What We Talk About When We Talk About Morality

Abstract

Morality is inherently social, yet much extant work in moral psychology ignores the central role of social processes in moral phenomena. To partly address this, this article examined the content of persuasive moral communication—the way people justify their moral attitudes in persuasive contexts. Across two studies, we explored variation in justification content (deontological, consequentialist, or emotive) as a function of moral foundations. Using justification selection techniques (Study 1) and open-ended justification production (Study 2), results demonstrate a preference (a) for deontological appeals in justifications for the sanctity foundation, (b) for consequentialist appeals for the individualizing foundations (care and fairness), and (c) for emotive appeals in justifications for the binding foundations (loyalty, authority and sanctity). The present research questions the generality of inferences about the primacy of emotions/intuition in moral psychology research and highlights the important role of reasons in persuasive moral communication.

Keywords

moral justifications deontology consequentialism emotion persuasive appeals moral foundations

Communication is fundamental to morality. Moral attitudes are learned, shaped, maintained, and changed via communicative practices, such as gossip (e.g., Feinberg, Willer, Stellar, & Keltner, 2012). Furthermore, communication plays a regulatory role in the moral domain: We communicate with others primarily to convey social information, to seek advice, and to police moral violations (Dunbar, 1995; Dunbar, 2004; Foster, 2004; Gibbard, 1990; Peters & Kashima, 2014). In addition, communication is necessary for organizing collective moral action, as demonstrated in studies that show increased cooperation when participants are allowed to communicate with each other in the resolution of resource dilemmas (e.g., Ostrom, Gardner, & Walker, 1994).

Despite being central to recent theories (e.g., DeScioli & Kurzban, 2013; Haidt, 2001), moral communication has been relatively neglected in empirical research. Most extant work focuses instead on intra-individual judgment processes (see Haidt & Kesebir, 2010, for a review). Although early work in moral psychology did elicit people’s reasons for their moral attitudes (e.g., Kohlberg, 1969), recent research, based on the assumption that such reasons are post hoc rationalizations, have moved away from reasons and reasoning as indicators of underlying psychological processes. In the present work, we focus on moral reasoning and communication as phenomena in themselves, not as windows into underlying psychological mechanisms. In two studies, we attempt to address this lack by exploring the way people engage in persuasive communication about moral violations across a variety of moral domains.

Moral Communication

DeScioli and Kurzban (2013) argue that moral judgments function as signals to coordinate side taking and reduce the frequency of costly conflicts. In other words, expressing a moral judgment signals to some third party that one has taken a particular stance on a moral issue or toward a moral actor. This allows the third party to side with the judge (or not, as the case may be) for support and solidarity and potential group-enforced punishment of a perpetrator.

Although not emphasized by DeScioli and Kurzban (2013), language and verbal communication (e.g., moral justifications) are important components of moral signaling. Language can be used creatively to challenge, confront, and even change our own moral beliefs (Pizarro, Detweiler-Bedell, & Bloom, 2006), or more importantly the beliefs of others (Mercier, 2011; Mercier & Sperber, 2011). Mercier (2011a) proposes that the primary function of reasoning (including moral reasoning) is argumentation (see also Mercier & Sperber, 2011). He posits that reasoning is not primarily for amending first impressions or developing beliefs (largely intra-individual functions), but rather for producing persuasive arguments to justify our beliefs to others (largely inter-individual functions).

When justifying moral beliefs or attitudes to others, different kinds of moral appeals may be used to support one’s moral position and to communicate one’s moral beliefs. It is this question—How do people talk about morality?—that forms the heart of the present research. To get a sense of the kinds of appeals open to people in the course of moral communication, we begin by considering how morality has been conceptualized in philosophical discourse.

Kinds of Moral Appeals

Moral philosophers in the modern era have generally relied on one of three kinds of reasons (or appeals) in their attempts to justify or describe normative moral systems: deontological, consequentialist, or emotivist. Deontology is a rule-based ethic that is derived from the Greek words for duty (deon) and the science or study of reason (logos). This system holds that certain actions are required (or permissible) versus forbidden, based not on the consequences of those actions but on the inherent wrongness of the acts themselves (Alexander & Moore, 2015; Kant, 1785/1993). As deontological ethics focus on moral obligations and adherence to moral principles and prescribed rules, deontological appeals are, therefore, appeals to duties, principles, and rules (e.g., “This kind of behavior goes against the rules we should live by”).

Consequentialism, on the other hand, emphasizes moral rightness or wrongness based on the outcomes or consequences of actions and contends that people should strive to maximize positive outcomes (Sinnott-Armstrong, 2014). The most well-known variety of consequentialism is utilitarianism. Utilitarian¹ philosophers, most notably Bentham (1789/1970) and Mill (1861/1998) argue that the morally right action is the one that brings about the most happiness. For Bentham, actions are not intrinsically right or wrong; rather, moral valence is determined on the basis of the effects of an action. Consequentialist appeals thus allude to outcomes, consequences, or effects and may incorporate notions of the greater good (e.g., “It’s wrong because it caused other to suffer”).

Philosophers such as Hume (1739-1740/1969), Ayer (1936/1946), Ogden and Richards (1923), and Stevenson (1944) propose that moral judgments are primarily, or nothing but, mere expressions of emotions, a position reflected in the spirit of Haidt’s (2001) social intuitionist model. For Stevenson especially, an attempt to change another person’s moral attitudes or beliefs requires a “nonrational” method, which is dependent on the emotional meaning of words and the emotional impact of these meanings on the listener. One option for people trying to justify their moral beliefs is to be lay emotivists—explicitly conveying the emotional substratum of their moral positions in attempts to persuade (e.g., “It’s wrong because it’s disgusting”). In such emotive appeals, the justification is nothing more than the moral-emotional reaction elicited by the stimulus.²

Content of Justifications: Empirical Support

Not only have moral philosophers sought justifications in these three broad approaches, but recent empirical research in moral psychology also suggests that people tend to spontaneously adopt these kinds of justifications when asked to ground their moral judgments. Piazza and Sousa (2014), for example, asked participants to justify why they considered two “harmless,” consensual violations (e.g., consensual cannibalism and consensual sibling incest) to be impermissible. In open-ended responses, the authors uncovered three categories of justifications that accounted for the majority of variation in responses: deontological appeals (accounting for the majority, 65.3% of responses), appeals to negative outcomes (12.8% of responses), and appeals to emotion (21.9% of responses).

Aguiar, Brañas-Garza, and Miller (2008) looked at the content of justifications for why participants chose to donate or keep money in an experiment in which they were given the option to distribute money between different charities or to keep the money for themselves. Seven categories of reasons for donating or keeping the money were identified, including consequentialism, deontology, randomness, lack of trust in the charities or experiment, legitimacy, egoism/hardship, and goodness. In this study, the majority of explanations appealed to outcomes, while deontological reasons (what should be done) were next frequent. Interestingly, no emotive reasons were given by participants for donating or keeping the money.

Although broadly supportive of the prevalence of deontological, consequentialist, and emotive appeals as prevalent types of moral justifications, these studies examined justification content in the context of a limited range of moral situations. Although people may use more consequentialist than deontological language when justifying donation behavior, for example, does such a pattern extend to other situations in which the alleviation of suffering (which is a primary purpose of charitable donation) is less salient?

The present research will address this shortcoming by systematically exploring justification content across a diverse range of moral contexts. To do this, we draw upon Moral Foundations Theory (MFT), one of the most expansive theories of moral content in contemporary moral psychology.

MFT

MFT (Graham et al., 2013; Haidt, 2012; Haidt & Graham, 2007; Haidt & Joseph, 2004) posits that there are five³ cognitive modules (or foundations) that drive moral intuitions: care/harm, fairness/cheating, loyalty/betrayal, authority/subversion, sanctity/degradation. Each of these modules is sensitive to a different set of triggers and gives rise to characteristic emotional responses. Although conceived of as distinct, modular mechanisms, much research has found that these five foundations cluster under two higher order groupings (individualizing and binding), which tend to more parsimoniously capture much of the variance in people’s moral attitudes (e.g., Laham & Corless, 2016; van Leeuwen & Park, 2009; Wright & Baril, 2013). The individualizing foundations, including care and fairness, focus on the rights and welfare of the individual, while the binding foundations, including loyalty, authority and sanctity, have been argued to focus on the cohesion of the group (Graham, Haidt, & Nosek, 2009).

To date, MFT has been applied to an abundance of empirical research, in a variety of different contexts, and in service of different research aims (see Graham et al., 2013, for a review). Despite ambiguities about the precise status of foundations (see Suhler & Churchland, 2011), it is clear that there is pragmatic utility in using MFT to systematically organize variability in moral content. We thus use the theory here.

The Current Studies

The present research investigates the content of persuasive moral justifications for three kinds of appeals (deontological, consequentialist, and emotive) as a function of moral foundation. Study 1 explores whether people choose different kinds of justifications when asked to justify their moral attitudes by selecting from three justification options. Study 2 provides a more realistic exploration of the content of moral communication by analyzing participants’ free-response justifications for deontological, consequentialist, and emotive language use.

Analytic strategy

As this is the first systematic attempt to characterize variability in justification content across moral foundations, we used an analytic approach that sought to maximize the chances of finding meaningful variation, while minimizing the likelihood of false positives. To this end, we approached analyses in both studies by first exploring variability in justification use between the broad, higher order clusters of individualizing (care, fairness) and binding (loyalty, authority, sanctity) clusters, and then exploring variation in justification use within each of these clusters using orthogonal contrasts. By prioritizing the individualizing versus binding difference, we capture that dimension of moral foundations variation that has explained the majority of variance in moral judgments both between and within cultures (e.g., Graham et al., 2009; Graham et al., 2011). By then comparing foundations within these broad clusters (thus producing a set of orthogonal contrasts), we more fully and specifically characterize variability in justification use, while controlling the type I error rate.

In addition, given the recent interest in moral psychology in the distinction between care and sanctity (e.g., Chakroff, Dungan, & Young, 2013; Chakroff & Young, 2015; Gray & Keeney, 2015; Gray, Schein, & Ward, 2014; Haidt, Koller, & Dias, 1993; Piazza et al., 2013; Rottman, Kelemen, & Young, 2014; Russell & Giner-Sorolla, 2011a, 2011b, 2011c; Russell & Piazza, 2015; Young & Saxe, 2011), we complemented this general analytic approach with targeted comparisons assessing justification use between these two foundations.

Hypotheses

In the context of this general approach, we tentatively hypothesized the following. First, we predicted more frequent appeals to consequentialism for the individualizing versus the binding cluster. Like the individualizing foundations, consequentialist appeals focus on outcome-based social justice concerns, including welfare maximization and harm minimization (e.g., Bentham, 1789/1970; Mill, 1861/1998). Based on the centrality of care/harm and social justice concerns to both consequentialism and the individualizing foundations (Haidt & Joseph, 2004), we expect to see a preference for consequentialist appeals in justifications of care and fairness violations. This is consistent with the findings of Aguiar et al. (2008) who showed that consequentialist appeals were dominant in justifications of donation decisions in the dictator game. For similar reasons, we expect more consequentialist justifications for care than sanctity, a prediction consistent with the relatively low frequency of consequentialist appeals for harmless sanctity violations found in Piazza and Sousa (2014).

Second, we expected emotive appeals to be greater in binding than the individualizing cluster. Although MFT posits that each foundation is characterized by a particular set of emotions (e.g., anger for care and fairness; Graham et al., 2013), we argue that the nature of the emotions typically elicited by the individualizing versus binding foundations differently lend themselves to elaborated, nonemotional justification.

Work by Russell and Giner-Sorolla, for example, shows that people are more able to generate elaborated justifications (those that reference “causes and consequences of the moral violation,” p. 637, Russell & Giner-Sorolla, 2011c) for experiences of anger than they are for disgust; when one is asked to justify a moral reaction grounded in disgust, cognitively elaborated reasons are not accessible, and one tends to fall back on unelaborated emotional justifications (a finding consistent with moral dumbfounding; Haidt et al., 1993). Given that disgust has been shown to be associated with wrongness judgments of binding foundation violations (e.g., Inbar, Pizarro, Iyer, & Haidt, 2012; especially sanctity; for example, Horberg, Oveis, Keltner, & Cohen, 2009) and that anger has been shown to be associated with individualizing violations (e.g., Rozin, Lowery, Imada, & Haidt, 1999; Tangney, Stuewig, & Mashek, 2007; Vasquez, Keltner, Ebenbach, & Banaszynski, 2001), we expect that people will more readily give nonemotional justifications to individualizing than binding violations. For similar reasons, we expect more emotive appeals for sanctity than for care violations.

Third, given that all violations involve the transgression of some rule or principle, we expected no substantial differences in deontological appeals across foundations. However, to the extent that simple appeals to principles (without elaboration in relation to consequences, for example) can be deemed unelaborated, we might expect, based on the findings of Russell and Giner-Sorolla (2011c), that binding foundations might elicit more deontological justifications. Consistent with this, and suggestive of greater deontological appeals especially for sanctity violations, people prefer to quickly voice their intuitive condemnation of taboo violations, as opposed to deliberately processing and reflecting on the violation before making their judgments (Merritt & Monin, 2011). Such a finding would also be consistent with Piazza and Sousa (2014) who demonstrated that deontological appeals were prevalent for harmless sanctity violations. Counting against this, however, is the fact that while most people consider violations of care and fairness violations to be morally wrong (i.e., transgressive of some moral norm), only a subset (e.g., political conservatives) consider binding violations so (Graham et al., 2013). This latter fact might lead to more deontological appeals, on the whole, for individualizing violations.

Study 1

Method

Participants

One hundred U.S. residents (59 male)⁴ were recruited using Amazon’s Mechanical Turk. Participants ranged in age from 18 to 66 (M_age = 33.29, SD_age = 10.46); 97% of participants listed English as their first language.

Materials and procedure

Moral violations

After providing consent, participants read 15 moral violations, three per foundation (e.g., bullying as a care violation; consensual incest as a sanctity violation), and were asked to make a moral judgment about each. Complete vignettes can be seen in Supplementary Material 1. Presentation order of the 15 moral violations was randomized. Violation vignettes were selected from a Mechanical Turk pilot study (n = 36; 16 female; M_age = 30.03 years) to ensure that violations adequately and uniquely tapped target foundations.

In the study proper, participants provided a moral wrongness judgment after reading each vignette (1 = extremely morally wrong, 6 = extremely morally right).

Moral justifications

Immediately after making the wrongness judgment, participants were instructed: “How would you justify your wrongness judgment to someone else to persuade him or her to agree with you? Choose one of the three options below to communicate your justification to someone else.”

Participants were then presented with three justification options that corresponded with the valence of their wrongness judgment.⁵ If participants rated the action as morally wrong (≤3), they were presented with the following three justification options:

Deontology: His action was morally wrong because it goes against the principles that we should live by. This act should be forbidden to all people. It violates moral rules and codes of conduct about what is permitted in our society. Everyone is obliged to follow these rules.

Consequentialism: His action was morally wrong because of the consequences it caused for others. People may have been hurt or their lives may have been affected by his behavior. He did not consider the repercussions of his action before he did this.

Emotive: His action was morally wrong because it was offensive. It made me feel emotional to think about what happened in the story. I cannot believe this kind of thing occurs. I find it appalling and upsetting. He should be ashamed of himself!

If a participant judged an action as morally positive (≥4), she saw the following three justification options:

Deontology: His action was not morally wrong because no moral principles were violated. No moral rules or codes of conduct were broken. His action is permissible. What he did was not forbidden.

Consequentialism: His action was not morally wrong because it caused no bad consequences. No one was hurt; no one’s life was affected by his behavior. There were no repercussions for him to consider. Because no one was affected, it was fine for him to behave this way.

Emotive: His action was not morally wrong because it was not offensive to me in any way. I had no emotional reaction to this story. This is not shocking. He has nothing to be ashamed of.

These justifications were piloted to ensure that each uniquely tapped the relevant appeal category (see Supplementary Material 2 for details of pilot study).

After reading each vignette and selecting a justification to communicate, participants provided demographic information, including age, gender, country of birth, country of residence, ethnic identity, first language, religiosity (1 = not at all, 5 = very religious), religious affiliation, political party, and political beliefs (1 = extreme left, 5 = centrist, 9 = extreme right).

Results and discussion

Frequency counts were computed for each justification type per moral foundation for each participant. These frequencies were then submitted to a (3)(Code: deontological vs. consequentialist vs. emotive) × (5)(Foundation: care vs. fairness vs. loyalty vs. authority vs. sanctity) repeated measures analysis of variance (ANOVA), which yielded a significant main effect⁶ of Code, F(2, 98) = 19.35, p < .01, $η_{p}^{2}$ = .28. This was qualified by a significant interaction effect, F(8, 92) = 21.18, p < .01, $η_{p}^{2}$ = .65, which demonstrates that justification selection differed as a function of Foundation (see Figure 1).

Figure 1.

Frequency of justification selection as a function of Foundation and Code.

To explore this two-way interaction, we conducted three separate (5)(Foundation: care vs. fairness vs. loyalty vs. authority vs. sanctity) level repeated measures ANOVAs, one per justification type. To test hypotheses, omnibus tests were followed by tests of contrasts exploring (a) the difference between individualizing and binding foundation clusters (3 3 −2 −2 −2), (b) the difference between care and sanctity (1 0 0 0 −1), (c) the difference within the individualizing cluster—between care and fairness (1 −1 0 0 0), and (d) differences within the binding cluster—loyalty and authority vs. sanctity (0 0 −1 −1 2)⁷ and loyalty vs. authority (0 0 1 −1 0).

There was a significant effect of Foundation on deontological justification choices, F(4, 96) = 3.89, p = .01, $η_{p}^{2}$ = .14. Contrast tests showed that the individualizing foundations did not differ from the binding foundations, F(1, 99) = 0.13, p = .72, 95% CI = [−0.16, 0.11], and, contrary to hypotheses, care did not differ significantly from sanctity, F(1, 99) = 2.68, p = .11, 95% CI = [−0.42, 0.04]. Within individualizing, care did not differ from fairness, F(1, 99) = 2.01, p = .16, 95% CI = [−0.07, 0.41]. However, deontological justifications were significantly more frequent for sanctity than for loyalty and authority, F(1, 99) = 11.78, p < .01, 95% CI = [0.16, 0.59]. They were also more frequent for loyalty compared with authority, F(1, 99) = 4.17, p = .04, 95% CI = [0.01, 0.41].

There was also a significant and much larger effect of Foundation on consequentialist choices, F(4, 96) = 22.18, p < .01, $η_{p}^{2}$ = .48. As expected, and consistent with the centrality of welfare/harm and justice concerns to consequentialist rhetoric, consequentialist justifications were more prevalent for the individualizing than binding foundations, F(1, 99) = 58.31, p < .01, 95% CI = [0.43, 0.72], and there were significantly more consequentialist appeals for care versus sanctity violations, F(1, 99) = 14.66, p < .01, 95% CI = [0.22, 0.68]. These justifications were also more prevalent for fairness than care, F(1, 99) = 22.06, p = .02, 95% CI = [−0.84, −0.34]. Within the binding cluster, sanctity was less frequently justified in consequentialist terms than were loyalty and authority, F(1, 99) = 6.24, p = .01, 95% CI = [−0.46, −0.05]; loyalty and authority did not significantly differ, F(1, 99) = .82, p = .37, 95% CI = [−0.29, 0.11].

Emotive justifications exhibited the largest Foundation effect, F(4, 96) = 36.93, p < .01, $η_{p}^{2}$ = .61. The emotive justification was chosen more often in the binding than individualizing foundations, F(1, 99) = 84.82, p < .01, 95% CI = [−0.67, −0.43], and more in the sanctity than care foundation, F(1, 99) = 7.85, p < .01, 95% CI = [−0.44, −0.08]. Within clusters, they were more dominant in the care than the fairness foundation, F(1, 99) = 34.68, p < .01, 95% CI = [0.28, 0.56], but there was no difference within the binding cluster (sanctity vs. loyalty/authority), F(1, 99) = 1.39, p = .24, 95% CI = [−0.32, 0.08], and (loyalty vs. authority), F(1, 99) = 1.34, p = .25, 95% CI = [−0.33, 0.09].

Results from Study 1 indicate that people choose to communicate different kinds of justifications for violations that exemplify different moral foundations. Specifically, deontological appeals were most often chosen for the moral foundation of sanctity, consequentialist justifications were preferred for individualizing (especially fairness) over the binding violations, and emotive justifications were chosen more often for binding than individualizing foundations.

Study 2

Because real-world moral communication does not involve selecting justifications among provided options, in Study 2, participants were asked to produce free-response justifications as a more realistic exploration of the content of moral communication. Broadly speaking, the aims of Study 2 were to replicate the findings of Study 1 (a) using free responses, (b) in a sample from a different country (Australia), and (c) using a different recruitment method.

Method

Participants

Participants were 165⁸ first-year psychology students (48 male) who ranged in age from 17 to 44 (M_age = 19.65, SD_age = 2.96). In all, 77.6% of participants listed English as their first language.

Procedure and materials

The same 15 vignettes from Study 1 were used in Study 2. After providing consent, participants were assigned to one of three vignette sets. In a between-participants design, participants responded to five of the 15 vignettes (one per foundation, so that foundation was a within-participants variable). Participants first read each violation vignette (presentation order was randomized) and provided a moral wrongness judgment for each violation (1 = extremely morally wrong, 7 = extremely morally right). They next provided open-ended justifications for each of their moral judgments after reading the following instructions: “How would you justify your judgment to someone else to persuade him or her to agree with you? Try to persuade this person of why you think X’s actions were morally right or morally wrong and why you think he should or should not be punished.”

After providing free-response justifications, participants completed the same demographics as used in Study 1.

Content analysis

Justifications were analyzed for deontological, consequentialist, and emotive content using the Linguistic Inquiry and Word Count (LIWC) program (see Pennebaker, Francis, & Booth, 2003). LIWC works by searching through uploaded texts for the presence of words contained in a “dictionary” (a set of pre-specified, thematically related words) and provides output of the percentage of words in those texts contained in each dictionary. As a first step, we developed dictionaries to measure deontology, consequentialism, and moral emotions following the steps taken in developing Pennebaker, Chung, Ireland, Gonzales, and Booth’s (2007) LIWC manual and the Moral Foundations LIWC dictionary (Graham et al., 2009).

This procedure had three phases: expansive, contractive, and validation. In the expansive phase, the authors and three other lab members generated as many associations, synonyms, and antonyms for the core concept words for each dictionary. We included full words or word stems when it was more appropriate. In the contractive phase, the authors deleted words that seemed too distantly related to the core concepts and words whose primary meanings were ambiguous. The deontology and consequentialism dictionaries were informed by reference to entries for these philosophical theories in the Stanford Encyclopedia of Philosophy (Alexander & Moore, 2015; Sinnott-Armstrong, 2014) and validated by comparing LIWC scores for each dictionary using (a) canonical deontological and consequentialist philosophy texts and (b) a set of contemporary philosophy papers that present either deontological and consequentialist arguments for particular issues. Deontological texts had significantly more deontological content than consequentialist texts, and vice versa (see Supplementary Material 3 for analysis and results). The moral emotions dictionary centered on the nine moral emotions proposed in Haidt (2003) including contempt, anger, disgust, shame, embarrassment, guilt, compassion, gratitude, and elevation. The finalized dictionaries are included in Supplementary Material 4.

Results and discussion

Prior to analysis, the 825 free-response justifications (five for each of the 165 participants) were spellchecked using Microsoft Word. Responses were content analyzed using LIWC.

As the three dictionaries differ in total number of words, standardized scores were computed to allow appropriate comparisons across Codes. For each Code, participants’ scores were standardized relative to the average of that Code across all Foundations. Thus, mean z-score for deontological language use for sanctity violations, for example, represents the mean deviation of deontological language use in sanctity violations from baseline deontological language use for all five Foundations.

A (3)(Code: deontological vs. consequentialist vs. emotive) × (5)(Foundation: care vs. fairness vs. loyalty vs. authority vs. sanctity) repeated measures ANOVA was conducted on these standardized scores.⁹ This yielded a significant main effect of Foundation, F(4, 161) = 4.63, p < .01, $η_{p}^{2}$ = .01, and a significant interaction effect of Foundation and Code, F(8, 157) = 8.39, p < .01, $η_{p}^{2}$ = .30.¹⁰

To explore the interaction, three separate (5)(Foundation: care vs. fairness vs. loyalty vs. authority vs. sanctity) repeated measures ANOVAs were conducted on the standardized scores, one for each Code. Following, Study 1, contrast analyses were used to explore variation as a function of Foundation (see Figure 2 for results).

Figure 2.

Standardized raw LIWC percentages for Foundation × Code interaction.

There was a marginal omnibus Foundation effect for deontological language use, F(4, 161) = 2.30, p = .06, $η_{p}^{2}$ = .05. As in Study 1, the individualizing and binding clusters did not differ, F(1, 164) = 1.78, p = .18, 95% CI = [−0.05, 0.24], nor did care and sanctity, F(1, 164) = .03, p = .86, 95% CI = [−0.26, 0.22] or care and fairness, F(1, 164) = 0.07, p = .80, 95% CI = [−0.21, 0.27]. Within the binding cluster, and consistent with Study 1, deontological justifications were more common for sanctity than loyalty and authority, F(1, 164) = 6.13, p = .01, 95% CI = [0.04, 0.36]. Loyalty and authority did not significantly differ, F(1, 164) = 0.06, p = .81, 95% CI = [−0.16, 0.20].

As in Study 1, analyses of standardized scores revealed a significant Foundation effect for consequentialism, F(4, 161) = 6.78, p < .01, $η_{p}^{2}$ = .14. Consistent with predictions, consequentialist language use was higher in the individualizing than binding clusters, F(1, 164) = 24.84, p < .01, 95% CI = [0.24, 0.55], and consequentialist appeals were significantly more prevalent for care than sanctity, F(1, 164) = 10.41, p < .01, 95% CI = [0.16, 0.65]. Care did not differ from fairness, F(1, 164) = 1.37, p = .24, 95% CI = [−0.10, 0.39], and there were no differences within the binding cluster, F(1, 164) = 1.31, p = .26, 95% CI = [−0.07, 0.25], and F(1, 164) = 0.02, p = .89, 95% CI = [−0.12, 0.11].

As in Study 1, emotive language use showed the largest Foundation effect, F(4, 161) = 13.89, p <.01, $η_{p}^{2}$ = .26. Also in keeping with Study 1, the use of moral emotions in justifications was more pronounced in the binding than in the individualizing foundations, F(1, 164) = 5.42, p = .02, 95% CI = [−0.27, −0.02], but contrary to hypotheses (and Study 1), there was no difference between care and sanctity, F(1, 164) = 2.64, p = .11, 95% CI = [−0.04, 0.41]. Furthermore, emotive language was more prominent in care than in fairness, F(1, 164) = 14.25, p < .01, 95% CI = [0.20, 0.65], and consistent with the trend observed in Study 1, emotive appeals were more frequent in loyalty and authority than in sanctity, F(1, 164) = 3.99, p = .05, 95% CI = [−0.35, −0.00], but loyalty and authority did not differ, F(1, 164) = 1.92, p = .17, 95% CI = [−0.38, 0.07].

General Discussion

This article reported two studies that found that the kinds of appeals used in persuasive moral justifications vary as a function of moral foundation. Specifically, deontological appeals were most prevalent for sanctity violations, appeals to consequences were more frequent for violations of the individualizing foundations (care and fairness), and participants communicated emotional appeals when justifying violations of the binding foundations (loyalty, authority, sanctity). The fact that results were broadly replicated using different samples (U.S. residents and Australian students) and using both justification selection techniques (Study 1) and open-ended justification production (Study 2) provides support for the robustness of the findings.

Deontological appeals showed the least variation across foundations; however, they were chosen for the sanctity foundation most frequently (although not significantly more frequently than for care). This prevalence of deontological appeals for the sanctity and care foundations cannot readily be explained in connection with such appeals’ hypothesized low degree of elaboration (such an account would predict that the unelaborated disgust of sanctity would generate more deontological appeals than the more easily elaborated anger of care). Nor can it be explained with reference to people’s more ready acceptance of rapid, intuitive appeals to principle for sanctity violations (Merritt & Monin, 2011). Rather, it appears that, on balance, care, fairness, and sanctity violations are more readily spoken about as violations of principles (and thus justified in reference to those principles), than are loyalty and authority violations (a pattern clearly indicated in Figure 2). For many, it may simply be that violations of loyalty and authority are not violations of clear rules that warrant justifications in relation to such rules.

Participants chose to justify individualizing over binding and care over sanctity violations with consequentialist appeals, which supports the hypothesized link between consequentialism and the individualizing foundations, in so far as both focus on welfare and justice concerns (Bentham, 1789/1970; Haidt & Joseph, 2004; Mill, 1861/1998) and is consistent with previous work on justifications of prosocial and fair actions (Aguiar et al., 2008) and on work showing that anger reactions (which typify emotion responses to individualizing violations) are more readily justified with elaborated responses (Russell & Giner-Sorolla, 2011c).

Results revealed a strong preference for emotive justifications in the binding versus the individualizing foundations. Although all foundations are theorized to have attendant emotional reactions (Haidt & Graham, 2007; Haidt & Joseph, 2004), it seems that appeals to such emotions as justifications are more common for violations of loyalty, authority, and sanctity. This pattern of results is consistent with the idea that binding foundation endorsement is associated with disgust sensitivity (e.g., Inbar et al., 2012) and that moral reactions grounded in disgust are often justified with appeals to unelaborated affective reactions (Russell & Giner-Sorolla, 2011c). On this account, we would have expected sanctity violations to be especially evocative of emotive justifications, but this is not strongly borne out by the data. Although emotive justifications were more frequent for sanctity than care in Study 1, this was not the case in Study 2, and, moreover, sanctity violations did not engender more frequent emotive appeals than the other binding foundations.

One interesting finding within the individualizing cluster was the relatively frequent emotive justification selection and appeal usage for the care foundation. Given that care violations are often linked to the emotions of empathy for a victim’s suffering (Miller & Cushman, 2013), anger (Gutierrez & Giner-Sorolla, 2007; Kam & Bond, 2009; Rozin et al., 1999) and compassion (Haidt & Joseph, 2004), it is perhaps unsurprising that emotive appeals were used frequently to justify care violations.

Taken together, the results suggest that, in comparison with other foundations, attitudes toward care violations tend to be justified primarily with reference to consequences, and secondarily with reference to principles and emotions.

Violations of fairness are also talked about with primary reference to consequences, but are markedly less often talked about with reference to emotions, than are care violations. Indeed, the difference between the consequentialist and emotive rhetoric for fairness violations is perhaps the most striking difference in the results. What is it about the fairness foundation that encourages consequentialist but not emotive arguments? First, by definition, fairness concerns cover moral issues of proportionality, reciprocity, and equality (Haidt, 2012; Haidt & Graham, 2007). To determine whether someone has been treated unfairly, one may often attempt to identify the consequences suffered by a perceived victim. Economic policy research argues that people make judgments when allocating resources based on both procedural fairness and “outcome” or “end-state” fairness criteria (Isaac, Mathieu, & Zajac, 1991). Thus, a consideration of consequences is likely to occur in the formation of the moral judgment and, as a result, may be more accessible for justification. Although ample research shows that fairness violations elicit negative emotions (especially anger), such affective reactions do not appear in people’s persuasive justifications. This may be (a) because such affective reactions are penetrable to cognitive elaboration in terms of consequences (e.g., this makes me angry because it involves an unfair distribution of outcomes; Russell & Giner-Sorolla, 2011c) or (b) because such justifications are deemed less legitimate for persuasive purposes in the fairness foundation.

Attitudes to sanctity violations seem to draw on both deontological and affective justifications, with comparatively less recourse to consequentialist reasoning. This can be interpreted as consistent with a general preference for cognitively less elaborated justifications that appeal to principle or emotion. The justifications for the loyalty and authority foundations also drew on a variety of content. No clear preference was shown for one appeal strategy over another in Study 1; in Study 2, emotive appeals were more often than average used to justify violations in these foundations.

One general observation that follows from this summary is that most violations are justified with recourse to a variety of appeals. This is consistent with Ditto, Pizarro, and Tannenbaum (2009), who suggest that most domains have more than one standard that can be perceived as appropriate. However, it appears that the majority of variation in appeal use occurs between violations, rather than within, as justification content scores are not significantly correlated within violation (shown in Supplementary Material 5). This suggests that although people may indeed tailor their justifications to suit moral content, they tend to focus on one kind of justification (deontological, consequentialist, or emotive) for each moral attitude they are asked to justify.¹¹

Although contemporary models of moral judgment, including Haidt’s (2001) social intuitionist model and Greene and colleagues’ dual process model (Greene, Nystrom, Engell, Darley, & Cohen, 2004; Greene, Sommerville, Nystrom, Darley, & Cohen, 2001), use stimuli with social content, less attention has been paid to social processes in the study of moral phenomena. In focusing specifically on persuasive communication about moral attitudes, we have taken a step toward reintroducing the social into the moral domain (Greenwood, 2011). In focusing on moral communication, rather than judgment, we also get a slightly different picture of the role of reason in the moral domain than is given by work on intra-individual processes. Consistent with the theorizing of Mercier (2011) and Mercier and Sperber (2011) that states that moral reasoning is for argumentation and persuasion, we find, in the persuasive context, that appeals are readily made (for some foundations more than others) to nonemotive justifications. So while emotion may drive moral judgment, it is not the case that appeals to emotion are the primary means of persuasive communication.

Our exclusive focus on the persuasive context of moral justifications, although consistent with recent theorizing about such contexts as key loci of moral reasoning, is, nevertheless, a limitation. In not considering justifications per se, we cannot clearly infer whether the patterns of justification use observed here are specific to persuasive communication or would be observed regardless of justification context (e.g., justifying to oneself). An interesting question to consider is whether justifications given in explicitly persuasive contexts would differ from “pure justifications” (for nonpersuasive purposes) and, moreover, whether persuasive justifications are tuned to different audiences (e.g., ingroup vs. outgroup; Higgins, 1992, 1999). Relational concerns, for example, have been shown to influence moral judgments (e.g., Simpson & Laham, 2015); future work might examine how justifications are tuned to different relational partners (e.g., spouse vs. coworker).

Another limitation of the current work is its focus on hypothetical, rather than in situ, persuasion. Although introducing moral communication goes some way to increasing the ecological validity of the study of moral psychological processes, it does not go far enough in the present study. Examining moral communication in ecologically valid, social interactions would enable researchers to explore the social, impassioned, and personal aspects of real moral debates that hypothetical, one-way communications cannot capture.

Two more minor limitations are worth noting and addressing in future work. First, although a pilot study demonstrated construct validity for the three justification options provided to participants in Study 1 (see Supplementary Material 2), there were a few points of ambiguity in justification wording. For example, there is a possibility that the emotive justification, which included statements of astonishment and surprise, may have been more relevant to some emotions (e.g., disgust) than over others (e.g., anger; Russell & Piazza, 2015). However, if the emotive justification was unduly tapping disgust rather than other emotions, we would have expected this category to be chosen most frequently for sanctity violations—which was not the case. Nevertheless, future research may benefit from more extensive pilot testing for the development of persuasive moral communication stimuli.

Second, although our focus on deontological, consequentialist, and emotive justifications is consistent with moral philosophical and psychological approaches to justifying moral attitudes, there are other ways of justifying one’s moral attitudes. Appeals to virtue or moral character, for example, might be readily made. Recent work has focused on the role of character assessments in moral judgments (e.g., Goodwin, Piazza, & Rozin, 2014); it remains a question for future research whether character is also appealed to in persuasive communication and whether it is differently so for violations of different moral foundations.

Much of everyday moral reasoning and judgment happens in thoroughly social contexts. Understanding how we talk about morality and the social, communicative processes that shape, change, and solidify our moral attitudes is of fundamental importance in gaining a comprehensive understanding of moral phenomena. The present studies, in providing a systematic survey of the appeals people make across a range of moral foundational violations, are an important first step in characterizing moral communication. Future work would do well to more fully explore the social processes central to moral psychology and to recognize that much of interest in the moral domain centers on what we say, rather than just on what we think.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Supplemental Material

The online supplemental material is available at .

Notes

References

Aguiar

Brañas-Garza

Miller

L. M.

(2008). Moral distance in dictator games. Judgement and Decision Making, 3, 344-354.

Alexander

Moore

(2015). Deontological ethics. In Zalta

E. N.

(Ed.), The Stanford encyclopedia of philosophy (Spring 2015 ed.). Retrieved from http://plato.stanford.edu/archives/spr2015/entries/ethics-deontological/

Ayer

A. J.

(1946). Language, truth and logic. London, England: Victor Gollancz. (Original work published 1936)

Bentham

(1970). An introduction of the principles of morals and legislation. Oxford, UK: The Athlone Press. (Original work published 1789)

Chakroff

Dungan

Young

(2013). Harming ourselves and defiling others: What determines a moral domain? PloS One, 8(9), e74434.

Chakroff

Young

(2015). Harmful situations, impure people: An attribution asymmetry across moral domains. Cognition, 136, 30-37.

DeScioli

Kurzban

(2013). A solution to the mysteries of morality. Psychological Bulletin, 139, 477-496.

Ditto

P. H.

Pizarro

D. A.

Tannenbaum

(2009). Motivated moral reasoning. Psychology of Learning and Motivation, 50, 307-338.

Driver

(2014). The history of utilitarianism. In Zalta

E. N.

(Ed.), The Stanford encyclopedia of philosophy (Winter 2014 ed.). Retrieved from http://plato.stanford.edu/archives/win2014/entries/utilitarianism-history/

10.

Dunbar

R. I.

(2004). Gossip in evolutionary perspective. Review of general psychology, 8(2), 100-110.

11.

Dunbar

R. I. M.

(1995). The trouble with science. Cambridge, MA: Harvard University Press.

12.

Feinberg

Willer

Stellar

Keltner

(2012). The virtues of gossip: Reputational information sharing as prosocial behavior. Journal of Personality and Social Psychology, 102, 1015-1030.

13.

Foster

E. K.

(2004). Research on gossip: Taxonomy, methods, and future directions. Review of General Psychology, 8(2), 78-99.

14.

Gibbard

(1990). Wise choices, apt feelings. Cambridge, UK: Cambridge University Press.

15.

Goodwin

Piazza

Rozin

(2014). Moral character information predominates in person perception and evaluation. Journal of Personality and Social Psychology, 106, 148-168.

16.

Graham

Haidt

Koleva

Motyl

Iyer

Wojcik

S. P.

Ditto

P. H.

(2013). Moral foundations theory: The pragmatic validity of moral pluralism. Advances in Experimental Social Psychology, 47, 55-130.

17.

Graham

Haidt

Nosek

B. A.

(2009). Liberals and conservatives rely on different sets of moral foundations. Journal of Personality and Social Psychology, 96, 1029-1046.

18.

Graham

Nosek

B. A.

Haidt

Iyer

Koleva

Ditto

P. H.

(2011). Mapping the moral domain. Journal of personality and social psychology, 101, 366-385.

19.

Gray

Keeney

J. E.

(2015). Disconfirming moral foundations theory on its own terms: Reply to Graham (2015). Social Psychological and Personality Science, 6, 874-877.

20.

Gray

Schein

Ward

A. F.

(2014). The myth of harmless wrongs in moral cognition: Automatic dyadic completion from sin to suffering. Journal of Experimental Psychology: General, 143, 1600-1616.

21.

Greene

J. D.

Nystrom

L. E.

Engell

A. D.

Darley

J. M.

Cohen

J. D.

(2004). The neural bases of cognitive conflict and control in moral judgment. Neuron, 44, 389-400.

22.

Greene

J. D.

Sommerville

R. B.

Nystrom

L. E.

Darley

J. M.

Cohen

J. D.

(2001). An fMRI investigation of emotional engagement in moral judgment. Science, 293, 2105-2108.

23.

Greenwood

J. D.

(2011). On the social dimensions of moral psychology. Journal for the Theory of Social Behaviour, 41, 333-364.

24.

Gutierrez

Giner-Sorolla

(2007). Anger, disgust, and presumption of harm as reactions to taboo-breaking behaviors. Emotion, 7, 853-868.

25.

Haidt

(2001). The emotional dog and its rational tail: A social intuitionist approach to moral judgment. Psychological Review, 108, 814-834.

26.

Haidt

(2003). The moral emotions. Handbook of Affective Sciences, 11, 852-870.

27.

Haidt

(2012). The righteous mind: Why good people are divided by politics and religion. New York, NY: Pantheon Books.

28.

Haidt

Graham

(2007). When morality opposes justice: Conservatives have moral intuitions that liberals may not recognize. Social Justice Research, 20(1), 98-116.

29.

Haidt

Joseph

(2004). Intuitive ethics: How innately prepared intuitions generate culturally variable virtues. Daedalus, 133(4), 55-66.

30.

Haidt

Kesebir

(2010). Morality. In Fiske

Gilbert

Lindzey

(Eds.), Handbook of social psychology (5th ed., pp. 797-832). Hobeken, NJ: Wiley.

31.

Haidt

Koller

S. H.

Dias

M. G.

(1993). Affect, culture, and morality, or is it wrong to eat your dog? Journal of Personality and Social Psychology, 65, 613-628.

32.

Higgins

E. T.

(1992). Achieving “shared reality” in the communication game: A social action that creates meaning. Journal of Language and Social Psychology, 11, 107-131.

33.

Higgins

E. T.

(1999). “Saying is believing” effects: When sharing reality about something biases knowledge and evaluations. In Thompson

L. L.

Levine

J. M.

Messick

D. M.

(Eds.), Shared cognition in organizations: The management of knowledge. Mahwah, NJ: Lawrence Erlbaum.

34.

Horberg

Oveis

Keltner

Cohen

A. B.

(2009). Disgust and the moralization of purity. Journal of Personality and Social Psychology, 97, 963-976.

35.

Hume

(1969). A treatise of human nature. London, England: Penguin. (Original work published 1739-1740)

36.

Inbar

Pizarro

Iyer

Haidt

(2012). Disgust sensitivity, political conservatism, and voting. Social Psychological and Personality Science, 3, 537-544.

37.

Isaac

R. M.

Mathieu

Zajac

E. E.

(1991). Institutional framing and perceptions of fairness. Constitutional Political Economy, 2, 329-370.

38.

Kam

C. C. S.

Bond

M. H.

(2009). Emotional reactions of anger and shame to the norm violation characterizing episodes of interpersonal harm. British Journal of Social Psychology, 48, 203-219.

39.

Kant

(1993). Grounding for the metaphysics of morals (3rd ed., Ellington

J. W.

,Trans.). Indianapolis, IN: Hackett. (Original work published 1785)

40.

Kohlberg

(1969). Stage and sequences: The cognitive-developmental approach to socialization. In Goslin

D. A.

(Ed.), Handbook of socialization theory and research (pp. 151-235). New York, NY: Academic Press.

41.

Laham

S. M.

Corless

(2016). Threat, morality and politics: A differentiated threat account of moral and political values. In Forgas

J. P.

Jussim

van Lange

P. A. M.

(Eds.), The social psychology of morality (pp. 56-70). New York, NY: Routledge.

42.

Mercier

(2011). What good is moral reasoning? Mind & Society, 10, 131-148.

43.

Mercier

Sperber

(2011). Why do humans reason? Arguments for an argumentative theory. Behavioral and Brain Sciences, 34(02), 57-74.

44.

Merritt

A. C.

Monin

(2011). The trouble with thinking: People want to have quick reactions to personal taboos. Emotion Review, 3, 318-319.

45.

Mill

J. S.

(1998). Utilitarianism, edited with an introduction by Roger Crisp: New York, NY: Oxford University Press. (Original work published 1861)

46.

Miller

Cushman

(2013). Aversive for me, wrong for you: First-person behavioral aversions underlie the moral condemnation of harm. Social and Personality Psychology Compass, 7, 707-718.

47.

Ogden

C. K.

Richards

I. A.

(1923). The meaning of meaning: A study of the influence of language upon thought and of the science of symbolism. New York, NY: Harcourt, Brace, & Company.

48.

Ostrom

Gardner

Walker

(1994). Rules, games, and common-pool resources. Ann Arbor: University of Michigan Press.

49.

Pennebaker

J. W.

Chung

C. K.

Ireland

Gonzales

Booth

R. J.

(2007). The development and psychometric properties of LIWC2007. Austin, TX: LIWC.net.

50.

Pennebaker

J. W.

Francis

M. E.

Booth

R. J.

(2003). Linguistic inquiry and word count: LIWC2001 manual. Mahwah, NJ: Lawrence Erlbaum.

51.

Peters

Kashima

(2014). Gossiping as moral social action: A functionalist account of gossiper perceptions. In Laszlo

Forgas

Vincze

(Eds.). Social cognition and communication. New York, NY: Psychology Press.

52.

Piazza

Russell

P. S.

Sousa

(2013). Moral emotions and the envisaging of mitigating circumstances for wrongdoing. Cognition & Emotion, 27, 707-722.

53.

Piazza

Sousa

(2014). Religiosity, political orientation, and consequentialist moral thinking. Social Psychological and Personality Science, 5, 334-342.

54.

Pizarro

D. A.

Detweiler-Bedell

Bloom

(2006). The creativity of everyday moral reasoning. In Kaufman

J. C.

Baer

(Eds.), Creativity and reason in cognitive development (pp. 81-98). Cambridge, UK: Cambridge University Press.

55.

Rottman

Kelemen

Young

(2014). Tainting the soul: Purity concerns predict moral judgments of suicide. Cognition, 130, 217-226.

56.

Rozin

Lowery

Imada

Haidt

(1999). The CAD triad hypothesis: A mapping between three moral emotions (contempt, anger, disgust) and three moral codes (community, autonomy, divinity). Journal of Personality and Social Psychology, 76, 574-586.

57.

Russell

P. S.

Giner-Sorolla

(2011a). Moral anger, but not moral disgust, responds to intentionality. Emotion, 11, 233-240.

58.

Russell

P. S.

Giner-Sorolla

(2011b). Moral anger is more flexible than moral disgust. Social Psychological and Personality Science, 2, 360-364.

59.

Russell

P. S.

Giner-Sorolla

(2011c). Social justifications for moral emotions: When reasons for disgust are less elaborated than for anger. Emotion, 11(3), 637-646.

60.

Russell

P. S.

Piazza

(2015). Consenting to counter-normative sexual acts: Differential effects of consent on anger and disgust as a function of transgressor or consenter. Cognition and Emotion, 29, 634-653.

61.

Shweder

Much

Mahapatra

Park

(1997). Divinity and the “big three” explanations of suffering. Morality and Health, 119, 119-169.

62.

Simpson

Laham

S. M.

(2015). Individual differences in relational construal are associated with variability in moral judgment. Personality and Individual Differences, 74, 49-54.

63.

Sinnott-Armstrong

(2014). Consequentialism. In Zalta

E. N.

(Ed.), The Stanford encyclopedia of philosophy (Spring 2014 ed.). Retrieved from http://plato.stanford.edu/archives/spr2014/entries/consequentialism/

64.

Stevenson

C. L.

(1944). Ethics and language. New Haven, CT: Yale University Press.

65.

Suhler

C. L.

Churchland

(2011). Can innate, modular “foundations” explain morality? Challenges for Haidt’s moral foundations theory. Journal of Cognitive Neuroscience, 23, 2103-2116.

66.

Tangney

J. P.

Stuewig

Mashek

D. J.

(2007). Moral emotions and moral behavior. Annual Review of Psychology, 58, 345.

67.

van Leeuwen

Park

J. H

. (2009). Perceptions of social dangers, moral foundations, and political orientation. Personality and Individual Differences, 42, 169-173.

68.

Vasquez

Keltner

Ebenbach

D. H.

Banaszynski

T. L.

(2001). Cultural variation and similarity in moral rhetorics voices from the Philippines and the United States. Journal of Cross-Cultural Psychology, 32(1), 93-120.

69.

Wright

J. C.

Baril

G. L.

(2013). Understanding the role of dispositional and situational threat sensitivity in our moral judgments. Journal of Moral Education, 42, 383-397.

70.

Young

Saxe

(2011). When ignorance is no excuse: Different roles for intent across moral domains. Cognition, 120, 202-214.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.49 MB