According to Baumard et al., cooperative behaviour is often impartial and fair. In particular, it exhibits a “common logic” of proportionality. This is most extensively discussed in relation to economic games. The authors argue that the behaviour elicited by such games is neither selfish nor altruistic, but governed by fairness. The initial suggestion is that the fair distribution of resources is proportional to the participants' contribution, where contribution is a relatively quantifiable function of effort and talent. However, this notion of contribution is replaced with that of “entitlement” or “right.” Dictators will give money to the extent that the recipient is deemed to be entitled to it. For example, children may exhibit miserly behaviour in economic games because they see themselves as fully entitled to whatever resource they are given. Similarly, cross-cultural variability in dictator allocation is attributed to different ways of understanding the respective rights or entitlements of the agent and recipient.
We are convinced by Baumard et al.'s basic account of the participants' behaviour; our concern is that nothing significant about fairness follows from this. Crucially, since what people consider to be a right or entitlement can vary (no principled limit or restriction on what can fall under this category is offered), the claim that a fair distribution will be proportional to the rights of participants is empty. It has no real predictive force since in any given scenario subjects can take radically different views of the rights in question. Of course, it may well be correct to say that what a particular person deems fair will be proportional to what rights she accords to relevant persons. However, since the mere idea of a right or entitlement implies that certain responses are fair or morally fitting, this sounds more like a virtual truism than a claim of any empirical significance. This is the core of our worry: that, far from being a surprising fact about human cooperation, the logic of proportionality turns out to be very meagre indeed, possibly deriving simply from the idea of fairness. It hardly seems like a problem to which an evolved sense of fairness is an adequate and illuminating solution.
Consequently, the existence of impartial proportionality does not appear to be a sufficient reason to postulate an evolved sense of fairness. Because it cannot explain or predict the behaviour that most people consider to be fair in particular circumstances, it fails the test of descriptive adequacy. Moreover, this account leaves open the possibility that people simply have different, culturally determined, and perhaps even mutually incompatible senses of fairness. In this case, there would be nothing much to say about how the mechanism for acquiring a sense of fairness evolved.
One way to circumvent this conclusion would be to show that the domain of fairness, manifest not only in peoples' behaviour but also in their considered moral judgements, is inherently structured and governed by non-trivial, substantive principles, the acquisition of which cannot be easily explained by appealing to individual experience. It is the search for such principles, the challenge of descriptive adequacy in the moral domain, which has underpinned the framework of Universal Moral Grammar. One basic insight is that human moral intuitions extend well beyond what can be derived from a mere proportionality principle or other similar formal principles of moral judgement, insofar as there appear to be significant generalisations about rights, duties, fairness, and morally acceptable conduct that these principles cannot predict. For example, most natural human moral systems appear to be deontic in their basic structure and to depend on distinctions such as act versus omission, mistake of norm versus mistake of fact, and intended versus foreseen effects, which appear to emerge early and reliably in child moral development (Mikhail Reference Mikhail2011). None of this is to deny the existence of substantial cross-cultural variation, only to suggest that this variation may be sharply limited. It is this kind of principled limitation to the domain – something hostage to further empirical inquiry – which can justify postulating an evolved sense of morality. Only in this context does it seem appropriate to ask the “ultimate how question”: How did morality evolve in the species, and what selective forces or other causes were responsible for this evolution?
Prioritising a structural account of the moral domain in this way would mirror the way inquiry has unfolded in the linguistic domain, where decades of research on the structural richness of languages aiming at descriptive adequacy have preceded the current claim that Universal Grammar “primarily constrains the ‘language of thought’” (Chomsky Reference Chomsky2007, p. 22; cf. Hinzen, in press; Kirkby & Mikhail, in preparation). This claim now also feeds into and constrains evolutionary theorizing in unforeseen ways, which presuppose 50 years of descriptive work in linguistic theory. It seems doubtful that research into the moral sense can bypass this stage.
According to Baumard et al., cooperative behaviour is often impartial and fair. In particular, it exhibits a “common logic” of proportionality. This is most extensively discussed in relation to economic games. The authors argue that the behaviour elicited by such games is neither selfish nor altruistic, but governed by fairness. The initial suggestion is that the fair distribution of resources is proportional to the participants' contribution, where contribution is a relatively quantifiable function of effort and talent. However, this notion of contribution is replaced with that of “entitlement” or “right.” Dictators will give money to the extent that the recipient is deemed to be entitled to it. For example, children may exhibit miserly behaviour in economic games because they see themselves as fully entitled to whatever resource they are given. Similarly, cross-cultural variability in dictator allocation is attributed to different ways of understanding the respective rights or entitlements of the agent and recipient.
We are convinced by Baumard et al.'s basic account of the participants' behaviour; our concern is that nothing significant about fairness follows from this. Crucially, since what people consider to be a right or entitlement can vary (no principled limit or restriction on what can fall under this category is offered), the claim that a fair distribution will be proportional to the rights of participants is empty. It has no real predictive force since in any given scenario subjects can take radically different views of the rights in question. Of course, it may well be correct to say that what a particular person deems fair will be proportional to what rights she accords to relevant persons. However, since the mere idea of a right or entitlement implies that certain responses are fair or morally fitting, this sounds more like a virtual truism than a claim of any empirical significance. This is the core of our worry: that, far from being a surprising fact about human cooperation, the logic of proportionality turns out to be very meagre indeed, possibly deriving simply from the idea of fairness. It hardly seems like a problem to which an evolved sense of fairness is an adequate and illuminating solution.
Consequently, the existence of impartial proportionality does not appear to be a sufficient reason to postulate an evolved sense of fairness. Because it cannot explain or predict the behaviour that most people consider to be fair in particular circumstances, it fails the test of descriptive adequacy. Moreover, this account leaves open the possibility that people simply have different, culturally determined, and perhaps even mutually incompatible senses of fairness. In this case, there would be nothing much to say about how the mechanism for acquiring a sense of fairness evolved.
One way to circumvent this conclusion would be to show that the domain of fairness, manifest not only in peoples' behaviour but also in their considered moral judgements, is inherently structured and governed by non-trivial, substantive principles, the acquisition of which cannot be easily explained by appealing to individual experience. It is the search for such principles, the challenge of descriptive adequacy in the moral domain, which has underpinned the framework of Universal Moral Grammar. One basic insight is that human moral intuitions extend well beyond what can be derived from a mere proportionality principle or other similar formal principles of moral judgement, insofar as there appear to be significant generalisations about rights, duties, fairness, and morally acceptable conduct that these principles cannot predict. For example, most natural human moral systems appear to be deontic in their basic structure and to depend on distinctions such as act versus omission, mistake of norm versus mistake of fact, and intended versus foreseen effects, which appear to emerge early and reliably in child moral development (Mikhail Reference Mikhail2011). None of this is to deny the existence of substantial cross-cultural variation, only to suggest that this variation may be sharply limited. It is this kind of principled limitation to the domain – something hostage to further empirical inquiry – which can justify postulating an evolved sense of morality. Only in this context does it seem appropriate to ask the “ultimate how question”: How did morality evolve in the species, and what selective forces or other causes were responsible for this evolution?
Prioritising a structural account of the moral domain in this way would mirror the way inquiry has unfolded in the linguistic domain, where decades of research on the structural richness of languages aiming at descriptive adequacy have preceded the current claim that Universal Grammar “primarily constrains the ‘language of thought’” (Chomsky Reference Chomsky2007, p. 22; cf. Hinzen, in press; Kirkby & Mikhail, in preparation). This claim now also feeds into and constrains evolutionary theorizing in unforeseen ways, which presuppose 50 years of descriptive work in linguistic theory. It seems doubtful that research into the moral sense can bypass this stage.