Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Children’s Reasoning About Empathy and Social Relationships

Children’s Reasoning About Empathy and Social Relationships REPORT Children’s Reasoning About Empathy and Social Relationships Alexis S. Smith-Flores , Gabriel J. Bonamy, and Lindsey J. Powell Department of Psychology, University of California, San Diego Keywords: empathy, relationships, joint reasoning an open access journal ABSTRACT Across the lifespan, empathic and counter-empathic emotions are shaped by social relationships. Here we test the hypothesis that this connection is encoded in children’s intuitive theory of psychology, allowing them to predict when others will feel empathy versus counter-empathy and to use vicarious emotion information to infer relationships. We asked 4- to 7-year-old children (N = 79) to make emotion predictions or relationship inferences in response to stories featuring two characters, an experiencer and an observer, and either a positive or negative outcome for the experiencer. In the context of positive outcomes, we found that children engaged in robust joint reasoning about relationships and vicarious emotions. When given information about the characters’ relationship, children predicted empathy from a friendly observer and counter-empathy from a rival observer. When given information about the observer’s response to the experiencer, children inferred positive and negative relationships from empathic and counter-empathic responses, respectively. In the context of negative outcomes, children predicted that both friendly and rival observers would feel empathy toward the experiencer, but they still used information about empathic versus counter-empathic responses to infer relationship status. Our results suggest that young children Citation: Smith-Flores, A. S., Bonamy, G. J., & Powell, L. J. (2023). Children’s in the US have a blanket expectation of empathic concern in response to negative outcomes, Reasoning About Empathy and Social but otherwise expect and infer that vicarious emotions are connected to social relationships. Relationships. Open Mind: Discoveries in Cognitive Science, 7, 837–854. Future research should investigate if children use this understanding to select social partners, https://doi.org/10.1162/opmi_a_00109 evaluate their own relationships, or decide when to express empathy toward others. DOI: https://doi.org/10.1162/opmi_a_00109 Supplemental Materials: https://doi.org/10.1162/opmi_a_00109 Received: 7 May 2023 INTRODUCTION Accepted: 12 September 2023 Think about the last few times you felt happy. Likely, some of those instances were the result of Competing Interests: The authors declare no conflicts of interest. a friend or family member’s good news. People’s emotions result not only directly from their own personal experiences, but also from witnessing the experiences of others. Emotions felt Corresponding Author: Alexis S. Smith-Flores as a result of someone else’s experiences are known as “vicarious emotions” (Wondra & alexis-smith@ucsd.edu Ellsworth, 2015). Vicarious emotions can vary both in valence and in their congruence with the likely emotions of the person who inspires them. We will use the term “empathy” to refer to Copyright: © 2023 congruent positive or negative emotions for those having good or bad experiences, respec- Massachusetts Institute of Technology tively (Batson et al., 2009). In contrast, counter-empathy involves feeling incongruent emo- Published under a Creative Commons Attribution 4.0 International tions toward others’ experiences, such as sadness in response to others’ good outcomes (CC BY 4.0) license (i.e., gluckschmerz) or happiness in response to others’ bad outcomes (i.e., schadenfreude) (R. H. Smith & van Dijk, 2018). The MIT Press Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. Empathy and counter-empathy tend to be applied selectively: People often feel more empa- thy toward those they care more about and feel counter-empathy toward disliked individuals or outgroups (Batson, 2009; Cikara & Fiske, 2012; Masten et al., 2010; Shamay-Tsoory et al., 2014). As a result, expressions of empathy and counter-empathy may be predicted by, and carry information about, a person’s social relationships. Understanding these connections could allow people to effectively seek social support, and to evaluate their own and others’ relationships based on others’ vicarious emotional expressions. Relatively little research has investigated how people think about the role of relationships in vicarious emotions, however, or how such reasoning develops. Here we consider how common aspects of intuitive psychol- ogy, available in early childhood, may allow children to engage in systematic reasoning about the connections between vicarious emotions and social relationships. We then test their ability to engage in this kind of reasoning. Reasoning about others’ emotions often involves considering their desires. Emotions arise from appraisals of the world relative to what the appraiser wants, needs, and expects. When goals are met and desires are fulfilled, observers expect positive responses, while failures and undesirable outcomes are often accompanied by negative responses (Moors, 2013; Oatley & Johnson-Laird, 2014). This dependence on individuals’ goals and preferences results in appraisals that are person-specific: One person’s good outcome may be someone else’s bad outcome, resulting in differing appraisals of the same situation (Manstead & Fischer, 2001; Siemer et al., 2007). For example, the outcome of a political election often excites some peo- ple and disappoints others, depending on preferences for different candidates or policies. An observer’s successful emotion prediction thus requires integrating the observed state of the world and the specific desires of the target person. An understanding of the appraisal process also allows for inverse reasoning about the causes of emotions. Based on an emotional expres- sion, an observer who knows the emoter’s desires can infer the experience that led to the expression, while an observer who knows what experience led to the expression could infer the emoter’s desires (Ong et al., 2015; Wu & Schulz, 2018). Even young children are able to use emotion vocalizations to infer unobserved causes (Smith-Flores & Feigenson, 2022; Wellman et al., 2000; Wu et al., 2017), suggesting that the ability to reason about emotion as an appraisal process appears early in development. Understanding emotions as the outcome of an appraisal process may also support reason- ing about vicarious emotions and their connection to social relationships (Smith-Flores & Powell, 2023). Concepts of social relationships involve the expectation that relationship part- ners care about one another’s goals and welfare (Afshordi & Liberman, 2021; Jern & Kemp, 2014; Powell, 2022). Friends and group members are generally expected to do more to help one another, and both adult and child observers infer relationships from the effort or partiality of offered support (Earp et al., 2021; Liberman & Shaw, 2017; Marshall et al., 2022; Olson & Spelke, 2008). Even infants expect affiliates to support each other when in need (Jin & Baillargeon, 2017; Pun et al., 2021). Conversely, concepts of negative relationships entail expectations of conflict and spite, reflecting a desire forthe otherpersontosufferorfail (Pietraszewski, 2021). Observers may expect these vicarious desires to serve as the basis for appraisals that lead to vicarious emotions. In addition to predicting vicarious emotions, an observer operating with this intuitive theory should also be able to infer what type of relation- ship led to an expressed vicarious emotion. Upon witnessing a counter-empathic response, indicating opposite appraisals for the experiencer and the emoter, an observer should infer a negative relationship between the two individuals. There is some evidence that both adults and children reason about the connection between vicarious emotions and social motivations, though most work focuses more on inferences of OPEN MIND: Discoveries in Cognitive Science 838 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. general social traits as opposed to specific social relationships. Heyman and Gelman (1998) found that children used a person’s pro- or antisocial motives to infer how pleased they were by positive or negative outcomes for the targets of those motives. Wang and Todd (2021) found that adults evaluate a person who empathizes with a valued target person (e.g., a children’s hospital worker) as being warmer, more likable, and more respectable than a non-empathizer. Three- to five-year-old children similarly find empathy for others’ suffering to be normative, which indicates that young children recognize and reason about others’ empathic concern (Paulus et al., 2020). When directed toward morally repugnant or disliked targets, however, adults may prefer those who are indifferent rather than empathic (Krems et al., 2023; Wang & Todd, 2021). This could reflect a dispreference for those who positively appraise outcomes one considers to be either morally wrong or disadvantageous, as well as a desire for partial- ity in friendships. In other work, adults use emotional expressions (e.g. of anger, sadness, or happiness), either alone or in social contexts, to infer emoters’ aggression, warmth, and affiliative tendencies (Hareli & Hess, 2010; Knutson, 1996). Similar findings extend to peo- ple’s evaluations of their own real-world social partners: physicians who empathize with their patients are viewed as more trustworthy (Kim et al., 2004), and perceived empathy between romantic partners is positively related to relationship satisfaction (Cramer, 2003; Cramer & Jowett, 2010; Gable et al., 2004), though it is unclear whether preferences for empathy in these social relationships reflect dispositional or relationship-based reasoning. Together, these findings suggest that adults and children make dispositional judgments about others using observations of their vicarious emotions. They fall short, however, of providing evidence for a systematic framework connecting vicarious emotions to relationship-based appraisals of outcomes. One set of looking time experiments tested 10- to 11-month-old infants’ expectations about vicarious emotions across different relationship contexts (Smith-Flores, Herrera-Guevara, & Powell, 2023). Infants were introduced to a pair of characters with either a positive or a neg- ative relationship, followed by trials in which one character attempted a goal-directed action and the other character responded happily or sadly to the outcome. In the context of a positive relationship and successful outcome, infants looked significantly less at happy than sad response trials, consistent with an expectation for the observer to emote happily to their friend’s success. In the context of a negative relationship, infants’ looking was not different across happy and sad response trials, suggesting no expectation of positive empathy for a rival. The interaction between relationship context and trial type was significant, supporting the con- clusion that infants around one year of age do take relationships into account in their expec- tations of vicarious emotions. However, there was no evidence that infants expected negative relationships to lead to counter-empathic emotions. This may reflect limited early reasoning about vicarious emotions, focused on predicting empathic responses. Alternatively, infants may struggle more generally with predicting negative emotions in such contexts; they also fail to expect that a friendly observer or even the actor will respond sadly to a failed attempt (Skerry & Spelke, 2014; Smith-Flores et al., 2023). By age 4, children are able to make robust, appraisal-based predictions about both positive and negative emotions (Fabes et al., 1991; Wellman et al., 2000), and appear to possess an emerging norm of empathic concern (Paulus et al., 2020). Four-year-old children are also able to reason about representations of emotions that are at odds with the current state of the world (Smith-Flores & Feigenson, 2021). Investi- gating children’s verbal predictions about vicarious emotions beginning at this age will thus allow us to explore reasoning about a more comprehensive combination of relationships, out- come valence, and vicarious responses. We also test children’s inferences about relationships from information about vicarious emotional responses. OPEN MIND: Discoveries in Cognitive Science 839 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. The current study investigates how 4- to 7-year-old children reason about relationships and expressions of empathy. Children were told four stories in which they were asked to predict how an observer would feel about an outcome for another person (i.e., their friend or rival; Empathy Block), and another four stories in which they were asked to infer the relationship between two characters after being given information about the observer’sempathicor counter-empathic response to the other person’s outcome (Affiliation Block). We preregistered three primary hypotheses on the Open-Science Framework (OSF; https://osf.io/d57ny). First, we hypothesized that children would predict that observers will respond empathically to their positive affiliates (i.e., friends). If this is true, then when children hear stories featuring two friends, they should be more likely to predict that the observing individual will respond to their friend’s outcome with a congruent emotion (i.e., happiness following a good outcome or sadness following a bad outcome) than an incongruent emotion. However, it is possible that children may expect that friends should respond with positive emotion toward their positive affiliates regardless of the outcome. If this is true, then children may predict that the observer will respond with a positive emotion regardless of the valence of their friend’s experience. Second, we hypothesized that children’s predictions of empathy will be weaker for rivals. If this is true, then children should predict less outcome-congruent emotion when making pre- dictions in stories about rivals as compared to friends, regardless of the outcome type. An alter- native hypothesis not anticipated in our preregistration is that children’s possession of strong norms about responding empathically toward others’ suffering may lead them to predict sim- ilar levels of empathy following negative outcomes, regardless of the characters’ relationship (Paulus et al., 2020). Young children tend to use norms rather than other factors, including preferences, to predict others’ behavior, and may even say that impermissible actions are impossible (Kalish & Shiverick, 2004; Shtulman & Phillips, 2018; R. H. Smith et al., 2009). Finally, we hypothesized that children will use vicarious emotions to infer social relation- ships. If so, children should be more likely to infer that two characters have a close positive relationship after hearing that one character responded empathically to the other character. In contrast, children should select responses that correspond to more distant, negative relation- ships after hearing about a character that responded counter-empathically to another charac- ter. Of course, it is also possible that children may use the valence of the emotion response to make inferences about others’ relationships. If this is true, then children may infer that when a character responds negatively to another character, regardless of the outcome of the event, that the two are negatively affiliated, while characters who respond positively are positively affiliated. We also conducted exploratory analyses on age-related changes in reasoning about vicar- ious emotions and relationships. Although we had no a priori hypotheses about developmen- tal change, several patterns were possible. First, there may be no change with age, either because children at all ages are at chance or because children make mature systematic pre- dictions from age 4. Next, there could be development both in terms of children’s differential prediction of vicarious emotion depending on relationships and in their inference of relation- ship valence from vicarious responses. This would indicate that younger children have a weaker understanding of how vicarious emotions are shaped by relationships, reflected in both their predictions of such emotions and inferences from them. Finally, we could observe developmental change in children’s predictions of vicarious emotions but not their inferences from them. This hypothesis is suggested by past results showing that young children often pre- dict that people will behave according to rules and norms, and sometimes even say that it is impossible to do otherwise (DeJesus et al., 2014; Kalish & Shiverick, 2004; Shtulman & OPEN MIND: Discoveries in Cognitive Science 840 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. Phillips, 2018; C. E. Smith et al., 2013). This tendency to predict strong adherence to norms diminishes with age. Because children view empathy for others’ pain as normative (Paulus et al., 2020), younger children may be less likely to predict counter-empathy, despite being able to reason about the meaning of counter-empathic responses for underlying social relationships. METHOD Participants Seventy-nine children, aged 4- to 7-years-old (M = 5.93 years; range = 3.98 years to 7.88 age years; 42 girls), participated at a science museum in Southern California. One additional child participated but was excluded for being outside the preregistered age range. Caregivers completed an optional demographic form at the time of participation. Fifty-one children were identified by their caregivers as White, seven as Asian, 1 as American Indian/Alaskan Native, 18 were identified as belonging to two or more races, and two caregivers declined to answer. Fifteen children were identified as Hispanic/Latinx and five caregivers declined to answer. Seventy-one children came from families where at least one caregiver had a college degree or higher, five had at least one caregiver who completed some college, one had at least one caregiver who had a high school diploma or GED, and two care- givers declined to answer. Our data collection plan was preregistered as follows: We ran data collection sessions at a local science museum until we had at least 72 participants with valid data for both blocks, while also planning to include data from participants with only one valid block and from par- ticipants collected in the final testing session required to reach the target sample size. This resulted in unequal numbers of participants included in the two blocks. Seventy-seven partic- ipants completed the Empathy Block and 74 participants completed the Affiliation Block. Participants received a small gift (i.e., temporary tattoos or stickers) to thank them for their participation. Materials & Procedures All materials including preregistrations, data, and analysis code are available on OSF (https:// osf.io/d57ny). Participants were read 4 stories per block across 2 blocks, for a total of 8 stories. Each of the four stories featured two characters, an actor and an observer. Stories in the Empa- thy Block presented participants with information about the relationship between two charac- ters and about a positive or negative outcome for one of the two characters (the actor), before asking participants to predict the emotion felt by the other character (the observer) following that outcome. Stories in the Affiliation Block presented participants with information about a positive or negative outcome for one character (the actor) and the emotion felt by the other character (the observer) following that outcome, before asking participants to infer the degree of affiliation (i.e., the closeness of the relationship) between the two characters. Each block also included an initial warm-up trial that familiarized participants with the relevant response scale (a positive to negative affect scale in the Empathy Block; a positive to negative relation- ship scale in the Affiliation Block) and tested their ability to use it in responding to questions posed by the experimenter. Participants were randomly assigned to 1 of 8 story orders that counterbalanced block order (Empathy Block first vs. second), outcome order (positive outcome first vs. second), and which affiliation or emotion type came first in each block (friendship/empathy vs. rivalry/counter-empathy). OPEN MIND: Discoveries in Cognitive Science 841 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. Empathy Block The researcher began by introducing a 4-point Likert scale, ranging from “very happy” to “very sad”, accompanied by simple, line-drawn faces depicting each emotion (Figure 1B). The researcher asked participants about their favorite and least favorite snack. Participants then rated, verbally or by pointing to the scale, how they would feel if they received a lot of their favorite or least favorite snack. As preregistered, participants who responded with the same emotion regarding their favorite and least favorite snack were excludedfromthisblock’s analysis (see Supplemental Information for number of children excluded from each block, Table S1). Participants were first introduced to the two characters by name and were told about their relationship (e.g., “Ryan and Chris are best friends. They like each other a lot.”). To ensure participants remembered the relationship between the two characters, the researcher then asked a manipulation check question (e.g., “Remind me, are Ryan and Chris friends?”). Par- ticipants who did not answer the manipulation check correctly were told the story again and had the manipulation check repeated (Table S2). Participants who did not answer the manip- ulation check correctly a second time were excluded from that question’s analysis (see Supplemental Information for number of children who failed the manipulation check once or twice for each story, Table S2). Following the manipulation check question, participants were then told about an event that happened to the actor while the observer watched (e.g., “At recess, Ryan races against another classmate and loses!”). The researcher then asked the test question regarding how the observer Figure 1. Stimuli from the Empathy Block. (A) Children heard 4 stories in total where the actor’s outcome (positive, negative) and characters’ affiliation (friends, rivals) varied across stories. (B) Children were asked to predict how the observer felt about the actor’s outcome using a 4- point scale ranging from “very happy” to “very sad”. OPEN MIND: Discoveries in Cognitive Science 842 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. felt about the actor’s outcome (e.g., “How does Chris feel about Ryan losing the race?”). The experimenter pointed to each item on the emotion scale, reading the options aloud to the par- ticipant. Participants’ verbal or pointed answers were recorded by a second researcher who observed from across the table. A 2 × 2 design was used so that each child heard one story each about a positive outcome for a friend, a negative outcome for a friend, a positive out- come for a rival, and a negative outcome for a rival. Affiliation Block The researcher started by introducing a 4-point Likert scale, used to ask how much people like one another and ranging from “a lot” to “not at all”, accompanied by pictures of two people close together or increasingly far apart (Figure 2B). To familiarize children with this affiliation scale, we asked participants about their favorite and least favorite characters. Participants then rated, verbally or by pointing to the scale, how much they like their favorite or least favorite character. As preregistered, participants who answered the same regarding their favorite and least favorite characters were excluded from this block’sanalysis(see Supplemental Information for number of children excluded from each block, Table S1). Participants were first introduced to the two characters by name and then were told about an event that happened to the actor while the observer watched (e.g., “At lunchtime, Kyle has a cookie in his lunch, but he drops it on the floor and has to throw it away!”; see Figure 2A). Participants were then told how the observer felt about what had happened to the actor (e.g., “Brent is very sad Kyle can’t eat his cookie”). The observer either felt very happy or very sad about the event that happened to the actor. Figure 2. Stimuli from the Affiliation Block. (A) Children heard 4 stories in total where the actor’s outcome (positive, negative) and the observer’s emotion (empathic, counter-empathic) changed across stories. (B) Children were asked to infer how much the observer and the actor liked each other using a 4-point scale ranging from “alot” to “not at all”. OPEN MIND: Discoveries in Cognitive Science 843 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. To ensure participants remembered the emotion felt by the observer, the researcher then asked a manipulation check question (e.g., “Remind me, how does Brent feel?”). Participants who did not answer the manipulation check correctly were told the story again and had the manipulation check repeated (Table S2). Participants who did not answer the manipulation check correctly a second time were excluded from that question’sanalysis (see Supplemental Information for number of children who failed the manipulation check once or twice for each story, Table S2). Finally, the researcher asked the test question regarding the affiliation between the two characters (e.g., “How much do Brent and Kyle like each other?”). The experimenter pointed to each item on the affiliation scale, reading the options aloud to the participant. Participants’ verbal or pointed answers were recorded by a second researcher who observed from across the table. A 2 × 2 design was used so that each participant heard one story each about a pos- itive outcome for the actor with a positive reaction from the observer, a negative outcome with a negative reaction, a positive outcome with a negative reaction, and a negative outcome with a positive reaction. Coding. All responses were coded by the second researcher at the time of the experiment. An additional coder coded 20% of participants from video recordings to ensure reliability. Agree- ment between coders was 100%. For analysis of data from the Empathy Block, participants’ raw emotion predictions were transformed to reflect the degree of expected empathy in the emotion prediction. Responses that indicated the greatest amount of predicted empathy (i.e., “very sad” to a negative outcome and “very happy” to a positive outcome) were coded as a 4, while responses that indicated the greatest predicted counter-empathy (i.e., “very happy” to a negative outcome and “very sad” to a positive outcome) were coded as a 1. Moderately empathic or counter-empathic responses were coded as a 3 or 2, respectively. In the Affiliation Block, responses that indicated the strongest positive relationship infer- ences (i.e., the characters like one another “a lot”) were coded as a 4, while responses that indicated the strongest negative relationship inferences (i.e., the characters like one another “not at all”) received a 1. Moderately positive or negative responses were coded as a 3 or 2, respectively. RESULTS Empathy Block First, to test our preregistered hypothesis that children predict that friends will respond empath- ically toward each other’s outcomes, we conducted separate, two-tailed one-sample t-tests from chance (2.5 on a 1–4 point scale) on children’s responses in the Empathy Block for each of the two stories that featured “best friends”. Regardless of the type of outcome (i.e., positive or negative), children predicted that the observing friend would respond empathically (positive outcome: M =3.662, SEM =0.092, t(76) = 12.674, p < .001, 95% CI [3.480, 3.845], d = 1.444.; negative outcome: M =3.342, SEM = 0.104, t(75) = 8.133, p < .001, [3.136, 3.548], d = 0.933; see Figure 3). To test the hypothesis that children have weaker expectations of empathy for rivals than friends, we fit a linear regression model of children’s responses in the Empathy Block. Affili- ation type (friends vs. rivals), outcome type (positive vs. negative), and the interaction between affiliation type and outcome type were included in the model as fixed effects. Affiliation type OPEN MIND: Discoveries in Cognitive Science 844 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. Figure 3. Children’s responses for each question in the Empathy Block. The means and standard errors are plotted in white. The dashed line represents the mid-point of the scale (2.5). Significance asterisks come from post-hoc t-tests with bonferroni adjustment as a result of a signif- icant interaction between affiliation type and outcome type. ***p < .001. and outcome type were dummy-coded and mean-centered. We preregistered a mixed-effect model that included participant as a random effect, however this model returned a singular fit convergence issue due to the participant-level random effect variance being close to zero. So, we removed the random effect from the full Empathy Block model. The resulting full model was as follows: response ∼ affiliation.type + outcome.type + affiliation.type * outcome.type Using nested model comparisons, we then compared this full model to models that excluded each factor individually to test how much the excluded factor contributed to explain- ing the variance in children’s predictions of empathy. We found a main effect of affiliation type, F(1, 301) = 107.46, p < .001 – children were more likely to predict empathy for friends (M = 3.503, SEM = 0.070) than for rivals (M = 2.349, SEM = 0.098). We also found a main effect of outcome type, F(1, 301) = 13.338, p < .001 – children were more likely to predict empathy following negative outcomes (M = 3.132, SEM = 0.087) than for positive outcomes (M = 2.727, SEM = 0.103). There was also an interaction between affiliation type and outcome type, F(1, 301) = 42.883, p < .001. Using post-hoc paired-sample t-tests with Bonferroni adjustments (α = .0125, .05 divided by 4 tests), we found that children’s predictions of empa- thy following negative outcomes did not differ significantly for friends versus rivals (M = 2.920, SEM = 0.138), once corrected for multiple comparisons, t(73) = 2.156, p = .034, 95% CI [0.032, 0.806], d = 0.251, but that children predicted significantly more empathy from friends (M = 3.662, SEM = 0.092) than rivals (M = 1.792, SEM = 0.105) following positive outcomes, t(76) = 12.236, p < .001, [1.566, 2.175], d = 1.394. Within affiliation type, children’s predic- tions of empathy for a friend were not significantly different across positive and negative out- come types, once corrected for multiple comparisons, t(73) = −2.269, p = .026, [−0.593, −0.039], d = −0.260, but children predicted significantly more empathy for a rival who expe- rienced a negative vs. positive outcome, t(74) = 6.770, p < .001, [0.790, 1.450], d = 0.781. Our data support the conclusion that children selectively predict empathic happiness for friends but not rivals following positive outcomes but expect empathic concern for both friends and rivals following negative outcomes. OPEN MIND: Discoveries in Cognitive Science 845 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. To investigate whether this pattern of reasoning changed across the age range in our sam- ple, we conducted an exploratory analysis of all data from the Empathy Block, adding a con- tinuous age variable and the associated two- and three-way interactions as fixed effects to the full model described above. A nested model comparison found a significant three-way inter- action between age, affiliation type, and outcome type, F(4, 297) = 4.438, p = .002 (Figure 4). Post-hoc pairwise comparisons with Bonferroni adjustments (α = .0125, .05 divided by 4 tests) showed that the only significant difference in slopes was between predictions of empathy for friends and rivals following negative outcomes, estimate = 0.440, t(297) = 3.389, p < .001, diff 95% CI [0.184, 0.695] – as children got older they were more likely to predict empathy for friends and counter-empathy for rivals after negative events. All other ts < 1.970 and ps > .050. We also preregistered an exploratory analysis of only the older half of the sample, children 6 or 7 years of age, to assess if joint reasoning about vicarious emotions and social relation- ships is more robust in older children. Using the same set of nested linear regression compar- isons as in the main analysis (i.e. without the continuous age factor), we again found a main effect of affiliation type on participants’ empathy predictions, F(1, 151) = 95.017, p < .001, and an interaction between affiliation type and outcome type such that empathy predictions varied more by affiliation type following positive vs. negative outcomes, F(1, 151) = 15.185, p < .001. However, post-hoc t-tests showed that in this sample subset, older children did predict more empathy following a negative outcome from a friend (M = 3.410, SEM = 0.120) than from a rival (M = 2.553, SEM = 0.202), t(37) = 3.141, p = .003, 95% CI [0.299, 1.385], d = 0.510. We followed up on this result by comparing older and younger children’s empathy predictions for each of the four story types and found that the only reliable difference was that older children predicted less empathy for rivals’ negative outcomes compared to younger children, t(73) = −2.82, p = .006, [−1.270, −0.219], d = 0.652. Younger children’s predictions of empathy for a rival’s negative outcome were greater than chance (2.5 on our scale), M = 3.297, SEM = 0.168, t(36) = 4.737, p < .001, [2.956, 3.639], d = 0.779, while older children’s predictions of empathy in the same situation were not significantly different from chance, M = 2.553, Figure 4. Children’s responses in the Empathy Block by age in years. OPEN MIND: Discoveries in Cognitive Science 846 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. SEM = 0.202, t(37) = 0.260, p = .796, [2.143, 2.962], d = 0.042. Thus children’s expectation of empathic concern for rivals appears to decrease with development. Affiliation Block To test if children use observations of empathic and counter-empathic responses to infer rela- tionships, we fit the preregistered mixed-effect model of participants’ responses in the Affilia- tion Block. Emotion type (empathic response vs. counter-empathic response), outcome type (positive vs. negative), and the interaction between emotion type and outcome type were included in the model as fixed effects. Participant was included as a random effect. Emotion type and outcome type were dummy-coded and mean-centered. The full model was as follows: response ∼ emotion.type + outcome.type + emotion.type*outcome.type + (1 | participant) We used nested model comparison to compare the full model to models that excluded each fixed effect factor individually to test how much the excluded factor contributed to explaining the variance in children’s relationship inferences. We found a main effect of emotion type, χ (1) = 219.09, p < .001 – children inferred a positive relationship more following an empathic response (M = 3.616, SEM = 0.061) than a counter-empathic response (M = 1.832, SEM = 0.084). We also found a main effect of outcome type, χ (1) = 11.136, p = .001 – children were more likely to infer a positive relationship following positive outcomes (M = 2.878, SEM = 0.098) than for negative outcomes (M = 2.582, SEM = 0.109). There was no interaction between emotion type and outcome type, χ (1) = 0.034, p = .853. In line with our predictions, children were more likely to infer a positive relationship between two characters when they observed one character respond empathically, rather than counter-empathically, to the other’s outcome, regardless of whether that outcome was a good or bad one (Figure 5). Figure 5. Children’s responses for each question in the Affiliation Block. The means and standard errors are plotted in white. The dashed line represents the mid-point of the scale (2.5). Significance asterisks come from nested model comparisons of mixed linear models testing for a main effect of emotion type. ***p < .001. OPEN MIND: Discoveries in Cognitive Science 847 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. As in the empathy block, we ran an exploratory analysis with a continuous age variable and related interactions as predictors for the Affiliation Block. There was a significant three-way interaction between empathic type, outcome type, and age, χ (4) = 21.850, p < .001 (Figure 6). Post-hoc pairwise comparisons with Bonferroni adjustments (α = .0125, .05 divided by 4 tests) showed that the only significant difference in slopes was between relationship inferences from empathy and counter-empathy following negative outcomes, estimate = 0.522, t(215) = diff 4.648, p < .001, 95% CI [0.301, 0.743]. As children got older they were more likely to infer positive relationships from empathy and negative relationships from counter-empathy after negative events. All other ts < 2.130 and ps > .034. By including the full set of age ranges, we found that children’s inferences about relationships from vicarious emotions following negative outcomes become stronger with age. However, there are several reasons to see the developmental change in this case as incre- mental, rather than qualitative. First, the preregistered analysis of the older half of the age group revealed the same pattern of effects as in the full dataset: In the older children’s responses to the Affiliation Block, we again found a main effect of empathic type, χ (1) = 190.51, p < .001 – children inferred a positive relationship more following an empathic response (M =3.753, SEM = 0.059) than a counter-empathic response (M = 1.662, SEM = 0.097). We also found a main effect of outcome type, χ (1) = 8.407, p = .004 – older children were still more likely to infer a positive relationship following positive outcomes (M = 2.859, SEM = 0.137) than for negative outcomes (M = 2.553, SEM = 0.149), as in the full dataset. There was no interaction between empathic type and outcome type, χ (1) = 2.894, p = .089. Second, an analysis of only the younger half of the sample also found the same pattern: a main effect of empathic type, χ (1) = 62.447, p < .001 – younger children inferred a positive relationship more following an empathic response (M = 3.464, SEM = 0.108) than a counter- empathic response (M = 2.030, SEM = 0.140); a main effect of outcome type, χ (1) = 4.189, p = .041 – children were still more likely to infer a positive relationship following positive outcomes (M =2.900, SEM = 0.141) than for negative outcomes (M =2.615, SEM = 0.162), and no interaction between empathic type and outcome type, χ (1) = 1.238, p = Figure 6. Children’s responses in the Affiliation Block by age in years. OPEN MIND: Discoveries in Cognitive Science 848 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. .266. In particular, 4- and 5-year-old children were more likely than chance to infer positive relationships from empathic responses to both positive outcomes, t(34) = 11.554, p < .001, 95% CI [3.501, 3.928], d = 1.953, and negative outcomes t(33) = 3.858, p < .001, [2.834, 3.578], d = 0.661, and more likely than chance to infer negative relationships from counter-empathic responses to both positive outcomes t(34) = −2.358, p = .023, [1.729, 32.443], d = −0.399, and negative outcomes t(30) = −2.367, p = .025, [1.509, 2.427], d = −0.425. Exploratory Preregistered Order Analyses. Across participants, outcome type order and affiliation type order (Empathy Block) or empathy type order (Affiliation Block) were counterbalanced. To examine any effect this may have had on children’s responses, we again used a nested model comparison to determine whether the addition of fixed order effects improved the model’s fit. There were no order effects observed in either the Empathy Block, F(2, 299) = 0.847, p = .430, or the Affiliation Block, χ (2) = 0.991, p = .609. Additional preregistered exploratory analyses can be found in Supplemental Information. DISCUSSION These data support the hypothesis that children can engage in systematic joint reasoning about social relationships and vicarious emotions (Smith-Flores & Powell, 2023). This includes both inferring social relationships from the information provided by empathic and counter- empathic responses (Affiliation Block) and making relationship-based predictions about vicarious emotions following positive outcomes (Empathy Block). The one context in which children did not display integrated reasoning about relationships and vicarious emotions across the full age range tested was when predicting one person’s response following someone else’s negative experience. In this context, children in the full sample expected others to express similar levels of empathy regardless of their relationships, though there was explor- atory evidence that this varied between younger and older participant age groups. This failure may reflect a lack of understanding about the role negative relationships play in schaden- freude. However, given children’s predictions of counter-empathy following rivals’ good outcomes and their robust tendency to infer negative relationships from counter-empathy regardless of outcome valence, we suspect this result instead reflects young children’s general tendency to predict that people will behave normatively, rather than counter-normatively (DeJesus et al., 2014; Kalish & Shiverick, 2004; Powell & Smith, 2013; Shtulman & Phillips, 2018). Both adults and children consider most expressions of schadenfreude to be inappropri- ate, even toward disliked individuals (e.g., Geraci et al., 2021; Paulus et al., 2020; Wang & Todd, 2021). Thus our participants may have predicted an observer would respond empath- ically to a rival’s suffering because this is what they believe the observer should do. Children’s responses in the Affiliation Block also changed with age, becoming less likely to fall into the “moderate” categories of relationships when the emotions were a result of negative outcomes, but from the youngest ages children still used empathy and counter-empathy to make opposite relationship inferences. We conclude that, in addition to supporting our three primary hypoth- eses, the data are most consistent with the third possible developmental pattern laid out in the introduction: from the earliest ages tested, children already provided evidence of understand- ing the link between relationships and vicarious emotions, but the expression of this knowl- edge interacts with other aspects of cognitive development, including changes in modal reasoning. Children’s ability to both predict vicarious emotion from relationships and infer relation- ships from vicarious emotion is consistent with the possession of an intuitive theory of OPEN MIND: Discoveries in Cognitive Science 849 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. psychology that connects relationship concepts to the causes of emotion. This proposal builds on work demonstrating that adults and children reason about emotions as the product of appraisals (Ong et al., 2019; Saxe & Houlihan, 2017; Wu et al., 2018), as well as the theory that early relationship concepts involve positing that relationship partners find the well-being of their affiliates rewarding (Hamlin et al., 2013; Powell, 2022). By connecting relationships to emotions through the reward that one person feels when good things happen to their friends (or bad things happen to their foes), or the sadness they feel in the opposite circumstances, this “adopted utility calculus” framework can support both forward and reverse inferences about how social relationships affect vicarious emotions (Smith-Flores & Powell, 2023). The unex- pected finding that children inferred somewhat less positive relationships following negative versus positive outcomes is also consistent with probabilistic reasoning over an intuitive the- ory: Because children expect empathy to be normative and thus common following negative outcomes, (1) this empathy is potentially less convincing evidence of a close relationship than positive empathy and (2) counter-empathy for the negative outcome (i.e., schadenfreude) is more surprising and thus better evidence of a highly negative relationship than counter- empathy for the positive outcome. However, this is not the only possible basis for the pattern of reasoning observed here. Future research should investigate the integration of vicarious emotion reasoning with other components of intuitive psychology, including reasoning about ignorance or false belief, to test this account. Relative to recent investigations of reasoning about vicarious emotions in preverbal infants, the current findings suggest both continuity and development. In a series of experiments, dif- ferences in 10- to 11-month-old infants’ looking time indicated that they were surprised if a friendly observer responded to an actor’s success with sadness, relative to a happy response. In contrast, when events depicted a rival observer watching the actor succeed, infants’ looking provided no evidence for an expectation of either a happy or a sad vicarious response (Smith- Flores et al., 2023). Thus, like the children in the current study, infants use social relationship information to inform their expectations about vicarious emotions after positive outcomes. However, infants did not use negative relationship information to predict counter-empathy from the observer following the actor’s success. In contrast, children in the current study, across all ages, predicted counter-empathy from rivals following positive events. Future research should more fully investigate the development of reasoning about counter-empathy. Infants’ lack of expectation may reflect a broader difficulty with reasoning about negative emo- tions, which extends to failures to predict negative emotional responses to direct experiences (Ruba et al., 2019, 2020; Skerry & Spelke, 2014). Alternatively, an understanding of empathic vicarious emotions may precede that of counter-empathic vicarious emotions, perhaps due to the difficulty of representing one person’s desire for the other to not achieve their goals (Feiman et al., 2015). These findings provide insight into children’s reasoning about others’ social interactions and highlight children’s sensitivity to others’ displays of vicarious emotions. Children’s own empathic behaviors have been shown to play an important role in their ability to form positive social relationships (Brown & Fredrickson, 2021;Denhametal., 2003). Several theories regarding the development of emotional competence posit that empathy and emotion reason- ing must develop together in order for children to achieve optimal social functioning (Hare et al., 2022; Hare & Parent, 2022; Taylor et al., 2017). Children with higher levels of emotion reasoning (i.e., the ability to correctly identify others’ emotions, also referred to as “cognitive empathy”; see Ruba & Pollak, 2020), have more positive social interactions and engage in more prosocial behaviors (Ensor et al., 2011). In clinical populations, training emotion recog- nition enhances empathic feelings and behavior (Teding van Berkhout & Malouff, 2016). OPEN MIND: Discoveries in Cognitive Science 850 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. Children’s developing understanding that vicarious emotions convey information about social relationships, displayed in the current findings, may also play a role in children’s own effective use of expressions of empathy. Our experiment finds that, at the group level, children readily predict empathy for friends’ successes and suffering, but it does not sensitively capture individual differences in children’s reasoning. Future research should explore connections to children’s own empathic tendencies and additional social–emotional competencies. Finally, empathic and counter-empathic responses are rich sources of information about others’ prosociality, such as their propensity to help or provide comfort to those in distress (Batson, 2013; Bruneau et al., 2017; Eisenberg et al., 1993; Morelli et al., 2015). In this exper- iment, we tested whether children could infer relationships from observations of empathic responses. Children inferred that empathy signaled positive relationships rather than negative ones. In other work, children have been found to prefer and reward characters who comforted another character, suggesting that children may evaluate individuals who display empathic behaviors more favorably than non-empathic individuals (Geraci et al., 2021). However, it remains unknown whether children’s positive evaluations carry over into their selection of social partners for themselves. In one case, children may select empathizers if they need someone’s help, someone to play with, or someone to learn from when no other information is available. However, counter-empathy may also be a sign of a potential social partner if the counter-empathy is directed toward someone the child does not like. Recent work found that adults do prefer those who are kind and trustworthy over those who are not, but also that adults want friends who will be more positive to them than others and even be less prosocial to their enemies (Krems et al., 2023). Future work may explore how children evaluate empathizers and counter-empathizers as potential social partners and the nuances of such evaluations. In sum, this experiment supports an early-developing social cognitive framework for joint reasoning about others’ relationships and vicarious emotions (Smith-Flores & Powell, 2023). This work gives rise to a number of questions about how children’s reasoning about other people carries over to their own social competencies and decision-making. AUTHOR CONTRIBUTIONS Alexis Smith Flores: Conceptualization, Methodology, Formal analysis, Visualization, Writing— Original draft, Writing—Review & editing. Gabriel Bonamy: Investigation, Methodology, Writing—Original draft. Lindsey Powell: Conceptualization, Funding acquisition, Methodology, Formal analysis, Supervision, Writing—Review & editing. ACKNOWLEDGMENTS We would like to thank the families who participated in these studies and Naomi Batarse, Ana Correa Avila Robb, and Marissa Garcia for assistance with data collection and coding. FUNDING STATEMENT This work was supported by a grant from the Sanford Institute for Empathy and Compassion and start-up funding provided by UC San Diego to Lindsey Powell, and a Predoctoral Ford Fellowship awarded to Alexis Smith-Flores. DATA AVAILABILITY STATEMENT This study was preregistered before data collection began. The research plan (including initial hypotheses and the analysis plan), materials and stimuli, data, and code used in the analysis are available on OSF: https://osf.io/d57ny. OPEN MIND: Discoveries in Cognitive Science 851 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. REFERENCES Afshordi, N., & Liberman, Z. (2021). Keeping friends in mind: Devel- reactions. Developmental Psychology, 27(5), 858–866. https:// opment of friendship concepts in early childhood. Social Devel- doi.org/10.1037/0012-1649.27.5.858 opment, 30(2), 331–342. https://doi.org/10.1111/sode.12493 Feiman, R., Carey, S., & Cushman, F. (2015). Infants’ representa- Batson, C. D. (2009). These things called empathy: Eight related but tions of others’ goals: Representing approach over avoidance. distinct phenomena. In The social neuroscience of empathy Cognition, 136,204–214. https://doi.org/10.1016/j.cognition (pp. 3–15). MIT Press. .2014.10.007, PubMed: 25498746 Batson,C.D.(2013). Of Empathy-Induced Altruism. Social Psy- Gable, S. L., Reis, H. T., Impett, E. A., & Asher, E. R. (2004). What chology and Economics, 281. do you do when things go right? The intrapersonal and interper- Batson, C. D., Ahmad, N., & Lishner, D. A. (2009). Empathy and sonal benefits of sharing positive events. Journal of Personality altruism. In Oxford handbook of positive psychology (2nd ed., and Social Psychology, 87(2), 228–245. https://doi.org/10.1037 pp. 417–426). Oxford University Press. https://doi.org/10.1093 /0022-3514.87.2.228, PubMed: 15301629 /oxfordhb/9780195187243.013.0039 Geraci, A., Rigo, P., Simonelli, A., Di Nuovo, S., & Simion, F. Brown, C. L., & Fredrickson, B. L. (2021). Characteristics and con- (2021). Preschoolers’ evaluations of comforting actions towards sequences of co-experienced positive affect: Understanding the third parties in different relationship contexts. Journal of Applied origins of social skills, social bonds, and caring, healthy commu- Developmental Psychology, 76, 101315. https://doi.org/10.1016 nities. Current Opinion in Behavioral Sciences, 39,58–63. /j.appdev.2021.101315 https://doi.org/10.1016/j.cobeha.2021.02.002 Hamlin, J. K., Ullman, T., Tenenbaum, J., Goodman, N., & Baker, C. Bruneau, E. G., Cikara, M., & Saxe, R. (2017). Parochial empathy (2013). The mentalistic basis of core social cognition: Experi- predicts reduced altruism and the endorsement of passive harm. ments in preverbal infants and a computational model. Develop- Social Psychological and Personality Science, 8(8), 934–942. mental Science, 16(2), 209–226. https://doi.org/10.1111/desc https://doi.org/10.1177/1948550617693064 , PubMed: .12017, PubMed: 23432831 29276575 Hare, M., & Parent, J. (2022). Child emotional competence: A uni- Cikara, M., & Fiske, S. T. (2012). Stereotypes and Schadenfreude: fied framework and assessment review of emotion reasoning, Affective and physiological markers of pleasure at outgroup mis- emotion stability, emotion regulation, and empathy. PsyArXiv. fortunes. Social Psychological and Personality Science, 3(1), https://doi.org/10.31234/osf.io/582wg 63–71. https://doi.org/10.1177/1948550611409245, PubMed: Hare, M., Parent, J., & Trucco, E. (2022). The Emotional Compe- 24416471 tence in Children (ECIC) Scale: A parent-report measure. PsyArXiv. Cramer, D. (2003). Facilitativeness, conflict, demand for approval, https://doi.org/10.31234/osf.io/snp5m self-esteem, and satisfaction with romantic relationships. The Hareli, S., & Hess, U. (2010). What emotional reactions can tell us Journal of Psychology, 137(1), 85–98. https://doi.org/10.1080 about the nature of others: An appraisal perspective on person /00223980309600601, PubMed: 12661706 perception. Cognition and Emotion, 24(1), 128–140. https://doi Cramer, D., & Jowett, S. (2010). Perceived empathy, accurate .org/10.1080/02699930802613828 empathy and relationship satisfaction in heterosexual couples. Heyman, G. D., & Gelman, S. A. (1998). Young children use motive Journal of Social and Personal Relationships, 27(3), 327–349. information to make trait inferences. Developmental Psychology, https://doi.org/10.1177/0265407509348384 34(2), 310–321. https://doi.org/10.1037/0012-1649.34.2.310, DeJesus, J. M., Rhodes, M., & Kinzler, K. D. (2014). Evaluations PubMed: 9541783 versus expectations: Children’s divergent beliefs about resource Jern, A., & Kemp, C. (2014). Reasoning about social choices and distribution. Cognitive Science, 38(1), 178–193. https://doi.org social relationships. Proceedings of the Annual Meeting of the /10.1111/cogs.12093, PubMed: 24117730 Cognitive Science Society, 36(36), 689–694. https://escholarship Denham, S. A., Blair, K. A., DeMulder, E., Levitas, J., Sawyer, K., .org/content/qt7752k71z/qt7752k71z.pdf Auerbach-Major, S., & Queenan, P. (2003). Preschool emotional Jin, K.-S., & Baillargeon, R. (2017). Infants possess an abstract competence: Pathway to social competence? Child Develop- expectation of ingroup support. Proceedings of the National ment, 74(1), 238–256. https://doi.org/10.1111/1467-8624 Academy of Sciences, U.S.A., 114(31), 8199–8204. https://doi .00533, PubMed: 12625448 .org/10.1073/pnas.1706286114, PubMed: 28716902 Earp, B. D., McLoughlin, K. L., Monrad, J. T., Clark, M. S., & Crockett, Kalish, C. W., & Shiverick, S. M. (2004). Children’s reasoning about M. J. (2021). How social relationships shape moral wrongness norms and traits as motives for behavior. Cognitive Develop- judgments. Nature Communications, 12(1), 5776. https://doi.org ment, 19(3), 401–416. https://doi.org/10.1016/j.cogdev.2004 /10.1038/s41467-021-26067-4, PubMed: 34599174 .05.004 Eisenberg,N., Fabes, R. A.,Carlo,G., Speer, A. L.,Switzer,G., Kim, S. S., Kaplowitz, S., & Johnston, M. V. (2004). The effects of Karbon, M., & Troyer, D. (1993). The relations of empathy- physician empathy on patient satisfaction and compliance. Eval- related emotions and maternal practices to children s comforting uation & the Health Professions, 27(3), 237–251. https://doi.org behavior. Journal of Experimental Child Psychology, 55(2), /10.1177/0163278704267037, PubMed: 15312283 131–150. https://doi.org/10.1006/jecp.1993.1007, PubMed: Knutson, B. (1996). Facial expressions of emotion influence inter- 8501423 personal trait inferences. Journal of Nonverbal Behavior, 20, Ensor, R., Spencer, D., & Hughes, C. (2011). ‘You feel sad?’ Emotion 165–182. https://doi.org/10.1007/BF02281954 understanding mediates effects of verbal ability and mother– Krems, J. A., Hahnel-Peeters, R. K., Merrie, L. A., Williams, K. E. G., child mutuality on prosocial behaviors: Findings from 2 years & Sznycer, D. (2023). Sometimes we want vicious friends: to 4 years. Social Development, 20(1), 93–110. https://doi.org People have nuanced preferences for how they want their friends /10.1111/j.1467-9507.2009.00572.x to behave toward them versus others. Evolution and Human Fabes, R. A., Eisenberg, N., Nyman, M., & Michealieu, Q. (1991). Behavior, 44(2), 88–98. https://doi.org/10.1016/j.evolhumbehav Young children’s appraisals of others’ spontaneous emotional .2023.02.008 OPEN MIND: Discoveries in Cognitive Science 852 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. Liberman, Z., & Shaw, A. (2017). Children use partial resource Ruba, A. L., Meltzoff, A. N., & Repacholi, B. M. (2019). How do sharing as a cue to friendship. Journal of Experimental Child Psy- you feel? Preverbal infants match negative emotions to events. chology, 159,96–109. https://doi.org/10.1016/j.jecp.2017.02 Developmental Psychology, 55(6), 1138–1149. https://doi.org .002, PubMed: 28285046 /10.1037/dev0000711, PubMed: 30829508 Manstead, A. S. R., & Fischer, A. H. (2001). Social appraisal: The Ruba, A. L., Meltzoff, A. N., & Repacholi, B. M. (2020). The social world as object of and influence on appraisal processes. In development of negative event-emotion matching in infancy: K. R. Scherer, A. Schorr, & T. Johnstone (Eds.), Appraisal pro- Implications for theories in affective science. Affective Science, cesses in emotion: Theory, methods, research (pp. 221–232). 1(1), 4–19. https://doi.org/10.1007/s42761-020-00005-x, Oxford: Oxford University Press. https://orca.cardiff.ac.uk PubMed: 36042945 /34660/ Ruba, A. L., & Pollak, S. D. (2020). The development of emotion Marshall, J., Gollwitzer, A., Mermin-Bunnell, K., Shinomiya, M., reasoning in infancy and early childhood. Annual Review of Retelsdorf, J., & Bloom, P. (2022). How development and culture Developmental Psychology, 2(1), 503–531. https://doi.org/10 shape intuitions about prosocial obligations. Journal of Experi- .1146/annurev-devpsych-060320-102556 mental Psychology: General, 151(8), 1866–1882. https://doi.org Saxe, R., & Houlihan, S. D. (2017). Formalizing emotion concepts /10.1037/xge0001136, PubMed: 35130013 within a Bayesian model of theory of mind. Current Opinion in Masten, C. L., Gillen-O’Neel, C., & Brown, C. S. (2010). Children’s Psychology, 17,15–21. https://doi.org/10.1016/j.copsyc.2017.04 intergroup empathic processing: The roles of novel ingroup iden- .019, PubMed: 28950962 tification, situational distress, and social anxiety. Journal of Exper- Shamay-Tsoory, S. G., Ahronberg-Kirschenbaum, D., & Bauminger- imental Child Psychology, 106(2–3), 115–128. https://doi.org/10 Zviely, N. (2014). There is no joy like malicious joy: Schaden- .1016/j.jecp.2010.01.002, PubMed: 20202649 freude in young children. PLOS ONE, 9(7), e100233. https:// Moors, A. (2013). On the causal role of appraisal in emotion. doi.org/10.1371/journal.pone.0100233, PubMed: 24988446 Emotion Review, 5(2), 132–140. https://doi.org/10.1177 Shtulman, A., & Phillips, J. (2018). Differentiating “could” from /1754073912463601 “should”: Developmental changes in modal cognition. Journal Morelli, S. A., Lieberman, M. D., & Zaki, J. (2015). The of Experimental Child Psychology, 165,161–182. https://doi emerging study of positive empathy. Social and Personality Psy- .org/10.1016/j.jecp.2017.05.012, PubMed: 28648467 chology Compass, 9(2), 57–68. https://doi.org/10.1111/spc3 Siemer, M., Mauss, I., & Gross, J. J. (2007). Same situation–different .12157 emotions: How appraisals shape our emotions. Emotion, 7(3), Oatley, K., & Johnson-Laird, P. N. (2014). Cognitive approaches to 592–600. https://doi.org/10.1037/1528-3542.7.3.592, PubMed: emotions. Trends in Cognitive Sciences, 18(3), 134–140. https:// 17683215 doi.org/10.1016/j.tics.2013.12.004, PubMed: 24389368 Skerry, A. E., & Spelke, E. S. (2014). Preverbal infants identify emo- Olson, K. R., & Spelke, E. S. (2008). Foundations of cooperation in tional reactions that are incongruent with goal outcomes. Cogni- young children. Cognition, 108(1), 222–231. https://doi.org/10 tion, 130(2), 204–216. https://doi.org/10.1016/j.cognition.2013 .1016/j.cognition.2007.12.003, PubMed: 18226808 .11.002, PubMed: 24321623 Ong, D. C., Zaki, J., & Goodman, N. D. (2015). Affective cognition: Smith, C. E., Blake, P. R., & Harris, P. L. (2013). I should but i won’t: Exploring lay theories of emotion. Cognition, 143, 141–162. Why Young Children endorse norms of fair sharing but do not https://doi.org/10.1016/j.cognition.2015.06.010,PubMed: follow them. PLOS ONE, 8(3), e59510. https://doi.org/10.1371 26160501 /journal.pone.0059510, PubMed: 23527210 Ong, D. C., Zaki, J., & Goodman, N. D. (2019). Computational Smith, R. H., Powell, C. A. J., Combs, D. J. Y., & Schurtz, D. R. models of emotion inference in theory of mind: A review and (2009). Exploring the when and why of schadenfreude. Social roadmap. Topics in Cognitive Science, 11(2), 338–357. https:// and Personality Psychology Compass, 3(4), 530–546. https://doi doi.org/10.1111/tops.12371, PubMed: 30066475 .org/10.1111/j.1751-9004.2009.00181.x Paulus, M., Wörle, M., & Christner, N. (2020). The emergence of Smith, R. H., & van Dijk, W. W. (2018). Schadenfreude and human altruism: Preschool children develop a norm for gluckschmerz. Emotion Review, 10(4), 293–304. https://doi.org empathy-based comforting. Journal of Cognition and Develop- /10.1177/1754073918765657 ment, 21(1), 104–124. https://doi.org/10.1080/15248372.2019 Smith-Flores, A. S., & Feigenson, L. (2021). Preschoolers represent .1693375 others’ false beliefs about emotions. Cognitive Development, 59, Pietraszewski, D. (2021). Toward a computational theory of social 101081. https://doi.org/10.1016/j.cogdev.2021.101081 groups: A finite set of cognitive primitives for representing Smith-Flores, A. S., & Feigenson, L. (2022). “Yay! Yuck!” toddlers any and all social groups in the context of conflict. Behavioral use others’ emotional responses to reason about hidden and Brain Sciences, 45, e97. https://doi.org/10.1017 objects. Journal of Experimental Child Psychology, 221, /S0140525X21000583, PubMed: 33902764 105464. https://doi.org/10.1016/j.jecp.2022.105464,PubMed: Powell, C. A. J., & Smith, R. H. (2013). Schadenfreude caused by 35644055 the exposure of hypocrisy in others. Self and Identity, 12(4), Smith-Flores, A. S., Herrera-Guevara, I. A., & Powell, L. J. (2023). 413–431. https://doi.org/10.1080/15298868.2012.687013 Infants expect friends, but not rivals, to be happy for each other Powell, L. J. (2022). Adopted utility calculus: Origins of a concept when they succeed. Developmental Science, e13423. https://doi of social affiliation. Perspectives on Psychological Science, 17(5), .org/10.1111/desc.13423, PubMed: 37312424 1215–1233. https://doi.org/10.1177/17456916211048487, Smith-Flores, A. S., & Powell, L. J. (2023). Joint reasoning about PubMed: 35549492 social affiliation and emotion. Nature Reviews Psychology, 1, Pun, A., Birch, S. A. J., & Baron, A. S. (2021). The power of allies: 374–383. https://doi.org/10.1038/s44159-023-00181-0 Infants’ expectations of social obligations during intergroup con- Taylor, R. D., Oberle, E., Durlak, J. A., & Weissberg, R. P. (2017). flict. Cognition, 211, 104630. https://doi.org/10.1016/j.cognition Promoting positive youth development through school-based .2021.104630, PubMed: 33636572 social and emotional learning interventions: A meta-analysis of OPEN MIND: Discoveries in Cognitive Science 853 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. follow-up effects. Child Development, 88(4), 1156–1171. https:// Psychological Review, 122(3), 411–428. https://doi.org/10.1037 doi.org/10.1111/cdev.12864, PubMed: 28685826 /a0039252, PubMed: 25961468 Teding van Berkhout, E., & Malouff, J. M. (2016). The efficacy of Wu, Y., Baker, C. L., Tenenbaum, J. B., & Schulz, L. E. (2018). empathy training: A meta-analysis of randomized controlled Rational inference of beliefs and desires from emotional expres- trials. Journal of Counseling Psychology, 63(1), 32–41. https:// sions. Cognitive Science, 42(3), 850–884. https://doi.org/10.1111 doi.org/10.1037/cou0000093, PubMed: 26191979 /cogs.12548, PubMed: 28986938 Wang, Y. A., & Todd, A. R. (2021). Evaluations of empathizers Wu, Y., Muentener, P., & Schulz, L. E. (2017). One- to depend on the target of empathy. Journal of Personality and four-year-olds connect diverse positive emotional vocaliza- Social Psychology, 121(5), 1005–1028. https://doi.org/10.1037 tions to their probable causes. Proceedings of the National /pspi0000341, PubMed: 32940513 Academy of Sciences, U.S.A., 114(45), 11896–11901. Wellman, H. M., Phillips, A. T., & Rodriguez, T. (2000). Young chil- https://doi.org/10.1073/pnas.1707715114, PubMed: dren’s understanding of perception, desire, and emotion. Child 29078315 Development, 71(4), 895–912. https://doi.org/10.1111/1467 Wu, Y., & Schulz, L. E. (2018). Inferring beliefs and desires from -8624.00198, PubMed: 11016555 emotional reactions to anticipated and observed events. Child Wondra, J. D., & Ellsworth, P. C. (2015). An appraisal theory of Development, 89(2), 649–662. https://doi.org/10.1111/cdev empathy and other vicarious emotional experiences. .12759, PubMed: 28271499 OPEN MIND: Discoveries in Cognitive Science 854 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Open Mind MIT Press

Children’s Reasoning About Empathy and Social Relationships

Loading next page...
 
/lp/mit-press/children-s-reasoning-about-empathy-and-social-relationships-Zab9xipVuG

References (70)

Publisher
MIT Press
Copyright
© 2023 Massachusetts Institute of Technology. Published under a Creative Commons Attribution 4.0 International (CC BY 4.0) license.
eISSN
2470-2986
DOI
10.1162/opmi_a_00109
Publisher site
See Article on Publisher Site

Abstract

REPORT Children’s Reasoning About Empathy and Social Relationships Alexis S. Smith-Flores , Gabriel J. Bonamy, and Lindsey J. Powell Department of Psychology, University of California, San Diego Keywords: empathy, relationships, joint reasoning an open access journal ABSTRACT Across the lifespan, empathic and counter-empathic emotions are shaped by social relationships. Here we test the hypothesis that this connection is encoded in children’s intuitive theory of psychology, allowing them to predict when others will feel empathy versus counter-empathy and to use vicarious emotion information to infer relationships. We asked 4- to 7-year-old children (N = 79) to make emotion predictions or relationship inferences in response to stories featuring two characters, an experiencer and an observer, and either a positive or negative outcome for the experiencer. In the context of positive outcomes, we found that children engaged in robust joint reasoning about relationships and vicarious emotions. When given information about the characters’ relationship, children predicted empathy from a friendly observer and counter-empathy from a rival observer. When given information about the observer’s response to the experiencer, children inferred positive and negative relationships from empathic and counter-empathic responses, respectively. In the context of negative outcomes, children predicted that both friendly and rival observers would feel empathy toward the experiencer, but they still used information about empathic versus counter-empathic responses to infer relationship status. Our results suggest that young children Citation: Smith-Flores, A. S., Bonamy, G. J., & Powell, L. J. (2023). Children’s in the US have a blanket expectation of empathic concern in response to negative outcomes, Reasoning About Empathy and Social but otherwise expect and infer that vicarious emotions are connected to social relationships. Relationships. Open Mind: Discoveries in Cognitive Science, 7, 837–854. Future research should investigate if children use this understanding to select social partners, https://doi.org/10.1162/opmi_a_00109 evaluate their own relationships, or decide when to express empathy toward others. DOI: https://doi.org/10.1162/opmi_a_00109 Supplemental Materials: https://doi.org/10.1162/opmi_a_00109 Received: 7 May 2023 INTRODUCTION Accepted: 12 September 2023 Think about the last few times you felt happy. Likely, some of those instances were the result of Competing Interests: The authors declare no conflicts of interest. a friend or family member’s good news. People’s emotions result not only directly from their own personal experiences, but also from witnessing the experiences of others. Emotions felt Corresponding Author: Alexis S. Smith-Flores as a result of someone else’s experiences are known as “vicarious emotions” (Wondra & alexis-smith@ucsd.edu Ellsworth, 2015). Vicarious emotions can vary both in valence and in their congruence with the likely emotions of the person who inspires them. We will use the term “empathy” to refer to Copyright: © 2023 congruent positive or negative emotions for those having good or bad experiences, respec- Massachusetts Institute of Technology tively (Batson et al., 2009). In contrast, counter-empathy involves feeling incongruent emo- Published under a Creative Commons Attribution 4.0 International tions toward others’ experiences, such as sadness in response to others’ good outcomes (CC BY 4.0) license (i.e., gluckschmerz) or happiness in response to others’ bad outcomes (i.e., schadenfreude) (R. H. Smith & van Dijk, 2018). The MIT Press Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. Empathy and counter-empathy tend to be applied selectively: People often feel more empa- thy toward those they care more about and feel counter-empathy toward disliked individuals or outgroups (Batson, 2009; Cikara & Fiske, 2012; Masten et al., 2010; Shamay-Tsoory et al., 2014). As a result, expressions of empathy and counter-empathy may be predicted by, and carry information about, a person’s social relationships. Understanding these connections could allow people to effectively seek social support, and to evaluate their own and others’ relationships based on others’ vicarious emotional expressions. Relatively little research has investigated how people think about the role of relationships in vicarious emotions, however, or how such reasoning develops. Here we consider how common aspects of intuitive psychol- ogy, available in early childhood, may allow children to engage in systematic reasoning about the connections between vicarious emotions and social relationships. We then test their ability to engage in this kind of reasoning. Reasoning about others’ emotions often involves considering their desires. Emotions arise from appraisals of the world relative to what the appraiser wants, needs, and expects. When goals are met and desires are fulfilled, observers expect positive responses, while failures and undesirable outcomes are often accompanied by negative responses (Moors, 2013; Oatley & Johnson-Laird, 2014). This dependence on individuals’ goals and preferences results in appraisals that are person-specific: One person’s good outcome may be someone else’s bad outcome, resulting in differing appraisals of the same situation (Manstead & Fischer, 2001; Siemer et al., 2007). For example, the outcome of a political election often excites some peo- ple and disappoints others, depending on preferences for different candidates or policies. An observer’s successful emotion prediction thus requires integrating the observed state of the world and the specific desires of the target person. An understanding of the appraisal process also allows for inverse reasoning about the causes of emotions. Based on an emotional expres- sion, an observer who knows the emoter’s desires can infer the experience that led to the expression, while an observer who knows what experience led to the expression could infer the emoter’s desires (Ong et al., 2015; Wu & Schulz, 2018). Even young children are able to use emotion vocalizations to infer unobserved causes (Smith-Flores & Feigenson, 2022; Wellman et al., 2000; Wu et al., 2017), suggesting that the ability to reason about emotion as an appraisal process appears early in development. Understanding emotions as the outcome of an appraisal process may also support reason- ing about vicarious emotions and their connection to social relationships (Smith-Flores & Powell, 2023). Concepts of social relationships involve the expectation that relationship part- ners care about one another’s goals and welfare (Afshordi & Liberman, 2021; Jern & Kemp, 2014; Powell, 2022). Friends and group members are generally expected to do more to help one another, and both adult and child observers infer relationships from the effort or partiality of offered support (Earp et al., 2021; Liberman & Shaw, 2017; Marshall et al., 2022; Olson & Spelke, 2008). Even infants expect affiliates to support each other when in need (Jin & Baillargeon, 2017; Pun et al., 2021). Conversely, concepts of negative relationships entail expectations of conflict and spite, reflecting a desire forthe otherpersontosufferorfail (Pietraszewski, 2021). Observers may expect these vicarious desires to serve as the basis for appraisals that lead to vicarious emotions. In addition to predicting vicarious emotions, an observer operating with this intuitive theory should also be able to infer what type of relation- ship led to an expressed vicarious emotion. Upon witnessing a counter-empathic response, indicating opposite appraisals for the experiencer and the emoter, an observer should infer a negative relationship between the two individuals. There is some evidence that both adults and children reason about the connection between vicarious emotions and social motivations, though most work focuses more on inferences of OPEN MIND: Discoveries in Cognitive Science 838 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. general social traits as opposed to specific social relationships. Heyman and Gelman (1998) found that children used a person’s pro- or antisocial motives to infer how pleased they were by positive or negative outcomes for the targets of those motives. Wang and Todd (2021) found that adults evaluate a person who empathizes with a valued target person (e.g., a children’s hospital worker) as being warmer, more likable, and more respectable than a non-empathizer. Three- to five-year-old children similarly find empathy for others’ suffering to be normative, which indicates that young children recognize and reason about others’ empathic concern (Paulus et al., 2020). When directed toward morally repugnant or disliked targets, however, adults may prefer those who are indifferent rather than empathic (Krems et al., 2023; Wang & Todd, 2021). This could reflect a dispreference for those who positively appraise outcomes one considers to be either morally wrong or disadvantageous, as well as a desire for partial- ity in friendships. In other work, adults use emotional expressions (e.g. of anger, sadness, or happiness), either alone or in social contexts, to infer emoters’ aggression, warmth, and affiliative tendencies (Hareli & Hess, 2010; Knutson, 1996). Similar findings extend to peo- ple’s evaluations of their own real-world social partners: physicians who empathize with their patients are viewed as more trustworthy (Kim et al., 2004), and perceived empathy between romantic partners is positively related to relationship satisfaction (Cramer, 2003; Cramer & Jowett, 2010; Gable et al., 2004), though it is unclear whether preferences for empathy in these social relationships reflect dispositional or relationship-based reasoning. Together, these findings suggest that adults and children make dispositional judgments about others using observations of their vicarious emotions. They fall short, however, of providing evidence for a systematic framework connecting vicarious emotions to relationship-based appraisals of outcomes. One set of looking time experiments tested 10- to 11-month-old infants’ expectations about vicarious emotions across different relationship contexts (Smith-Flores, Herrera-Guevara, & Powell, 2023). Infants were introduced to a pair of characters with either a positive or a neg- ative relationship, followed by trials in which one character attempted a goal-directed action and the other character responded happily or sadly to the outcome. In the context of a positive relationship and successful outcome, infants looked significantly less at happy than sad response trials, consistent with an expectation for the observer to emote happily to their friend’s success. In the context of a negative relationship, infants’ looking was not different across happy and sad response trials, suggesting no expectation of positive empathy for a rival. The interaction between relationship context and trial type was significant, supporting the con- clusion that infants around one year of age do take relationships into account in their expec- tations of vicarious emotions. However, there was no evidence that infants expected negative relationships to lead to counter-empathic emotions. This may reflect limited early reasoning about vicarious emotions, focused on predicting empathic responses. Alternatively, infants may struggle more generally with predicting negative emotions in such contexts; they also fail to expect that a friendly observer or even the actor will respond sadly to a failed attempt (Skerry & Spelke, 2014; Smith-Flores et al., 2023). By age 4, children are able to make robust, appraisal-based predictions about both positive and negative emotions (Fabes et al., 1991; Wellman et al., 2000), and appear to possess an emerging norm of empathic concern (Paulus et al., 2020). Four-year-old children are also able to reason about representations of emotions that are at odds with the current state of the world (Smith-Flores & Feigenson, 2021). Investi- gating children’s verbal predictions about vicarious emotions beginning at this age will thus allow us to explore reasoning about a more comprehensive combination of relationships, out- come valence, and vicarious responses. We also test children’s inferences about relationships from information about vicarious emotional responses. OPEN MIND: Discoveries in Cognitive Science 839 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. The current study investigates how 4- to 7-year-old children reason about relationships and expressions of empathy. Children were told four stories in which they were asked to predict how an observer would feel about an outcome for another person (i.e., their friend or rival; Empathy Block), and another four stories in which they were asked to infer the relationship between two characters after being given information about the observer’sempathicor counter-empathic response to the other person’s outcome (Affiliation Block). We preregistered three primary hypotheses on the Open-Science Framework (OSF; https://osf.io/d57ny). First, we hypothesized that children would predict that observers will respond empathically to their positive affiliates (i.e., friends). If this is true, then when children hear stories featuring two friends, they should be more likely to predict that the observing individual will respond to their friend’s outcome with a congruent emotion (i.e., happiness following a good outcome or sadness following a bad outcome) than an incongruent emotion. However, it is possible that children may expect that friends should respond with positive emotion toward their positive affiliates regardless of the outcome. If this is true, then children may predict that the observer will respond with a positive emotion regardless of the valence of their friend’s experience. Second, we hypothesized that children’s predictions of empathy will be weaker for rivals. If this is true, then children should predict less outcome-congruent emotion when making pre- dictions in stories about rivals as compared to friends, regardless of the outcome type. An alter- native hypothesis not anticipated in our preregistration is that children’s possession of strong norms about responding empathically toward others’ suffering may lead them to predict sim- ilar levels of empathy following negative outcomes, regardless of the characters’ relationship (Paulus et al., 2020). Young children tend to use norms rather than other factors, including preferences, to predict others’ behavior, and may even say that impermissible actions are impossible (Kalish & Shiverick, 2004; Shtulman & Phillips, 2018; R. H. Smith et al., 2009). Finally, we hypothesized that children will use vicarious emotions to infer social relation- ships. If so, children should be more likely to infer that two characters have a close positive relationship after hearing that one character responded empathically to the other character. In contrast, children should select responses that correspond to more distant, negative relation- ships after hearing about a character that responded counter-empathically to another charac- ter. Of course, it is also possible that children may use the valence of the emotion response to make inferences about others’ relationships. If this is true, then children may infer that when a character responds negatively to another character, regardless of the outcome of the event, that the two are negatively affiliated, while characters who respond positively are positively affiliated. We also conducted exploratory analyses on age-related changes in reasoning about vicar- ious emotions and relationships. Although we had no a priori hypotheses about developmen- tal change, several patterns were possible. First, there may be no change with age, either because children at all ages are at chance or because children make mature systematic pre- dictions from age 4. Next, there could be development both in terms of children’s differential prediction of vicarious emotion depending on relationships and in their inference of relation- ship valence from vicarious responses. This would indicate that younger children have a weaker understanding of how vicarious emotions are shaped by relationships, reflected in both their predictions of such emotions and inferences from them. Finally, we could observe developmental change in children’s predictions of vicarious emotions but not their inferences from them. This hypothesis is suggested by past results showing that young children often pre- dict that people will behave according to rules and norms, and sometimes even say that it is impossible to do otherwise (DeJesus et al., 2014; Kalish & Shiverick, 2004; Shtulman & OPEN MIND: Discoveries in Cognitive Science 840 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. Phillips, 2018; C. E. Smith et al., 2013). This tendency to predict strong adherence to norms diminishes with age. Because children view empathy for others’ pain as normative (Paulus et al., 2020), younger children may be less likely to predict counter-empathy, despite being able to reason about the meaning of counter-empathic responses for underlying social relationships. METHOD Participants Seventy-nine children, aged 4- to 7-years-old (M = 5.93 years; range = 3.98 years to 7.88 age years; 42 girls), participated at a science museum in Southern California. One additional child participated but was excluded for being outside the preregistered age range. Caregivers completed an optional demographic form at the time of participation. Fifty-one children were identified by their caregivers as White, seven as Asian, 1 as American Indian/Alaskan Native, 18 were identified as belonging to two or more races, and two caregivers declined to answer. Fifteen children were identified as Hispanic/Latinx and five caregivers declined to answer. Seventy-one children came from families where at least one caregiver had a college degree or higher, five had at least one caregiver who completed some college, one had at least one caregiver who had a high school diploma or GED, and two care- givers declined to answer. Our data collection plan was preregistered as follows: We ran data collection sessions at a local science museum until we had at least 72 participants with valid data for both blocks, while also planning to include data from participants with only one valid block and from par- ticipants collected in the final testing session required to reach the target sample size. This resulted in unequal numbers of participants included in the two blocks. Seventy-seven partic- ipants completed the Empathy Block and 74 participants completed the Affiliation Block. Participants received a small gift (i.e., temporary tattoos or stickers) to thank them for their participation. Materials & Procedures All materials including preregistrations, data, and analysis code are available on OSF (https:// osf.io/d57ny). Participants were read 4 stories per block across 2 blocks, for a total of 8 stories. Each of the four stories featured two characters, an actor and an observer. Stories in the Empa- thy Block presented participants with information about the relationship between two charac- ters and about a positive or negative outcome for one of the two characters (the actor), before asking participants to predict the emotion felt by the other character (the observer) following that outcome. Stories in the Affiliation Block presented participants with information about a positive or negative outcome for one character (the actor) and the emotion felt by the other character (the observer) following that outcome, before asking participants to infer the degree of affiliation (i.e., the closeness of the relationship) between the two characters. Each block also included an initial warm-up trial that familiarized participants with the relevant response scale (a positive to negative affect scale in the Empathy Block; a positive to negative relation- ship scale in the Affiliation Block) and tested their ability to use it in responding to questions posed by the experimenter. Participants were randomly assigned to 1 of 8 story orders that counterbalanced block order (Empathy Block first vs. second), outcome order (positive outcome first vs. second), and which affiliation or emotion type came first in each block (friendship/empathy vs. rivalry/counter-empathy). OPEN MIND: Discoveries in Cognitive Science 841 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. Empathy Block The researcher began by introducing a 4-point Likert scale, ranging from “very happy” to “very sad”, accompanied by simple, line-drawn faces depicting each emotion (Figure 1B). The researcher asked participants about their favorite and least favorite snack. Participants then rated, verbally or by pointing to the scale, how they would feel if they received a lot of their favorite or least favorite snack. As preregistered, participants who responded with the same emotion regarding their favorite and least favorite snack were excludedfromthisblock’s analysis (see Supplemental Information for number of children excluded from each block, Table S1). Participants were first introduced to the two characters by name and were told about their relationship (e.g., “Ryan and Chris are best friends. They like each other a lot.”). To ensure participants remembered the relationship between the two characters, the researcher then asked a manipulation check question (e.g., “Remind me, are Ryan and Chris friends?”). Par- ticipants who did not answer the manipulation check correctly were told the story again and had the manipulation check repeated (Table S2). Participants who did not answer the manip- ulation check correctly a second time were excluded from that question’s analysis (see Supplemental Information for number of children who failed the manipulation check once or twice for each story, Table S2). Following the manipulation check question, participants were then told about an event that happened to the actor while the observer watched (e.g., “At recess, Ryan races against another classmate and loses!”). The researcher then asked the test question regarding how the observer Figure 1. Stimuli from the Empathy Block. (A) Children heard 4 stories in total where the actor’s outcome (positive, negative) and characters’ affiliation (friends, rivals) varied across stories. (B) Children were asked to predict how the observer felt about the actor’s outcome using a 4- point scale ranging from “very happy” to “very sad”. OPEN MIND: Discoveries in Cognitive Science 842 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. felt about the actor’s outcome (e.g., “How does Chris feel about Ryan losing the race?”). The experimenter pointed to each item on the emotion scale, reading the options aloud to the par- ticipant. Participants’ verbal or pointed answers were recorded by a second researcher who observed from across the table. A 2 × 2 design was used so that each child heard one story each about a positive outcome for a friend, a negative outcome for a friend, a positive out- come for a rival, and a negative outcome for a rival. Affiliation Block The researcher started by introducing a 4-point Likert scale, used to ask how much people like one another and ranging from “a lot” to “not at all”, accompanied by pictures of two people close together or increasingly far apart (Figure 2B). To familiarize children with this affiliation scale, we asked participants about their favorite and least favorite characters. Participants then rated, verbally or by pointing to the scale, how much they like their favorite or least favorite character. As preregistered, participants who answered the same regarding their favorite and least favorite characters were excluded from this block’sanalysis(see Supplemental Information for number of children excluded from each block, Table S1). Participants were first introduced to the two characters by name and then were told about an event that happened to the actor while the observer watched (e.g., “At lunchtime, Kyle has a cookie in his lunch, but he drops it on the floor and has to throw it away!”; see Figure 2A). Participants were then told how the observer felt about what had happened to the actor (e.g., “Brent is very sad Kyle can’t eat his cookie”). The observer either felt very happy or very sad about the event that happened to the actor. Figure 2. Stimuli from the Affiliation Block. (A) Children heard 4 stories in total where the actor’s outcome (positive, negative) and the observer’s emotion (empathic, counter-empathic) changed across stories. (B) Children were asked to infer how much the observer and the actor liked each other using a 4-point scale ranging from “alot” to “not at all”. OPEN MIND: Discoveries in Cognitive Science 843 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. To ensure participants remembered the emotion felt by the observer, the researcher then asked a manipulation check question (e.g., “Remind me, how does Brent feel?”). Participants who did not answer the manipulation check correctly were told the story again and had the manipulation check repeated (Table S2). Participants who did not answer the manipulation check correctly a second time were excluded from that question’sanalysis (see Supplemental Information for number of children who failed the manipulation check once or twice for each story, Table S2). Finally, the researcher asked the test question regarding the affiliation between the two characters (e.g., “How much do Brent and Kyle like each other?”). The experimenter pointed to each item on the affiliation scale, reading the options aloud to the participant. Participants’ verbal or pointed answers were recorded by a second researcher who observed from across the table. A 2 × 2 design was used so that each participant heard one story each about a pos- itive outcome for the actor with a positive reaction from the observer, a negative outcome with a negative reaction, a positive outcome with a negative reaction, and a negative outcome with a positive reaction. Coding. All responses were coded by the second researcher at the time of the experiment. An additional coder coded 20% of participants from video recordings to ensure reliability. Agree- ment between coders was 100%. For analysis of data from the Empathy Block, participants’ raw emotion predictions were transformed to reflect the degree of expected empathy in the emotion prediction. Responses that indicated the greatest amount of predicted empathy (i.e., “very sad” to a negative outcome and “very happy” to a positive outcome) were coded as a 4, while responses that indicated the greatest predicted counter-empathy (i.e., “very happy” to a negative outcome and “very sad” to a positive outcome) were coded as a 1. Moderately empathic or counter-empathic responses were coded as a 3 or 2, respectively. In the Affiliation Block, responses that indicated the strongest positive relationship infer- ences (i.e., the characters like one another “a lot”) were coded as a 4, while responses that indicated the strongest negative relationship inferences (i.e., the characters like one another “not at all”) received a 1. Moderately positive or negative responses were coded as a 3 or 2, respectively. RESULTS Empathy Block First, to test our preregistered hypothesis that children predict that friends will respond empath- ically toward each other’s outcomes, we conducted separate, two-tailed one-sample t-tests from chance (2.5 on a 1–4 point scale) on children’s responses in the Empathy Block for each of the two stories that featured “best friends”. Regardless of the type of outcome (i.e., positive or negative), children predicted that the observing friend would respond empathically (positive outcome: M =3.662, SEM =0.092, t(76) = 12.674, p < .001, 95% CI [3.480, 3.845], d = 1.444.; negative outcome: M =3.342, SEM = 0.104, t(75) = 8.133, p < .001, [3.136, 3.548], d = 0.933; see Figure 3). To test the hypothesis that children have weaker expectations of empathy for rivals than friends, we fit a linear regression model of children’s responses in the Empathy Block. Affili- ation type (friends vs. rivals), outcome type (positive vs. negative), and the interaction between affiliation type and outcome type were included in the model as fixed effects. Affiliation type OPEN MIND: Discoveries in Cognitive Science 844 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. Figure 3. Children’s responses for each question in the Empathy Block. The means and standard errors are plotted in white. The dashed line represents the mid-point of the scale (2.5). Significance asterisks come from post-hoc t-tests with bonferroni adjustment as a result of a signif- icant interaction between affiliation type and outcome type. ***p < .001. and outcome type were dummy-coded and mean-centered. We preregistered a mixed-effect model that included participant as a random effect, however this model returned a singular fit convergence issue due to the participant-level random effect variance being close to zero. So, we removed the random effect from the full Empathy Block model. The resulting full model was as follows: response ∼ affiliation.type + outcome.type + affiliation.type * outcome.type Using nested model comparisons, we then compared this full model to models that excluded each factor individually to test how much the excluded factor contributed to explain- ing the variance in children’s predictions of empathy. We found a main effect of affiliation type, F(1, 301) = 107.46, p < .001 – children were more likely to predict empathy for friends (M = 3.503, SEM = 0.070) than for rivals (M = 2.349, SEM = 0.098). We also found a main effect of outcome type, F(1, 301) = 13.338, p < .001 – children were more likely to predict empathy following negative outcomes (M = 3.132, SEM = 0.087) than for positive outcomes (M = 2.727, SEM = 0.103). There was also an interaction between affiliation type and outcome type, F(1, 301) = 42.883, p < .001. Using post-hoc paired-sample t-tests with Bonferroni adjustments (α = .0125, .05 divided by 4 tests), we found that children’s predictions of empa- thy following negative outcomes did not differ significantly for friends versus rivals (M = 2.920, SEM = 0.138), once corrected for multiple comparisons, t(73) = 2.156, p = .034, 95% CI [0.032, 0.806], d = 0.251, but that children predicted significantly more empathy from friends (M = 3.662, SEM = 0.092) than rivals (M = 1.792, SEM = 0.105) following positive outcomes, t(76) = 12.236, p < .001, [1.566, 2.175], d = 1.394. Within affiliation type, children’s predic- tions of empathy for a friend were not significantly different across positive and negative out- come types, once corrected for multiple comparisons, t(73) = −2.269, p = .026, [−0.593, −0.039], d = −0.260, but children predicted significantly more empathy for a rival who expe- rienced a negative vs. positive outcome, t(74) = 6.770, p < .001, [0.790, 1.450], d = 0.781. Our data support the conclusion that children selectively predict empathic happiness for friends but not rivals following positive outcomes but expect empathic concern for both friends and rivals following negative outcomes. OPEN MIND: Discoveries in Cognitive Science 845 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. To investigate whether this pattern of reasoning changed across the age range in our sam- ple, we conducted an exploratory analysis of all data from the Empathy Block, adding a con- tinuous age variable and the associated two- and three-way interactions as fixed effects to the full model described above. A nested model comparison found a significant three-way inter- action between age, affiliation type, and outcome type, F(4, 297) = 4.438, p = .002 (Figure 4). Post-hoc pairwise comparisons with Bonferroni adjustments (α = .0125, .05 divided by 4 tests) showed that the only significant difference in slopes was between predictions of empathy for friends and rivals following negative outcomes, estimate = 0.440, t(297) = 3.389, p < .001, diff 95% CI [0.184, 0.695] – as children got older they were more likely to predict empathy for friends and counter-empathy for rivals after negative events. All other ts < 1.970 and ps > .050. We also preregistered an exploratory analysis of only the older half of the sample, children 6 or 7 years of age, to assess if joint reasoning about vicarious emotions and social relation- ships is more robust in older children. Using the same set of nested linear regression compar- isons as in the main analysis (i.e. without the continuous age factor), we again found a main effect of affiliation type on participants’ empathy predictions, F(1, 151) = 95.017, p < .001, and an interaction between affiliation type and outcome type such that empathy predictions varied more by affiliation type following positive vs. negative outcomes, F(1, 151) = 15.185, p < .001. However, post-hoc t-tests showed that in this sample subset, older children did predict more empathy following a negative outcome from a friend (M = 3.410, SEM = 0.120) than from a rival (M = 2.553, SEM = 0.202), t(37) = 3.141, p = .003, 95% CI [0.299, 1.385], d = 0.510. We followed up on this result by comparing older and younger children’s empathy predictions for each of the four story types and found that the only reliable difference was that older children predicted less empathy for rivals’ negative outcomes compared to younger children, t(73) = −2.82, p = .006, [−1.270, −0.219], d = 0.652. Younger children’s predictions of empathy for a rival’s negative outcome were greater than chance (2.5 on our scale), M = 3.297, SEM = 0.168, t(36) = 4.737, p < .001, [2.956, 3.639], d = 0.779, while older children’s predictions of empathy in the same situation were not significantly different from chance, M = 2.553, Figure 4. Children’s responses in the Empathy Block by age in years. OPEN MIND: Discoveries in Cognitive Science 846 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. SEM = 0.202, t(37) = 0.260, p = .796, [2.143, 2.962], d = 0.042. Thus children’s expectation of empathic concern for rivals appears to decrease with development. Affiliation Block To test if children use observations of empathic and counter-empathic responses to infer rela- tionships, we fit the preregistered mixed-effect model of participants’ responses in the Affilia- tion Block. Emotion type (empathic response vs. counter-empathic response), outcome type (positive vs. negative), and the interaction between emotion type and outcome type were included in the model as fixed effects. Participant was included as a random effect. Emotion type and outcome type were dummy-coded and mean-centered. The full model was as follows: response ∼ emotion.type + outcome.type + emotion.type*outcome.type + (1 | participant) We used nested model comparison to compare the full model to models that excluded each fixed effect factor individually to test how much the excluded factor contributed to explaining the variance in children’s relationship inferences. We found a main effect of emotion type, χ (1) = 219.09, p < .001 – children inferred a positive relationship more following an empathic response (M = 3.616, SEM = 0.061) than a counter-empathic response (M = 1.832, SEM = 0.084). We also found a main effect of outcome type, χ (1) = 11.136, p = .001 – children were more likely to infer a positive relationship following positive outcomes (M = 2.878, SEM = 0.098) than for negative outcomes (M = 2.582, SEM = 0.109). There was no interaction between emotion type and outcome type, χ (1) = 0.034, p = .853. In line with our predictions, children were more likely to infer a positive relationship between two characters when they observed one character respond empathically, rather than counter-empathically, to the other’s outcome, regardless of whether that outcome was a good or bad one (Figure 5). Figure 5. Children’s responses for each question in the Affiliation Block. The means and standard errors are plotted in white. The dashed line represents the mid-point of the scale (2.5). Significance asterisks come from nested model comparisons of mixed linear models testing for a main effect of emotion type. ***p < .001. OPEN MIND: Discoveries in Cognitive Science 847 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. As in the empathy block, we ran an exploratory analysis with a continuous age variable and related interactions as predictors for the Affiliation Block. There was a significant three-way interaction between empathic type, outcome type, and age, χ (4) = 21.850, p < .001 (Figure 6). Post-hoc pairwise comparisons with Bonferroni adjustments (α = .0125, .05 divided by 4 tests) showed that the only significant difference in slopes was between relationship inferences from empathy and counter-empathy following negative outcomes, estimate = 0.522, t(215) = diff 4.648, p < .001, 95% CI [0.301, 0.743]. As children got older they were more likely to infer positive relationships from empathy and negative relationships from counter-empathy after negative events. All other ts < 2.130 and ps > .034. By including the full set of age ranges, we found that children’s inferences about relationships from vicarious emotions following negative outcomes become stronger with age. However, there are several reasons to see the developmental change in this case as incre- mental, rather than qualitative. First, the preregistered analysis of the older half of the age group revealed the same pattern of effects as in the full dataset: In the older children’s responses to the Affiliation Block, we again found a main effect of empathic type, χ (1) = 190.51, p < .001 – children inferred a positive relationship more following an empathic response (M =3.753, SEM = 0.059) than a counter-empathic response (M = 1.662, SEM = 0.097). We also found a main effect of outcome type, χ (1) = 8.407, p = .004 – older children were still more likely to infer a positive relationship following positive outcomes (M = 2.859, SEM = 0.137) than for negative outcomes (M = 2.553, SEM = 0.149), as in the full dataset. There was no interaction between empathic type and outcome type, χ (1) = 2.894, p = .089. Second, an analysis of only the younger half of the sample also found the same pattern: a main effect of empathic type, χ (1) = 62.447, p < .001 – younger children inferred a positive relationship more following an empathic response (M = 3.464, SEM = 0.108) than a counter- empathic response (M = 2.030, SEM = 0.140); a main effect of outcome type, χ (1) = 4.189, p = .041 – children were still more likely to infer a positive relationship following positive outcomes (M =2.900, SEM = 0.141) than for negative outcomes (M =2.615, SEM = 0.162), and no interaction between empathic type and outcome type, χ (1) = 1.238, p = Figure 6. Children’s responses in the Affiliation Block by age in years. OPEN MIND: Discoveries in Cognitive Science 848 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. .266. In particular, 4- and 5-year-old children were more likely than chance to infer positive relationships from empathic responses to both positive outcomes, t(34) = 11.554, p < .001, 95% CI [3.501, 3.928], d = 1.953, and negative outcomes t(33) = 3.858, p < .001, [2.834, 3.578], d = 0.661, and more likely than chance to infer negative relationships from counter-empathic responses to both positive outcomes t(34) = −2.358, p = .023, [1.729, 32.443], d = −0.399, and negative outcomes t(30) = −2.367, p = .025, [1.509, 2.427], d = −0.425. Exploratory Preregistered Order Analyses. Across participants, outcome type order and affiliation type order (Empathy Block) or empathy type order (Affiliation Block) were counterbalanced. To examine any effect this may have had on children’s responses, we again used a nested model comparison to determine whether the addition of fixed order effects improved the model’s fit. There were no order effects observed in either the Empathy Block, F(2, 299) = 0.847, p = .430, or the Affiliation Block, χ (2) = 0.991, p = .609. Additional preregistered exploratory analyses can be found in Supplemental Information. DISCUSSION These data support the hypothesis that children can engage in systematic joint reasoning about social relationships and vicarious emotions (Smith-Flores & Powell, 2023). This includes both inferring social relationships from the information provided by empathic and counter- empathic responses (Affiliation Block) and making relationship-based predictions about vicarious emotions following positive outcomes (Empathy Block). The one context in which children did not display integrated reasoning about relationships and vicarious emotions across the full age range tested was when predicting one person’s response following someone else’s negative experience. In this context, children in the full sample expected others to express similar levels of empathy regardless of their relationships, though there was explor- atory evidence that this varied between younger and older participant age groups. This failure may reflect a lack of understanding about the role negative relationships play in schaden- freude. However, given children’s predictions of counter-empathy following rivals’ good outcomes and their robust tendency to infer negative relationships from counter-empathy regardless of outcome valence, we suspect this result instead reflects young children’s general tendency to predict that people will behave normatively, rather than counter-normatively (DeJesus et al., 2014; Kalish & Shiverick, 2004; Powell & Smith, 2013; Shtulman & Phillips, 2018). Both adults and children consider most expressions of schadenfreude to be inappropri- ate, even toward disliked individuals (e.g., Geraci et al., 2021; Paulus et al., 2020; Wang & Todd, 2021). Thus our participants may have predicted an observer would respond empath- ically to a rival’s suffering because this is what they believe the observer should do. Children’s responses in the Affiliation Block also changed with age, becoming less likely to fall into the “moderate” categories of relationships when the emotions were a result of negative outcomes, but from the youngest ages children still used empathy and counter-empathy to make opposite relationship inferences. We conclude that, in addition to supporting our three primary hypoth- eses, the data are most consistent with the third possible developmental pattern laid out in the introduction: from the earliest ages tested, children already provided evidence of understand- ing the link between relationships and vicarious emotions, but the expression of this knowl- edge interacts with other aspects of cognitive development, including changes in modal reasoning. Children’s ability to both predict vicarious emotion from relationships and infer relation- ships from vicarious emotion is consistent with the possession of an intuitive theory of OPEN MIND: Discoveries in Cognitive Science 849 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. psychology that connects relationship concepts to the causes of emotion. This proposal builds on work demonstrating that adults and children reason about emotions as the product of appraisals (Ong et al., 2019; Saxe & Houlihan, 2017; Wu et al., 2018), as well as the theory that early relationship concepts involve positing that relationship partners find the well-being of their affiliates rewarding (Hamlin et al., 2013; Powell, 2022). By connecting relationships to emotions through the reward that one person feels when good things happen to their friends (or bad things happen to their foes), or the sadness they feel in the opposite circumstances, this “adopted utility calculus” framework can support both forward and reverse inferences about how social relationships affect vicarious emotions (Smith-Flores & Powell, 2023). The unex- pected finding that children inferred somewhat less positive relationships following negative versus positive outcomes is also consistent with probabilistic reasoning over an intuitive the- ory: Because children expect empathy to be normative and thus common following negative outcomes, (1) this empathy is potentially less convincing evidence of a close relationship than positive empathy and (2) counter-empathy for the negative outcome (i.e., schadenfreude) is more surprising and thus better evidence of a highly negative relationship than counter- empathy for the positive outcome. However, this is not the only possible basis for the pattern of reasoning observed here. Future research should investigate the integration of vicarious emotion reasoning with other components of intuitive psychology, including reasoning about ignorance or false belief, to test this account. Relative to recent investigations of reasoning about vicarious emotions in preverbal infants, the current findings suggest both continuity and development. In a series of experiments, dif- ferences in 10- to 11-month-old infants’ looking time indicated that they were surprised if a friendly observer responded to an actor’s success with sadness, relative to a happy response. In contrast, when events depicted a rival observer watching the actor succeed, infants’ looking provided no evidence for an expectation of either a happy or a sad vicarious response (Smith- Flores et al., 2023). Thus, like the children in the current study, infants use social relationship information to inform their expectations about vicarious emotions after positive outcomes. However, infants did not use negative relationship information to predict counter-empathy from the observer following the actor’s success. In contrast, children in the current study, across all ages, predicted counter-empathy from rivals following positive events. Future research should more fully investigate the development of reasoning about counter-empathy. Infants’ lack of expectation may reflect a broader difficulty with reasoning about negative emo- tions, which extends to failures to predict negative emotional responses to direct experiences (Ruba et al., 2019, 2020; Skerry & Spelke, 2014). Alternatively, an understanding of empathic vicarious emotions may precede that of counter-empathic vicarious emotions, perhaps due to the difficulty of representing one person’s desire for the other to not achieve their goals (Feiman et al., 2015). These findings provide insight into children’s reasoning about others’ social interactions and highlight children’s sensitivity to others’ displays of vicarious emotions. Children’s own empathic behaviors have been shown to play an important role in their ability to form positive social relationships (Brown & Fredrickson, 2021;Denhametal., 2003). Several theories regarding the development of emotional competence posit that empathy and emotion reason- ing must develop together in order for children to achieve optimal social functioning (Hare et al., 2022; Hare & Parent, 2022; Taylor et al., 2017). Children with higher levels of emotion reasoning (i.e., the ability to correctly identify others’ emotions, also referred to as “cognitive empathy”; see Ruba & Pollak, 2020), have more positive social interactions and engage in more prosocial behaviors (Ensor et al., 2011). In clinical populations, training emotion recog- nition enhances empathic feelings and behavior (Teding van Berkhout & Malouff, 2016). OPEN MIND: Discoveries in Cognitive Science 850 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. Children’s developing understanding that vicarious emotions convey information about social relationships, displayed in the current findings, may also play a role in children’s own effective use of expressions of empathy. Our experiment finds that, at the group level, children readily predict empathy for friends’ successes and suffering, but it does not sensitively capture individual differences in children’s reasoning. Future research should explore connections to children’s own empathic tendencies and additional social–emotional competencies. Finally, empathic and counter-empathic responses are rich sources of information about others’ prosociality, such as their propensity to help or provide comfort to those in distress (Batson, 2013; Bruneau et al., 2017; Eisenberg et al., 1993; Morelli et al., 2015). In this exper- iment, we tested whether children could infer relationships from observations of empathic responses. Children inferred that empathy signaled positive relationships rather than negative ones. In other work, children have been found to prefer and reward characters who comforted another character, suggesting that children may evaluate individuals who display empathic behaviors more favorably than non-empathic individuals (Geraci et al., 2021). However, it remains unknown whether children’s positive evaluations carry over into their selection of social partners for themselves. In one case, children may select empathizers if they need someone’s help, someone to play with, or someone to learn from when no other information is available. However, counter-empathy may also be a sign of a potential social partner if the counter-empathy is directed toward someone the child does not like. Recent work found that adults do prefer those who are kind and trustworthy over those who are not, but also that adults want friends who will be more positive to them than others and even be less prosocial to their enemies (Krems et al., 2023). Future work may explore how children evaluate empathizers and counter-empathizers as potential social partners and the nuances of such evaluations. In sum, this experiment supports an early-developing social cognitive framework for joint reasoning about others’ relationships and vicarious emotions (Smith-Flores & Powell, 2023). This work gives rise to a number of questions about how children’s reasoning about other people carries over to their own social competencies and decision-making. AUTHOR CONTRIBUTIONS Alexis Smith Flores: Conceptualization, Methodology, Formal analysis, Visualization, Writing— Original draft, Writing—Review & editing. Gabriel Bonamy: Investigation, Methodology, Writing—Original draft. Lindsey Powell: Conceptualization, Funding acquisition, Methodology, Formal analysis, Supervision, Writing—Review & editing. ACKNOWLEDGMENTS We would like to thank the families who participated in these studies and Naomi Batarse, Ana Correa Avila Robb, and Marissa Garcia for assistance with data collection and coding. FUNDING STATEMENT This work was supported by a grant from the Sanford Institute for Empathy and Compassion and start-up funding provided by UC San Diego to Lindsey Powell, and a Predoctoral Ford Fellowship awarded to Alexis Smith-Flores. DATA AVAILABILITY STATEMENT This study was preregistered before data collection began. The research plan (including initial hypotheses and the analysis plan), materials and stimuli, data, and code used in the analysis are available on OSF: https://osf.io/d57ny. OPEN MIND: Discoveries in Cognitive Science 851 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. REFERENCES Afshordi, N., & Liberman, Z. (2021). Keeping friends in mind: Devel- reactions. Developmental Psychology, 27(5), 858–866. https:// opment of friendship concepts in early childhood. Social Devel- doi.org/10.1037/0012-1649.27.5.858 opment, 30(2), 331–342. https://doi.org/10.1111/sode.12493 Feiman, R., Carey, S., & Cushman, F. (2015). Infants’ representa- Batson, C. D. (2009). These things called empathy: Eight related but tions of others’ goals: Representing approach over avoidance. distinct phenomena. In The social neuroscience of empathy Cognition, 136,204–214. https://doi.org/10.1016/j.cognition (pp. 3–15). MIT Press. .2014.10.007, PubMed: 25498746 Batson,C.D.(2013). Of Empathy-Induced Altruism. Social Psy- Gable, S. L., Reis, H. T., Impett, E. A., & Asher, E. R. (2004). What chology and Economics, 281. do you do when things go right? The intrapersonal and interper- Batson, C. D., Ahmad, N., & Lishner, D. A. (2009). Empathy and sonal benefits of sharing positive events. Journal of Personality altruism. In Oxford handbook of positive psychology (2nd ed., and Social Psychology, 87(2), 228–245. https://doi.org/10.1037 pp. 417–426). Oxford University Press. https://doi.org/10.1093 /0022-3514.87.2.228, PubMed: 15301629 /oxfordhb/9780195187243.013.0039 Geraci, A., Rigo, P., Simonelli, A., Di Nuovo, S., & Simion, F. Brown, C. L., & Fredrickson, B. L. (2021). Characteristics and con- (2021). Preschoolers’ evaluations of comforting actions towards sequences of co-experienced positive affect: Understanding the third parties in different relationship contexts. Journal of Applied origins of social skills, social bonds, and caring, healthy commu- Developmental Psychology, 76, 101315. https://doi.org/10.1016 nities. Current Opinion in Behavioral Sciences, 39,58–63. /j.appdev.2021.101315 https://doi.org/10.1016/j.cobeha.2021.02.002 Hamlin, J. K., Ullman, T., Tenenbaum, J., Goodman, N., & Baker, C. Bruneau, E. G., Cikara, M., & Saxe, R. (2017). Parochial empathy (2013). The mentalistic basis of core social cognition: Experi- predicts reduced altruism and the endorsement of passive harm. ments in preverbal infants and a computational model. Develop- Social Psychological and Personality Science, 8(8), 934–942. mental Science, 16(2), 209–226. https://doi.org/10.1111/desc https://doi.org/10.1177/1948550617693064 , PubMed: .12017, PubMed: 23432831 29276575 Hare, M., & Parent, J. (2022). Child emotional competence: A uni- Cikara, M., & Fiske, S. T. (2012). Stereotypes and Schadenfreude: fied framework and assessment review of emotion reasoning, Affective and physiological markers of pleasure at outgroup mis- emotion stability, emotion regulation, and empathy. PsyArXiv. fortunes. Social Psychological and Personality Science, 3(1), https://doi.org/10.31234/osf.io/582wg 63–71. https://doi.org/10.1177/1948550611409245, PubMed: Hare, M., Parent, J., & Trucco, E. (2022). The Emotional Compe- 24416471 tence in Children (ECIC) Scale: A parent-report measure. PsyArXiv. Cramer, D. (2003). Facilitativeness, conflict, demand for approval, https://doi.org/10.31234/osf.io/snp5m self-esteem, and satisfaction with romantic relationships. The Hareli, S., & Hess, U. (2010). What emotional reactions can tell us Journal of Psychology, 137(1), 85–98. https://doi.org/10.1080 about the nature of others: An appraisal perspective on person /00223980309600601, PubMed: 12661706 perception. Cognition and Emotion, 24(1), 128–140. https://doi Cramer, D., & Jowett, S. (2010). Perceived empathy, accurate .org/10.1080/02699930802613828 empathy and relationship satisfaction in heterosexual couples. Heyman, G. D., & Gelman, S. A. (1998). Young children use motive Journal of Social and Personal Relationships, 27(3), 327–349. information to make trait inferences. Developmental Psychology, https://doi.org/10.1177/0265407509348384 34(2), 310–321. https://doi.org/10.1037/0012-1649.34.2.310, DeJesus, J. M., Rhodes, M., & Kinzler, K. D. (2014). Evaluations PubMed: 9541783 versus expectations: Children’s divergent beliefs about resource Jern, A., & Kemp, C. (2014). Reasoning about social choices and distribution. Cognitive Science, 38(1), 178–193. https://doi.org social relationships. Proceedings of the Annual Meeting of the /10.1111/cogs.12093, PubMed: 24117730 Cognitive Science Society, 36(36), 689–694. https://escholarship Denham, S. A., Blair, K. A., DeMulder, E., Levitas, J., Sawyer, K., .org/content/qt7752k71z/qt7752k71z.pdf Auerbach-Major, S., & Queenan, P. (2003). Preschool emotional Jin, K.-S., & Baillargeon, R. (2017). Infants possess an abstract competence: Pathway to social competence? Child Develop- expectation of ingroup support. Proceedings of the National ment, 74(1), 238–256. https://doi.org/10.1111/1467-8624 Academy of Sciences, U.S.A., 114(31), 8199–8204. https://doi .00533, PubMed: 12625448 .org/10.1073/pnas.1706286114, PubMed: 28716902 Earp, B. D., McLoughlin, K. L., Monrad, J. T., Clark, M. S., & Crockett, Kalish, C. W., & Shiverick, S. M. (2004). Children’s reasoning about M. J. (2021). How social relationships shape moral wrongness norms and traits as motives for behavior. Cognitive Develop- judgments. Nature Communications, 12(1), 5776. https://doi.org ment, 19(3), 401–416. https://doi.org/10.1016/j.cogdev.2004 /10.1038/s41467-021-26067-4, PubMed: 34599174 .05.004 Eisenberg,N., Fabes, R. A.,Carlo,G., Speer, A. L.,Switzer,G., Kim, S. S., Kaplowitz, S., & Johnston, M. V. (2004). The effects of Karbon, M., & Troyer, D. (1993). The relations of empathy- physician empathy on patient satisfaction and compliance. Eval- related emotions and maternal practices to children s comforting uation & the Health Professions, 27(3), 237–251. https://doi.org behavior. Journal of Experimental Child Psychology, 55(2), /10.1177/0163278704267037, PubMed: 15312283 131–150. https://doi.org/10.1006/jecp.1993.1007, PubMed: Knutson, B. (1996). Facial expressions of emotion influence inter- 8501423 personal trait inferences. Journal of Nonverbal Behavior, 20, Ensor, R., Spencer, D., & Hughes, C. (2011). ‘You feel sad?’ Emotion 165–182. https://doi.org/10.1007/BF02281954 understanding mediates effects of verbal ability and mother– Krems, J. A., Hahnel-Peeters, R. K., Merrie, L. A., Williams, K. E. G., child mutuality on prosocial behaviors: Findings from 2 years & Sznycer, D. (2023). Sometimes we want vicious friends: to 4 years. Social Development, 20(1), 93–110. https://doi.org People have nuanced preferences for how they want their friends /10.1111/j.1467-9507.2009.00572.x to behave toward them versus others. Evolution and Human Fabes, R. A., Eisenberg, N., Nyman, M., & Michealieu, Q. (1991). Behavior, 44(2), 88–98. https://doi.org/10.1016/j.evolhumbehav Young children’s appraisals of others’ spontaneous emotional .2023.02.008 OPEN MIND: Discoveries in Cognitive Science 852 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. Liberman, Z., & Shaw, A. (2017). Children use partial resource Ruba, A. L., Meltzoff, A. N., & Repacholi, B. M. (2019). How do sharing as a cue to friendship. Journal of Experimental Child Psy- you feel? Preverbal infants match negative emotions to events. chology, 159,96–109. https://doi.org/10.1016/j.jecp.2017.02 Developmental Psychology, 55(6), 1138–1149. https://doi.org .002, PubMed: 28285046 /10.1037/dev0000711, PubMed: 30829508 Manstead, A. S. R., & Fischer, A. H. (2001). Social appraisal: The Ruba, A. L., Meltzoff, A. N., & Repacholi, B. M. (2020). The social world as object of and influence on appraisal processes. In development of negative event-emotion matching in infancy: K. R. Scherer, A. Schorr, & T. Johnstone (Eds.), Appraisal pro- Implications for theories in affective science. Affective Science, cesses in emotion: Theory, methods, research (pp. 221–232). 1(1), 4–19. https://doi.org/10.1007/s42761-020-00005-x, Oxford: Oxford University Press. https://orca.cardiff.ac.uk PubMed: 36042945 /34660/ Ruba, A. L., & Pollak, S. D. (2020). The development of emotion Marshall, J., Gollwitzer, A., Mermin-Bunnell, K., Shinomiya, M., reasoning in infancy and early childhood. Annual Review of Retelsdorf, J., & Bloom, P. (2022). How development and culture Developmental Psychology, 2(1), 503–531. https://doi.org/10 shape intuitions about prosocial obligations. Journal of Experi- .1146/annurev-devpsych-060320-102556 mental Psychology: General, 151(8), 1866–1882. https://doi.org Saxe, R., & Houlihan, S. D. (2017). Formalizing emotion concepts /10.1037/xge0001136, PubMed: 35130013 within a Bayesian model of theory of mind. Current Opinion in Masten, C. L., Gillen-O’Neel, C., & Brown, C. S. (2010). Children’s Psychology, 17,15–21. https://doi.org/10.1016/j.copsyc.2017.04 intergroup empathic processing: The roles of novel ingroup iden- .019, PubMed: 28950962 tification, situational distress, and social anxiety. Journal of Exper- Shamay-Tsoory, S. G., Ahronberg-Kirschenbaum, D., & Bauminger- imental Child Psychology, 106(2–3), 115–128. https://doi.org/10 Zviely, N. (2014). There is no joy like malicious joy: Schaden- .1016/j.jecp.2010.01.002, PubMed: 20202649 freude in young children. PLOS ONE, 9(7), e100233. https:// Moors, A. (2013). On the causal role of appraisal in emotion. doi.org/10.1371/journal.pone.0100233, PubMed: 24988446 Emotion Review, 5(2), 132–140. https://doi.org/10.1177 Shtulman, A., & Phillips, J. (2018). Differentiating “could” from /1754073912463601 “should”: Developmental changes in modal cognition. Journal Morelli, S. A., Lieberman, M. D., & Zaki, J. (2015). The of Experimental Child Psychology, 165,161–182. https://doi emerging study of positive empathy. Social and Personality Psy- .org/10.1016/j.jecp.2017.05.012, PubMed: 28648467 chology Compass, 9(2), 57–68. https://doi.org/10.1111/spc3 Siemer, M., Mauss, I., & Gross, J. J. (2007). Same situation–different .12157 emotions: How appraisals shape our emotions. Emotion, 7(3), Oatley, K., & Johnson-Laird, P. N. (2014). Cognitive approaches to 592–600. https://doi.org/10.1037/1528-3542.7.3.592, PubMed: emotions. Trends in Cognitive Sciences, 18(3), 134–140. https:// 17683215 doi.org/10.1016/j.tics.2013.12.004, PubMed: 24389368 Skerry, A. E., & Spelke, E. S. (2014). Preverbal infants identify emo- Olson, K. R., & Spelke, E. S. (2008). Foundations of cooperation in tional reactions that are incongruent with goal outcomes. Cogni- young children. Cognition, 108(1), 222–231. https://doi.org/10 tion, 130(2), 204–216. https://doi.org/10.1016/j.cognition.2013 .1016/j.cognition.2007.12.003, PubMed: 18226808 .11.002, PubMed: 24321623 Ong, D. C., Zaki, J., & Goodman, N. D. (2015). Affective cognition: Smith, C. E., Blake, P. R., & Harris, P. L. (2013). I should but i won’t: Exploring lay theories of emotion. Cognition, 143, 141–162. Why Young Children endorse norms of fair sharing but do not https://doi.org/10.1016/j.cognition.2015.06.010,PubMed: follow them. PLOS ONE, 8(3), e59510. https://doi.org/10.1371 26160501 /journal.pone.0059510, PubMed: 23527210 Ong, D. C., Zaki, J., & Goodman, N. D. (2019). Computational Smith, R. H., Powell, C. A. J., Combs, D. J. Y., & Schurtz, D. R. models of emotion inference in theory of mind: A review and (2009). Exploring the when and why of schadenfreude. Social roadmap. Topics in Cognitive Science, 11(2), 338–357. https:// and Personality Psychology Compass, 3(4), 530–546. https://doi doi.org/10.1111/tops.12371, PubMed: 30066475 .org/10.1111/j.1751-9004.2009.00181.x Paulus, M., Wörle, M., & Christner, N. (2020). The emergence of Smith, R. H., & van Dijk, W. W. (2018). Schadenfreude and human altruism: Preschool children develop a norm for gluckschmerz. Emotion Review, 10(4), 293–304. https://doi.org empathy-based comforting. Journal of Cognition and Develop- /10.1177/1754073918765657 ment, 21(1), 104–124. https://doi.org/10.1080/15248372.2019 Smith-Flores, A. S., & Feigenson, L. (2021). Preschoolers represent .1693375 others’ false beliefs about emotions. Cognitive Development, 59, Pietraszewski, D. (2021). Toward a computational theory of social 101081. https://doi.org/10.1016/j.cogdev.2021.101081 groups: A finite set of cognitive primitives for representing Smith-Flores, A. S., & Feigenson, L. (2022). “Yay! Yuck!” toddlers any and all social groups in the context of conflict. Behavioral use others’ emotional responses to reason about hidden and Brain Sciences, 45, e97. https://doi.org/10.1017 objects. Journal of Experimental Child Psychology, 221, /S0140525X21000583, PubMed: 33902764 105464. https://doi.org/10.1016/j.jecp.2022.105464,PubMed: Powell, C. A. J., & Smith, R. H. (2013). Schadenfreude caused by 35644055 the exposure of hypocrisy in others. Self and Identity, 12(4), Smith-Flores, A. S., Herrera-Guevara, I. A., & Powell, L. J. (2023). 413–431. https://doi.org/10.1080/15298868.2012.687013 Infants expect friends, but not rivals, to be happy for each other Powell, L. J. (2022). Adopted utility calculus: Origins of a concept when they succeed. Developmental Science, e13423. https://doi of social affiliation. Perspectives on Psychological Science, 17(5), .org/10.1111/desc.13423, PubMed: 37312424 1215–1233. https://doi.org/10.1177/17456916211048487, Smith-Flores, A. S., & Powell, L. J. (2023). Joint reasoning about PubMed: 35549492 social affiliation and emotion. Nature Reviews Psychology, 1, Pun, A., Birch, S. A. J., & Baron, A. S. (2021). The power of allies: 374–383. https://doi.org/10.1038/s44159-023-00181-0 Infants’ expectations of social obligations during intergroup con- Taylor, R. D., Oberle, E., Durlak, J. A., & Weissberg, R. P. (2017). flict. Cognition, 211, 104630. https://doi.org/10.1016/j.cognition Promoting positive youth development through school-based .2021.104630, PubMed: 33636572 social and emotional learning interventions: A meta-analysis of OPEN MIND: Discoveries in Cognitive Science 853 Children’s Reasoning About Empathy and Social Relationships Smith-Flores et al. follow-up effects. Child Development, 88(4), 1156–1171. https:// Psychological Review, 122(3), 411–428. https://doi.org/10.1037 doi.org/10.1111/cdev.12864, PubMed: 28685826 /a0039252, PubMed: 25961468 Teding van Berkhout, E., & Malouff, J. M. (2016). The efficacy of Wu, Y., Baker, C. L., Tenenbaum, J. B., & Schulz, L. E. (2018). empathy training: A meta-analysis of randomized controlled Rational inference of beliefs and desires from emotional expres- trials. Journal of Counseling Psychology, 63(1), 32–41. https:// sions. Cognitive Science, 42(3), 850–884. https://doi.org/10.1111 doi.org/10.1037/cou0000093, PubMed: 26191979 /cogs.12548, PubMed: 28986938 Wang, Y. A., & Todd, A. R. (2021). Evaluations of empathizers Wu, Y., Muentener, P., & Schulz, L. E. (2017). One- to depend on the target of empathy. Journal of Personality and four-year-olds connect diverse positive emotional vocaliza- Social Psychology, 121(5), 1005–1028. https://doi.org/10.1037 tions to their probable causes. Proceedings of the National /pspi0000341, PubMed: 32940513 Academy of Sciences, U.S.A., 114(45), 11896–11901. Wellman, H. M., Phillips, A. T., & Rodriguez, T. (2000). Young chil- https://doi.org/10.1073/pnas.1707715114, PubMed: dren’s understanding of perception, desire, and emotion. Child 29078315 Development, 71(4), 895–912. https://doi.org/10.1111/1467 Wu, Y., & Schulz, L. E. (2018). Inferring beliefs and desires from -8624.00198, PubMed: 11016555 emotional reactions to anticipated and observed events. Child Wondra, J. D., & Ellsworth, P. C. (2015). An appraisal theory of Development, 89(2), 649–662. https://doi.org/10.1111/cdev empathy and other vicarious emotional experiences. .12759, PubMed: 28271499 OPEN MIND: Discoveries in Cognitive Science 854

Journal

Open MindMIT Press

Published: Oct 27, 2023

There are no references for this article.