Culture and Awe: Understanding Awe as a Mixed Emotion

Abstract

Recent work is establishing awe as an important positive emotion that offers physical and psychological benefits. However, early theorizing suggests that awe’s experience is often tinged with fear. How then, do we reconcile emergent positive conceptualizations of awe with its more fearful elements? We suggest that positive conceptualizations of awe may partially reflect modern Western experiences of this emotion, which make up the majority of participant samples when studying awe. To test whether awe contains more fearful qualities outside of Western cultures, we compared participants’ experiences of this emotion in China to those in the United States. In a two-week daily diary study (Study 1), Chinese participants reported greater fear than American participants during experiences of awe, but not a comparison positive emotion. In response to a standardized awe induction (Study 2), Chinese participants reported more fear, whereas American participants reported more positive emotions. Physiological changes in autonomic activity differed by culture only for heart rate, but not skin conductance or respiratory sinus arrhythmia. These findings reveal that awe may be experienced as a more fearful, mixed emotion in China than in the United States and suggest that current positive conceptualizations of awe may reflect a disproportionate reliance on modern Western samples.

Supplementary Information

The online version contains supplementary material available at 10.1007/s42761-024-00243-3.

Awe is an emotion that arises when individuals encounter something vast that challenges their current knowledge structures (Keltner & Haidt, 2003). Awe is quickly emerging as an important positive emotion that offers numerous benefits for one’s well-being (Monroy & Keltner, 2023), physical health (e.g., Stellar et al., 2015), and social interactions (e.g., Piff et al., 2015). However, this positive conceptualization of awe stands in contrast to early definitions that describe it as an emotion tinged with fear (McDougall, 1908), a refined form of fear (Hall, 1897), or a blend of fright and amazement (Lazarus, 1991). Awe’s etymology further highlights its origins in fear. Awe comes from the words aw or ahe, meaning terror, dread, and extreme reverence. Awe also has a long tradition in philosophical, aesthetic, and religious discussions of the sublime and numinous, which evoke a feeling akin to awe that is described as a sense of delightful horror that is both terrifying and fascinating (Burke, 1757/1990; Otto, 1970/1923). How then, do we reconcile emergent positive conceptualizations of awe with these more fearful elements?

We suggest that current positive conceptualizations of awe by researchers may reflect modern Western experiences of this emotion, which are quite pleasant, not particularly fearful, and often accompanied by other positive emotions like compassion, gratitude, and love (Gordon et al., 2017; Nelson-Coffey et al., 2019; Tong, 2015). The majority of research on awe’s experience relies on recently collected Western samples. Only a handful of studies examine awe in other cultures and none of them focuses on the extent to which awe is accompanied by feelings of fear (e.g., Bai et al., 2017; Dong & Ni, 2019; Guan, Chen et al., 2019; Guan, Xiang et al., 2019; Razavi et al., 2016). One study suggests that experiences of awe may be more mixed in non-Western cultures, finding that Japanese individuals are more dispositionally prone to feel the threatening aspects of awe (Nakayama et al., 2020). However, this study did not examine actual experiences of awe in the moment. Therefore, we aimed to identify whether awe may be experienced as more of a mixed emotion, accompanied by greater fear, in an Eastern culture like China compared to a Western culture like the United States.

Inherent to our claims is the notion that awe experiences have become less fearful over time in Western cultures. Some scholars suggest that the fearful component of awe has been deemphasized in modern, and we would argue more specifically modern Western, cultures (Bonner & Friedman, 2011). This claim is supported by a recent linguistic analysis of natural language use among historical corpora of predominantly Western texts in English and French from the late 1890s to the 1990s (Xu et al., 2021). This analysis found that over the past 100 years, the meaning of the word awe in these languages has transformed from more negative to positive in valence.

Different cultural norms about emotions, societal values, and historical movements may have encouraged awe experiences to diverge over time, reducing awe’s fearful components in Western cultures. For instance, the idealization of positive affect and discomfort with feeling contradictory emotions at the same time may have diminished fear during awe experiences in the United States. Positive affect is strongly valued and negative affect is strongly devalued in the United States, whereas these norms are comparatively weaker in countries like China (Eid & Diener, 2001; Koopmann-Holm et al., 2014; Uchida et al., 2004). Critically, cultural differences in what emotions are idealized translate into differences in emotion experience (Tsai, 2007). For example, one study found that American participants more frequently experienced positive than negative emotions, whereas Japanese participants experienced both with roughly equal frequency (Kitayama et al., 2000).

Relatedly, Western cultures are less inclined to experience mixed emotions (e.g., happiness and sadness) at the same time. While positive and negative emotions are perceived to fall on opposite ends of the same continuum in Western and Individualist cultures, they are perceived to be more orthogonal in Asian and Collectivist cultures (Bagozzi et al., 1999). As a result, studies of the structure of self-reported experience among Americans reveal negative correlations between positive and negative emotions, whereas Asian cultures have weak or sometimes positive correlations between the two (e.g.,Schimmack et al., 2002). Greater comfort with mixed emotions in Eastern cultures also appears to be rooted in dialectical philosophical traditions that originated in China, which embrace contradictory perspectives (Peng & Nisbett, 1999). As a result, individuals from Western cultures are less likely to feel positive and negative emotions simultaneously compared to those from Asian cultures, especially during pleasant experiences (Larsen et al., 2001; Miyamoto et al., 2010; Shiota et al., 2010; Spencer-Rodgers et al., 2010).

In addition, the endorsement of more egalitarian values in the United States may have reduced the fearful elements of awe that arise from submissiveness to another in power. Theoretical accounts consider awe to be a status-based emotion, claiming it emerged as a response to powerful leaders and that it helps individuals subvert their own interests for those of the leader or group (Keltner & Haidt, 2003). Fear during awe may partially stem from social appraisals of reduced rank and submissiveness vis-à-vis another more powerful person or entity (Stellar et al., 2017). People in the United States tend to endorse egalitarianism and nonhierarchical relations, whereas people in China tend to endorse hierarchies in which individuals accept the roles and obligations associated with their place in the social strata (Hofstede, 2016). Different cultural values may shift the kinds of experiences that evoke awe, leading to fewer status-based experiences that evoke fearful submissiveness in the United States compared to China.

Finally, historical movements in Western cultures may have shifted toward experiencing awe more frequently in the context of extraordinary beauty, which lacks fearful qualities. Transcendentalism and Romanticism, which became popular in Western cultures in the early 1800s, focused heavily on reverence for beauty in nature and art. Today, beauty represents one of awe’s most common themes, defining roughly 60% of awe experiences generated by Americans (Yaden et al., 2019). This positive conceptualization of awe may have been further solidified by Maslow’s (1962) introduction of peak experiences, which are akin to awe, and quite positive. Together, these converging lines of reasoning set the stage for the central hypothesis tested in this investigation—awe experiences are accompanied by greater fear in China than in the United States.

Present Research

To test our predictions, we conducted two studies, a two-week daily experience sampling method that maximized ecological validity and a psychophysiological laboratory study that presented participants with a standardized and controlled awe induction. Study 1 allowed us to document everyday experiences of awe while reducing retrospective bias in reporting (e.g., Kahneman et al., 1993). Study 2 ensured that our effects from Study 1 were not the result of different eliciting events or antecedents of awe in each culture.

We aimed to reduce linguistic concerns about whether fear is inherently part of the semantics of awe in Chinese more so than in English by avoiding using the word awe to elicit awe in Study 2. We also gathered measures of autonomic physiology in Study 2, extending our hypothesis to a more implicit realm of measurement and addressing concerns that American participants were experiencing a fearful response, but simply under-reporting it.

Across our two studies, we measured the extent to which awe represented a mixed emotion by focusing on the degree to which fear was reported during the experience of awe. However, we also measured the extent to which other positive emotions accompanied awe’s experience (Study 2) since part of identifying awe as a positive experience has been the cooccurrence of other positive states (Nelson-Coffey et al., 2019). Therefore, we also predicted that during the awe induction American participants would report higher levels of positive affect than Chinese participants. Although these data were collected in 2014 before pre-registration was an established best practice, to aid in transparency and replicability, we have made all our data files and syntax available to other researchers on Open Science Framework (https://osf.io/tvgzc/?view_only=5d308471ddce4b81940d4d26450340fe).

Study 1

To test our hypothesis, we chose an ecologically valid method that allowed participants to select culturally appropriate experiences of awe and joy that were meaningful to them. Participants reported on an experience that made them feel awe or joy every day for two weeks. We chose to ask about joy as a comparison emotion to ensure that differences in fear would be localized to awe and not be present for any positive emotion. We also predicted that joy experiences would be relatively easy for participants to think of, affording us more instances to compare with awe than other emotions like compassion or pride. We hypothesized that Chinese participants would report more fear during experiences of awe than American participants, but that there would be no differences in reported fear during experiences of joy.

Method

Sample

A total of 166 participants were included in the final analyses. We aimed to collect 100 participants in each condition based on similar methods from previously published work (Bai et al., 2017; Gordon et al., 2017). One hundred participants in China also represented the maximum number of participants we could collect given the research participant pool in that location.

In China, 88 undergraduates from an undergraduate class at a major public university in Beijing participated in this study for compensation. Chinese students born in an East Asian country and proficient in Mandarin participated. We excluded one participant who finished less than half of the diary and four who failed to follow directions and filled out the diary incorrectly, leaving a final sample of 83 participants (28 males and 55 females). The average age of the sample was 18.84 years (SD = 0.99).

In the United States, 120 undergraduates from a large West Coast university took part in this study for compensation. American students born in the United States who had not lived in a foreign country for more than 2 years were recruited. We excluded 32 participants who finished less than half of the diary and five who failed to follow directions and filled out the diary incorrectly, leaving a final sample of 83 participants (16 males and 67 females). The average age of the sample was 19.96 years old (SD = 1.56). The ethnicities of this sample were as follows: 3% African-American, 41% Asian-American, 36% European-American, 18% Latin-American, and less than 1% other ethnicities.

We conducted sensitivity analyses using G*Power (Version 3.1; Faul et al., 2007) to determine the smallest effects our study was powered to detect at 80% power with a two-tailed test (alpha = .05). To account for the nonindependence of daily diaries nested within individuals, we calculated the effective sample size using the design-effect equation: N_effective = N/(1 + (n_cluster − 1) × ρ), where N represents the total number of observations, n_cluster represents the cluster size (in our case, the average cluster size was 13.67), and ρ represents the intraclass correlation coefficient (ICC; i.e., the measure of nonindependence between daily diaries within individuals). Using an ICC of .18 (based on a null model in which we did not account for autoregressive correlations), our effective sample size was 691.7. Rounding down to 691, we were powered to detect effects as small as r = .11.

Procedure

We used a two-stage translation process in which all materials from the United States were translated into Mandarin and then back-translated into English by two bilingual research assistants. During an initial orientation session with participants (conducted in class by the authors fluent in Mandarin in China and online in English in the United States), awe (敬畏) was defined as an emotion and by its unique appraisals—a sense of vastness and the perception that one’s current understanding of the world has been challenged (Keltner & Haidt, 2003). In addition, an established facial expression by an Asian-American woman (Shiota et al., 2003) and a vocal burst of awe in which ethnicity and nationality are hidden were also provided to help participants understand this emotion with nonverbal representations of the state (Cordaro et al., 2016).

Each evening for 14 days, participants were sent a link with an online survey. Participants reported on either their awe or joy experiences. Because awe was our focal emotion, and because it is typically experienced with lower frequency, participants were first asked how many times they experienced awe (敬畏) that day ranging from 0 (None) to 5 (More than 5 times that day). If they experienced awe, they were asked to describe their strongest experience in three to five sentences, a technique used by Yaden et al. (2019) when developing a state measure of awe experience. For those who did not report having any experiences of awe that day, they were asked to report how many times they had experienced joy (喜悦). Regardless of their answer to this question, they were asked to describe their most joy-inducing moment that day in three to five sentences. This was done to induce compliance by preventing participants from stating they felt no awe or joy to shorten their time taking the survey. However, if a participant indicated that they had no joy experiences that day, their data was not included in the joy condition, even if they wrote something in the text box as instructed. After reporting on their awe or joy experience, participants also indicated how much awe or joy they felt from 1 (Not at all) to 5 (Extremely). Then, participants reported how scared (害怕的) they felt during the awe or joy experience ranging from 1 (Not at all) to 5 (Extremely). This question was embedded within other measures collected for another project (e.g., worries, control, and interconnection; see supplement for all other measures included in this study and where they are published).

Content Analysis

There were 801 experiences of awe reported by participants. Chinese responses were translated into English using Google Translate. Three English-speaking coders were then trained to identify the elicitor of the awe experience. To identify larger categories of elicitors, the first author coded 50% of responses and noted any categories that appeared for more than three participants. This process resulted in ten categories, which were later condensed into six categories1: (1) Art, music, literature, and film; (2) Nature, monuments, and architecture; (3) Other people; (4) Ideas or human innovation; (5) Oneself (the self); (6) Other (did not clearly fit into other categories). Some participants’ responses included two elicitors (e.g., a character in a book, which would be coded as both art, music, literature, and film as well as other people). In these instances, coders selected the primary elicitor based on which elicitor was the focus of the reported experience (e.g., if the awe was primarily about the character in the book and their qualities, the primary elicitor was other people; if the awe was primarily about the book itself it would be coded as art, music, literature, film). Coders were given the chance to identify a second elicitor in these cases.

For the primary elicitors, coders showed strong reliability with an average measure ICC (two-way random, with absolute agreement) of .78. In creating the final category assignment for the elicitor, when there was disagreement among coders, the elicitor was designated as the one that at least two of the three coders selected. If all three coders had identified unique elicitors, the first author determined the elicitor, this was done for 64 cases (8% of responses). In addition, seven cases could not be identified as having a clear elicitor or were not classified as an awe experience, despite being written about by the participant, and therefore received no code.

Results and Discussion

Descriptive Statistics and Data Analytic Strategy

Chinese participants completed 1,247 diaries (awe experiences: n = 492, joy experiences: n = 502, neither: n = 253). American participants completed 1,271 diaries (awe experiences: n = 312, joy experiences: n = 519, neither: n = 440). We included diaries filled out in the evening, after midnight, and into the next morning. On average, Chinese participants had six awe experiences (SD = 3.91) and six joy experiences (SD = 3.11) across the two weeks. American participants had four awe experiences (SD = 3.16) and six joy experiences (SD = 3.08) on average across the two weeks. Mean levels of reported intensity of awe during awe experiences were slightly lower for Chinese (M = 3.07, SE = 0.04) than for American participants (M = 3.23, I = 0.05; F(1, 304.38) = 5.45, p = .02, 95% CI_difference: 0.03, 0.29, r = .13). On the other hand, mean levels of reported intensity of joy during joy experiences were greater for Chinese (M = 3.25, SE = 0.06) than for American participants (M = 2.96, SE = 0.05; F(1, 440.22) = 15.28, p < .001, 95% CI_difference: 0.15, 0.44; r = .18). It is unclear why Chinese participants experienced awe more frequently than American participants. One potential reason may be that Chinese participants reported more interpersonal forms of awe (see the “Exploratory Analyses” section), which may allow for more frequent opportunities to experience this emotion.

We used multilevel modelling (Mixed Models, SPSS v22), nesting days within individuals. To account for correlations between days, we used an autoregressive correlation matrix with heterogeneous variances for the residuals. Intercepts and slopes were allowed to vary and we used an unstructured covariance matrix for the random effects, modelling all random variances and covariances. We entered culture (coded as − 1 = China and 1 = USA) as a between-subjects variable, emotion experience (coded as − 1 = joy, 1 = awe) as a within-subjects variable, and the cross-level interaction between the two.

Main Analyses

We measure the mixed nature of awe by examining participants’ reports of fear, which is equivalent to the MIN score statistical approach that is traditionally used to assess mixed emotions.2 Our hypothesis was that participants in China would report more fear than Americans during awe experiences, but not during joy experiences. Analyses of the main effects revealed that reports of fear did not differ across cultures (F(1, 147.51) = 0.31, p = .58, 95% CI_difference: − 0.15, 0.09; r = .05) and that fear was greater during awe experiences than joy experiences (F(1, 142.20) = 59.08, p < .001, 95% CI_difference: 0.33, 0.55; r = .54). Importantly, there was an interaction between culture and emotion experience (see Fig. 1; F(1, 141.63) = 4.79, p = .03, 95% CI_interaction: 0.02, 0.47; r = .18), such that Chinese participants reported more fear during awe experiences (M = 1.88, SD = 0.08) than did American participants (M = 1.63, SD = 0.08; F(1, 138.4) = 4.61, p = .03, 95% CI_difference: 0.02, 0.47; r = .18), but equivalent levels of fear during joy experiences (Chinese: M = 1.32, SD = 0.05; Americans: M = 1.33, SD = 0.04; F(1, 137.9) = 0.002, p = .97, 95% CI_difference: − 0.13, 0.12; r = .004). Further, for both Chinese (F(1, 124.84) = 52.41, p < .001, 95% CI_difference: − 0.71, − 0.40; r = .54) and American participants (F(1, 158.13) = 13.50, p < .001, 95% CI_difference: − 0.47, − 0.14; r = .28), awe contained more fear than joy.

Fig. 1.

Interaction between culture and type of emotion experience in predicting self-reported fear in Study 1 (measured from 1 to 5)

Exploratory Analyses

We compared the frequency of awe elicitors across cultures to ascertain whether greater fear during awe experiences among Chinese participants simply reflected a greater proportion of fear-inducing awe elicitors among Chinese participants. We conducted a binomial logistic mixed model with culture as a predictor and the presence or absence of each elicitor as an outcome. We found that Chinese participants were more likely to state that other people elicited awe, whereas American participants were more likely to state that nature, monuments, and architecture and themselves elicited awe (see Table 1 for frequencies and significance tests). Art, music, literature, and film; ideas and human innovation; and the other categories were equally common elicitors in both cultures.

Table 1.

Fear rating, percent (frequency), and test for difference in frequency of awe elicitors by culture in Study 1

Awe elicitors	Fear rating	Percent (frequency)		Significance test for elicitor frequency predicted by culture
		China	US
Nature, monuments, architecture	1.51a	10% (49)	18.3% (57)	F(1, 131) = 8.55, p = .004, 95% CIdiff: .03, .15, r = .25
Art, music, literature, film	1.72a	10.2% (50)	7.1% (22)	F(1, 144) = 1.18, p = .28, 95% CIdiff: − .06, .02, r = .09
Other people	1.69a	59.3% (291)	49.8% (155)	F(1, 101) = 4.67, p = .03, 95% CIdiff: − .20, − .01, r = .21
Ideas, human innovations	1.77ac	7.1% (35)	5.8% (18)	F(1, 171) = 0.05 p = .82, 95% CIdiff: − .03, .02, r = .02
Oneself	2.11bc	3.1% (15)	8.7% (27)	F(1, 174) = 7.99, p = .005, 95% CIdiff: .01, .06, r = .21
Other	2.16b	9.2% (45)	9% (28)	F(1, 126) = 0.04, p = .84, 95% CIdiff: − .04, .05 r = .02

Elicitors with fear ratings that differ significantly by culture are denoted with different superscripts

Next, we examined how much fear each elicitor predicted. We ran a mixed model predicting fear from elicitor type, entered as a within-subjects fixed effect. Type of elicitor predicted fear ratings (F(5, 646.89) = 6.43, p < .001, r = .10). Only oneself and the other categories elicited significantly greater fear across all participants (see Table 1 for pairwise comparisons and mean fear ratings for each elicitor). All other categories elicited equivalent and lower levels of fear. In addition, oneself was the only elicitor that showed differences in frequency by culture, and interestingly, it was more common in the American sample. These results suggest it is unlikely that greater fear during awe among Chinese participants simply reflected a greater proportion of elicitors that induced fear among participants from this culture. Rather these findings suggest that awe is experienced as a more mixed state for Chinese compared to American participants.

After conducting this study, we noticed that a larger portion than we had anticipated of our American sample was Asian-American (40.96%). Therefore, we conducted exploratory analyses in which we divided our sample into three groups: Asian-Americans, non-Asian-Americans, and Chinese (see supplement for analyses). These analyses offered an opportunity to examine cultural differences in experiences of awe that included a group of individuals who are influenced by both Asian and American cultures. In sum, we found that Chinese participants reported significantly more fear than non-Asian-American participants and Asian-American participants fell between these two groups and were not significantly different from either group.

Overall, these findings demonstrate that daily awe experiences are accompanied by greater fear for Chinese than for American participants. This was not the case for joy. Though importantly, fear was still greater during experiences of awe than joy, even in American samples, suggesting fear is not entirely absent during awe even in a Western culture. In addition, these effects were not likely explained by a greater proportion of fearful elicitors in China versus the United States. This finding provides initial evidence that experiences of awe may be more mixed outside of Western cultures.

Study 2

Building upon Study 1’s initial evidence for greater fear during daily experiences of awe in China than in the United States, we brought participants into the laboratory and presented them with an awe-inducing film clip. This method addressed some limitations from Study 1. First, we used a standardized induction in which participants saw the same awe-inducing film clip, which allowed us to control for the nature of the elicitor that gave rise to the experience of awe. We chose a nature film clip since nature is a common elicitor of awe and it is less subject to cultural variation in meaning than other elicitors of awe, such as admired individuals, collective behavior, cultural artifacts, music, or art.

Second, we measured a variety of negative and positive emotions to examine how much they were elicited during experiences of awe. In keeping with Study 1, we predicted that Chinese participants would report greater fear during the awe induction than Americans, but also that Americans would report greater positive emotions than Chinese participants. We expected that greater negative emotion would be localized to fear specifically, but to confirm this prediction we also collected other negative emotions (e.g., anger, jealousy). Although our main focus was fear, demonstrating that American participants experienced more positive emotions (e.g., appreciation) during awe experiences compared to Chinese participants would further lend support to our broader claim that awe is experienced more positively within a Western culture. It would also be in keeping with previous claims that awe is often accompanied by other positive emotions (Nelson-Coffey et al., 2019).

Third, we measured peripheral physiological responses. Gordon et al. (2017) found that heart rate (HR), skin conductance (SCL), and respiratory sinus arrhythmia (RSA) differentiated more fearful forms of awe. They found that during experiences of awe reports of fear were associated with increases in HR and SCL, whereas reports of positive emotions were associated with increases in RSA. Greater RSA has also been documented by Chirico et al. (2017) during immersive, presumably positive, experiences of awe. These findings are in keeping with past work that documented that SCL (a marker of sympathetic activation) and HR are associated with greater fear, anxiety, and stress (Ax, 1953; Dawson et al., 2000; Levenson, 2003; Stemmler et al., 2001), and RSA (a marker of parasympathetic activity; Mendes, 2016) is associated with greater positive emotionality (e.g., Kok et al., 2013; Oveis et al., 2009). In light of these literatures on peripheral physiological response and emotion, we employed these measures, predicting that Chinese participants would exhibit greater increases in HR and SCL than American participants, converging with the expectation that they would experience greater fear. On the other hand, consistent with more positive experiences of awe, we predicted that American participants would exhibit greater increases in RSA than Chinese participants. These findings represent an important extension beyond self-report since peripheral physiological measures are less controllable and therefore considered to be more genuine indicators of a person’s internal state.

Method

Sample

A total of 137 participants are included in the final analyses. In China, 64 undergraduates (31 males, 33 females) from a major public university in Beijing participated in this study for compensation. In the United States, 73 undergraduates (19 males, 54 females) from a large West Coast university took part in this study for credit in a psychology course. The ethnicities that made up the American sample were as follows: 6.8% African-American, 39.7% Asian-American, 28.8% European-American, 15.1% Latin-American, and 9.6% other ethnicities. The samples were recruited using the same criteria from Study 1.

Power calculations for an independent samples t-test of fear, our primary variable of interest, suggested a sample size of 200 would allow us to detect a medium to large effect size (d = 0.4) at 80% power. However, we fell short of this sample size given that we only had access to 100 participants in the research pool in China. A power sensitivity analysis suggests that with 80% power we could detect a medium effect size (d = 0.48) for our primary variable of interest, fear, using an independent samples t-test given our sample size.

Procedure

We used the same two-stage translation process from Study 1 in which all study materials were translated into Mandarin and then back-translated into English by two bilingual research assistants. Chinese research assistants ran the lab sessions. In China, participants arrived at the lab in groups of one to four and were brought into a large testing room with individual cubicles and seated in front of a computer. We connected participants to the MP 150 data acquisition and analysis systems (Biopac Systems, Inc.) to collect physiological data. Sensors were placed on participants’ skin in a Lead II configuration to gather Electrocardiogram (ECG) signals, a belt was placed on the torso to assess respiration frequency, and two sensors were placed on the palmar surface of the left hand to measure skin conductance.

Participants first reported on their demographic information and completed some personality surveys. This period was also intended to allow participants to habituate to wearing the physiological sensors. With headphones on, participants then watched two videos that were each approximately two minutes in duration. The first was a neutral video so that baseline physiology could be assessed. This video depicted people building a fence and has been used in past experiments to induce more neutral states (e.g., Stellar et al., 2012). We overlaid the speaking portions of this video with music so that no language was presented to participants. The second was an awe-inducing video, depicting images of Earth taken from the film Planet Earth set to music (with no words). The video showed giant waterfalls and areal views over mountain ranges set to powerful orchestral music and has been used before to elicit awe in American samples (Gordon et al., 2017; Stellar et al., 2018).

After watching the awe-inducing video, participants reported on awe (敬畏) and fear (恐惧) as well as a number of other emotions/states: amused, angry, appreciation, embarrassed, compassion/sympathy, content, curiosity, envy, jealousy, guilt, happy, humility, inspiration, love, pride, sad, shame, and surprise using a scale from 1 (Not at all) to 5 (Very Strong). In our American sample, we had different numeric anchors, from 1 (Not at all) 10 (Very Strong), which was an oversight. Therefore, all scores from China were converted to the 10-point anchor scale. To make the minimum and maximum values comparable, we converted the Chinese scale to 1–10 before comparing the means using the following formula: 2.25 * x − 1.25. Positive emotions (compassion, appreciation, inspiration, pride, happy, amusement, love) showed high reliability (α = .80) and were aggregated. Negative emotions that did not include fear (envy, shame, embarrassed, anger, sadness) showed acceptable reliability and were aggregated (α = .67). After, participants watched other movies for an unrelated study (see supplement for other measures in this study). At the end of the session, participants were debriefed and released.

The method for this study was identical in the United States except for the following differences. Participants arrived individually and were seated in their own testing rooms, whereas in China it was a large testing room with separate cubicles. The list of emotions participants filled out was slightly different (included wonder, amazement, anxious, relaxed/comfortable; excluded content); only emotions that were presented in both samples were used in analyses. This study was followed by different variables for another experiment (see supplement).

Physiological Measures

We were unable to record clean signals from three participants (two from China, one from the US); therefore, these individuals are not included in any analyses of our physiological measures. All physiological measures were normally distributed; therefore, we did not log transform them as is sometimes the case with these measures.

Heart Rate (HR)

For each video, Electrocardiogram (ECG) recordings were sampled at a rate of 1 kHz for the entire two minutes of each video. Heart rate (HR) was calculated by turning ECG signals into beats per minute using Biopac’s analysis program. HR was aggregated over the whole two minutes for each video. Artifacts in the signal, which resulted from movement, were corrected manually in less than 5% of all data files. If we needed to impute more than three R-spikes during any given video, those participants’ data files were not included in the analysis; however, this was not the case for any participants in this study. We had two criteria for excluding physiological data from analyses of HR. First, if we were unable to collect a viable signal from participants; this was the case for four participants (two from China, two from the US). Second, if a participant’s HR was identified as an outlier (more than 3 SD above or below the mean HR); this was not the case for any participants. As a result of these exclusion criteria, the final sample was 130.

Skin Conductance (SCL)

SCL was computed by averaging unfiltered signals from the entire two minutes of each video. Ten participants (seven from China, three from the US) did not produce a viable SCL signal and three (from China) were more than three standard deviations from the mean. These participants were excluded from analyses of SCL, leaving a final sample of 124.

Respiration Rate

Respiration signals were collected without any filtering, but a post-collection digital filter was added to smooth the signals (low-pass filter: 1 Hz, high-pass filter: 0.05 Hz). Across each two-minute video induction, this signal was then converted into respiration rate by calculating the breaths per minute using Biopac’s analysis program. Three participants (one from China, two from the US) did not produce a viable respiration rate signal; none was identified as outliers. These participants were excluded from analyses of respiration rate, leaving a final sample of 137.

Respiratory Sinus Arrhythmia (RSA)

To calculate RSA, we used Biopac’s analysis program, which uses the ECG and respiration channels. This system calculates the variability in the time between R-spikes in the QRS complex in the ECG channel as a function of exhalation and inhalation from the respiration channel, resulting in a continuous time-series output for RSA. We then aggregated RSA across the two minutes of each video. Since we were unable to collect a HR signal for four participants (two from China, two from the US), we could not calculate RSA from these participants. In addition, two people (from the US) were more than three standard deviations from the mean. Therefore, they were not included in analyses using RSA, leaving a final sample of 128.

Results and Discussion

Main Analyses

Chinese and American participants reported equivalent levels of awe in response to watching the Planet Earth video (t(132) = 1.41, p = .16, d = 0.24; see Table 2 for means). Awe was the most intensely experienced of all the emotions reported. In keeping with our hypothesis, and conceptually replicating the results from Study 1, in viewing the awe-inspiring video, participants in China reported greater fear than participants in the American sample (t(132) = 2.54, p = .01, d = 0.44). In addition, American participants reported more positive emotions during the awe induction than Chinese participants (t(132) = 2.51, p = .01, d = 0.44).

Table 2.

Means (SD) for emotions in Study 2

	China	US
Awe	7.93 (2.44)	7.37 (2.20)
Fear	3.69 (2.56)	2.64 (2.23)
Negative emotions	1.79 (.95)	1.76 (1.19)
Positive emotions	3.32 (1.49)	4.03 (1.73)

Emotions were on a 1 to 10 scale. Significant differences (p < .05) are bolded

Follow-up analyses revealed these differences in positive emotions during awe were driven primarily by greater amusement and appreciation (using family-wise correction for testing twelve emotions with p < .004 as significance threshold; see supplement for individual emotion means and comparison tests). There were no differences in our aggregate measure of negative emotions besides fear (t(132) = 0.14, p = .89, d = 0.02) across cultures. Examining individual negative emotions revealed no clear patterns. Chinese participants reported more envy, but less embarrassment, with no differences in shame, sadness, or anger with a family-wise correction for multiple tests.

Next, we examined patterns of physiological activity. We conducted separate repeated measures analyses using culture as the between-subjects variable and each physiological measure’s (HR, SCL, and RSA) change from the neutral baseline to the awe video as the within-subjects measures. For RSA, we controlled for changes in respiration rate, as past work has suggested respiration should be included as a covariate to isolate the unique variance of RSA (Grossman & Kollai, 1993).

There was a significant effect of culture on changes in HR from the baseline to the awe video (F(1, 128) = 14.08, p < .001, η² = 0.10). Chinese participant’s HR increased from the neutral to the awe video, but not significantly (F(1, 59) = 2.05, p = .16, η² = 0.03), whereas American participant’s HR decreased significantly (F(1, 69) = 14.63, p ≤ .001, η² = 0.18). There was no effect of culture on SCL from the baseline to the awe video (F(1, 119) = 0.13 p = .72, η² = 0.001), or on RSA (F(1, 124) = 2.15, p = .15, η² = 0.02).

Exploratory Analyses

Again, our sample from the United States was comprised of a large portion of Asian-American participants (39.72%). Therefore, we conducted exploratory analyses comparing Asian-Americans, non-Asian-Americans, and Chinese participants (see full analyses in the supplement). To summarize, we found that Chinese participants reported significantly more fear than non-Asian-Americans, and Asian-Americans fell between these two groups and were not significantly different from either (see supplement for physiology findings). Although caution should be taken in interpreting these effects given the small group sizes, they are in the predicted pattern.

In sum, the results from Study 2 replicated and extended in important ways those from Study 1, showing that awe was accompanied by greater fear in China than in the US. It is important to note that we did not measure fear during the neutral video since it was intended as a non-emotional control. Thus, it is possible that levels of fear may have been elevated for that video suggesting a heightened level of fear at baseline. Physiological results were more mixed. Culture did affect changes in HR, but not SCL or RSA. The effect on HR was driven by notable deceleration among American participants. However, our failure to document differences in patterns of physiological activation between Chinese and American participants for respiratory sinus arrhythmia (RSA) and skin conductance (SCL) may reflect that we were underpowered to detect smaller effect sizes. Importantly, reports of awe did not differ by culture, suggesting that these results could not be attributed to cultural differences in the intensity of awe.

General Discussion

Our studies suggest that awe’s current conceptualization as a positive emotion may more heavily reflect its experience in modern Western cultures. Outside of Western cultures awe experiences may be more fearful. Across two studies, participants from China reported greater fear during daily experiences (Study 1) and a standardized laboratory induction (Study 2) of awe compared to American participants. These cultural differences in subjective experience were accompanied by diverging patterns of heart rate activation, but not SCL or RSA. Our results from Study 1 suggest that the cultural differences we documented were not likely due to different kinds of awe experiences, a finding that Study 2 confirmed using a standardized awe induction. Interestingly, for both studies, Chinese participants reported more fear than American participants who were not Asian-American, and Asian-American participants fell between these two groups. While these findings fall in line with what would be predicted based on exposure to Asian culture, caution should be taken as they are post hoc exploratory tests and the sample sizes for each group are small.

This work has important implications for research on awe. Experiences of awe enhance well-being (e.g., Rudd et al., 2012), unless the awe experience is more fearful (Gordon et al., 2017). This finding raises the question, would these established well-being effects of awe be as robust in other cultures like China? In addition, recent work has identified that Chinese participants are less prosocial when experiencing negative compared to positive awe (Guan, Chen et al., 2019), though interestingly this is not the case for Americans (Piff et al., 2015). These results suggest that awe experiences in China compared to the United States may generate different effects on a variety of outcomes impacted by fear such as curiosity, creativity, exploration, or stress.

In addition, a growing number of studies suggest that differences between more and less fearful forms of awe may be deeply embedded in biological systems. For example, gray matter volume (Guan, Chen et al., 2019; Guan, Xiang et al., 2019) and neural activation (Takano & Nomura, 2022) distinguish between positive and negative forms of awe. Therefore, awe experiences may have different biological underpinnings in the United States and China that reflect differences in the role of fear during this emotion.

There are important limitations to our studies. First, we collected data from only two cultures. We chose the United States since the majority of research has been conducted in this culture, and China due to its diverging cultural norms about emotions, values, and social movements. However, future work should extend the research on awe beyond China and the United States to make broader claims or identify additional cultural dimensions that may explain differences in experiences of awe. Second, we compared awe to a neutral state (Study 2) and joy (Study 1); however, future work should consider comparing awe to low arousal positive states like contentment that are valued more highly in China (Tsai, 2007) or gratitude, which can represent more of a mixed emotional experience (Oh & Tong, 2023). Third, future work should include additional measures that assess fear. For example, electroencephalography, or the small movements of muscles in the face, has been used in previous work on awe (Chirico et al., 2017) and may better differentiate between more fearful versus more positive experiences of awe.

This work points to the promise of studying more complex emotions cross-culturally and offers insights into how culture shapes discrete emotional experience and physiology (e.g., Grossmann et al., 2014; Mesquita et al., 2016; Shiota et al., 2010; Tsai, 2007). It also suggests that generating a complete portrait of awe is best served by examining it across multiple cultural and even temporal contexts.

Supplementary Information

Below is the link to the electronic supplementary material.

Additional Information

Funding

Partial financial support was received from the John Templeton Foundation (95524 and 88210) and the Social Sciences and Humanities Research Council (503200).

Competing interests

The authors declare no competing interests.

Availability of data and material

All data files and syntax available to other researchers on Open Science Framework: https://osf.io/tvgzc/?view_only=bf02d7b4bc3b4a55b9023d5b46845513

Code availability

None (syntax is included in above link).

Authors’ contributions

All authors contributed to the design of the studies; JS, YB, and GM collected the data in China and the US; CA and AG collected the data in the US only; JS and AG analyzed the results; JS wrote the manuscript; and all authors provided feedback.

Ethics approval

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of the University of California, Berkeley and Tsinghua University (2012–05-4358).

Consent to participate

Informed consent was obtained from all individual participants included in the study.

Consent for publication

The authors affirm that human research participants provided informed consent for publication of this data.