To address the shortage of cross-cultural research on putative, panhuman features of moral judgement, Fessler et al.  conducted a study with samples drawn from seven different societies. There is much to be praised in their efforts, which advance the recent debate initiated by Kelly et al.  regarding whether people view harmful transgressions as independent of authority (wrong regardless of the view of any legitimate authority) and universally wrong (wrong in all places and times), as argued by Turiel and his colleagues over the last four decades, and by ourselves in the context of this debate [3–6].
Fessler et al. claim that people do not conceive harmful transgressions as authority independent and universally wrong because people's third-party moral judgements evolved to ‘increase individual fitness within local culturally constructed social arenas’, which implies that their judgements should be parochial: they should not be sensitive to wrongdoings distant in space and time and they should be sensitive to the opinion of local authorities. Moreover, Fessler et al. claim that their new study supports their moral parochialism hypothesis, providing a ‘powerful challenge’ to positions like Turiel's and ours. Here, we argue that Fessler et al.'s findings can be interpreted in a way that is quite consistent with our position, and that provides instead a powerful challenge to their evolutionary, moral parochialism hypothesis.
We [3,4] entered the aforementioned debate by offering an empirically guided methodological critique of the research of Kelly et al. We also proposed a deflationary reformulation of Turiel's original hypothesis in which harmful transgressions are understood as authority independent and universally wrong when they are perceived to involve injustice and basic-rights violations [5,6]. One major criticism we had of Kelly et al.'s study is that it presented participants cases of harmful actions, such harm as military training or as punishment, that many participants viewed as justifiable [3,4]. Fessler et al. have taken a large step in addressing this earlier criticism by employing vignettes depicting harmful actions that appear to involve ‘clear and substantial harm, violations of rights and/or injustice’. Indeed, in this respect, their new study provides an excellent test of our hypothesis.
Their study included seven cases of harm ostensibly involving injustice, such as a woman being raped or a man battering his wife without provocation. They presented participants with such cases, and assessed their moral judgements of the harmful acts on a 5-point badness/goodness scale: ‘How good or bad is what A did?’ (‘Extremely bad’; ‘Bad’; ‘Neither good or bad’; ‘Good’; ‘Extremely good’). After answering this first question, participants were provided with probes concerning authority dependence, temporal distance and spatial distance (for details, see ). For each of these questions, participants were assessed again with the same 5-point badness/goodness scale. The aim of the task is to probe whether participants will change their initial judgements of wrongdoing, given the approval of a local authority or the fact that the action occurred in a distant time or place.
In the context of the task, evidence consistent with our hypothesis are instances where a participant initially thinks the harmful acts are wrong (i.e. ‘Extremely Bad’ or ‘Bad’), and then does not reverse their position to not-wrong (i.e. ‘Neither good nor bad’, ‘Good’ or ‘Extremely good’) following the authority, temporal and spatial distance probes. If the majority of responses involve retention of the initial judgement of wrongdoing, this would be consistent with our account, but not with the parochialist account. Fessler et al. do not describe their results in a manner that could test our hypothesis, i.e. that present the amount of responses that involve non-reversals of the initial judgement of wrongdoing. Table 1 presents such a breakdown of Fessler et al.'s results. As can be seen, the vast majority of responses from all seven field sites involve non-reversals of the initial judgement of wrongdoing, and this was true across all three probes.
Obviously, there is still a non-negligible minority that did reverse their initial judgement of wrongdoing. However, there are several ways of explaining these minority responses that are compatible with our hypothesis [3–6]. For example, it is possible that, in response to the authority's approval of the act, participants inferred that the authority possessed some deeper insight about the event (e.g. additional reasons why the man slapped his wife), which led them to reappraise of the injustice of the act. As the authors did not measure the perceived injustice of the act before or after the presentation of the authority dependence probe, it is unclear whether participants who reversed their judgement also changed their construal of the injustice of the event.
When Fessler et al.'s data are viewed in this alternative manner, it becomes apparent that moral-parochialist responses represent a tiny minority cross-culturally. This provides a powerful challenge to their evolutionary argument about moral parochialism, since one cannot support an evolutionary argument about the nature of moral judgements with a cross-cultural minority. Instead, their findings become more consistent with an alternative evolutionary hypothesis, based on mutualism , which argues that intuitions about authority independence and universalism follow from the panhuman disposition to think in terms of reciprocal social contracts that obligate people to respect the basic interests of others by not selfishly harming one another .
However, Fessler et al. may reply that their perspective is predicated on the idea that people's moral judgements cannot be dichotomized in terms of judging that an action is wrong (i.e. ‘Extremely bad’ and ‘Bad’) or not-wrong (i.e. ‘Neither good nor bad’, ‘Good’ and ‘Extremely good’) as we did in our interpretation of their results, since from their perspective these judgements should be understood in terms of a ‘graded continuum’ of condemnation . Moreover, they may argue that their statistical analysis shows that the authority dependence, temporal and spatial distance factors explain a substantial amount of the graded reduction of condemnation in participants' judgements when you take into account the entire 5-point scale.
We are sceptical about modelling normative judgements simply on a graded continuum (in terms of psychological validity) . We would argue that it is plausible to suppose that participants parse the 5-point badness/goodness scale categorically in terms of the act being wrong or not-wrong. We do not see much psychological significance in shifts from ‘Extremely bad’ to ‘Bad’ in the context of their scale. This is supported by the fact that a comparable number of responses increased in their degree of condemnation as that decreased in their degree of condemnation, among those responses that retained an ‘Extremely bad’ or ‘Bad’ judgement: in this group, averaging across the three probes, 10% shifted from ‘Bad’ to ‘Extremely bad’, while 16% shifted from ‘Extremely bad’ to ‘Bad’ (74% retained the same level of badness).
Even setting aside this conceptual issue, we argue that a graded-continuum approach to the data would still challenge their evolutionary hypothesis. To support their evolutionary hypothesis, they would have to show that the majority of responses in most field sites reduced the initial judgement. Yet, again, Fessler et al. do not describe their results in a way that addresses the issue of majority responses. Table 2 presents the amount of responses that did not reduce in condemnation as a result of the authority dependence, temporal and spatial distances probes, either because the initial badness judgement was maintained (e.g. ‘Bad’/’Bad’) or because there was an increase in the level of condemnation (‘Bad’/’Extremely bad’). In other words, this table represents the amount of responses that are inconsistent with the parochialism hypothesis, under the graded-continuum approach.
As can be seen, even when accepting the graded-continuum approach, the clear majority of responses in all field sites, and across all three probes, are inconsistent with the parochialism hypothesis. These results again provide a powerful challenge to their hypothesis about the evolution of parochial morality, as such a claim depends on showing that most people across societies are inclined to reduce their condemnation of harmful acts when a local authority approves or the actions occur in another place or time.
In sum, although we praise Fessler et al.'s use of cross-cultural samples to test competing models of moral judgement, we question their analysis of the data. When analysed in the manner we outlined here, their findings become consistent with our theoretical proposal , not with theirs . We leave it to the broader scientific community to decide which is the most appropriate approach to their data.
The accompanying reply can be viewed at http://dx.doi.org/10.1098/rspb.2015.2628.
- Received August 23, 2015.
- Accepted October 8, 2015.
- © 2016 The Author(s)
Published by the Royal Society. All rights reserved.