Differentiating positive emotions across cultures

Differentiating positive emotions across cultures

Research Master’s Thesis

Yong-Qi Cong 2015

Supervised by:

Dr. Disa Sauter (University of Amsterdam)

Prof. dr. Dacher Keltner (University of California, Berkeley)

Second assessor: Prof.dr. Agneta Fischer

Abstract

Emotion research in the past has largely focused on negative emotions. Despite the important role positive emotions play in well-being and interpersonal relationships. The current research investigates the expressive behaviours and appraisal patterns of ten positive emotions (amusement, awe, compassion, desire, gratitude, interest, love, pride, relief and embarrassment). Data were collected from two cultures (China and the US) to investigate the universality of these positive emotions. Participants recalled personal events involving each positive emotion in turn and shared their memories of these events in detail while being video-recorded. Participants’ expressive behaviours during the recalling including facial expressions, upper-body movements and vocalisations were coded.

Participants also filled in appraisal questionnaires about each emotional experience including both general appraisal dimensions and emotion- specific appraisal dimensions targeting at the appraisal profiles of the 10 emotions. Results show that most of these positive emotions are associated with unique patterns of expressive behaviours and appraisal patterns. Specific expressive behaviours of the 10 emotions were mostly culturally consistent, with a few cultural variations for some emotions when individual behaviours occurred within one culture but not the other. Using spontaneous emotion expressions, we replicated findings of previous research on the expressive behaviours of amusement, awe, interest, love, pride and embarrassment, but not compassion and desire, possibly due to

differences in the emotion-elicitation procedure. For the general appraisal

dimension, cultural variations emerged from awe, compassion and gratitude. For the emotion-specific appraisal dimensions, cultural variation emerged for every emotion, suggesting that emotion-specific appraisal dimensions are better at capturing aspects of appraisal that vary across cultures.

Differentiating Positive Emotions across Cultures

The ratio of people’s experiences of positive to negative emotions in daily life has been shown to predict the overall level of subjective well-being (Diener, 2000; Fredrickson, 2008), the stability of their marriages (Carstensen, Gottman, & Levenson, 1995), and their performance in business teams (Losada, 1999). Positive emotions have also been shown to promote physical health (Fredrickson, 2000; Pressman & Cohen, 2005), to fuel psychological resiliency (Fredrickson, 2001; Tugade, Fredrickson, Barrett, 2004) and to enhance social relationships (Shiota, Campos, Keltner, & Hertenstein, 2004). But despite the numerous benefits positive emotions have and their effects on both subjective well-being and interpersonal relationships, research has largely focused on negative emotions.

One of the reasons why discrete positive emotions have not yet been extensively studied may be the suggestion that pleasant experiences are less emotionally differentiated than unpleasant experiences (e.g., de Rivera, Possel, Verette, & Weiner, 1989; Ekman, 1992; Ellsworth & Smith, 1988). There are considerably more words to describe unpleasant feelings than pleasant ones in most languages (Dodds et al., 2014). Most theories and research on emotion reflect this asymmetry. Ekman’s theory of basic emotions includes only one positive emotion: happiness (Ekman& Friesen, 1986; Ekman, 1992), in addition to anger, sadness, fear, disgust, and surprise. Both Izard (1977; Izard & Buechler, 1980) and Tompkins (1984), proposed interest as another positive basic emotion, but they also add other negative emotions to the list, such as contempt, distress, and shame. Researchers tended to consider all positive emotional states as within the same family (e.g., Davidson, 1993; Ekman, 1992; Isen, Niedenthal, & Cantor, 1992), and many studies therefore did not focus on possible distinctions among positive emotions (e.g., Folkman, 2008; Fredrickson, 2001; Tugade & Fredrickson, 2004). In contrast, great efforts have been made to make very fine-grained differentiations between different negative emotions (e.g., distinguishing contempt, anger, and disgust from each other, Rozin, Lowery, Imada & Haidt, 1999; Hutcherson & Gross, 2011; distinguishing shame, guilt, and embarrassment from each other, Tangney, Miller, Flicker, Barlow, 1996).

In recent years, emerging research into positive emotions has provided evidence for the existence of multiple discrete positive emotions (e.g., Campos, Shiota, Keltner, Gonzaga, & Goetz, 2013; Ellsworth & Smith, 1988; Mortillaro, Mehu,& Scherer, 2011; Shiota, Campos, & Keltner, 2003; Shiota, Neufeld, Yeung, Moser, & Perea, 2011). For example, Shiota and colleagues (2011) observed

qualitatively different profiles of autonomic nervous activation in response to stimuli eliciting enthusiasm, love, amusement and awe. Ellsworth and Smith (1988) found that challenge, interest, love, hope, playfulness and tranquility could be differentiated from each other by their appraisal profiles. Campos and colleagues (2013) found both distinctive expressive display and appraisal patterns for amusement, awe, interest, and pride. These findings suggest that there are important differentiations among positive emotions. However, limited cross-cultural studies on positive emotions have been conducted, although research has shown that culture can have a significant influence on both expressive behaviours and appraisals of emotions (e.g., elfenbein, 2013; Elfenbein, Beaupré, Lévesque, & Hess, 2007; Scherer, 1997; Tsai, Chentsova-Dutton, & Freire-Bebeau, Przymus, 2002).

Researchers have long debated whether emotions are universal or culturally specific. Although extreme positions have been taken (e.g., Russell, 1994; Ekman, 1994), most recent theories have tried to account for both universality and cultural specificity by studying which particular aspects of emotion show similarities or differences across cultures (e.g., Elfenbein & Ambady, 2002; Mesquita & Frijda, 1992; Scherer & Wallbott, 1994).

The current study aimed to differentiate 10 positive emotions on two core features of emotional responding: expressive behaviours and appraisals. The similarities and differences of the expressive displays and subjective experience of positive emotions were compared between two cultures, Previous research on non-verbal expressions and appraisals of positive emotion will be reviewed next with reference to cross-cultural work.

Non-verbal expressions

Basic emotions are characterized by specific facial expressions involving different facial muscular movements (Ekman & Friesen, 1976). For example, Happiness is associated with a Duchenne smile (Ekman, Friesen, & Ancoli, 1980). Recent research attempting to dissect the global category “happiness” into several discrete positive emotions has found that displays of positive emotions can be

characterized by expressions beyond the Duchenne smile, including upper face actions and upper body movements. For example, pride is indicated by a small Duchenne smile with pressed lips and straight shoulders and back (Tracy & Robins, 2004); Awe is indicated by raised head and eyes, widened eyes, and slightly raised inner eyebrows (Shiota, Campos & Keltner, 2003). Other positive emotions that have been associated with unique facial expressions include amusement (Sarra & Otta, 2001), compassion (Keltner, 1995), desire (Gonzaga, Turner, Keltner, Altemus, & Campos, 2002), interest, pride (Mortillaro, Mehu & Scherer, 2011), love, joy and contentment (Campos et al., 2013). Table 1 summarizes these findings.

Table 1 Expressive displays associated with discrete positive emotions

Emotion Associated expressive bahaviours Reference Amusement Drop-jaw Duchenne smile, head

movement back, head bounce and gaze direction upwards

Shiota, Campos & Keltner (2003); Campos et al. (2003), Keltner (1995)_

Awe Raised head and eyes, widened eyes, and (slightly) raised inner eyebrows

Shiota, Campos & Keltner (2013); Campos et al. (2013) Compassion Oblique eyebrows, concerned or fixed

gaze, and head movement forward

Keltner (1995) Contentment Duchenne smile, lip press, head nod Campos et al. (2003)

Desire A variety of lip-related actions including lip licks, wipes and tongue protrusions

Campos et al. (2013) Joy Duchenne smile, head bounce, jaw drop Campos et al. (2013) Interest Raised and contracted brows, lip presses,

head tilts, head moves forward and forward leans

Campos et al. (2013); Mortilaro, Mehu, & Scherer (2011)

Love Open-mouthed Duchenne smile, with head tilts and closed eyes

Campos et al. (2013) Pride Duchenne smile with compressed lips, Shiota, Campos & Keltner

back straight, shoulders back and head tilts back

(2003); Mortillaro, Mehu, & Scherer (2011); Tracy & Robins (2004)

The prototypical expressive behaviours associated with positive emotions listed above were mostly established based on studies conducted in the US. Cross-cultural research on other emotions suggest that some emotions and their

corresponding expressive displays are universal, meaning that they can be interpreted reliably and similarly across cultures (e.g., Ekman, 1994; 1989). This has been termed the “Universality Thesis”. For example, facial expressions associated with the six basic emotions proposed by Ekman have been shown to be universal (Ekman & Friesen, 1971; Elfenbein & Ambady, 2002; Scherer & Wallbott, 1994). Prototypical expressions of happiness, anger, fear, sadness, disgust and surprise can be recognized at above chance rates in different cultures around the world including remote

preliterate cultures (see Ekman, 1994 for a review of these findings and Elfenbein & Ambady, 2002 for a meta-analysis).

For example, Elfenbein and colleagues (2007) asked participants from Canada and Gabon to pose a range of emotions using facial expressions (among the 10 tested emotions, only one was a positive emotion). They then analysed these expressions using the Facial Action Coding System (FACS, Ekman & Friesen, 1978), which objectively codes facial expressions from facial muscle movements. Reliable cultural differences were found beyond the prototypical facial expressions of emotions and idiosyncratic differences of individual posers.

The “Universality Thesis” states that emotion expressions can be interpreted reliably and similarly across cultures (Ekman, 1994). No study has specifically compared the non-verbal expressions of positive emotions across cultures. A cross-cultural investigation will help us identify which positive emotions are associated with specific expressions consistently across cultures and what are the cultural variations, if any. This is important in its own right as it will enrich our knowledge about emotion in general, considering the unbalanced focus on negative emotions in existing research. It is also specifically relevant for the universality thesis. which has been established based on research using predominantly negative emotions.

Additionally, previous research has mostly used posed expressions of emotions (e.g., Elfenbein et al., 2007; Mortillaro, Mehu, Scherer, 2011; Shiota, Campos,

Keltner, 2003), which have limited ecological validity (Cohn & Schmidt, 2004; Davis & Gibson, 2000; Russell, 1994; Valstar, Pantic, Ambadar, & Cohn, 2006). It has been suggested that posed expressions originate in a different region of the brain than do spontaneous facial expressions (Rinn, 1984), and posed expressions have also been found to be more asymmetric and lateralized than spontaneous expressions (Ekman, Hager, & Friesen, 1981; Skinner & Mullen, 1991). Therefore, posed expressions may not be accurate representations of real-life emotional expressive behaviours. A cross-cultural study examining positive emotions using spontaneous emotion expressions is thus needed.

Appraisals

An appraisal is an individual’s evaluation of an emotion-eliciting event or situation. Research shows that specific (configuration of) appraisals can be used to differentiate emotions (Frijda, Kuipers, & Ter Schure, 1989; Roseman, 2001; Siemer,

Mauss, & Fross, 2007; Scherer, Schorr & Johnstone, 2001; Smith & Ellsworth, 1985; van Dijk & Zeelenberg, 2002). For example, Smith and Ellsworth (1985) compared 15 emotions (among which 5 were positive) on six appraisal dimension: pleasantness, anticipated effort, certainty, attentional activity, self-other responsibility and situation control. Results showed that most of these emotions varied systematically along each of the appraisal dimensions and could be discriminated from each other by their appraisal patterns.

A few studies have specifically looked at the appraisal patterns of positive emotions. For example, Ellsworth and Smith (1988) examined the differentiation of pleasant emotional experiences by appraisals of effort, agency and certainty. The three appraisal dimensions were systematically manipulated to describe participants’

emotional experience and the authors concluded that appraisal patterns of 6 positive emotions (interest, hope, challenge, tranquility, playfulness and love) show

considerable differentiation in their appraisal patterns.

Given the highly subjective nature of the evaluation of a situation, one would expect significant cultural differences in cognitive appraisals of emotional events (Mesquita & Ellsworth, 2001; Mesquita & Frijda, 1992; Scherer, 1997). Scherer (1997) compared appraisals for 7 emotions (joy, anger, fear, sadness, disgust, shame and guilt) across 37 countries, which were grouped into 6 geographical regions. Results showed more cultural similarities than differences across geographical regions for emotion-specific appraisal profiles. However, cultural differences of general appraisal tendencies emerged. For example, participants in Africa were generally more likely to appraise events as more immoral, more unfair and more externally caused than participants from other regions. The appraisal dimensions used in this study were standard appraisal dimensions that have been used in many other studies (e.g., Smith & Ellsworth, 1985; Ellsworth & Scherer, 2003; Mauro, Sato, Tucker, 1992). However, most previous appraisal studies considered mainly negative emotions and the relevant appraisal dimensions were also developed for these negative emotions. Therefore, when these appraisal dimensions are applied to positive emotions, it is difficult to judge how well they capture the appraisal profiles.

Only one study to date has specifically examined appraisals of positive emotions cross-culturally. Tong (2015) tried to differentiate 13 positive emotions (amusement, awe, challenge, compassion, contentment, gratitude, hope, interest, joy, pride, relief, romantic love and serenity) by their appraisals using data from US and Singaporean participants. Using similar appraisal dimensions as in Scherer (1997), Tong (2015) identified specific appraisal patterns for most of the positive emotions.

However, contrary to Scherer’s (1997) findings on primarily negative

emotions, Tong (2015) did not find significant cultural differences for any emotion or any specific appraisal dimension. There are two possible interpretations for the lack of cultural differences in Tong’s (2015) study. One possibility is that the 13 emotions he examined all had culturally consistent appraisal patterns, meaning that events

triggering the same emotions are appraised in the same manner in both cultures. This would be remarkable given the high number of emotions and appraisals examined, and the distance between the two cultures compared, especially given that studies on mainly negative emotions or both negative and positive emotions have consistently found cultural differences in appraisal patterns (e.g., Borke & Su, 1972; Matsumoto et al., 1988; Mauro et al., 1992; Scherer, 1997; Scherer & Wallbott, 1994), even with much smaller cultural distances (e.g., Scherer, Wallbott, & Summerfield, 1986

compared eight European countries). Therefore, it would appear unlikely that there are no cultural differences between individuals from different cultures in their emotional appraisal patterns of positive emotions.

Another possibility is that there are meaningful cultural differences in the appraisal patterns of (at least some of) the emotions, which Tong’s (2015) study failed to detect. The appraisal dimensions used in Tong’s (2015) study targeted very general characteristics of the emotional experience. Some theorists suggest that there are fewer cultural differences for the more “basic” appraisal dimensions such as pleasantness, certainty, coping ability and goal conduciveness (Mauro, Sato & Tucker, 1992). In contrast, cross-cultural differences are more likely to be found for complex appraisal dimensions requiring cultural schemata (Scherer, 1997). The appraisal items used in Tong’s study were adapted from previous appraisal studies, which have mostly focused on negative emotions. Being rather general appraisal dimensions made it possible to use them to test positive emotions, too. However, this might be the reason why they were not able to capture cultural variations in positive emotions. In order to test this, a cross-cultural study should be conducted using both the general appraisal-items and more complex appraisal appraisal-items specific for positive emotions.

The current study

The goal of the current study was to better differentiate and understand positive emotions by examining both non-verbal expressions and appraisals. A story-telling paradigm was used to elicit spontaneous emotion expressions and the stories were served as stimuli for the appraisals. Participants talked about a personal event related to a specific emotion while being video recorded (for non-verbal expressions coding) and then filled in appraisal questionnaires about their experiences. The appraisal questionnaires included both items targeting basic appraisal dimensions and those targeting more complex appraisal dimensions developed specifically for positive emotions.

Ten emotions were investigated in the current study: they are: amusement, awe, compassion, desire, embarrassment, gratitude, interest, love, pride, and relief.

Embarrassment was included here despite being considered a negative emotion because it is often associated with positive signals– for example, smiling or giggling (Keltner, 1995; 1996) and seen as a sign of appeasement (Keltner & Buswell, 1997). This makes the emotion relevant for a comparison with other positive emotions, especially when it comes to non-verbal expressions. Data were collected from two countries, the US and China. We used community samples to maximize the generalizability of the findings.

Method

Participants

Participants consisted of 50 Chinese adults (76% females, Mage= 25.6, SDage =

7.6) from Dalian, a medium-sized coastal city in the northeast of China and 30

American adults (56% females, Mage= 33.7, SDage= 14.1) from Berkeley, a comparable

city on the west coast of the US. Participants from both cultures were recruited from the local community through announcements on social media. None of the participants had lived for more than 2 years outside of their home country. All participants

received monetary compensation for taking part in the study. Design

Participants recalled a specific time in their life when they had experienced a target emotion and told their experience to the experimenter in as much detail as possible while being video-recorded. Participants also filled in two appraisal questionnaires measuring both general appraisal dimensions and emotion-specific appraisal dimensions. We used a within-subject design so that each participant recalled emotional events and filled in appraisal questionnaires for all emotions.1 The order of emotions was randomized across participants.

Materials

Two appraisal questionnaires were used. The first questionnaire included 13 general appraisal items taken from Tong (2015). The general appraisal dimensions measured were Pleasantness, Relevance, Problem, Goal attainment, Agency-self, Agency–other, Agency –circumstances, Control-self, Control-other, Control-

circumstances, Certainty, Predictability, and Effort. See Appendix A for the full list of the general appraisal items.

The second appraisal questionnaire included 20 emotion-specific appraisal items (2 items per emotion). These items aimed to capture the appraisal patterns of the positive emotions examined in this study and were developed based on previous research on the specific features of positive emotions. For example, awe has been shown to be associated with the feeling that a situation challenges one’s worldview and feeling that one is small relative to the environment or others (Keltner & Haidt, 2003; Shiota, Keltner, & Mossman, 2007). The two items developed to capture these two features were “Understanding” (Did you feel that you have difficulty

understanding what was happening in the situation?) and “Perceived size of self” (To what extent did you feel small or large relative to the environment?). As another example, gratitude has been shown to be associated with one’s perception of having been benefited by another’s action and wanting to reciprocate or benefit another (Algoe et al., 2008; McCullough et al., 2001). The two items developed to capture these two features were “Benefit – passive” (Did you feel you had benefitted from someone else’s actions?) and Benefit – active (Did you feel that you wanted to do something that would benefit others?). See Appendix B for a full list of the specific appraisal items

1Five participants did not finish all the emotions within the scheduled time (i.e., two hours). They were allowed to stop after two hours.

All items were rated on a 0 to 10 Likert-scale (0=not at all, 5= mid-point, 10 = extremely). An NA (not applicable) option was also included for each item to be selected if participants thought the item was not relevant for their experience. English questionnaires were translated into Chinese by bilingual research assistants in

Berkeley who have experience with translating research materials between the two languages.

In addition to the appraisal questionnaires, participants also rated the extent to which they experienced the target emotion during the event (How strongly did you experience the emotion during the event?) as well as during the recalling (How

strongly did you experience the emotion while recalling the event?). Emotion intensity was rated on a scale of 0 to 10 (0=not at all, 10 = extremely).

Procedure

Participants listened to a recording of the name and definition of a target emotion in their own language. They were then asked to explain the emotion in their own words to make sure they understood the emotion. When participants seemed to have misunderstood the emotion (as judged by the experimenter), the experimenter played the recording of the emotion definition again and explained the emotion in other words to the participants until they understood. Participants were then asked to recall a specific event involving the emotion and talked about it in detail. After participants finished talking about each experience, the experimenter asked them to describe the moment when they experienced the target emotion most strongly. The experimenter then briefly summarized the experience of the participant (with

clarification from the participant if necessary) before asking the participant to pose an expression of the emotion.2 Participants then filled in the appraisal questionnaires before moving on to the next emotion. The procedure was repeated for all 10 emotions or until 2 hours elapsed since start of the session.

Results and discussion Expressive Behaviours

Analysis

A coding scheme was developed for the non-verbal expressions of emotions. The scheme was adapted from FACS by simplifying the codes, that is, coding descriptive facial actions instead of fine muscle movements. In addition, more upper-body actions and some vocalisations were added. Instead of trying to map out the specific physiological changes in facial muscle actions, we were more interested in how the expressive behaviours appear to an average observer. See Appendix C for a list of all the behaviours coded.

Participants’ description of the moment when they experienced the target emotion most strongly was used for expressive behaviour coding. This moment was chosen for coding for two reasons. First, the entire story-telling session lasted long and

2_{The posed expressions were subsequently not used for analysis for two reasons. Firstly, many} participants had difficulty posing – they either did not want to act out the emotion or did not know how to. Secondly, the posed expressions that we did get were very different from the spontaneous expressions and thus did not seem to be representative of the real-life expressive behaviours.

contained many moments not necessarily related to the target emotion. To code the entire session would be both impractical and partly likely to not represent the target emotion. Secondly, answering the question “what was the moment when you

experienced this emotion most strongly?” required the participants to create a moment of particularly deep re-living. This moment would arguably be the moment when the participant experienced the target emotion most strongly during the recalling

procedure and therefore the expressive behaviours displayed should also be most clear. Most participants’ answers to this question lasted within half a minute. To control for the length of the video clips used for coding, answers that were longer than half a minute were only coded for the first 30 seconds. A binary coding was applied to each coded behaviour (i.e., each behaviour was judged to either have occurred or not to have occurred). An occurrence frequency was then calculated for each action by counting the total number of times the action occurred and dividing that by the number of video clips coded. The frequency was calculated for each emotion separately for each culture.

Two frequency levels were determined for behaviours associated with each emotion based on previous research (Campos et al., 2013). Behaviours that occurred more than 50% of the time were considered as “high frequency” behaviours.

Behaviours that occurred between 30% and 50% were considered as “moderately high frequency” behaviours.

Results and discussion

The emotion-specific patterns of expressive displays were largely consistent across cultures, with only minor cultural variations. Table 2 summarizes the results.

Smiles occurred either very frequently or moderately frequently for all 10 emotions. Laughter occurred very frequently for amusement and moderately

frequently for embarrassment for both cultures. A few actions frequently occurred for many emotions, such as head tilt (5 emotions both culture, 2 emotions single culture), gaze down (7 emotions both culture, 3 emotions single culture), and gaze to side (6 emotions both cultures, 4 emotions single culture). The frequent occurrence of these behaviours might either result from the story-telling procedure, as observed in multiple emotions, or indicate shared functions of these emotions.

There were also two seemingly culture-specific actions - Chinese nodded frequently and Americans often raised their eyebrows. Nodding was a frequently occurring action in 9 out of 10 emotions (all except amusement) for Chinese participants, while eyebrow raise was a frequently occurring action for 9 out of 10 emotions (all except love) for American participants. These two actions might thus be specific behaviour associated with the story-telling or as part of the culture-specific communication styles.

Amusement was associated with frequent displays of smiles and laughter for participants from both cultures, which confirms previous findings (e.g., Campos et al., 2003; Shiota, Campos & Keltner, 2003). In addition, we found head tilt, gaze down, and gaze to side to be moderately frequently associated with amusement in both cultures. However, these actions also occurred frequently in other emotions examined in this study, such as awe, compassion, gratitude, interest, and love. Gaze upwards, as

Table 2 Frequently occurring expressive behaviours by emotion and culture

Amusement Awe Compassion Desire Gratitude

CN (N=29) US (N=22) CN (N=28) US (N=23) CN (N=29) US (26) CN (N=29) US (N=24) CN (N=26) US (N=26) High frequency (>50%) Wide smile (79.3%) Eyebrow raise (59.1%) Small smile (71.4%) Eyebrow raise (65.3%) Gaze down (51.7%) Eyebrow raise (57.7%) None Eyebrow raise (66.7%) Nod (65.4%) Eyebrow raise (53.8%) Vocalisation – laughter (75.8%) Vocalisation – laughter (59.1%) Gaze to side (51.7%) Gaze to side (53.8%) Gaze down (53.8%) Gaze to side (53.8%) Small smile (68.9%) Small smile (54.6%) Head turn (51.7%) Lip press (53.8%) Small smile (53.8%) Moderate frequency (30-50%) Gaze down (48.2%) Wide smile (50%)

Head tilt (50.0%) Gaze up (43.5%) Furrowed brows (41.4%) Gaze down (50.0%) Head back/up (48.3%) Widened eyes (33.3%) Gaze to side (38.5%) Small smile (46.2%) Head tilt (48.2%) Gaze to side (50.0%) Furrowed brows (46.4%) Small smile (43.5%) Small smile (37.9%) Head shake (46.2%) Small smile (44.8%) Gaze down (33.3%) Lip press (38.5%) Head back/up (38.5%) Eyebrow raise (31.0%) Head tilt (45.5%) Gaze up (42.9%) Head back/up (39.1%)

Nod (34.5%) Head tilt (42.3%) Gaze to side (41.4%) Closed eyes (33.3%) Wide smile (34.6%) Gaze down (34.6%) Gaze to side (31.0%) Gaze down (31.8) Eyebrow raise (35.7%) Gaze to side (39.1%) Lip press (34.5%) Furrowed brows (34.6%) Eyebrow raise (37.9%) Wide smile (33.3%) Head tilt (34.6%) Head tilt (34.6%) Head down (31.0%) Gaze up (31.8&) Gaze down (35.7%) Widened eyes (34.8%) Vocalisation – tut (31.0%) Closed eyes (34.6%) Gaze down (37.9%) Wide smile (30.8%) Nod (35.7%) Wide smile (34.8%) Small smile (30.8%) Furrowed brow (31.0%) Nod (30.8%) Gaze to distance (32.1%) Head tilt (34.8%) Wide smile (31.0%) Gaze up (30.8%) Furrowed brow (30.4%) Nod (31.0%) Lip press (30.4%)

Interest Love Pride Relief Embarrassment CN (N=29) US (N=24) CN (N=27) US (N=20) CN (N=28) US (N=26) CN (N=29) US (N=21) CN (N=29) US (N=23) High frequency (>50%) Nod (65.5) Eyebrow raise (54.2%) Small smile (63.0%) Small smile (60%) Small smile (60.7%) Gaze to side (61.5%) Small smile (62.1%) Eyebrow raise (66.7%) Gaze down (65.5%) Eyebrow raise (65.2%) Small smile (55.2%) Gaze to side (54.2%) Head tilt (59.3%) Gaze to side (60%) Eyebrow raise (57.7%) Wide smile (55.2%) Small smile (57.1%) Lip press (51.7%) Gaze to side (65.2%) Nod (51.9%) Head tilt (55%) Small smile (57.7%) Moderate frequency (30%-50%) Head tilt (48.3%) Head tilt (50.0%) Gaze down (40.7%) Gaze down (45%) Gaze down (50.0%) Nod (50%) Nod (48.3%) Nod (47.6%) Small smile (48.3%) Wide smile (47.8%) Eyebrow raise (41.4%) Small smile (50.0%) Lip press (40.7%) Wide smile (45%) Gaze to side (50.0%) Wide smile (42.3%) Gaze down (41.4%) Gaze to side (47.6%) Wide smile (41.4%) Vocalisation – Laughter (47.8%) Gaze down (41.4%) Gaze up (41.7%) Gaze up (37.0%) Closed eyes (35%) Nod (50.0%) Head tilt (38.5%) Head back/up (37.9%) Head back/up (42.9%) Head down (41.4%) Lip press (43.5%) Wide smile (34.5%) Widened eyes (33.3%) Head back/up (33.3%) Lip press (35%) Head back/up (35.7%) Gaze down (34.6%) Gaze to side (34.5%) Head shake (42.9%) Head turn (41.4%) Widened eyes (39.1%) Gaze up (31.0%) Nod (35%) Wide smile (32.1%) Lip press (30.8%) Head forward (31.0%) Wide smile (38.1%) Vocalisation- Laughter (41.4%) Gaze down (39.1%) Gaze to side (37.9%) Closed eyes (34.8%) Face touch (37.9%) Head turn (34.8%) Furrowed brows (31.0%) Furrowed brows (30.4%) Eyebrow raise (31.0%) Small smile (30.4%) Nod (31.0%) Head shake (30.4%) Gaze up (30.4%)

found to be a characteristic behaviour of amusement in previous studies (e.g., Campos et al., 2003; Keltner, 1995), was found to occur only for US participants. This result is consistent with findings from previous research, which has only been conducted with US samples. The current results suggest that this may not be a cross-culturally consistent behaviour associated with amusement.

Awe has been associated with raised head and eyes, widened eyes, and (slightly) raised inner eyebrows in previous research (e.g., Campos et al., 2013; Shiota, Campos & Keltner, 2013). We found that, in additional to smiling, participants displayed eyebrow raise, gaze up, furrowed brows and head tilt, either with high frequency or moderate frequency when recalling an awe experience. The frequent occurrence of eyebrow raise and gaze up both confirm previous findings. In addition, head back and widened eyes were found to occur moderately frequently only for US participants. Widened eyes has also been found to be characteristic expressive display of awe in previous studies using US samples (e.g., Campos et al., 2013; Shiota, Campos & Keltner, 2013). Furrowed brows and head tilt have not been previously associated with awe. However, previous research on awe has defined it to be the emotion experienced during cognitive accommodation or schema formation, as a response to novel, complex stimuli or simply having difficulty understanding the situation (Campos et al., 2013; Keltner & Haidt, 2003; Shiota, Keltner, & Mossman, 2007; Shiota Campos, Keltner, Hertenstein, 2004). Furrowed brows and head tilt might be an expression of this mental process. Two other culture-specific actions that occurred with moderate frequency was gaze to distance for Chinese participants and lip press for US participants. These are possibly culture-specific behaviour characteristic of awe, but more future studies are needed to confirm this.

Actions that occurred frequently or moderately frequently when participants recalled a compassion experience were: gaze down, gaze to side, furrowed brows, small smile, and lip press for both cultures. Chinese participants additionally showed (moderately) frequent lip press and tutting, while US participants showed (moderately) frequent head shake, and closed eyes. Many of these actions seem to come from an attempt to avoid looking, which could be explained by the discomfort that participants experienced caused by the suffering of another. Furrowed brows, head shake, lip press and tutting could also be the reflection of the discomfort (Mpleasantness = 2.00 across

cultures on a scale from 0-10) and the dissatisfaction with the situation (Msatisfaction = 3.15 across

cultures on a scale from 0-10) experienced during the recalling. The frequent behaviours for compassion found in the current study look quite different from the pattern found by previous studies (e.g., Keltner, 1995). An attempt to avoid looking has not been previously associated with compassion. This could be due to the difference in emotion elicitation procedure used. Most previous studies use posed emotion expressions (e.g,, Elfenbein et al., 2007; Keltner, 1995; Mortillaro, Mehu, Scherer, 2011; Shiota, Campos, Keltner, 2003), while the current study used spontaneous emotion expressions. In the context of posing, participants may be less likely to avoid looking. Therefore, the expressive behaviours for compassion found in the current study provide insights into spontaneous emotion expressions and point future research into a new direction.

For desire, the frequent or moderately frequent behaviours for both cultures were: smile, eyebrow raise, and gaze down. Additionally, Chinese participants showed head back, gaze to side, and furrowed brows while US participants showed widened eyes and closed eyes. We did not find any lip-related actions as did previous research (e.g., Campos et al., 2013). There are two possible reasons why the patterns we found differed from previous findings. The first is that, as noted previously, the current study employed a different elicitation procedure aiming to target

spontaneous emotion expressions. The second is that previous studies on desire used a narrower definition of the emotion – namely sexual desire, while the current study used a broader definition of the emotion to refer to all types of desires. The two types of desires could be the same emotion, or they could have significant differences. Future research could try to compare the expressive behaviours and appraisal patterns of general desire to sexual desire.

Interest was found to be associated with smile, eyebrow raise, head tilt, and gaze up for both cultures. Chinese participants additionally showed gaze down and US participants additionally showed widened eyes. Eyebrow raise and head tilt both confirm previous findings (e.g., Campos et al., 2013; Morilaro, Mehu & Scherer, 2011). We did not find head forward and forward leans to be frequently occurring actions for interest, as did by previous research.

Love was associated with (moderately) frequent display of smile, head tilt, nod, gaze down, and lip press. Additionally, Chinese participant also frequently showed gaze up and head back, while US participants also frequently showed closed eyes and nod. Smile, head tilt and closed eyes have been previously found to be characteristic expressions of love. Consistent with previous research, we found all three actions to occur frequently in the current study, although head tilt and closed eyes were only for US participants. This again suggests that previous findings on expressive behaviours of love (and possibly other emotions) based on mainly US samples do not fully

generalize to other cultures.

Expressive displays of pride were characterized by smile, nod, gaze down and gaze to side for both cultures. Additionally, Chinese participants showed head back and US participants showed head tilt and lip press. Smile, lip press and head back were found in previous research to be

associated with pride (e.g., Tracy & Robins, 2004). All three actions occurred frequently in the current study, too, although not for both cultures. We also found new actions not recognized before, including nod, gaze down and gaze to side. The two actions with an attempt to avoid looking might have resulted from participants’ feeling shy talking about their pride experience. This could be tested in future research by adding an appraisal item targeting this feeling.

Previous studies did not find any specific patterns of expressive behaviours associated with gratitude (e.g., Campos et al., 2013). We did, however, find many, including smile, nod, gaze down, gaze to side and head tilt for both cultures, and additionally gaze to side and lip press for Chinese participants and head back and gaze up for US participants. This could be due to the differences in the emotion-elicitation procedure we used.

Relief was shown to be associated with high frequency or moderately high frequency display of smile, nod, head back and gaze to side for both cultures, and additionally gaze down and head forward for Chinese participants and head shake for US participants. Facial expressions of relief have not been specifically studied before, but research has found relief to be associated with signs (Sauter & Scott, 2007; Soltysik & Jelen, 2005). We did not observe any frequently-occurring vocalisations during the recall of relief experiences. As noted before, this could be due to the emotion elicitation procedure used in the current study. As participants were constantly talking during the recall, vocalisations were less likely to occur.

Lastly, embarrassment was characterized by smile, laughter, gaze down, gaze to side, lip press, eyebrow raise, head turn and furrowed brow for both cultures, and additionally head down, face touch and nod for Chinese participants, and widened eyes, closed eyes, head shake and gaze up for US participants. Many of the actions involve gaze aversion, which together with smile, head turn and occasional face touch have been marked as the distinct display of embarrassment by previous research (Haidt & Keltner, 1999; Keltner, 1995; Keltner & Harker, 1998). The current study with spontaneous emotion expressions thus confirms the distinct display of embarrassment as defined in previous research.

Appraisals

Emotion intensity

In order to find out whether the emotion intensity rating differed between the past and present (i.e., when the event occurred and during recalling), between the two cultures and across emotions, a 10 (emotion) x 2 (intensity past and present) x 2 (culture) repeated measure mixed

analysis of variance (MANOVA) was performed with emotion and intensity as within-subject factors and culture as between-subject factor. There was a main effect of time, F (1, 71) = 155.12 , p < 0.001 , η2 = 0.68, meaning the intensity reported for the past was significantly different from that for the moment of recalling. Although the intensity ratings for both time points were rather high, participants experienced the target emotion more strongly during the actual event than during the recalling, with a higher emotion intensity reported, Mpast= 8.61, SDpast= 0.12 versus Mpresent= 6.74, SDpresent= 0.18. There were no significant differences between cultures or across emotions, and no interactions with these factors.

Participants recalled emotional events of high intensity, both during the actual event in the past (M =8.61 on a 10-point scale) and during the recalling/ re-living in the present (M = 6.74 on a 10-point scale). Although emotions experienced during recalling were less intense than during the original events, the reported intensity was still very high compared to other emotion-eliciting methods such as watching film clips or reading/listening to emotion stories of someone else. For example, Gross and Levenson (1995) evaluated 78 emotion-eliciting films and selected 14 out of these, which were most effective at eliciting 7 target emotions. For amusement, the two films elicited emotion intensity of 5.54 and 5.86 respectively (on a 9-point scale), compared to intensity ratings of 8.43 (past) and 6.93 (present) for the same emotion in the present study. The average rating of emotion intensity for all selected films in Groaa and Levenson’s studies was 5.02. Other films that were not selected as they were less effective at eliciting emotions than the selected movies, as shown by participants giving lower intensity ratings. This indicates that the story-telling method employed in the current study is very effective in emotion elicitation, at least for positive emotions.

Missing data and NA responses

There were a total of 26400 possible data points (33 questions per questionnaire x 10 emotions per participant x 80 participants). Five participants did not complete all 10 emotions due to time constraints, with a total of 7 emotions missed (1 x amusement, 1 x love, 1 x interest, 2 x desire, and 2 x embarrassment). This resulted in 231 missing data points (33 questions per emotion questionnaire x 7 emotions). In addition, participants sometimes skipped questions or did not give answers for some questions for unknown reasons, which resulted in 335 additional missing data points. Consequently, the analyses were conducted with a total of 25834 data points.

There were a total of 939 “not applicable” (NA) responses (3.6% of all responses), in which 608 came from Chinese participants and 331 from American participants. The majority of NA responses came from the general appraisal items. Although there were less general appraisal questions than specific appraisal questions, a total of 692 NA responses were counted for the general appraisal items (6.7%), compared to only 84 (0.5%) for the specific appraisal items. There was no significant difference in the number of NA responses across emotions and cultures.

The NA option was provided so that participants would have the possibility to skip an appraisal item if they did not or would not appraise the event on that dimension. The NA response rate in the current study (3.6% overall) was very low compared to previous studies (e.g. 16.3% by Scherer et al., 1997; 10.01% by Tong, 2015). One key difference between the current study and previous studies is that new appraisal items were used, adding to the frequently used general appraisal items and resulting in a wider range of appraisal dimensions measured. This may have contributed to a different response distribution in the current study from previous studies. The NA response rate was also much lower for the specific appraisals (0.5%) than for the general appraisals (6.7%). The specific appraisal items were in fact rarely perceived as irrelevant for participants’ emotional experiences. This is unsurprising because these items were developed specifically for positive emotions, as an attempt to capture the appraisal profiles of the 10 particular emotions

examined in the current study. The low NA response rate for these items indicates their suitability for use with the current set of positive emotions. Future research could test whether these items, or a subset therein, are also suitable for measuring appraisals of other emotions beyond the 10 studied here.

General appraisals

In order to find out whether appraisal patterns differed across emotions and between the two cultures, a 10 (emotion) x13 (general appraisal items) x 2 (culture) mixed analysis of variance (MANOVA) was performed with emotion and appraisal as within-subject factors and culture as between-subject factor3. There were a significant main effect of emotion, F (9, 23) = 8.19, p < 0.001, η2 = 0.76 and a significant main effect of appraisal, F (12, 20) = 9.50, p < 0.001, η2 = 0.85. There were also significant interaction effect between emotion and culture, F (9, 279) = 2.06, p = 0.033, η2 = 0.06, between appraisal and culture F (12, 372) = 3.87, p < 0.001, η2 = 0.11, and

between emotion and appraisal, F (108, 3348) = 9.42, p< 0.001, η2 = 0.23, qualified by a three-way interaction of emotion, appraisal and culture, F (108, 3348) = 1.68, p = 0.034, η2 = 0.11. The interaction effects between emotion and appraisal suggest that appraisal profiles differ among the 10 emotions. The interaction effects involving culture suggest that appraisal profiles differ between the two cultures.

Next, to find out for which emotions did appraisal patterns differ between the cultures, a 13 (general appraisal) x 2 (culture) MANOVA was performed for each of the emotion correcting for inflated error rate resulting from multiple testing using a Bonferroni-Holm correction4. Culture had a significant main effect on appraisals for awe, F (13, 39) = 4.34 , p < 0.001, η2 = 0.59 ,

compassion, F (13,40) = 4.69 , p < 0.001, η2 = 0.60 and gratitude, F (13, 50) = 3.11, p= 0.002, η2 = 0.45, meaning that appraisal profiles for these emotions differ between the two cultures.

To look more closely at these cultural differences, posthoc tests were conducted, comparing the cultural groups for each individual appraisal items for each of the emotions for which a main effect of culture was found (awe, compassion and gratitude).

For awe, significant cultural differences emerged on two appraisal dimensions (Figure 1.1), pleasantness, F (1, 51) = 20.2, p < 0.001, η2 = 0.28 and effort, F (1, 51) = 12.07 , p = 0.001 , η2 = 0.19. Chinese participants perceived their awe experience as being less pleasant (MCN = 5.93, SDCN = 3.35 versus MUS= 9.17, SDUS = 0.94) and more effortful (MCN = 4.93, SDCN = 3.62 versus MUS =

1.83, SDUS = 2.62) than did American participants. One possible explanation is that the Chinese word for awe has a stronger connotation of fear, as the Chinese word literally is a combination of fear and respect. Fear is typically a negative emotion, and so an experience involving more fear might naturally be less pleasant, and may require more effort to cope with. This mismatch between the two cultures could potentially mean two things. The first possibility is that the concept of awe for Chinese people involves more fear, meaning the lower pleasantness of the emotion reflects a real difference in the emotional experience. The other possibility is that the translation of awe in Chinese used so far is not an accurate translation, referring to a slightly different emotional state than the English “awe” does. The concept of awe in China thus might exist as the same experience with a different verbal label or without any verbal label. This could be tested by devising a task where emotion or emotional experiences are communicated without using any verbal labels. A task

3 _{Mauchly’s test of Sphericity showed that the sphericity assumption was violated for the factor appraisal. Therefore, a}

Greenhouse-Geisser correction was applied to all effects involving appraisal.

4_{Bonferroni-Holm method, or Holm’s sequential Bonferroni method (1979) is a more powerful version and a} commonly used variant of the simple Bonferroni correction. It counteracts the problem of multiple comparisons by adjusting the rejection criteria of each of the individual comparisons sequentially instead of uniformly.

independent of language should be able to detect whether the current effect reflects cultural differences in emotional experience, concepts, or language.

For compassion, significant cultural differences emerged on three appraisal items (Figure 1.2), pleasantness, F (1, 52) = 13.65, p = 0.001, η2 = 0.21, problem, F (1, 52) =32.35, p < 0.001, η2 = 0.38, and goal-attainment, F (1, 52) = 4.28, p = 0.044, η2 = 0.08. Chinese participants perceived their compassion experience to be less pleasant (MCN = 0.85, SDCN = 1.99) than did American participants (MUS = 3.41, SDUS = 2.99), but also less problematic (MCN = 3.56, SDCN = 3.04 versus MUS = 7.93, SDUS = 2.59), suggesting that the Chinese participants were more touched by another’s suffering (perhaps leading to more emotional suffering themselves) but that they did not see this as a problem that needed to be solved (by them). This is counterintuitive as China has been considered a collectivistic society (Hofstede, 1993; Trompenaars 1993; Tung 1991), which values harmony and relationships with in-group memners. One might thus expect Chinese individuals to be more likely to help others in need. However, this is not what our data suggest. One possibility is that Chinese participants did not view those they feel compassion towards as in-group members. This would be consistent with the results of the specific appraisal item “connection”, where US participants felt significantly more connected to others in the situation when they were recalling the compassion experience than did the Chinese (MUS = 5.76, SDUS = 3.49 versus MCN = 3.32, SDCN = 3.38).

Chinese participants were also less likely to think that they had achieved any important goals during their compassion experience (MCN = 2.93, SDCN = 3.35 versus MUS = 4.89, SDUS = 3.62). This could be explained by Americans seeing the situation as more problematic and therefore being more motivated to solve the problem, and thus being more likely to take action. If they did act to help solve the problem they perceived to exist, they may then rate goal-attainment as higher. This is also suggested by the responses on one specific appraisal item, “wish to change”, where US participants less wanted to change the situation when they were recalling than did Chinese participants (MUS = 5.14, SDUS = 4 versus MCN = 7.18, SDCN = 2.88). If they had already taken action to help, they would less likely want to change the situation again while recalling. This would also moderate the unpleasant emotional reaction of compassion and explain why Americans rated the experience of compassion to be more pleasant than did Chinese. This could be confirmed in future research by including questions specifically asking whether participants have taken actions related to their appraisals or looking at action tendencies related to these appraisals.

Figure 1 Significant cultural differences in general appraisal items

For gratitude, significant cultural differences emerged on three appraisal items (Figure 1.3), agency-other, F (1, 62) = 5.17, p = 0.026, η2 = 0.08, effort F (1, 62) = 4.27, p = 0.043, η2 = 0.06, and certainty F (1, 62) = 13.4, p = 0.001, η2 = 0.18. Chinese participants perceived their gratitude experience to be less caused by others as compared to US participants (MCN =6.08, SDCN = 4.2 versus MUS = 8.19, SDUS = 2.64). This is not to say that they did not perceive the gratitude

experience to have been caused by others at all, but that they did so less than did Americans (MCN =6.08 versus MUS = 8.19). One possibility is that the Chinese participants thought that the

experience of gratitude was to a large extent also caused by themselves, as they needed or wanted to actively appreciate what others had done for them. Consistent with this notion, Chinese participants did score higher on the agency - self item than American participants did (MCN =6.87, SDCN = 3.54 versus MUS =5.38, SDUS =3.30), although the difference did not reach significance.

Chinese participants also perceived the situation to be more effortful than did American participants (MCN =7.05, SDCN = 2.88 versus MUS = 5.38, SDUS = 3.56) while recalling a gratitude experience. This could be related to the high self-agency. If Chinese participants felt they played an active role in their gratitude experience themselves, they might be more likely to think about take actions, to express their gratitude or to reciprocate for example, which all take extra effort.

Lastly, Chinese participants felt more certain about what had happened than did their American counterparts (MCN = 8.58, SDCN = 1.99 versus MUS = 6.42, SDUS = 2.72). This could be related to the higher perceived self-agency – as we are generally more certain of things we did ourselves than things other people did. Comparing with other emotions on this dimension, Chinese participants scored higher on certainty for every other emotion, too, but the differences were not big enough to reach significance. Therefore, it is also possible that Chinese people are more certain about what they report, which was reflected for one of the emotions. This could be tested by including questions on certainty not directly related to emotion (e.g., how certain are you of when this event happened?).

Appraisals specific to positive emotions

To find out whether the appraisal patterns of the emotion-specific dimensions differ across emotions and between the two cultures, a 10 (emotion) x 21 (specific appraisal questions) x 2 (culture) mixed analysis of variance (MANOVA) was performed with emotion and appraisal as

within-subject factors and culture as between-subject factor5. There was a significant main effect of emotion, F (9, 21) = 7.82, p < 0.001, η2 = 0.77 and a significant main effect of appraisal, F (20, 10) = 10.22, p < 0.001, η2 = 0.95. There were also significant interaction effects between appraisal and culture, F (20, 10) = 4.78, p = 0.007, η2 = 0.91, between emotion and appraisal, F (21, 608) = 7.95, p< 0.001, η2 = 0.22, and a three-way interaction of emotion, appraisal and culture, F (21,608) = 1.88, p = 0.01, η2

= 0.06. The interaction effects between emotion and appraisal suggest that appraisal profiles differ among the 10 emotions. The interaction effects involving culture suggest that appraisal profiles differ between the two cultures.

Next, a 21 (specific appraisal questions)6 x 2 (culture) MANOVA was performed for each of the emotions using Bonferroni-Holm correction for multiple comparisons. Culture had a significant main effect on appraisals for all emotions (all ps ≤ 0.005). Table 3 shows the test statistics of the main effect for each emotion and Table 4 summarizes on which appraisal items cultural differences emerge.

Table 3 Main effect of culture on appraisals for each emotion

Emotion df error df F p η2 Amusement 21 25 3.22 0.003 0.73 Awe 21 26 4.08 < 0.001 0.77 Compassion 21 27 4.58 < 0.001 0.79 Desire 21 30 5.15 < 0.001 0.78 Gratitude 21 29 3.27 0.002 0.70 Interest 21 26 2.95 0.005 0.71 Love 21 27 8.82 < 0.001 0.87 Pride 21 32 3.47 0.001 0.70 Relief 21 20 6.79 < 0.001 0.88 Embarrassment 21 29 3.26 0.002 0.70

Table 4 Significant cultural differences per appraisal by emotion (all dfs = 1)

Appraisal F p η2 MCN (SD) MUS (SD) Amusement 3. Strength 7.69 0.008 0.15 2.07 (3.08) 5.06 (4.28) 4. Social value 13.02 0.001 0.22 2.10 (2.94) 5.53 (3.34) 16. Other-awareness 19.54 < 0.001 0.30 1.53 (2.73) 5.71 (3.70) 18. Relaxation 4.58 0.038 0.09 8.77 (1.94) 7.18 (3.17) Awe 1. Size of self (L) 5.51 0.023 0.11 1.23 (2.94) 3.50 (3.75) 3. Strength 24.83 < 0.001 0.35 1.88 (2.88) 6.41 (3.42) 6. Incongruity 5.68 0.021 0.11 1.12 (1.86) 2.91 (3.26) 14. Wish to change 4.51 0.039 0.09 2.31 (3.16) 0.73 (1.61) 16. Other-awareness 10.23 0.002 0.18 1.58 (2.87) 4.41 (3.26) 17. Satisfaction 6.47 0.014 0.12 6.19 (3.07) 8.18 (2.17) Compassion 1. Size of self (L) 18.08 < 0.001 0.28 1.21 (1.69) 3.76 (2.51) 3. Strength 43.67 < 0.001 0.48 1.39 (2.13) 5.81 (2.54) 10. Connection 6.09 0.017 0.12 3.32 (3.38) 5.76 (3.49)

5 _{Mauchly’s test of Sphericity showed that the sphericity assumption was violated for both emotion and} appraisal. Therefore, a Greenhouse-Geisser correction was applied to all effects reported.

6 _{The appraisal dimension “size of self” had two sub-ratings – large and small. Therefore the analysis covered 21}

12. Curiosity 8.85 0.005 0.16 1.07 (1.88) 3.05 (2.77) 14. Wish to change 4.29 0.044 0.08 7.18 (2.88) 5.14 (4.00) 16. Other-awareness 5.18 0.028 0.10 2.00 (3.16) 4.05 (3.06) 19. Yearning 4.28 0.044 0.08 6.96 (3.79) 4.86 (3.15) Desire 1. Size of self (L) 6.20 0.016 0.11 2.32 (2.72) 4.46 (3.46) 6. Incongruity 10.03 0.003 0.16 0.93 (1.54) 3.50 (3.97) 9. Vulnerability 7.03 0.011 0.12 3.07 (3.28) 5.71 (3.90) 10. Connection 5.18 0.027 0.09 3.82 (3.42) 6.08 (3.74) 11. Wish to learn 10.56 0.002 0.17 6.82 (3.81) 3.46 (3.61) 18. Relaxation 13.78 0.001 0.22 7.57 (2.77) 4.25 (3.67) 20. Pleasure-focus 7.87 0.007 0.14 7.04 (3.45) 4.17 (3.93) Gratitude 1. Size of self (L) 12.07 0.001 0.20 1.54 (2.44) 4.20 (3.01) 3. Strength 9.88 0.003 0.17 2.81 (3.18) 5.68 (3.35) 11. Wish to learn 5.19 0.027 0.10 5.77 (3.83) 3.36 (3.72) 12. Curiosity 4.59 0.037 0.09 1.19 (2.17) 2.84 (3.24) 16. Other-awareness 12.32 0.001 0.20 2.38 (3.07) 5.52 (3.31) Interest 1. Size of self (S) 7.50 0.009 0.14 1.54 (2.65) 3.70 (2.77) 1. Size of self (L) 7.51 0.009 0.14 2.21 (3.12) 4.55 (2.59) 2. Understanding 8.85 0.005 0.16 1.04 (2.13) 3.20 (2.91) 3. Strength 12.81 0.001 0.22 2.50 (3.35) 5.80 (2.84) 4. Social value 7.29 0.010 0.14 3.29 (3.68) 5.85 (2.50) 6. Incongruity 4.41 0.041 0.09 0.96 (1.77) 2.55 (3.41) 9. Vulnerability 6.35 0.015 0.12 1.18 (1.98) 2.95 (2.89) 16. Other-awareness 20.78 < 0.001 0.31 1.68 (2.92) 5.50 (2.78) 18. Relaxation 6.03 0.018 0.12 8.75 (1.88) 6.95 (3.19) Love 1. Size of self (L) 17.74 < 0.001 0.27 1.52 (2.21) 4.86 (3.33) 3. Strength 14.72 < 0.001 0.24 2.44 (3.17) 5.91 (3.12) 5. Playfulness 9.07 0.004 0.16 2.33 (3.09) 5.18 (3.53) 6. Incongruity 6.79 0.012 0.13 0.48 (1.01) 2.27 (3.40) 8. Benefit – active 7.92 0.007 0.14 7.74 (3.36) 4.82 (3.91) 9. Benefit – passive 9.41 0.004 0.17 2.70 (3.16) 5.55 (3.31) 11. Wish to learn 4.82 0.033 0.09 5.52 (3.65) 3.18 (3.78) 12. Curiosity 6.25 0.016 0.12 1.67 (2.02) 3.45 (2.97) 14. Wish to change 4.18 0.046 0.08 3.59 (3.82) 1.59 (2.82) 16. Other-awareness 7.54 0.009 0.14 2.15 (3.05) 4.59 (3.16) Pride 1. Size of self (L) 9.93 0.003 0.16 4.38 (3.79) 7.08 (2.14) 2. Understanding 4.16 0.047 0.07 0.69 (1.58) 2.00 (3.01) 6. Incongruity 7.46 0.009 0.13 1.14 (1.90) 3.16 (3.42) 11. Wish to learn 11.11 0.002 0.18 6.93 (3.58) 3.80 (3.28) 12. Curiosity 8.47 0.005 0.14 1.48 (2.77) 3.88 (3.28) 18. Relaxation 8.48 0.005 0.14 8.52 (2.29) 5.20 (3.87) 19. Yearning 11.80 0.001 0.19 8.28 (2.67) 5.20 (3.87) Relief 1. Size of self (S) 20.84 < 0.001 0.34 1.38 (2.67) 5.63 (3.30) 8. Benefit – active 4.24 0.046 0.10 5.27 (3.67) 3.00 (3.10) 9. Benefit – passive 4.92 0.032 0.11 2.58 (3.30) 5.06 (3.87) 11. Wish to learn 6.65 0.014 0.14 6.12 (3.33) 3.11 (3.57) 13. Helplessness 5.09 0.030 0.11 1.88 (2.49) 4.06 (3.79) 15. Self-awareness 5.30 0.027 0.12 8.15 (1.69) 6.56 (2.80) 17. Satisfaction 11.56 0.002 0.22 7.54 (1.82) 5.25 (2.54) 18. Relaxation 35.46 < 0.001 0.47 9.04 (1.31) 5.44 (2.61) 19. Yearning 5.97 0.019 0.13 6,54 (3.44) 3.81 (3.62)

20. Pleasure-focus 30.53 < 0.001 0.43 6.23 (3.63) 0.94 (1.53) Embarrass- ment 1. Size of self (S) 20.83 < 0.001 0.30 3.00 (3.37) 7.32 (3.32) 3. Strength 7.11 0.010 0.13 1.03 (1.80) 3.09 (3.61) 11. Wish to learn 6.20 0.016 0.11 4.86 (3.83) 2.32 (3.30) 18. Relaxation 14,94 < 0.001 0.23 6.03 (3.37) 2.59 (2.84) 20. Pleasure-focus 8.22 0.006 0.14 4.28 (4.05) 1.41 (2.70)

General cultural tendencies in Specific appraisals

Some appraisal items tend to show cultural differences across emotions. For example, there were significant differences of “perceived size of self” for 9 out of 10 emotions tested. American participants felt larger than did Chinese participants in awe, compassion, desire, gratitude, love, and pride; they felt smaller than the Chinese participants in relief, and embarrassment; and they felt both larger and smaller than the Chinese participants in interest. It appears that the perception of size of self is an appraisal dimension that is likely to vary across cultures, at least for positive emotion.

American participants scored significantly higher than Chinese participants on “strength” in 7 out of 10 emotions, which are amusement, awe, compassion, gratitude, interest, love, and

embarrassment. This suggests that American participants had a general tendency to feel stronger than did Chinese participants. There were significant cultural differences on “other-awareness” in six emotions, with Chinese participants scoring lower on “other-awareness” for amusement, awe, compassion, gratitude, interest and love. This implies that Chinese participants report being less aware of others when they are experiencing positive emotions.

The item “relaxation” also showed cultural differences in 6 emotions: Chinese participants scored higher in “relaxation” than American participants for amusement, desire, interest, pride, relief and embarrassment. This indicates that Chinese participants generally feel more relaxed when experiencing these (positive) emotions, or at least they report so. Tsai (2002) compared the physiological reactivity to 7 emotions (happiness, pride, love, anger, disgust, and sadness) between Hmong American and European Americans and did not find any cultural difference in skin

conductivity during recalling of the emotions. The seemingly more relaxed experience of Chinese participants in the current study could be real, but it could also reflect differences in ideal affect state.

“Incongruity” showed cultural differences for 5 emotions: Chinese participants scored lower in incongruity for awe, desire, interest, love, pride, meaning that they were less likely to recognize incongruity in the event they recalled. This resonates with results from the general appraisal item “certainty”, where Chinese participants scored higher than Americans for every emotion – if they are more certain about what happened, they would also less likely see incongruity in the situation. Cultural variations in Specific appraisals per emotion

For amusement, US participants felt and more socially valued than did Chinese participants. This could be a general cultural tendency as detailed above rather than an emotion specific appraisal profile. US participants also felt way more aware of other people’s opinion than did Chinese

participants during the recalling. Perhaps US participants were more conscious of their image in front of others in general, especially a stranger. Consistent with this notion, US participants also reported to feel less relaxed during the recalling than did Chinese participants.

For awe, in addition to feeling larger in size and stronger, US participants also felt more aware of other’s opinions. Moreover, US participants felt more satisfied when recalling the awe experience than did the Chinese. And they felt less of a wish to change. These results are consistent with the response to the “pleasantness” item in the general appraisal dimensions, where US

For compassion, next to feeling larger and stronger, US participants also felt more connected with other people in the situation than did the Chinese, which could explain why they were more likely to perceive a problem in the situation – as people who they feel connected with were suffering, something has to be done. While for Chinese participants, they feel bad for seeing someone else suffering (less pleasantness rating of the event), but because they don’t feel very connected with the people suffering, they do not perceive it as their problem to solve.

For desire, US participants again felt larger than did Chinese participants, but they also felt more vulnerable and more connected to others. This could be due to differences in the type of events participants named – it’s possible that US participants recalled more desire events that involved people while Chinese participants’ recalled events were more materialistic desire.

Moreover, US participants felt less relaxed and less pleasure-focused than did Chinese participants while recalling their desire experience. This could again be attributed to differences in the events – US participants could have more often named events involving more serious desires relating to career or study, while Chinese participants recalled events relating to less meaningful desires. A follow up event analysis could be conducted in the future to find these out.

For gratitude, US participants felt more curious but had less wish to learn than did Chinese participants. This finding seems very contradicting. If US participants did feel curious, there must have been some other reason why they did not have a high wish to learn. Going back to the general appraisal items, Chinese participants scored lower on agency-other and higher on agency-self, which is an indicator that Chinese participants felt more like they could or should take or have taken actions themselves in the gratitude situation. Feeling of curiosity does not require agency but to lean something does require agency. Future research could find this out by adding more

questions to this appraisal dimension.

For interest, there were quite some cultural variation – the two cultures differed on eight appraisal dimensions. The first interesting finding is that US participants felt both larger and smaller than did Chinese participants while recalling their interest experience. Because there were two sub-ratings for the appraisal item “size of self”, this is technically possible. However, it’s difficult to explain what this means. More over, US participants again felt stronger, more socially valued, and more aware of other people’s opinion. Perhaps their interest experience was more related to something they could perform and potentially poform well., thus increasing their social value. Alternatively, if they had pursued or been pursuing their interest, they could simply feel socially valued from knowing something more. Future research could add questions about whether participants had already taken action regarding their experience. Furthermore, US participants recognized more incongruity in the situation and had more difficulty understand what as going on. This is another un-explainable finding until further research is done.

For love, there were also considerable culture variation – the two cultures differed on ten appraisal dimensions. First, the US participants felt larger, stronger and more aware of other people’s opinions while recalling their love experience. US participants also recognized more incongruity in the situation and felt more playful. It could be that love is a more light-hearted and pleasant experience for US participants than for Chinese participants. This is reflected in the responses to the appraisal item “pleasantness”, where US participants scored higher than the Chinese (MUS= 8.75, SDUS =1.77 versus MCN = 5.18, SDCN= 4.45).

Another interesting finding is that although US participants felt more that they have

benefited from someone else when they were recalling their love experience, they scored lower on wanting to benefit someone else. There has to be some reason that US participants are less likely to reciprocate, despite their strong feeling of having benefited from another. One possibility is that this stems from the cultural differences in self-construal. People from the US are thought to construct and maintain an independent self-construal, whereas individuals from East Asian societies construct