., is it inconsistent with regard to the clinical definition of paedophilia), and so cannot be justified as a reason for one’s negative attitude (i.e., if one could engage in virtual sex with one’s virtual child bride then, presumably, that act of virtual paedophilia would be considered acceptable). Therefore, despite resulting in the same de re attitude towards virtual paedophilia as other members of the community, closer scrutiny reveals why it should not contribute to the norm’s robustness. Such a rejection would be problematic for a norm based solely on this or a limited number of reasons for a negative attitude towards virtual paedophilia (i.e., an already less robust norm would be weakened further). Where a norm comprises many more reasons, however, CEE’s claim is that it is more likely to withstand the loss of one or more of its scaffolds (should they be shown to be unjustified, as was the case with the last example). Undermining the norm would therefore be harder to do, owing to its robustness; although, as I note in Young (2019): “[CEE] does not rule out a change of de re attitude—there is therefore a degree of fluidity inherent within [the approach]—but it does make any change less capricious” (p. 471).

Local and global normativity

Continuing my response to Dennison’s objection: CEE may be accused of favouring the majority view, although this is not strictly true. As I have discussed, CEE advocates for a more robust norm, rather than a greater number of voices, even though the former typically coincides with the latter. Therefore, if it seems as if CEE favours a majority view, then this is only because, typically, where more people subscribe to a norm, there is a greater diversity of reasons for this view within their number, thereby making the norm a more robust construction. Where this is not the case, however, CEE would not favour the majority view. Moreover, in accordance with CEE, where the majority in society A feel secure in their moral attitude because they have not been exposed (let us allow) to what they would recognize as sufficient evidence and/or argument to persuade them to revise their beliefs and therefore their reasons for having, say, a negative de re attitude towards X (whatever X happens to be), then, from their point of view, and therefore in accordance with their local normative standard, this is the attitude one ought to have. Those who oppose this view of X would therefore be wrong (relative to society A’s norm) to do so.

Society A satisfies the following requirements of CEE:

• (i) The constructed norm is based on the majority’s shared de re attitude which is grounded on a greater diversity of de dicto attitudes than other expressed de re attitudes.

• (ii) The majority members are not resistant to new evidence that they would recognize as contradicting or otherwise challenging their respective reasons (i.e., their diverse de dicto attitudes) for the de re attitude they hold.

• (iii) The majority members have not yet been exposed to the kind of evidence mentioned in (ii), nor has an internal inconsistency in any of the member’s respective de dicto attitudes been revealed to them in a manner that they are capable of recognizing.

• (iv) Each majority member is therefore presently secure in the reasons they have for their negative de re attitude towards X.

For society A, their local normative claim that “X ought not to be permitted because it is immoral” is justified relative to the specific level of exposure to evidence and reasoning outlined in points (i)-(iv). However, where a different society—society B—opposes the assertion that X is immoral, and does so based on a more robust set of reasons than are found in Society A (which, recall, holds the opposing view), then not only is society B’s local normative claim that “X ought to be permitted because it is moral” justified relative to the specific level of exposure to evidence and reasoning outlined in points (i)-(iv), but also justified is describing its norm as morally superior to A’s, whenever a comparison is made.

In situations where one is examining the justification for a moral claim within a particular society in isolation (so to speak), CEE provides a means for local normative claims to be justified based on the level of robust relativism found within that particular society. However, CEE also enables these local normative claims to be extended whenever norms are compared and there are justificatory grounds for declaring one society’s norm superior to another’s. In such a situation, a global normative claim can be justified that is still relative in its scope, rather than absolute or in some other way fixed. A global normative claim is relative to the competing (although indeterminate number of) norms under scrutiny, and subsequently justified relative to the most robust de re attitude from among those available for comparison. It remains fluid, however (as noted above), and therefore subject to change relative to the effects of exposure to future constructed norms. It is for this reason that, in a society whose members are open to new ways of improving existing beliefs (i.e., whose de dicto and de re attitudes are recognized as defeasible), CEE would require a change to the morally inferior norm (e.g., society A’s attitude towards X), otherwise, it would be vulnerable to the charge that it was grounded on dogmatism or wilful ignorance.

One might object, however, that such a description of normative claims (i.e., local versus global) could potentially—quite easily, in fact—produce the following contradiction: Where X is considered in relation to society A in isolation, relative to that society, the norm that X is immoral is justified, but when considered in a wider context (i.e., when compared with society B), the same norm is said to be unjustified. In reply, I would say that this is indeed the case because what is being examined, here, is the justification for the norm, and not its truth (qua its independent reality). When viewed in the context of society A alone, relative to that society and the robustness of the de re attitude that constitutes the norm, it is justified. To be clear: justified, not true. When the context changes, however, then relative to the new context, the norm (as it turns out, here) is not justified. The justification for what ought to be the case, as expressed by a particular norm, is therefore relative to the robustness of the de dicto attitudes that construct that norm (that de re attitude). As the relativity changes (and, of course, one’s awareness of this changes), so too (potentially) does the justification.

Applying CEE to attempts to dissolve and resist the dilemma

Attempts to dissolve the dilemma have challenged the lack of context present in the original formulation of the gamer’s dilemma, and therefore whether it is in fact the case that we do intuit that all cases of virtual murder are permissible and all cases of virtual paedophilia are impermissible. Ali (2015), for example, argues that whether we intuit that either of these virtual activities is permissible depends both on the type of video game or other type of online environment we are discussing and the context in which it occurs.¹⁴ Nader (2020) makes a similar point in the content of esports.¹⁵ For Ramirez (2020), however, whether virtual murder and/or virtual paedophilia is considered immoral depends on the extent to which the activity produces a virtually real experience. (i.e., an experience equivalent to that which one would experience if engaging in the act for real). Where the activity risks producing such an experience, Ramirez argues, it should not be permitted.

Attempts to dissolve the dilemma have been of benefit to the ongoing debate because they have made those interested in how best to respond to the dilemma aware of the importance of context; although specifying this does not necessarily dissolve the dilemma; rather, it simply limits those contexts in which it occurs (Luck, 2018). That said, for CEE, such attempted dissolutions, like attempted resolutions, conform to the way we construct norms, even when refining the context. Each attempted dissolution can therefore be mapped onto CEE’s underlying structure. Thus, where S declares that “X is immoral when played in that video game but not this one”:

(CEE_c) S disapproves of P, and believes that X realizes P in context C1 but not in context C2.

Using Ramirez’s (2020) explanation as an example, this means that where S disapproves of P—the production of virtually real experiences in the context of virtual enactments of actual immoral activities (such as actual murder and actual paedophilia)—and believes X realizes P in context C1, then S will hold a negative attitude towards X in this context but not in C2, because S does not believe that P is realized by X in this latter context (irrespective of which of the two target virtual activities X represents). Where enough people share this contextualized de re attitude—although, again, not necessarily for the same reason (e.g., following Luck (2022), some may believe that X fails to take the wrongdoing represented by the virtual activity seriously in C1 but not C2, and so on)—a norm will be established that is context dependent. The robustness of the norm will, of course, depend on the number of other viable reasons available to scaffold the de re attitude on which the norm is based.

CEE is also able to integrate Montefiore and Formosa’s (2022) attempt to resist the dilemma into its own expressivist approach. In brief, Montefiore and Formosa argue that our contrasting intuitions about virtual murder and virtual paedophilia, despite their prima facie appearance, are not in fact moral intuitions. Instead, they reflect certain non-moral properties, more akin to aesthetics (e.g., S intuits that X is impermissible because it is in poor taste) or psychological biases (e.g., learned associations with innate disgust responses; Haidt, 2001). Our intuitions about these virtual activities do not therefore contradict the claim that there are no moral differences between virtual paedophilia and virtual murder, because our intuitions are not about moral properties. The converse must also be true, of course: If a moral difference is established, then our intuitions do not reflect this difference because, to reiterate, they are not about moral properties.

The idea that there are moral properties of actions and events which, in the case of those relating to the gamer’s dilemma, our intuitions happen not to reflect suggests that these properties exist independently of our attitude towards them. But for CEE, it is precisely our approbation or disapprobation of these properties (to borrow from Hume) and therefore the actions/events from which they are instantiated that makes them—because of our attitude, and therefore relative to us—moral or immoral properties, and subsequently moral or immoral actions/events.

Recall that I argued that our intuitions are able to contribute to a constructed norm. For this norm to be robust, however, it must comprise an indefinite number of reasons (the greater the number of distinct reasons that satisfy the Requirement, the more robust the norm). Therefore, if our intuitions are to have any normative force, they must acquire this from their alignment with, and therefore their contribution to, the constructed norm, not independently of it (i.e., individual reasons contribute to the norms overall normative strength—its robustness—which they are then able to draw from). And, as the objectified norm concerns what should be permissible or impermissible, its normativity is moral (albeit socially constructed) in nature. The fact, then, that I disapprove of X because it disgusts me (for example), draws its normative strength (regarding the permissibility of X ) not from the merits of this reason alone, but from of the fact that this reason aligns with many other reasons in producing a socially constructed moral norm.

Conclusion

I have sought to support and clarify both CEE and my previous use of it to resolve the gamer’s dilemma. In the former case by arguing that, in the absence of independently existing moral properties (an anti-realist stance fundamental to CEE and other forms of expressivism), robust relativism provides the means by which a society’s constructed norm acquires its normative credentials such that, if more robust, it can justifiably claim its own norm is morally superior to another’s. In addition, by demonstrating how the respective approaches taken by the various other attempts to resolve, dissolve or resist the dilemma can be incorporated within, and explained by, CEE’s approach to morality, which includes describing how our intuitions are able (along with an indeterminate number of other reasons) to contribute to the construction of a society’s norm, I have defended CEE attempted resolution of the gamer’s dilemma: namely, there is a moral difference between virtual murder and virtual paedophilia in virtue of different constructed norms about these distinct virtual activities, which our intuitions align with and, in so doing, help scaffold.