The role of social status in sarcasm interpretation: Evidence from the United Kingdom and China

ABSTRACT We investigated the effect of culture and social status on sarcasm interpretation. Two hundred U.K. participants and 200 Chinese participants read scenarios in which the final comment could be either literal or sarcastic criticism and the speaker had equal, higher, or lower social status compared to the recipient. Comments were rated on degree of sarcasm, amusement, politeness, and aggression. Results showed that compared to literal criticism, sarcasm was viewed as being more polite and less aggressive in the United Kingdom but more aggressive in China. When the speaker had higher social status, sarcasm was rated less amusing and marginally more aggressive in the United Kingdom but less aggressive and more polite in China. Results have theoretical implications, in that theories designed to explain the emotional impact of sarcasm (such as the Tinge Hypothesis) need to be modified to take cultural differences into account, as well as practical implications for successful cross-cultural communication.


Introduction
Imagine that someone was 30 minutes late for a workshop and another delegate said to them, "You are so early!"How would you interpret this comment?How amusing would you find it?Would you think that the speaker intended to be aggressive?The answers to these questions may depend on who the speaker is and what their relative social status is compared with the recipient.For example, imagine this comment was made by a professor to a student, or by a student to a professor, or by a professor to another professor.Do you think your answers would depend on your own cultural background?The aim of the current research is to answer these questions by manipulating the relative social status of the speaker and recipient and examining the perception of sarcastic criticism across the United Kingdom and China.
Although there is some debate in the literature regarding precise definitions, sarcasm is often described as being a type of verbal irony in which the speaker expresses his or her intended meaning by saying the opposite and for which there is typically a victim (see e.g., Kreuz, 2020, for recent discussion).Researchers who have examined irony in political practice have stated that in a broader sense, ironic interactions involve a power difference between the speaker and the listener (Hutcheon, 1994).The theory of psychological distance (Benign Violation Theory, e.g., McGraw & Warren, 2010) suggests that the power differences between the speaker and the listener may affect the perception of the social distance between them and subsequently influence the interpretation of jokes (Kant & Norman, 2019).In addition, research on language production has demonstrated that the speaker is more likely to use polite language if the recipient has higher social status (Brown & Levinson, 1987;Holtgraves, 1994).Thus, characters' relative social status is a factor that is likely to be important in sarcasm perception.Indeed, previous research suggests that certain speaker characteristics can influence sarcasm comprehension, such as speakers' age (Jared & Pandolfo, 2021), occupation (Katz & Pexman, 1997;Pexman & Olineck, 2002a), and communicative style (e.g., Regel et al., 2010) and whether the speaker was known to be sarcastic before (e.g., Țurcan et al., 2020).However, there has been little research on the role of social status in sarcasm interpretation.
Results from cross-cultural studies suggest that people from different cultures may consider inequality acceptable to varying degrees.That is, people from individualist cultures are more likely to see themselves as equal to others, whereas people from collectivist cultures tend to accept a hierarchical society (Hofstede, 2001;Schwartz, 1994).The Benign Violation Theory (e.g., Kant & Norman, 2019;McGraw & Warren, 2010) also suggests that culture may influence people's perception of power differences.That is, in some cultures, the social distance between lower power and higher power people might be smaller.Zhu and Filik (2023) found cultural differences in sarcasm interpretation; that is, Chinese participants rated sarcastic criticism as being more aggressive than literal criticism, whereas U. K. participants perceived sarcastic criticism as more amusing and polite than literal criticism.Moreover, the Benign Violation Theory (e.g., Kant & Norman, 2019;McGraw & Warren, 2010) also suggests that culture may influence the shared opinions on a good or bad behavior in a society.For example, culture may have an impact on how appropriate people consider the behavior of individuals who have higher or lower power, which then may influence their perception of a humorous utterance as being either kind or malign.Thus, it is of interest to examine in more detail the interaction between cultural background and relative social status in the perception of critical comments.

Social functions of sarcasm
Regarding the social functions of sarcasm, there have been mixed findings as to whether sarcasm mutes the negativity of criticism as suggested by the Tinge Hypothesis (e.g., Dews et al., 1995;Dews & Winner, 1995) or instead enhances the perceived negativity (e.g., Colston, 1997, see Filik, 2023, for a recent overview).There are a number of possible reasons for these mixed findings.For example, the differences may be due to different modes of presentation; that is, Colston (1997) examined written irony, whereas Dews and Winner (1995) used recordings of ironic conversations with intonation, which may dilute the negativity of a critical utterance.Alternatively, Pexman and Olineck (2002b) suggested that the differences in sarcasm interpretation might depend on the perspective that the raters adopted.For example, Bowes and Katz (2011) found that participants taking the speaker perspective found sarcasm more humorous than when they were taking the recipient perspective.However, some research showed no or small significant effects of perspective in sarcasm interpretation.For instance, Toplak and Katz (2000) found that participants viewed sarcasm as more impolite than literal criticism when taking both the speaker and the recipient perspectives.Finally, a recent study that investigated sarcasm perception in participants from the United Kingdom and China found that whereas U.K. participants rated sarcasm as being more amusing and polite than literal criticism, Chinese participants rated sarcasm as being more amusing yet also more aggressive (Zhu & Filik, 2023).Thus, whether or not sarcasm mutes or enhances the perceived negativity of criticism, and the factors that influence this, are yet to be established.

Speaker characteristics and sarcasm interpretation
Previous research on the comprehension of indirect language (in this case, indirect requests) suggests that speaker characteristics are important (Holtgraves, 1994).Holtgraves proposed that knowledge of the speaker and the context could make the nonliteral interpretation of an utterance possible.For example, when interpreting the negative statement, "It's noisy in here," spoken by a person with higher social status, the hearer could quickly gather that a noisy room is not desirable and the speaker does not like a noisy room.Then the hearer may take action to shut the door.The results showed that the speaker's status could influence the hearer's accessibility of the intended meaning; that is, when the speaker has a higher status, the hearer develops an expectation of how a statement is to be interpreted, and the literal meaning of the indirect request may not be activated.
As for the role of speaker characteristics in sarcasm comprehension, Țurcan et al. (2020) examined whether knowledge that the speaker had been sarcastic before influenced ease of sarcasm processing in an eye-tracking study.The results showed that when the speaker had been introduced as being sarcastic, readers had no difficulty in processing sarcasm compared with literal criticism.In contrast, when the speaker was not previously sarcastic, the sarcastic remark was more difficult to process than literal language.Jared and Pandolfo (2021) investigated the role of speaker age in irony comprehension using rating tasks (Experiment 1) and eye-tracking during reading (Experiment 2).They manipulated speaker age by using character names that were typical of older and younger individuals.Results showed that the speaker age did not affect ratings of irony interpretation but did have an impact on reading times when the ironic statements were not familiar.That is, for statements that are less often used as ironic criticism, knowledge that the speaker was old resulted in extra processing time when reading the comment compared with when knowing that the speaker was young.Katz and Pexman (1997) investigated the effect of speaker's occupation in metaphor and irony interpretation and found that speaker's occupation, which may be related to social status, influenced participants' interpretation of irony.Specially, comments such as "Children are precious gems" uttered by a cab driver were rated as being more sarcastic than when uttered by a scientist.Pexman et al. (2000) further investigated this issue with self-paced reading in which passages were presented in a word-by-word moving-window format.They manipulated speaker's occupation in three conditions: high irony occupation (e.g., comedian), high metaphor occupation (e.g., artist), and no occupation mentioned.The results showed that comments made by speakers who had high irony occupations and high metaphor occupations resulted in longer reading times than those made by speakers with no occupation mentioned, especially for unfamiliar statements.Pexman and Olineck (2002a) then investigated the effect of occupation stereotypes in the interpretation of ironic intent when statements were literal criticism or sarcastic criticism (e.g., "That sounds pretty dull/exciting.")rather than metaphors.Results showed that in the presence of contextual information which supported a literal or sarcastic interpretation, speaker occupation did not affect ratings (Experiment 1).However, in more "neutral" contexts (Experiment 2), speakers with more sarcastic occupations were rated as being more sarcastic, suggesting that occupation stereotypes may be more influential when contextual information is less informative.Pexman and Olineck (2002a) further investigated the aspects of occupation stereotypes that may be related with sarcasm use in Experiment 3. The dimensions included tendency to be humorous, tendency to be critical, tendency to be insincere, tendency to have good relationships, tendency to be aggressive, education level, and social status.Results showed that social status was positively correlated with education level, which was negatively correlated with sarcasm use tendency.That is, sarcastic statements were judged as being more likely to be made by people with lower perceived education levels (who may in turn have lower social status).The authors explained the results with reference to the Implicit Display Theory (Utsumi, 2000), which suggests that sarcastic speakers have a negative attitude and this attitude is expressed indirectly.Since sarcasm may have a face-saving function, speakers with lower social status, who are concerned about directly insulting recipients with higher social status, may use sarcasm to hedge the offense.

Social status and sarcasm interpretation
In a more direct investigation of the influence of social status on sarcasm interpretation, Gucman (2016) manipulated the relative social rank of the speaker and recipient.The social rank manipulation involved three conditions: speaker superior to recipient, speaker inferior to recipient, and speaker equal to recipient.The scenarios were events that happened between students and their professors or between colleagues, for example, being late for a meeting or trouble with writing/rating an assignment.Their rating items included smartness, criticality, humorousness, and offensiveness.The experiment was conducted in Polish, and participants were Polish-speaking university students.Results showed that when the speaker had a higher social status (e.g., professor), ironic criticism was rated as being less critical than literal criticism.In addition, when the speaker had higher social status, ironic criticism was rated as less critical and offensive than when the speaker had lower social status.Gucman (2016) suggested that criticizing ironically serves a face-saving function and protects relations between the speaker and the recipient, especially when the speaker has higher social status.This argument seems to go against Pexman and Olineck's (2002a) discussion that people with lower social status were more likely to use sarcasm to save face than people with higher social status, although note that this study did not directly investigate relative social status.It is also important to note that Gucman (2016) did directly investigate the role of relative social status but used different scenarios across conditions, and the experiment was a between-subject design that only focused on one type of relationship (student and professor).Thus, it is of interest to further investigate the influence of relative social status on sarcasm perception to control for extraneous variables and examine whether effects generalize across a greater range of social relationships.

Current study
In the current study, we investigated the role of social status in sarcasm interpretation by manipulating the relative social status of speakers and recipients; that is, speakers had equal social status with recipients, higher social status than recipients, or lower social status than recipients.Data were collected in both the United Kingdom and China, with participants being presented with the same type of stimuli across the two countries (presented in English for participants the United Kingdom and in Chinese for participants in China).Specifically, participants were required to read scenarios in which the final sentence was either a sarcastic or literal criticism made by one of the characters.After each scenario, participants were asked to rate the final comment on four dimensions: How sarcastic is the comment?How amusing is the comment?Does the speaker intend to be aggressive?Does the speaker intend to be polite?

Predictions
Regarding the overall comparison between perceptions of sarcastic versus literal criticism, we predicted that if sarcasm enhances negativity (Colston, 1997), sarcastic comments would be rated as less amusing than literal criticism and speaker intent would be judged as being less polite and more aggressive.In contrast, if sarcasm mutes negativity as suggested by the Tinge Hypothesis (e.g., Dews & Winner, 1995), sarcastic comments would be rated as being more amusing and the speaker intent would be rated as being more polite and less aggressive.It is important to note that these findings may differ across participants in the United Kingdom and China.That is, Zhu and Filik (2023) found that participants in the United Kingdom rated sarcastic comments as being more sarcastic, amusing, and polite than literal criticism, whereas participants in China rated sarcastic comments as being more sarcastic, amusing, and aggressive than literal criticism.
In relation to the effect of social status, Gucman (2016) found that ironic criticism was viewed as being less critical and offensive when the speaker had a higher social status.Thus, following Gucman, we expected participants to find sarcastic comments to be less aggressive when the speaker had a higher status compared to when the speaker had a lower status or an equal status with the recipient.Gucman (2016) also found that when the speaker had a higher social status, ironic criticism was rated as less critical than literal criticism; thus, we predicted that sarcastic comments would be rated as being less aggressive than literal criticism especially when the speaker had a higher status.In contrast, Pexman and Olineck (2002a) suggested that low status individuals might use sarcasm rather than literal criticism to save face.Following this, the alternative prediction was that speaker intent would be rated as being more polite following sarcastic than literal criticism when the speaker had a lower social status.However, it is important to note that Gucman's (2016) participants were undergraduate students in Poland and Pexman and Olineck's (2002a) participants were undergraduate students in Canada.Thus, it is informative to examine whether data from participants in the United Kingdom and China will pattern in a similar manner.
As for the interaction between social status effects and culture, according to cross-cultural studies (Hofstede, 2001;Schwartz, 1994), people who have lower social status tend to avoid potential confrontation with people who have higher social status more so in a hierarchical society (such as China) compared to an egalitarian society (such as the United Kingdom).Following this, it would be regarded as being more unacceptable for a lower social status person to criticize a higher status person in China.Sarcasm was also considered as more aggressive than literal criticism in China (Zhu & Filik, 2023).Therefore, we predicted that for participants in China, it would be perceived as being more aggressive when a lower social status person criticizes a person who has a higher social status, especially in a sarcastic way.In the United Kingdom, people are more likely to see themselves as equal to others and consider sarcasm as being more amusing than literal criticism.Thus, we predicted that the rating differences between equal (or lower) and higher conditions in measures of sarcasm, aggression, amusement, and politeness might be less great than in China.

Participants
Two hundred English native speakers, aged 18 to 25 years (M = 19.51,SD = 1.62) from the United Kingdom were included in this study, with 62 men and 138 women.Within this sample, 170 U. K. participants were recruited from the University of Nottingham through the School of Psychology's research participation scheme, from which first year undergraduates could gain course credits by taking part in the experiment.Another 30 U.K. participants (who were all university students) were recruited from Prolific (Prolific, London) to balance the gender split with the China sample.Two hundred Chinese native speakers, aged 18 to 28 years (M = 20.34,SD = 1.72) were included, with 70 men, 127 women, and 3 not reported.Within this sample, 113 Chinese participants were recruited through tutors sending the study link to their first-year undergraduate students and a further 87 Chinese participants were recruited via a WeChat group consisting of university students.

Materials and design
The experiment had a mixed design, with participants' cultural background as a between-subjects factor (United Kingdom vs. China), and two within-subject factors; that is, two comment types (literal vs. sarcastic) by three levels of social status of speaker versus recipient (higher status vs. lower status, equal status vs. equal status, lower status vs. higher status).Forty-eight experimental scenarios were created (see Table 1 for an example).There were six within-subject and withinscenario conditions.To counterbalance, participants saw each experimental scenario in only one of the six conditions; thus, seeing eight scenarios in each condition.We also included 16 filler scenarios (either neutral or positive comments) to balance the potential negative impression of the critical comments.Thus, there were six versions of the survey, each of which included 48 experimental scenarios and 16 filler scenarios.
All the scenarios were either interactions between individuals of equal status, such as the man and his brother, the student and their classmate, the boss and their business partner, the professor and another professor, the lawyer and their business partner, or interactions between individuals of dissimilar status, such as professors and students, bosses and employees, workers and managers, doctors and patients/nurses, university officers and students, police officers and drivers, lawyers and assistants/employees, and parents and sons/daughters (adults).Scenarios between bosses and employees, lawyers and assistants, doctors and patients, and police officers and drivers included events that happened in the workplace, such as tasks and meetings.Scenarios between parents and their grown-up children were related to events happening in daily life.Scenarios between professors and students included events relating to courses, assignments, meetings, and so on.
Materials for Chinese participants were principally translated versions of the materials that were shown to U.K. participants, but some scenarios were excluded if they failed the pretest (below) with suitable replacements created for the Chinese version (see Table 2 for an example).All materials are available at https://osf.io/43yun/.

Pretest of social status and relationship between character pairings
The aim of the pretest was twofold.First, since we aimed to manipulate the relative social status of the speaker versus the recipient, the hierarchical status of the character pairings was determined from the pretest (that is, we adopted materials in which pretest participants reported a status difference in the unequal condition and not the equal condition).Second, we assessed the closeness between the speaker and the recipient in a scenario and controlled it across conditions as much as possible.Previous research has showed that closeness between the speaker and the recipient may influence the interpretation of sarcasm, since a close relationship between the speaker and the recipient may make the speaker's attitude more transparent and thus serve as a cue for sarcastic intent (Pexman & Zvaigzne, 2004).
We first recruited 39 U.K. participants through survey swap websites, such as Survey Circle, to rate the social status of one character compared with the other character in 50 scenarios on Table 1.Example material (English Version).
Speaker higher social status Literal The employee was going on business with their boss to another city.However, the employee didn't arrive on time and they missed the flight.The boss said to the employee: "You are so late."Sarcastic The employee was going on business with their boss to another city.
However, the employee didn't arrive on time and they missed the flight.The boss said to the employee: "You are so punctual." Speaker lower social status Literal The employee was going on business with their boss to another city.However, the boss didn't arrive on time and they missed the flight.The employee said to the boss: "You are so late."Sarcastic The employee was going on business with their boss to another city.
However, the boss didn't arrive on time and they missed the flight.The employee said to the boss: "You are so punctual."Speaker equal social status Literal The boss was going on business with their business partner to another city.However, the partner didn't arrive on time and they missed the flight.The boss said to their partner: "You are so late."Sarcastic The boss was going on business with their business partner to another city.
However, the partner didn't arrive on time and they missed the flight.The boss said to their partner: "You are so punctual."a seven-point rating scale, from Much lower (1 point) to About the same (4 point) to Much higher (7 point), and their relationships, from Extremely distant (1 point) to Neither close nor distant (4 point) to Extremely close (7 point).For example, after reading a scenario, participants were asked to rate "the social status of the manager, compared with the worker" and "the relationship between the manager and the worker" using the above rating scales.We selected scenarios following the criteria that the difference between ratings of social status for the unequal (e.g., the manager, compared with the worker) versus equal (e.g., the worker, compared with their coworker) conditions was more than 1 point and the difference between ratings of closeness of relationship for the unequal and equal conditions was below 1 point.After this filtering process, 37 scenarios remained.We then revised and created another 14 scenarios for a second pretest.We recruited another 40 U.K. participants from survey swap websites to complete the same pretest rating tasks, resulting in another 11 scenarios.
We conducted paired-samples t-tests on these 48 experimental scenarios to compare the relative social status between character pairings across unequal (e.g., the social status of the professor compared with the student) and equal (e.g., the social status of the professor compared with another professor) conditions.Results showed that the social status difference between the speaker and the recipient of the unequal condition was higher than the equal condition, M unequal = 5.59, SEM = 0.05, M equal = 4.09, SEM = 0.02, t (47) = 29.96,p < .001.That is, in the unequal condition, the speaker (or recipient) had a (more than) slightly higher (Slightly higher = 5) status than the counterpart.In the equal condition, the speaker (or recipient) had a (generally) the same (About the same = 4) status as the counterpart.
The relationship between the speaker and the recipient in the equal condition was closer than that in the unequal condition, M equal = 3.69, SEM = 0.11, M unequal = 3.49, SEM = 0.14, t (47) = 2.87, p = .006.As we can see, although there was a statistically significant difference in relationship closeness rating between the two conditions, the closeness score of the equal condition was 0.2 higher than the unequal condition, both of which were near Neither close nor distant (4 point).
We pretested the Chinese stimuli using the same method.We first recruited 40 Chinese participants to rate 49 scenarios that were translated from the English versions.Following the above selection criteria, this resulted in 28 usable scenarios.We then created another 23 scenarios and recruited another 38 Chinese participants to rate the relative social status and relationship closeness between characters to end up with a final set of 48 scenarios for Chinese participants.We conducted paired-samples t-tests, and results showed that the social status difference between the speaker and the recipient of the unequal condition was higher than the equal condition, M unequal = 5.68, SEM = 0.06, M equal = 4.09, SEM = 0.03, t (47) = 28.39,p < .001.There was no significant difference between the equal and unequal conditions regarding the closeness of the relationship between the speaker and the recipient, M unequal = 4.11, SEM = 0.10, M equal = 4.20, SEM = 0.09, t (47) = 1.28, p = .21.

Procedure
The study was approved by the Ethics Committee of the School of Psychology, University of Nottingham (Ref.S1303).Participants were randomly assigned to one of the six versions of the experiment through Qualtrics (Qualtrics, Provo, UT).Participants were required to read the information sheet carefully and then consent to participate in the study.If they did not consent, the experiment would terminate.For all participants who agreed to take part, their age, gender, and native language data were collected.We also included a question to filter participants who had diagnoses of autism, schizophrenia, or dyslexia, as these diagnoses may influence their performance in the reading task (e.g., Kalandadze et al., 2018;Rapp et al., 2013;Reis et al., 2020).Before the scenario reading task, participants were given instructions to read each scenario and complete the ratings carefully.After reading each scenario, participants were required to rate four items: "How sarcastic is the comment?""How amusing is the comment?""Does the speaker intend to be polite?""Does the speaker intend to be aggressive?"After they finished the task, they were debriefed.

Results
Results for the four rating measures 1   Since we investigated sarcasm interpretation from a cross-cultural perspective, we conducted measurement invariance analysis to assess whether the scenario reading task in the U.K. sample and the Chinese sample measured the same constructs across both cultural groups (e.g., Hu & Bentler, 1999;Van de Schoot et al., 2012).Specifically, we used Multigroup Confirmatory Factor Analysis in R (R core team, 2021; version 4.2.2;lavaan package, Rosseel, 2012; semTool package, Jorgensen et al., 2022) to test measurement invariance for rating measures of sarcasm, amusement, aggression, and politeness.Generally, models of the rating measures had acceptable fit indices, with comparative fit index (CFI) and Tucker-Lewis index (TLI) close to .95,root mean square error of approximation (RMSEA) < .06.All the measurement invariance analysis scripts, data, and results are available at https://osf.io/43yun/.
The main analyses of the data from the four rating scales were conducted using linear mixed models in R (R Core Team, 2021; version 4.0.3;lme4 package, version 1.2.1335).We entered cultural group, comment type, and status (with interactions) as fixed-effects factors into the model.The random-effect structure included intercepts and slopes for the fixed effects of within-subject factors (i.e., comment type and status) across participants and scenarios (Barr et al., 2013;Zhu & Filik, 2023).All the fixedeffects factors (i.e., comment type, cultural group, and status) were coded using sum contrasts.To establish the structure, we started with the maximal model; that is, cultural group × comment type × status + (1 + comment type + status + comment type: status | subject) + (1 + comment type + status + comment type: status | scenario).If the model did not converge, we removed components following the data analysis in Zhu and Filik (2023).Descriptive statistics for the four rating scales are presented in Tables 3 and 4.
The main effects of the fixed factors were obtained through an analysis of variance in the mixedeffects model (Type III with Satterthwaite approximation for degrees of freedom).P values were obtained using the package LmerTest (Kuznetsova et al., 2017) in R. The values of fixed-effects parameters are presented in Table 5.
From Table 5, we can see three-way interaction effects among comment type, status, and cultural group in ratings of sarcasm, aggression, and politeness.There were two-way interaction effects between comment type and cultural group and between social status and cultural group in ratings of amusement.We examined the effects of each rating measure through breaking down the interactions using simple contrasts (package emmeans in R, Lenth, 2022) with an "mvt" adjustment method (multivariate t distribution) for the p value.The degrees of freedom in the emmeans tests were calculated using the Satterthwaite approximation.

Results for the sarcasm rating measure
There was a three-way interaction between comment type, social status, and cultural group in ratings of sarcasm, F (2, 18311) = 8.67, p < .001.We plotted the interaction to see how factors influenced each other (Figure 1).From the plot and the multiple comparisons (Table 6), we can see that there was a main effect of comment type in ratings of sarcasm; that is, sarcastic comments were rated as being more sarcastic than literal comments across social statuses in both the U.K. and China samples.However, the comment type effect was stronger in U.K. participants than in Chinese participants.That is, the rating difference between sarcastic and literal comments in the U.K. sample was bigger than in the Chinese sample.Chinese participants rated literal comments as being more sarcastic than U. K. participants (Table 6).The multiple comparisons also showed a social status effect in literal comments in the Chinese sample only (Table 6); that is, in China, literal criticism made by speakers who had a lower social status was considered as being more sarcastic than that made by speakers who had a higher or equal status compared with recipients.

Results for the aggression rating measure
There was a three-way interaction between comment type, social status, and cultural group in ratings of aggression (Figure 2), F (2, 18303) = 9.31, p < .001.The multiple comparisons showed that the interaction effect between comment type and social status varied across cultural groups.In the Chinese sample, the effect of social status depended on comment type.In the literal condition, there were significant differences between the higher and the equal condition, between the higher and lower condition, and between the equal and lower condition.However, in the sarcastic condition, there were only significant differences between the higher and the equal condition, and between the higher and lower condition (Table 7).That is, when speakers had higher social status, their intent when using both literal and sarcastic criticism was considered as being less aggressive than the other two conditions.When speakers had lower social status, only literal criticism was considered as being more aggressive than when speakers had equal status with the recipients.There were comment type effects across all social status conditions in the Chinese sample; that is, sarcastic comments were rated as being more aggressive than literal criticism, but this effect was smaller when speakers had lower social status (Figure 2).In contrast, in the U.K. sample, there was only a marginally significant effect in the sarcastic condition between the higher and equal conditions (Table 7).That is, in the United Kingdom, the intention behind using sarcasm was considered to be marginally more aggressive when speakers had higher social status than when speakers had equal status with recipients.There was a comment type effect across all social status conditions in the U.K. sample; that is, sarcastic comments were rated as being less aggressive than literal criticism, but this effect was smaller when speakers had higher social status (Figure 2).

Results for the amusement rating measure
There were two-way interaction effects between comment type and cultural group (Figure 3), F (1, 399) = 68.24,p < .001,and between social status and cultural group (Figure 4), F (2, 18317) = 7.91, p < .001, in ratings of amusement.The multiple comparisons showed a significant effect of comment

DISCOURSE PROCESSES
type in ratings of amusement across the United Kingdom and China (Table 8).That is, sarcastic comments were rated as being more amusing than literal language in both the U.K. and Chinese samples.However, this difference was smaller for Chinese participants (Figure 3).As for the cultural difference in ratings of amusement, U.K. participants considered sarcastic criticism to be more amusing than Chinese participants (Table 8).In relation to the interaction between social status and cultural group, there was only a significant effect in the U.K. sample between the higher and equal or lower conditions (Table 8).That is, in the United Kingdom, criticism (literal and sarcastic) made by speakers who had higher social status was considered as being less amusing than when speakers had lower or equal social status compared with recipients.

Results for the politeness rating measure
There was a three-way interaction between comment type, social status, and cultural group in ratings of politeness (Figure 5), F (2, 18297) = 4.73, p = .009.The multiple comparisons showed that the interaction between comment type and social status varied across the United Kingdom and China.In the U.K. sample, there was a significant effect of comment type in ratings of politeness across all conditions of social status.That is, in the United Kingdom, speakers' intent when using sarcastic comments was considered as being more polite than when using literal criticism.However, in the Chinese sample, the speaker intent behind using sarcasm was considered as being significantly more polite than literal criticism only when speakers had lower social status, and when speakers had higher and equal social status  compared with recipients, the speaker intent was considered as being marginally more polite than literal language (Table 9).In relation to social status effects, there were significant effects in the Chinese sample only.That is, Chinese participants rated speakers' intent when using either literal or sarcastic criticism as being more polite when speakers had higher social status than when speakers had lower or equal social status compared with recipients (Table 9).As for the interaction between cultural group and social status in ratings of politeness, there was only a significant difference in the literal condition when speakers had higher social status (Table 9).That is, when speakers had higher social status, Chinese participants considered the intention behind making a literal criticism as being more polite than U.K. participants.

Discussion
This study investigated how social status influenced sarcasm interpretation by manipulating two types of comment and three levels of relative social status and examining the responses of participants in both the United Kingdom and China.The results showed that sarcasm interpretation varied across cultures.While sarcastic comments were considered as being more sarcastic and amusing than literal criticism in both the U.K. and China samples, for the perception of speaker intent, U.K. participants considered sarcastic comments as being more polite and less aggressive than literal criticism, whereas Chinese participants perceived sarcastic comments as being more aggressive than literal criticism.The results also showed effects of social status in criticism interpretation and that the effects depended on culture.That is, both literal and sarcastic criticism were viewed as less amusing when speakers had higher social status than when speakers had lower or equal status compared with recipients in the U. K. sample but not in the Chinese sample.Sarcastic comments were perceived as being more aggressive by U.K. participants when speakers had higher social status than when speakers had equal status with recipients.In contrast, both literal and sarcastic comments were viewed as less aggressive and more polite by Chinese participants when speakers had higher social status than when speakers had lower or equal status compared with recipients.The three-way interactions also showed that in China, when speakers had lower social status, literal criticism was considered as being more sarcastic, aggressive, and less polite than when speakers had higher status or equal status compared with recipients.

Cultural differences in sarcasm interpretation
Our findings that sarcasm interpretation varied across cultures support and extend the findings of Zhu and Filik (2023).Specifically, similar to Zhu and Filik's findings, in the current study, U.K. participants perceived sarcastic comments as being more amusing and less aggressive than literal criticism, whereas Chinese participants perceived sarcastic comments as being more amusing but also more aggressive than literal criticism.Of relevance here is previous research on cultural differences in humor (see Jiang et al., 2019 for a review), which showed that being humorous or funny is not always viewed positively in China, as people in Eastern cultures perceived humor as being more negative than people from Western cultures and showed less tendency to use humor to resolve difficulty.Thus, in China, some people may use sarcasm to be funny; however, the perceiver may nevertheless interpret it as an aggressive behavior.This could potentially explain the apparent discrepancy between Blasko et al.'s (2021) findings that Chinese participants reported using sarcasm to be funny and findings from both this study and Zhu and Filik (2023) that Chinese participants rated sarcasm as more aggressive (yet also more amusing).

Cultural differences in the role of social status in sarcasm interpretation
Whereas Zhu and Filik (2023) investigated individual differences in sarcasm interpretation and how these may be moderated by cultural background, in the current study the focus was instead on effects of social status in sarcasm interpretation and how these may differ across cultures.The results showed social status effects in sarcasm interpretation and the effects varied across cultures.In the U.K. sample, sarcastic comments made by speakers who had higher social status were rated as being less amusing and the speaker intent was viewed as marginally more aggressive compared with the other two conditions.This is contrary to Gucman's (2016) findings that when speakers had higher status, sarcastic comments were perceived as being less critical and offensive.However, our findings that Chinese participants considered the speaker intent of sarcastic comments as being less aggressive and more polite when speakers had higher status to some extent are consistent with Gucman's (2016) results.It might be that in the United Kingdom, which belongs to an individualist culture with a lower power distance index (the extent to which people from a society could accept the unequal distribution of power, Hofstede, 2001), sarcastic comments made by speakers of higher social status were rated as being less amusing and more aggressive than the other conditions.This finding in the United Kingdom is in line with the argument suggested by the Benign Violation Theory (e.g., Kant & Norman, 2019;McGraw & Warren, 2010) that humorous comments uttered by people who have higher social status are more likely to be considered as inappropriate and less funny than humorous utterances by people with lower social status.In contrast, in China, which is a country from a collectivist culture with higher power distance index, people who have higher social status might be more likely to be considered as the authority and lower social status people are less inclined to challenge people who have higher social status (Bochner & Hesketh, 1994).Thus, it may be more tolerable for higher social status people to criticize those with lower status in either a literal or sarcastic way.Even the intention behind both literal and sarcastic criticism by speakers who had higher social status was perceived as being more polite than the other two conditions in the China sample.It may not be surprising that the results of the China sample were in line with Gucman's (2016) study, which was conducted in Poland, as Poland is a country with an individualist culture but has a higher power distance index (https://www.hofstede-insights.com/country/poland/).That is, Poland is a relatively hierarchical society that has common features with China.Therefore, the differences in the social status effects between the United Kingdom and Poland and the similarities in effects between Poland and China provided further evidence that the role of social status in sarcasm interpretation was influenced by culture.
The results also showed that in the Chinese sample, when the speaker had a lower social status, the speaker intent behind literal criticism was perceived as being more aggressive and less polite than when the speaker had equal status with the recipient.However, there was no such significant effect in the sarcastic condition.It seems that, when lower social status people are criticizing sarcastically, their negative intentions (i.e., to be more aggressive and less polite) were viewed as being to some extent reduced, compared with when speakers had higher or equal status with the recipients.Although the speaker intent of sarcastic comments was rated as being more aggressive than literal criticism in the China sample, sarcasm to some extent saved the face of speakers who had lower social status.The reason might be that in China when lower social status people criticize sarcastically, the sarcastic comments might be perceived as the superficially positive meaning and their intention might be misinterpreted in some cases (Anolli et al., 2001).

Implications
The finding of cultural and social status effects in sarcasm interpretation has important implications for theoretical accounts.First, our results suggest that sarcasm interpretation varied across the United Kingdom and China.U.K. participants rated sarcastic comments as being more sarcastic and amusing and the speaker intent as being less aggressive and more polite than literal criticism, which supports the Tinge Hypothesis (Dews et al., 1995).However, in the Chinese sample, sarcastic comments were considered as being more sarcastic and amusing but the speaker intent as being more aggressive than literal language, which is contrary to the predictions of the Tinge Hypothesis.As noted above, these results both support and extend the findings from previous research (Zhu & Filik, 2023).In the current study, questions more directly probed speaker intent (i.e., "Does the speaker intend to be polite/ aggressive?"), whereas Zhu and Filik (2023) asked more general questions about comment perception ("How polite/aggressive is the comment?"), but a similar pattern of results emerged.Given that the cultural differences regarding perceived levels of aggression are relatively consistent across studies, we suggest this constitutes strong evidence that whether or not sarcasm mutes the negativity of criticism depends on the perceiver's cultural background.That is, when a perceiver is from a culture that views sarcasm as being amusing and people tend to use sarcasm more often in daily life (i.e., the United Kingdom; Zhu & Filik, 2023), sarcasm serves the function of diluting the negativity of criticism.In contrast, when a perceiver is from a culture that considers sarcasm as a more eccentric expression in a conversation (i.e., China), sarcasm enhances the perceived aggression of criticism.However, this conclusion would benefit from further research from a larger number of different cultures (not limited to the United Kingdom and China).
Second, finding social status effects enriches our knowledge of the contextual factors that may influence sarcasm interpretation.Interestingly, different patterns of effects were found depending on whether the participants were from the United Kingdom or from China.When speakers had higher social status than recipients, U.K. participants considered sarcasm as being less amusing and marginally more aggressive.In contrast, Chinese participants rated sarcasm as being less aggressive when the speaker had higher social status.Thus, it is clear that there is a complex interplay between relative social status and the perception of sarcastic comments and that the socioemotional impression varies across cultures with different perceptions of social hierarchy.
This study contributes to new approaches to investigating figurative language comprehension by examining issues from a cross-cultural perspective and taking the interaction between relevant factors into consideration.We suggest that there might be a dynamic framework in which both discourse factors and cultural factors may work together in influencing sarcasm interpretation.For example, the United Kingdom and Poland are countries that are considered to be from an individualist culture and China is a country from a collectivist culture.While the social functions of sarcasm were perceived as being more amusing and less critical than literal criticism in participants from the United Kingdom (Zhu & Filik, 2023) and Poland (Gucman, 2016), participants in China considered sarcasm as being more aggressive than literal criticism (see also Zhu & Filik, 2023).Moreover, the United Kingdom has a lower power distance index, whereas China and Poland have a higher power distance index, which indicates that China and Poland are relatively hierarchical societies (Hofstede, 2001).Participants from China and Poland perceived sarcasm as being less critical when speakers had higher social status compared with when speakers had lower or equal status with recipients.However, the situation was somewhat reversed in U.K. participants.The social status effects in the U.K. sample were not as great as the effects in the China sample, as there were only significant effects in dimensions of amusement and marginally significant effects in aggression in U.K. participants.
Sarcasm is a type of figurative language commonly used in daily life by people from different cultures (Blasko et al., 2021).Therefore, the results of the current study also have practical implications.With the prevalence of online communication, such as e-mail and social media, there are more opportunities for people from different cultures to communicate.However, we may not be sure whether the sarcastic utterances we make in certain contexts could be interpreted as intended by people who have a different social status to us and who come from another culture.The findings of the current study suggest that sarcastic utterances could mute the negativity of criticism only in some cultures (e.g., the United Kingdom).In other cultures (e.g., China), sarcastic utterances may instead enhance the negativity of criticism.Moreover, in societies of higher power distance index (e.g., Poland and China), sarcasm may be viewed as more acceptable when uttered by people who have higher social status even though the sarcastic comments are considered as more negative than straightforward criticism.Therefore, this study may provide some guidelines for cross-cultural communication.That is, it may be unwise to use sarcasm with people from a culture you are not familiar with, especially when you are not sure of your relative social status compared to your conversational partner.

Limitations
This study has a number of potential limitations.For example, we investigated the role of social status in sarcasm interpretation by manipulating the relative social status of characters in a scenario, instead of recruiting participants who have higher or lower social status.Thus, social status, as a social factor, was included in this study as a contextual factor and the socioemotional impact was investigated from an observer perspective.However, in communication, sarcastic conversation happens between people who may have different social status; thus, it might be important to examine the perception of sarcasm in participants who have hierarchical status as well as examining their own personal tendency to use sarcasm.In addition, it should be acknowledged that the participant pool was predominantly young adults from a university population, who were mostly women.There may be age and gender differences in sarcasm interpretation across the different cultures that were not possible to detect with the current sample.Additional analyses suggested that there were no gender differences, but the number of participants was not balanced across genders.The narrow age range of the participant pool precluded any analysis of age differences.

Conclusion
The results of the current study have important theoretical and practical implications.Specifically, findings suggest that the socioemotional perception of sarcasm may depend on culture; that is, sarcasm may mute the negativity of criticism in some cultures (e.g., the United Kingdom) but enhance negativity in others (e.g., China).In addition, patterns of social status effects varied across cultures.In the Chinese sample, when the speaker had higher social status, their sarcastic comments were perceived as more positive than when speakers had lower or equal status compared with recipients.In contrast, the situation was the opposite in the U.K. sample.Thus, theoretical accounts such as the Tinge Hypothesis (Dews & Winner, 1995) need to be reconsidered in light of the finding that cultural and discourse factors interact with each other in the perception of the emotional impact of sarcasm.In terms of practical implications, this study provides evidence to inform successful intercultural communication, which is particularly important given the ongoing increase in intercultural communication via text message, e-mail, and social media.

Figure 1 .
Figure 1.Three-way interaction plot of ratings of sarcasm.Note: Error bars represent ±1 SE.

Figure 2 .
Figure 2. Three-way interaction plot of ratings of aggression.Note: Error bars represent ±1 SE.

Figure 3 .
Figure 3. Interaction plot between comment type and cultural group in ratings of amusement.Note: Error bars represent ±1 SE.

Figure 4 .
Figure 4. Interaction plot between social status and cultural group in ratings of amusement.Note: Error bars represent ±1 SE.

Figure 5 .
Figure 5. Three-way interaction plot of ratings of politeness.Note: Error bars represent ±1 SE.

Table 4 .
Descriptive statistics of rating measures in the China sample (N = 200).

Table 3 .
Descriptive statistics of rating measures in the U.K. Sample (N = 200).

Table 5 .
Results of the linear mixed models and the fixed-effects parameters.
DISCOURSE PROCESSES

Table 6 .
Simple contrasts of three-way interaction effects for rating measures of sarcasm.

Table 7 .
Simple contrasts of three-way interaction effects for rating measures of aggression.

Table 8 .
Simple contrasts of two-way interaction effects for rating measures of amusement.

Table 9 .
Simple contrasts of three-way interaction effects for rating measures of politeness.