Abstract
Dignity is prominently endorsed in health care, organizations, and law. However, humanities research casts doubt over the utility of this concept, disputing that “dignity” captures any unique ethical value distinguishable from “respect”. Here, we test a recent proposal that dignity entails special regard for humiliation. We created a set of vignettes describing a “perpetrator” making offensive remarks or gestures to a “victim”, and asked lay participants to rate how well each offense corresponds to a dignity violation as opposed to a respect violation. We manipulated the victims’ resulting emotions within-subject, across “humiliation”, “anger”, and “baseline” conditions, while controlling for other factors (e.g., victim’s gender, vulnerability). We found substantial evidence (BF = 4.11) for our preregistered hypothesis that humiliation increased dignity-relatedness ratings as compared with the baseline condition, and very strong evidence when compared with the anger condition (BF = 84.75). These findings provide empirical backing to dignitarian ethics with applications in health care, organizations, and law.
The concepts “dignity and respect” play an important role in defining human rights and responsibilities and related legal discourses (Kretzmer & Klein, 2021; Waldron, 2013). Commitment to dignity and respect is widely espoused by providers of health, welfare, and other social services, as well as by employers across diverse organizations.
Although “dignity” and “respect” are prominently endorsed as values in community settings, dignity, unlike respect, has received comparatively little attention from social scientists and psychologists and has only recently been extensively discussed by philosophers. In fact, some philosophers have expressed skepticism that “dignity” picks out a unique value at all, arguing that it is merely a faddish term for existing values, particularly respect for persons or their autonomy (Macklin, 2003). Adding to this skepticism is the fact that “dignity” and “respect” are often used interchangeably. It is common to hear phrases such as “respect for dignity” or “dignity is to be treated with respect”, which creates confusion around whether dignity is different to respect.
Ethnographic studies on dignity in health care (Baillie, 2009; Chochinov, 2002; Franklin et al., 2006; Gallagher et al., 2008; Gallagher & Seedhouse, 2002; Oosterveld-Vlug et al., 2014; Seedhouse, 2002; van Gennip et al., 2013) and workplace management (Crowley, 2013; Lucas et al., 2017; Thomas & Lucas, 2019) have not dispelled such skepticism, yielding often unwieldy and conceptually disunified lists of behaviors said to confer or undermine dignity: these can include respect for privacy, a safe and calm environment, respect for individuality, and so on. The use of dignity in many such studies is so broad and all-encompassing as to raise the suspicion that “dignity” collapses into “good quality health care” or “good workplace management” (Barclay, 2016; Thomas & Lucas, 2019).
Overall, there are reasons for skepticism about dignity, raising the question whether the espousals of dignity as a guiding value, across numerous contexts globally, are in fact distinctive and meaningful.
In response to skepticism about dignity, recent theoretical accounts argue that dignity is in fact a unique value not captured by other values, including respect. These philosophically diverse accounts share the core idea that dignity is connected to our social vulnerabilities; specifically, that violations of dignity humiliate or degrade the victims of such violations (Barclay, 2018; Etinson, 2020; Killmister, 2020; Luban, 2009; Margalit, 1996; Statman, 2000). On these accounts, victims of dignity violations can be degraded or humiliated in different ways, including expressions of disrespectful attitudes such as contempt, disgust, disdain, or ridicule (Barclay, 2018; Etinson, 2020; Killmister, 2020). The expressions of such attitudes degrade and humiliate by sending a disparaging message that the victim is beneath the perpetrator (Etinson, 2020), lowered in the eyes of the community (Killmister, 2020), or lacking in the attributes that qualify one as a social equal (Barclay, 2018).
These recent philosophical accounts of dignity deny that dignity is the same as what we usually understand by respect. Unlike “dignity”, “respect” has been subject to extensive philosophical analysis for some time (S. Buss, 1999; Darwall, 1977; Dean & Sensen, 2021; Hill, 1998). A common conception of violations of respect traces to Immanuel Kant, namely, treating a person as a mere means rather than an end in themselves; violations of a person’s autonomous agency are paradigm examples, such as a failure of consent, or lying and stealing. This conception of respect comes apart from recent philosophical accounts of dignity. For example, there is a difference between stealing someone’s bicycle forcing them to walk to the shop, and stealing someone’s wheelchair and forcing them to crawl around in public; both actions are morally wrong, and disrespect the agents as ends in themselves, but only the latter involves the distinctive violation of the person’s dignity (Etinson, 2020). It is also common to conceive of respect as acknowledgment of the equal moral status of all human beings, which is denied when people or groups are stripped of equal opportunities or rights. Yet not all discrimination is adequately explained as solely violations of respect: the concept of dignity allows us to appreciate the specific humiliations and degradations of, for example, a caste society where some people are treated as “untouchable.” Finally, more everyday instances of rudeness and failure of manners (e.g., queue jumping) may be disrespectful but are merely annoying and frustrating, rather than humiliating violations of dignity. The new theoretical accounts of dignity therefore imply a conceptual distinction from broader, common understandings of respect—however this claim is yet to be empirically tested.
These philosophical accounts use the term humiliation to designate a factual/true/objective state of being humiliated (henceforth the “humiliated state”), whereby a person can be humiliated even if they do not experience humiliation subjectively (e.g., as an emotion). Psychology research, however, conceives of humiliation as a subjective experience, a self-conscious emotion that arises when a person perceives themself to be demeaned, exposed or put-down (Fernández et al., 2015), where the humiliated state is associated with reports of feeling a cluster of self-conscious emotions including humiliation, shame, and embarrassment (Elison & Harter, 2007). The humiliated state can also make the victim feel anger, annoyance, or irritation, which are not self-conscious emotions; according to appraisal theory, this happens when victims perceive belittlement as unwarranted or unjust, rejecting it altogether (De Cremer et al., 2008; Fernández et al., 2015; Frijda et al., 1989). Indeed, when people are given descriptions of a victim’s reaction to receiving disparaging feedback, such as that the victim rejected the feedback and qualified it as unjust, they can identify the likely emotional state of the victim as anger (Fernández et al., 2015). Conversely, when victims are described as internalizing the disparagement, people identify self-conscious emotions like humiliation, shame and embarrassment. Specifically, when the victim internalized the disparagement despite seeing it as unjust, participants identify humiliation (Fernández et al., 2015). These findings suggest that people readily distinguish between emotions victims might experience when perceiving mistreatment as subjectively devaluing or not. Therefore, people can easily integrate information about a victim’s emotional state into their beliefs about whether the victim is devalued or not, which could in turn shape their evaluation of mistreatments as uniquely violating dignity. However, whether laypeople account for a victim’s subjectivity in this way has not been empirically scrutinized.
Overall, dignity is a concept claimed to have significant practical relevance in law, health care, and workplace relations. However, given that understanding of what dignity is has been so broad and disparate within these domains, attempts to promote dignity have been obscure, and it is as yet unclear whether public understanding of dignity maps onto any of the various philosophical theories. Given that no psychological science investigations contrasting laypeople’s perceptions of dignity against respect exist so far, it is still an open question whether people distinguish dignity from respect in any systematic way.
Here, we accordingly test the key hypothesis, derived from new theories of dignity, that dignity is different from respect because dignity uniquely involves considerations about whether the treatment involves humiliation. We operationalize humiliation in line with psychological accounts as a subjective emotion and specifically test whether subjective emotional outcomes of the treatment affect perceptions that victim’s dignity is being threatened, such that self-aware emotions (e.g., humiliation, shame, embarrassment) are expected to increase perceptions that negative treatment is related to dignity. Other negative emotions such as anger, which is typically other-directed and typically entails less self-conscious reflection on self (Tracy et al., 2007), are expected to either have no impact on or decrease perceptions that the negative treatment is related to dignity because they may provide information that could potentially disambiguate what the victim is feeling (compared with instances where emotional state of the victim is unknown). Considering that this is the first empirical inquiry into differences between dignity and respect, and that psychological theorizing about dignity is only at its inception, we tried to accommodate the pressing need for further theorization by, in addition, conducting pre-planned exploration of several other factors, such as a degradation theme, the victim’s gender, mode of degradation delivery, and the victim’s vulnerability to the degradation.
Current Study
In the current study, we asked participants to read textual vignettes describing offenses and rate how well these offenses correspond to violations of dignity as opposed to violations of respect (henceforth referred to as the “dignity-relatedness ratings”). Specifically, given the large overlap between dignity, respect and good conduct in general (Barclay, 2016; Thomas & Lucas, 2019), we wanted to ensure participants report even the subtlest intuitions about the way the offenses preferentially relate to dignity or respect. We designed a rating scale that entails a forced choice between dignity and respect for each vignette, to detect any patterns of preferential responding across participants and groups of vignettes. A novel set of 44 vignettes was developed, with each vignette describing one actor (i.e., the “perpetrator”) using speech or gestures to treat another actor (i.e., the “victim”) offensively. The offenses were inspired by existing examples of dignity and respect violations discussed in the philosophical literature. Since some of these examples contained obvious elements of ridicule, while others did not, and given that ridicule is a feature that may be relevant for perceptions of humiliation and thus dignity, we incorporated elements of ridicule into half of the vignettes (e.g., the perpetrator “made a joke about” the victim), and not the other half (e.g., perpetrator “yelled at” the victim), to test for effects of ridicule presence. The vignette content additionally varied across several dimensions (victim and perpetrator gender, the nature of their relationship, the topic of the offense, and factors that make victims more vulnerable to the offense (e.g., disability)), which were held balanced across ridicule conditions. Several considerations guided our choices of these additional dimensions: we endeavored to cover a range of offense topics emphasized in research on domains of self-esteem (Crocker & Wolfe, 2001; Harter & Bukowski, 2012; von Soest et al., 2016), which also ensured the compatibility of the vignettes with self-conscious emotions; we wanted our inferences to generalize across genders and different relationship types; finally, we wanted to incorporate dimensions we noticed frequently appearing in examples used in philosophy and ethics discourses (gestures, disability/vulnerability factors). To test if the perceptions of dignity depend on the emotional outcomes of the treatment, we asked participants to rate each vignette three times across three blocked emotion conditions: (a) base condition with vignettes presented without any reference to the victim’s emotional states; (b) humiliation condition, in which information describing the victim experiencing feelings of humiliation or shame was added to each vignette; and (c) anger condition, in which information describing the victim experiencing anger or irritation was added to each vignette. The target sample of N = 71 was determined using sensitivity analyses, on the basis of pilot data, as a part of the preregistration procedure. We hypothesized that the humiliation and ridicule conditions would shift the ratings toward the dignity end of the scale, whereas the anger condition would either have no effect on the ratings or shift them away from the dignity end of the scale. Using a preregistered model-comparison approach, we investigated, without raising any particular hypotheses, if additional dimensions we varied vignettes by (i.e., theme of the offense, communication mode, the nature of the relationship, victim’s gender, presence of a vulnerability factor) also affect dignity-relatedness ratings. In addition, we conducted a follow-up study (N = 71) with a slight variation of the method, wherein participants made the ratings on separate dignity-relatedness and respect-relatedness scales, to test a hypothesis that humiliation condition increases the ratings of dignity-relatedness (independently from respect-relatedness) relative to the baseline ratings and more than it increases respect-relatedness. We report on the follow-up study in the supplementary materials.
Methods
Transparency and Openness
We follow the Journal Article Reporting Standards (Kazak, 2018) and fully report all details about how we determined our sample size, data exclusions, manipulations, and measures we took. All data, analysis code, and research materials are available at https://osf.io/c639k/. Data were analyzed using R, version 4.3.0 (R Core Team, 2023) and the packages ggplot, version 3.4.2 (Wickham, 2016), BayesFactor (Morey & Rouder, 2011), brms (Bürkner, 2017), and rStan (Stan Development Team, 2023). The hypotheses, design (i.e., conditions, randomization strategies, sample size and rationale, exclusion criteria, manipulated variables, and measured variables), and statistical analyses (t-test and general linear model [GLM]) were preregistered using the Open Science Foundation standard preregistration form and are available at: https://osf.io/7b4ax.
Participants
The study was approved by the Monash University Human Research Ethics Committee (Project ID 38180). Participants were recruited using Prolific, a remote study recruitment platform. The recruitment was constrained to individuals fitting the following criteria based on their answers to the Prolific screening questions: adults (age higher than 18), fluent in English, residing in the United Kingdom, no ongoing mental illness or condition, no medication use, with normal or corrected-to-normal vision. We stopped recruiting once our preregistered target sample size of 71 participants was attained, after applying exclusion criteria. Importantly, this target sample was determined using sensitivity analyses as a part of the preregistration procedure and was geared at 80% power to detect a small-to-medium effect with a Cohen’s d of 0.30 for a paired samples t-test at the .05 alpha error probability. We determined our expectations for the effect size based on the lower bounds of the 80% confidence intervals of the effects observed while piloting, which is aimed at correcting for potential effect size inflation that small samples can produce (Albers & Lakens, 2018; Perugini et al., 2014). Eighty-two participants (41 male and 41 female, Mage= 45.52, SD = 14.75, ranging from 21 to 77) completed the study and were remunerated with £4 GBP upon completion. Eleven participants were excluded based on the following exclusion criteria: eight participants failed at least one of the three attention check questions randomly placed among the vignettes (i.e., prompts to select a particular option on the scale), and three participants had extremely low variance (SD < 0.3) in their 6-point scale dignity-relatedness rating responses within at least one of the three experimental blocks. The final sample included 71 individuals (35 male; 36 female, Mage= 45.93, SD = 14.92, ranging from 21 to 77).
Vignettes
We developed a set of 44 textual vignettes describing one actor (the “perpetrator”) using speech or gestures to treat another actor (the “victim”) in an offensive manner (Table 1). Half of the offenses contained elements of ridicule (e.g., the perpetrator “mocked”, “made fun of”, “laughed at”, “made a joke about”, or “pretended to dance like” the victim), while the other half did not (e.g., perpetrator “told [something to]”, “yelled at”/“explained [something to]”, or “ignored” the victim). The vignette content was additionally varied across multiple dimensions:
Victim’s gender using popular female and male (22 each) names (and frequently also additionally gender pronouns) to refer to the victim. To ensure the male and female names are recognizable and roughly matched for familiarity and frequency of use, all the names were taken from the list of 100 most popular names in the United States over the last century (www.ssa.gov/oact/babynames/decades/century.html).
The mode of communication with the 36 core vignettes describing offenses delivered via speech (e.g., the perpetrator “yelled”, “insisted explaining”, “joked that [..]”, “asked”), and 12 additional vignettes involving gestures (e.g., “sighed”, “pointed finger and laughed”, “sniffed air”).
The nature of the relationship between the perpetrator and victim was varied by referring to the perpetrator using relational nouns belonging to four broad categories: professional hierarchical relationships (e.g., “boss”, “teacher”, “supervisor”, “doctor”), professional nonhierarchical relationships (e.g., “colleague”, “classmate”, “co-worker”), friends and family (“father”, “daughter”, “sister”, “friend”), and strangers (e.g., “passer-by”, “stranger”, “receptionist”, “waiter”). Within speech vignettes there were eight vignettes of each category, while gesture vignettes were constrained to professional relationships (six vignettes of each professional category).
To choose the offense topics that most people care about, we relied on research on domains of self-esteem (Crocker & Wolfe, 2001; Harter & Bukowski, 2012; von Soest et al., 2016) to define four offense themes: cognitive competence (including topics like technological competence, scholastic performance, mental performance speed), physical competence (e.g., talent for sports, clumsiness, physical abilities), physical appearance (e.g., comments and reactions regarding looks and smells of the victim), and behavioral conduct (e.g., social competencies, conscientiousness, and nurturance of the victim).
Some vignettes also presented the victim as having a disability (i.e., dyslexia, dwarfism, autism), or an otherwise medicalized feature which relates to the topic of the offense that could make them vulnerable to marginalization and bullying (i.e., facial scars, needing walking aids, large size). The addition of these vulnerability factors was relatively evenly spread across ridicule conditions, victim genders, and themes among speech items, but less so among gesture items.
The sets of vignettes belonging to two ridicule conditions (“absent” and “present”) were balanced for victim’s gender, the nature of the relationship between the perpetrator and the victim; the theme of the offense, and whether the action victim performed was a spoken or gestural as well as matched for word length (Mridicule-present = 19.32, Mridicule-absent = 19.27, MD = 0.05, 95% CI [−1.93, 1.84]) and character length (Mridicule-present = 106.00, Mridicule-absent = 106.55, MD = 0.05, 95% CI [−8.11, 9.20]).
New Dignity & Respect Vignettes.
Note. Comm. mode = communication mode, CC = cognitive competence, PC = physical competence, PA = physical appearance, BC = behavioral competence, FF = friends or family, prof H = professional hierarchical relationship, prof NH = professional nonhierarchical relationship.
Emotion Conditions
The vignettes were used to prepare three emotion conditions: (a) base condition, in which vignettes were presented as described in the previous section, and did not contain any descriptors of the victim’s emotional states; (b) humiliation condition, in which victims were described as feeling humiliation-related emotions as a consequence of the offense, by adding the phrase “which made her feel” and one of the following emotion terms to the end of the sentence: “ashamed”, “humiliated”, or “embarrassed” (added to 11 vignettes each, each assigned to every theme—ridicule combination within the speech items), and “belittled”/“disgraced” (used across 11 vignettes interchangeably, collectively assigned to every theme—ridicule combination within the speech items); (c) anger condition, in which victims were using the same phrase at the end of the sentence described as feeling anger-related emotions, using the following terms: “angry”, “irritated”, “annoyed” (added to 11 vignettes each, each assigned to every theme—ridicule combination within the speech items), and “enraged”/“furious” (used across 11 vignettes interchangeably, collectively assigned to every theme—ridicule combination within the speech items).
Procedure
Participants were instructed to carefully read examples of people experiencing negative treatment from others (i.e., the vignettes) and indicate whether they consider that there is a lack of dignity or a lack of respect in how the person is being treated on a 6-point scale: much more about respect (1), moderately more about respect (2), slightly more about respect (3), slightly more about dignity (4), moderately more about dignity (5), and much more about dignity (6) (for a variation on this approach with separate dignity-relatedness and respect-relatedness ratings, a follow-up study was conducted which is reported in the supplementary materials). They were told that each example is unique and asked to read the entire sentence carefully before they decided on their response. The vignettes across three emotion conditions were presented in blocked fashion with the order of vignettes randomly shuffled for each participant. The entire list of 44 baseline condition vignettes was presented within the first block (i.e., on the first page). The humiliation and anger blocks were presented in two order conditions: (a) humiliation on the second page and anger on the third page; and (b) anger on the second page, humiliation on the third page. Each participant was randomly assigned to one of these order conditions using block randomizations, which ensured the distribution of participants was as even as possible (Nhumiliation-anger = 35) as in the other (Nanger-humiliation = 36). Once the participant progressed forward to the next page, they could not return to pages which they had completed. Next, participants were asked to complete a short version of (Clifford et al., 2015) moral foundation vignettes (Crone et al., 2021) with 36 items; report their political affiliation with most popular U.K. political parties; as well as report their fusion with the political party using an online version of the identify fusion diagram (Swann et al., 2009), which we plan to analyze and report on in a separate paper, once we achieve an adequate sample size for individual difference analyses.
Statistical Analyses
t-tests
To investigate the effects of the emotion condition on dignity-relatedness ratings, we tested for rating shifts across vignettes in different conditions. We calculated the average rating shift for every subject and every emotion condition pair by subtracting raw dignity-relatedness ratings for each vignette across relevant pairs of emotion conditions (humiliation—base, anger—base, and humiliation—anger) and averaging across 44 vignettes for every subject. We used a Bayes Factor one sample t-tests to compare average shifts across conditions to zero (null hypothesis). Within the base condition (vignettes without information about the emotions felt by the victim), we calculated participant-level mean dignity-relatedness ratings for vignettes that belong to the ridicule-absent group and for those that belong to the ridicule-present group and then compared these participant-level means using a Bayesian paired-samples t-test. For both tests, we used the BayesFactor toolbox in R programming environment, and the default options for priors: variance of the normal population was assigned the noninformative Jeffrey’s prior distribution, and the standardized effect size was assigned a Cauchy distribution with r = 1/√2 (Morey & Rouder, 2011).
General Linear Models
Dignity-relatedness ratings were also modeled using Bayesian general linear mixed modeling, an analysis approach suitable for hierarchical data (e.g., data of multiple individuals across several conditions). We used this approach to test for effects of any other design dimensions (theme of the offense, communication mode, the nature of the relationship, victim’s gender, presence of a vulnerability factor) on dignity-relatedness ratings, as well as to check if the results obtained by t-test analyses (any effects of emotion condition and ridicule presence) change when other vignette feature dimensions are accounted for in the model. The mixed-level modeling approach additionally allowed us to specify and test for random effects across subjects and items. We fitted the models using the brms package (Bürkner, 2017), which uses Hamiltonian Monte Carlo samplers of the Stan statistical computing software to compute Bayesian posterior estimates (Stan Development Team, 2023). For each model we used the default priors automatically selected within the brms package: for intercepts, we used a Student’s t distribution centered around four with 3 degrees of freedom and a scale factor of 2.5; for standard deviations of group-level effects we used a half Student’s t distribution (>0) with 3 degrees of freedom with the peak at 0; for correlations of group-level random effects we used Lewandowski–Kurowicka–Joe distribution with uniform densities; and for b estimates we used flat priors. Models were estimated using five sampling chains, each of which had 20,000 warmup samples and 20,000 postwarmup samples, thus totalling 80,000 postwarmup samples. We took a forward selection approach, testing whether addition of terms improves the posterior predictive accuracy of the model (i.e., model fit). Using posterior out-of-sample simulations and the leave-one-out (LOO) cross-validation approach, we calculated expected log pointwise predictive density (ELPD) estimates, which have been shown to be more robust than other within-sample model fit approximations (e.g., the Watanabe–Akaike Information Criterion) in cases where priors are weak and individual observations are influential (Vehtari et al., 2017). The first model consisted of random intercepts for participants and items, and from there we took forward steps by (a) adding a fixed slope term and (b) a random slope term for different participants, for each variable in the following order: emotion condition, ridicule, presence, theme, communication mode, relation type, victim’s gender, and presence of a vulnerability factor (see Table 2). In every step in our forward model comparison procedure, when ELPD showed an increase (indicating improved posterior predictive accuracy), we calculated Bayes Factors for model comparison (Vehtari et al., 2017) to test the hypothesis that the model containing the additional term was better than the reference model. We only accepted the additional term when evidence for improvement was substantial (BF > 3) and, for new fixed effects terms, when the 95% credibility intervals for group-level effect estimates of the added term did not overlap with zero.
Forward Model Selection Path: Model Fit and Comparison Statistics
Note. ELPD = expected log pointwise predictive density, SE = standard error, Interc. = intercept, BF = Bayes Factor for the hypothesis that the current model is better than the reference model, Ref. M. ID = Reference model ID, FE = fixed effect, CI = credibility intervals.
Results
Emotion Conditions
The pooled mean of dignity-relatedness ratings in the base condition (Mbase = 3.45 ± 0.07) was lower than in the humiliation condition (Mhumiliation = 3.67 ± 0.08), and slightly higher than in the anger condition (Manger = 3.39 ± 0.08). The pooled mean rating shift from the baseline condition to the humiliation condition (i.e., the difference in ratings first averaged across vignettes and then participants, also shown in Figure 1) was positive Mpooled(humiliation-base) = 0.22 ± 0.08, 95% CI [0.06, 0.37], and the Bayes Factor t-test indicated substantial evidence (BF10 = 4.113% ± 0.01%) that the shift was greater than zero. The effect size estimate indicated a medium effect (dCohen = 0.443). The pooled mean rating shift from the baseline condition to the anger condition was negative but small (Mpooled(anger-base) = −0.06 ± .08, 95% CI [−0.22, 0.10]), and the t-test Bayes Factor indicated substantial evidence in favor of the null hypothesis (BF10 = 0.275% ± 0.09%). The pooled mean rating difference between the humiliation condition and the anger condition was positive (Mpooled(humiliation-anger) = 0.27 ± 0.07, 95% CI [0.13, 0.41]), and the Bayes Factor t-test indicated very strong evidence (BF10 = 84.753% ± 0%) that the difference was greater than zero. The effect size estimate indicated a medium effect (dCohen = 0.562).

Dignity-Relatedness Rating Shifts Across Emotion Conditions.
Ridicule Presence
Mean dignity ratings and mean shifts are plotted in Figure 2. The pooled mean of dignity-relatedness ratings in the ridicule-present condition (Mpresent = 3.49 ± 0.09) was slightly higher than in the ridicule-absent condition (Mabsent = 3.40 ± 0.07). However, Bayesian paired sample t-test, comparing the pooled mean of the paired mean differences against zero (MDpresent-absent = 0.09 ± 0.07, 95% CI [−0.03, 0.20]), indicated weak evidence in favor of the null hypothesis that there is no difference (BF10 = 0.398% ± .05%).

Dignity-Relatedness Ratings Across Ridicule Conditions.
Offense Theme and Vulnerability Factor Presence
The best-fitting model in our general linear mixed-effects model selection procedure (presented in Table 2) included main effect terms for the emotion condition, the theme of the offense, and vulnerability factor presence. The slopes of these three main effects were allowed to vary across individuals, and the intercepts were allowed to vary across individuals and items. Addition of each of these terms was warranted by extreme evidence (BF > 100) for improvements in the model fit. Addition of these terms did not lead to any substantial changes in estimates for other terms (see Table 3). Humiliation condition led to an average rating shift of 0.22 points toward dignity on the dignity-respect scale, 95% CI [0.06, 0.38], whereas anger-related terms did not substantially differ from the base ratings (βanger-base = −0.06, 95% CI [−0.22, 0.10]). Physical appearance themed vignettes had significantly higher dignity-relatedness ratings than vignettes of the remaining three themes (0.60 scale points higher 95% CI [0.25, 0.95] than the second-ranked behavioral competencies theme). Vignettes containing mentions of potential vulnerabilities to offenses (e.g., disability), had dignity-relatedness ratings that were .50 scale-point higher 95% CI [0.15, 0.84] than other vignettes. Mean dignity ratings across themes and vignettes are shown in Figure 3.
Fixed Effect Estimates for All Candidate Models.
Note. M. ID = model ID. Predictors whose estimates do not cross zero are marked with “*”, the winning model is bolded.

Dignity-Relatedness Ratings Across Themes and in the Presence of Vulnerability Factors.
Discussion
In the current study, we investigated if people’s perceptions of offensive treatments can distinguish dignity violations from respect violations, and if perceptions of dignity violations depend on the offense victim’s resulting emotional states as well as the presence of ridicule in the offense. We did this by asking participants to rate how well vignette descriptions of offenses correspond to violations of dignity as opposed to violations of respect across three conditions manipulating the emotions that victims felt (baseline, humiliation, and anger), as well as across two conditions manipulating the ridicule content of the vignettes (present and absent). Humiliation and ridicule were both hypothesized to shift the ratings toward the dignity end of the scale, whereas anger was hypothesized to shift the ratings toward the respect end of the scale. Moreover, we investigated the impact of additional variables including offense theme, communication mode, the nature of the relationship, victim’s gender, and presence of vulnerability factors.
Our results show substantial evidence that the baseline ratings shifted toward the dignity end of the scale in the humiliation condition, and very strong evidence that dignity-relatedness ratings in the humiliation condition were higher than in the anger condition. This indicates that information about a victim’s subjective experience of humiliation increases perceptions of offensive treatments as dignity violations, as opposed to respect violations. These findings were replicated in our follow-up study in which participants made separate dignity-relatedness and respect-relatedness ratings, which showed extreme evidence that humiliation condition increases dignity-relatedness (independently from respect-relatedness), and very strong evidence that these shifts are stronger than shifts in respect-relatedness (see Supplementary Figure 1). Conversely, our results show no evidence that the anger condition shifted the baseline ratings toward the respect end of the scale, with substantial evidence for the null hypothesis, suggesting that anger-related emotions do not play an important role in distinguishing dignity from respect. As for the ridicule manipulation, our results show no evidence that vignettes containing ridicule had higher dignity-relatedness ratings, with weak, inconclusive, evidence in favor of the null hypothesis. These effects remained stable when controlling for additional design variables. As for the effects of other variables, there was extreme evidence for the effects of the physical appearance theme and presence of vulnerability factors on increases in dignity-relatedness ratings, suggesting these factors increase perceptions of offenses as violations of dignity.
Our findings that victims’ humiliation-related emotions increased perceptions of offensive treatments as dignity violations (as compared with both the baseline and anger conditions) is consistent with the idea that perceptions of dignity violations uniquely involve considerations about the victims’ humiliation and thus support new philosophical accounts of violations of dignity as uniquely involving humiliation (Barclay, 2018; Etinson, 2020; Killmister, 2020; Luban, 2009; Margalit, 1996; Statman, 2000). This finding suggests that people rely on emotional cues about the subjective experience of self-conscious emotions to infer whether violations target dignity or respect. This inference may to some extent, and in line with appraisal theories (Fernández et al., 2015), involve theory of mind about the victims’ appraisals of the offense (i.e., thinking whether the victim may be internalizing or accepting the humiliated state). Alternatively, it could be that people have a lexical representation or hold explicit knowledge about the relations between constructs of dignity and the self-conscious emotion tags. This alternative account is, however, relatively unlikely because ethnographic investigations which tap into explicit and declarative knowledge so far have done little to identify unique features of dignity (Barclay, 2016; Thomas & Lucas, 2019). For example, asking people open-ended questions about dignity often yields unwieldy and conceptually disunified lists of behaviors said to confer or undermine dignity (Barclay, 2016; Thomas & Lucas, 2019). Given that people are often unaware of what exactly shapes their subtle moral preferences (Haidt, 2001), it may be that it was necessary, as in this study, to adopt a psychological methods approach and elicit the difference by engaging people in social reasoning tasks probing their implicit knowledge of these constructs. This finding is, to our knowledge, the first piece of evidence that laypeople share intuitions about dignity that are different from respect.
We found no evidence that victims’ experience of anger-related emotions systematically shifts dignity-relatedness ratings as compared with the baseline. From a philosophical perspective this finding is however unsurprising because, unlike humiliation, anger does not feature in philosophical discourses around differences between respect and dignity. Thus, anger does not appear to have a unique relationship with respect that would shift the ratings toward either end of the scale. These results also do not provide support for the idea that informing participants that victim feels anger would (compared with instances where the emotional state of the victim is unknown) disambiguate what the victim is feeling and therefore decrease perceptions that the negative treatment is related to dignity. It could be that participants did not make any assumptions about the victim’s emotional states in the baseline condition. It could also be that participants assumed anger-related emotions more frequently than humiliation-related emotions, so that the disambiguation of the emotions the offense caused, in the other two conditions, would only shift the dignity score in the humiliation condition. Importantly, the anger condition served as an important alternative control for humiliation, demonstrating that dignity has a special relation to a specific cluster of self-conscious emotions, feelings of humiliation, shame and embarrassment, and not indiscriminately to all negatively valenced emotions.
We did not find support for our hypothesis that vignettes containing ridicule should have higher dignity-relatedness ratings. However, our weak evidence in favor of the null hypothesis is also relatively inconclusive. It may be the case that ridicule presence has a small effect that we could not detect with our sample geared for detecting small-to-medium effects (Cohen’s d > 0.30). However, in the context of our other findings, it is safe to say that ridicule presence does not have as strong an effect as some other factors (i.e., subjective humiliation, physical appearance, and vulnerability presence). This does not negate philosophical accounts that emphasize the role of ridicule in humiliating victims, along with other expressions of disparagement such as disgust and contempt. It may suggest that accounts of the humiliated state which de-emphasize the role of subjective humiliation may be in tension with lay conceptions of dignity.
We further found extreme evidence that the physical appearance theme and the presence of vulnerability factors increased dignity-relatedness ratings, which is somewhat surprising for philosophical accounts, which do not highlight these factors above others as specially connected to dignity. It is possible that vulnerability factors and physical appearance are pertinent for participants’ assessment of victims’ humiliated state. For example, it may be that information about disabilities, illnesses and other vulnerability factors included in our vignettes signal to participants that the offense could be an instance of ableism (i.e., a view that disabled individuals are inferior), undermining the victim’s equal social status, and leads them to deduce a humiliated state or humiliated feelings. Although disability segregation and discrimination is now outlawed in many countries and internationally (Breslin & Yee, 2002; Sabatello & Schulze, 2014), discrimination against individuals with disabilities persists (Allen et al., 2021; Equality and Human Rights Commission, 2016; Wilson, 2021), and the discourse that disability is pitiful (which implies a lowered social standing) may still be pervasive (Hayes & Black, 2003; Hughes, 2013). This may feed into people’s inferences about the humiliated state. Similarly, widespread negative attitudes around people with obesity or other disvalued physical characteristics (Kennedy, 1990; Rahal et al., 2021) might signal to participants that the victims with stigmatized physical characteristics are likely to be relegated to inferior social status or related humiliated state.
More broadly, it could be that the awareness that the victim durably or immutably possesses the characteristics portrayed as negative in the offense (or otherwise commonly seen as negative), plays an important role in distinguishing uniquely dignitarian violations. Our vignette offenses targeting cognitive, physical, and behavioral competences mostly described perpetrators’ reactions to a particular event (e.g., playing a sports game, working on a project together) in which the offense may provide limited insight into victim’s enduring characteristics (i.e., the victim may have lacked a competency just in that particular situation), whereas offenses targeting physical appearance or vulnerabilities were often not prompted by transitory events: as such, vignettes may be perceived as more diagnostic of more durable and immutable characteristics of the victim. Thus, it could be that perceptions of dignity depend on the inferences that observers make about victims’ characteristics and the transitoriness of these characteristics. Offenses targeting more durable and immutable characteristics of victims may be considered more threatening to dignity if perceived to be attacking the victim’s identity (Elison & Harter, 2007). This view is consistent with our main findings that presence of humiliated emotions increases dignity relatedness, because these emotions are characteristic for situations in which people’s immutable attributes are tainted or viewed negatively (Tangney, 1996). Whatever the causes may be, our findings suggest that dignity considerations (as perceived by lay people) are especially relevant when addressing or treating individuals with particular vulnerabilities, and when communicating about issues regarding physical appearance.
Implications and Future Directions
These findings have important implications for philosophical discourses about the ethics of dignity and respect, as well as for implementation of these concepts in organizations, health care and law. This study dispels skepticism about whether dignity captures a unique value (Macklin, 2003), recognizable by laypeople (Barclay, 2016; Thomas & Lucas, 2019), by demonstrating that laypeople share intuitions about dignity that are different from respect in ways congruent with the new theories of dignity. This implies that the concept of dignity is defensible in ethics and provides empirical foundations for theoretical ethical frameworks based on dignity. The implication is that the public use of this concept is defensible too, and that it should be possible to implement health care and organization policies that seek to promote the dignity of clients, patients, and employees. If laypeople can identify these unique features in our experiment, they should be able to do so in their ongoing interactions in health care or workplace settings without any prior training. Ill-advised training that conflates dignity with other values could, however, interfere with people’s intuitions about these constructs, and lead to poorer differentiation between them. More effort should be put into devising appropriate education and intervention strategies that incorporate this new understanding of dignity.
Our findings suggest that victims’ subjective feelings and their identity are incorporated into lay conceptions of dignity. This may generate ethically challenging cases, such as, to illustrate, a racist patient who may feel humiliated when treated by a doctor belonging to a minority, but of whom most philosophical accounts would deny that he suffers a loss of dignity. Our study does not speak to whether or not such subjectivity trumps the influence of consensus-views; future research may consider teasing apart the relative contributions of these two factors to dignity perceptions, helping to identify potential subjectivity biases that could be targeted by ethics education interventions.
Recent philosophical accounts operate with a broader conception of humiliation as a state where a person is said to be lowered, demeaned or belittled, whether they actually experience humiliation or not (Barclay, 2018; Etinson, 2020; Killmister, 2020; Waldron et al., 2012). This accommodates important cases such as, for example, a person with a serious cognitive impairment who can be humiliated and degraded and thus suffer a dignity violation when she is left naked and exposed to public view, even if she herself does not feel humiliated (Barclay, 2018). Further research is needed to investigate if lay people perceive such cases specifically as dignity violations, in spite of the absent cue for self-conscious emotion.
Our findings imply that for laypeople, dignity considerations are particularly relevant when interacting with others who have specific vulnerabilities, and when communicating about physical appearance. While we offer various interpretations relating dignity to discrimination, stigmatization and offense immutability, further research is needed to investigate these dimensions, leaving large scope for psychological research to inform the ethics of dignity, with potential to advance our understanding of bullying, ostracism and broadly construed dehumanization (Bastian & Haslam, 2010; Haslam, 2006; Haslam & Loughnan, 2014; Kteily & Landry, 2022). Our methods can be adopted to study differences in notions of dignity between individuals and cultures, providing possibilities for future research to advance our understanding of dignity-undermining and conferring conditions, and provide targeted recommendations for policy or interventions in specific settings. Future research could also fine-tune our understanding of links between self-conscious emotions and dignity. Specifically, it is possible that more nuanced hypotheses about subtler differences between emotion labels can be formulated. For instance, shame is posited to be more durable and less transient than embarrassment (A. H. Buss, 1980; Kristjánsson, 2010), which might indicate to observers that the victim’s durable and immutable characteristics are targeted, and thus, the offense may be more related to dignity. Contrastingly, embarrassment is more closely related to public as opposed to private offenses (Keltner & Buswell, 1996; Tracy & Robins, 2004), which we could speculate can indicate to the observer that the offense entailed a loss of social standing, and make it more related to dignity. More broadly, there is a multitude of theories about the distinctions between these emotions and little consensus on differentiating criteria (Crozier, 2014), so further work is needed to map these possible differences to distinctions between dignity and other values. Further research is also needed to investigate links between dignity and emotions outside the context of dignified treatment. For example, guilt is a self-conscious emotion that is rarely felt as a result of a mistreatment by others, that is often felt as a result of one’s own wrongdoings and subsequent reparation motives, and that has weaker links to immutability than humiliated emotions (Tangney, 1996), and thus may relate to dignity less than other self-evaluative emotions.
More broadly, our findings suggest that people have moral concerns that are uniquely dignitarian, and which are shaped by the emotional content of social interactions and by people’s inferences about social hierarchy dynamics. These insights carve a new space for basic morality psychology investigations into the cultural or evolutionary origins of these concerns, and the cognitive and emotional processes that underpin them. The psychology of dignity should also seek to characterize individual differences in the importance people place on these concerns and understand how they shape people’s everyday interactions with others.
Conclusion
What it means to treat and be treated with “dignity” is relevant for all humans. We show that dignity uniquely involves accounting for the self-conscious emotions of those subject to violations and provide an empirical backing to dignitarian ethics with applications in health care, organizations and law.
Supplemental Material
sj-docx-1-psp-10.1177_01461672251319063 – Supplemental material for Dignity Is Distinct From Respect: How Treating Others With Dignity Entails Accounting for Their Self-Conscious Emotions
Supplemental material, sj-docx-1-psp-10.1177_01461672251319063 for Dignity Is Distinct From Respect: How Treating Others With Dignity Entails Accounting for Their Self-Conscious Emotions by Milan Andrejević, Jakob Hohwy and Linda Barclay in Personality and Social Psychology Bulletin
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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by an Australian Research Council grant (ARC DP190100734) to L.B. We have no competing interests.
Supplemental Material
Supplemental material is available online with this article.
Author Note
The hypotheses, design, and analyses were preregistered: https://osf.io/7b4ax. Data of all participants, materials including the Dignity & Respect Vignettes, as well as the analyses scripts that support the findings of this study are publicly available on an Open Science Framework (OSF) repository (
). All authors contributed to conception and design. M.A. programmed the experiment, collected and analyzed data. L.B. and M.A. drafted the article. All authors reviewed and revised the manuscript.
References
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.
