Dimensions of wisdom perception across twelve countries on five continents

The study addresses whether perceptions of wisdom are culturally variable or share common dimensions across societies. Drawing on diverse philosophical traditions that variously emphasize analytic reflection versus socio-emotional attunement, the authors note competing predictions: individualist cultures may prioritize reflective/analytic skills, whereas collectivist cultures may prioritize socio-emotional competencies. Conversely, work in cognitive, developmental, and social psychology suggests potential cross-cultural convergence in social judgment and mind perception, often structured by agency/competence and experience/communion dimensions. The research asks: What latent dimensions guide evaluations of wisdom-related characteristics in others and oneself? Are these dimensions invariant across cultures, and how do they relate to explicit attributions of wisdom, knowledge, and understanding? The focus is on judgments under uncertainty—contexts where wisdom is typically invoked.

Prior research on mind perception has identified one, two, or three dimensions (e.g., agency/intentional control and experience/consciousness; sometimes adding a social/physiological dimension), largely in WEIRD samples. Limited cross-cultural work suggests both similarities (e.g., differentiating mind- vs. body-related capacities) and culture-specific nuances (e.g., heart/socio-emotional dimensions, separate social relationships in Fiji). Social judgment research similarly proposes general dimensions (agency/competence and communion/warmth) that may organize person perception globally, though most studies focus on stereotypes of groups rather than concrete decision contexts. Scholarship on lay conceptions of wisdom shows mixed evidence about Western emphasis on cognitive traits versus inclusion of socio-emotional traits. There are also reasons to expect that judgments about wisdom differ from judgments about general mental states or group stereotypes. Finally, even with invariant dimensions, cultures may differ in how these dimensions combine (e.g., one dimension necessary, others secondary) when attributing wisdom.

Design and instrument: Participants compared ten human targets (self, scientist, doctor, teacher, fair person, politician, religious person, and 75-, 45-, and 12-year-old) in pairwise fashion regarding the likelihood of engaging in 19 behaviors relevant to making a difficult choice without a clear right or wrong answer (e.g., think logically; care for others’ feelings). Items covered mind, heart, and body themes (e.g., metacognition, emotion regulation, perspective-taking, humility, attention to emotions and bodily expressions, nature/divinity). A five-point comparative scale ranged from “much less likely” to “much more likely,” with “equally likely” as the midpoint. After comparisons, participants provided explicit 5-point ratings of each target’s wisdom, knowledge, and understanding (order randomized). To reduce fatigue, each participant was randomly assigned one reference target to compare against all others. Materials included culturally specific names and brief role descriptions for each target. Translations underwent expert-led adaptation and back-translation across sites, with discussions to align key terms (e.g., “understanding”). Sampling and sites: Convenience samples from 16 sites across 12 countries on five continents (11 languages) were collected between 2019 and 2021 via Qualtrics or paper-and-pencil (Slovakia, Morocco). Samples included university students (Canada, Ecuador, Peru, US) and broader community samples elsewhere; Japan and two Indian samples used a shortened version (five targets). Total N ≈ 2650. For multilevel modeling power, sites were grouped into eight cultural regions: India; China; Korea and Japan; South Africa; Morocco; Slovakia; North America (Canada, US); South America (Ecuador, Peru). Demographics varied by site (pooled mean age ≈ 30; 55% female; religiosity varied widely). Ethics: Approved by institutional ethics boards at all sites; informed consent obtained. Data handling and analysis: Pairwise design reduces acquiescent response bias. Within-person level captured comparisons of one reference with multiple comparison targets; between-person level captured differences across individuals. Missing data were handled via full information maximum likelihood; comparisons of a target to itself were imputed as “equally likely.” Explicit ratings were transformed to pairwise differences aligned with the comparison structure; variables were grand-mean centered and standardized for pooled/region models. Three items were excluded from focal models due to conceptual/measurement issues (one inaction item, one reverse-coded pride item, one bodily tension item); analyses with all 19 items produced similar results. Modeling strategy: Multilevel exploratory and confirmatory factor analyses identified latent dimensions. Isomorphic models (equal loadings across within- and between-person levels) were preferred for parsimony after fit comparisons (e.g., isomorphic CFI=0.956, RMSEA=0.022 vs. non-isomorphic CFI=0.963, RMSEA=0.021; SRMRwithin=0.028/0.024; SRMRbetween=0.082/0.036). A method factor for response tendencies improved fit negligibly and was omitted. Multigroup-multilevel CFA tested measurement invariance across eight cultural regions; partial metric invariance was supported. Multilevel SEMs (MLR; Bayesian for latent interactions) related latent dimensions to explicit wisdom, knowledgeability, and understanding ratings, controlling for participant covariates (age, gender, education, religiosity) and target gender. Software: Mplus 8.8; summaries in R 4.1.

- Two latent dimensions: Reflective Orientation (thinking before acting, logical and flexible thinking, recognizing change, emotion control, applying knowledge/experience) and Socio-Emotional Awareness (caring for others’ feelings, attending to own emotions, considering others’ perspectives, intellectual humility, reliance on neutral third parties, sense of humor, attention to nature/divinity, awareness of bodily expressions). Acceptable multigroup-multilevel CFA fit (e.g., CFI≈0.912, RMSEA≈0.033; SRMRwithin≈0.032; SRMRbetween≈0.078). - Invariance: Partial metric invariance across eight cultural regions; small fit differences between configural and metric models (ΔCFI=0.010; ΔRMSEA=0.008; ΔSRMRwithin=0.006; ΔSRMRbetween=0.007). Three items showed cross-cultural variability in loadings: attention to nature/divinity; others’ perspective; awareness of bodily expressions. - Association between dimensions: Positive correlation at the within-person level across regions, r=0.69 [0.66–0.71], t=63.0, p<0.001; stronger in East Asia and South Africa (rs≈0.75–0.88) than Americas and North Africa (rs≈0.33–0.77). - Links to explicit ratings (pooled zero-order): Wisdom—Reflective r=0.47 [0.46–0.49], Socio-Emotional r=0.23 [0.22–0.25] (difference=0.24, t=19.7, p<0.001). Knowledgeability—Reflective r=0.50 [0.49–0.52], Socio-Emotional r=0.21 [0.19–0.23] (difference=0.29, t=34.6, p<0.001). Understanding—Reflective r=0.43 [0.41–0.44], Socio-Emotional r=0.33 [0.31–0.35] (difference=0.10, t=8.06, p<0.001). - Multilevel SEM with latent interaction (Bayesian): Reflective positively predicted wisdom (b=0.51, 95% CI [0.48–0.53], p<0.001); Socio-Emotional negatively predicted wisdom (b=−0.15, [−0.18, −0.12], p<0.001); positive interaction (b=0.09, [0.08, 0.10], p<0.001). Socio-Emotional reduced wisdom at low/mid Reflective levels; highest wisdom attributed when both dimensions were high. - Cross-cultural stability of effects: Reflective positively predicted wisdom, knowledgeability, understanding in all regions (except non-significant for understanding in China). Socio-Emotional’s effect on wisdom was non-significant in North America, South America, Morocco; significant negative in five other regions. For knowledgeability, Socio-Emotional was negative in all regions except Morocco. For understanding, Socio-Emotional was negative in India and South Africa, positive in most other regions, non-significant in South America and Morocco. - Target profiles: 12-year-old lowest on both dimensions. Doctor and scientist highest on Reflective; fair person and teacher highest on Socio-Emotional; 75-year-old high on both. Target positions on Reflective were highly stable across cultures (avg intercorrelation r=0.97), but more variable on Socio-Emotional (avg r=0.81); mean difference r=0.16 (d=0.33). Higher-status targets consistently rated higher on Reflective, but inconsistently on Socio-Emotional. - Gender effects (targets): Female targets rated slightly lower on Reflective (b=−0.02, p=0.004), comparable on Socio-Emotional (b=0.01, p=0.156). - Self-views: Participants saw themselves as less reflective than six targets (ps<0.001), more reflective than religious person, 12-year-old, and 45-year-old; more socio-emotionally aware than six targets, but less than fair person and teacher; not different from 75-year-old. Across regions (except South Africa), self-rated understanding exceeded self-rated knowledgeability and wisdom. - Robustness: Excluding the 12-year-old or self-ratings minimally altered loadings; model fit decreased slightly; inter-factor correlation remained moderate-high (≈0.63–0.68).

The findings indicate that judgments of wisdom in difficult, ambiguity-laden decisions rely on two latent dimensions—Reflective Orientation and Socio-Emotional Awareness—that are largely invariant across diverse cultural regions. Reflective Orientation emerges as a primary, necessary foundation: after controlling for reflection, higher socio-emotional awareness alone can reduce perceived wisdom, especially at low reflection levels, suggesting that caring or emotion-driven responses without reflective control are not seen as wise. Conversely, when reflection is high, socio-emotional awareness contributes positively, yielding the highest wisdom attributions when both are strong. Cross-culturally, targets’ reflective standing is perceived consistently, whereas socio-emotional standing varies more, consistent with the idea that socio-emotional behaviors are more culturally scripted and context-dependent. The self-other asymmetry shows domain-specific self-enhancement: people rate themselves higher on socio-emotional qualities and lower on reflective ones compared to exemplars, echoing patterns in personality self-judgment (agreeableness vs. conscientiousness) and suggesting that socio-emotional traits may be valued and/or more subjectively defined. The results extend mind and social perception research beyond WEIRD contexts into decision-making scenarios typical of wisdom, and help clarify philosophical distinctions between theoretical (sophia) and practical (phronesis) wisdom: reflective competencies map onto theoretical wisdom more strongly, while socio-emotional competencies contribute to practical wisdom, particularly when built on a reflective base. The work suggests potential universals in the structure of wisdom perception while acknowledging culturally variable applications of socio-emotional criteria.

This study identifies two robust, cross-culturally invariant dimensions shaping wisdom perception—Reflective Orientation and Socio-Emotional Awareness—and shows how they combine to inform attributions of wisdom, knowledge, and understanding in concrete decision contexts. Reflective Orientation is a necessary foundation for perceived wisdom; socio-emotional awareness enhances wisdom mainly in tandem with strong reflection. Target profiles and self-assessments reveal consistent cross-cultural structure with culturally variable socio-emotional attributions. These insights advance folk psychology and social judgment research beyond the Global North and clarify how people globally conceptualize desirable cognitive and socio-emotional qualities in wise decision-making. Future research should test these dimensions with broader, mutually exclusive target sets, employ open-ended/natural language methods, examine domain-specific trust and decision outcomes linked to each dimension, and explore whether similar dimensions organize perceptions of moral exemplars.

- Convenience sampling across sites limits representativeness; some regional groupings (e.g., South Africa’s sub-samples; Japan vs. South Korea) may mask heterogeneity. - Limited, partly overlapping target set may inflate interdependence between dimensions; however, robustness checks with exclusions and random subsets yielded similar structures. - Questionnaire format and pairwise design may promote cross-cultural consistency; translation ambiguities (e.g., “understanding”) pose semantic challenges despite back-translation and expert review. - Only configural and metric invariance were testable at the within level; three items showed cross-cultural loading variability. - No preregistration; methodological shift from planned MDS to multilevel SEM occurred post hoc due to advantages for bias control and measurement modeling. - Some sites had smaller samples (e.g., indigenous/minority groups); merging into broad regions was necessary for power and may obscure site-level differences.