Cultural diversity in unequal societies sustained through cross-cultural competence and identity valuation

The study addresses when and how cultural diversity in a specific domain of coordination can be sustained within structurally unequal yet integrative societies—societies that allow cultural diversity while encouraging mutually beneficial inter-group interaction. Prevailing theories often imply that diversity cannot persist under such conditions because coordination requires convergence on a single norm. The paper proposes that cross-cultural competence (CCC)—the ability to coordinate using multiple groups’ norms—combined with valuation of in-group identity can sustain diversity. Motivated by ethnographic observations among minority Matsigenka and majority Mestizos in Amazonian Peru, the author formulates a model to examine whether, despite unequal power and incentives, diversity of coordination norms can persist without sacrificing inter-group coordination. The work is important for informing strategies to maintain minority cultural forms while enabling productive inter-ethnic engagement, countering political fears of cultural loss.

The paper situates its contribution within models of cultural change that typically preclude coordination across distinct norms, making persistent diversity incompatible with efficient interaction. Prior work explains mixed equilibria via migration, learning errors, strong identity benefits that outweigh cross-group coordination, limited assortment across groups, network fragmentation, or complementary (rather than correlative) coordination. However, these mechanisms often reduce inter-group coordination or rely on external inflows. Empirical and theoretical discussions of CCC span medicine, international business, and inter-ethnic relations, noting its prevalence and potential to reduce conflict and foster cohesion. The author highlights gaps: most models ignore CCC as a distinct developmental path (retaining in-group preference while learning out-group norms, or adopting out-group preference while retaining in-group knowledge). This omission fosters a perceived tradeoff between diversity and coordination. The study aims to escape this tradeoff by explicitly integrating CCC and identity valuation into the dynamics.

Empirical component: Field data from Amazonian Peru include minority indigenous Matsigenka (group S) and majority Mestizos (group L). Participants answered a norm vignette on inheritance division: need-based (norm 1) versus equal division (norm 2). Of 77 Matsigenka, 75% preferred norm 1; of 82 Mestizos, 68% preferred norm 2. A subset of 103 participants guessed the most common response in both in- and out-groups. Phenotypes were assigned: CCC phenotypes (1X, 2X) for individuals who preferred norm 1 or 2 and correctly guessed both groups’ modal norms; UCC phenotypes (11, 22) for those who preferred a norm and (incorrectly) believed both groups mostly preferred their own preferred norm. Bayesian item-response theory (IRT) models estimated phenotype frequencies and uncertainty (Appendices B.1.6–B.1.8). Phenotype distributions were examined overall and for subgroups with higher inter-ethnic engagement (e.g., Matsigenka educated with Mestizos; Mestizo employers of Matsigenka). Other observed phenotypes existed but were excluded from direct model comparison due to classification ambiguity.

Theoretical model: A population has minority group S and majority group L. Two mutually exclusive norms (1, 2) govern coordination. Each time step, individuals interact once; successful coordination yields payoff 1, miscoordination 0. Individuals interact with in-group with probability αx and out-group with probability 1−αx. CCC individuals can coordinate with either norm but incur learning cost m≥0 and, when coordinating on the non-preferred norm, cognitive dissonance cost c∈[0,1]. UCC individuals coordinate only with their preferred norm. Out-group coordination yields an extra benefit β≥0, with structural inequality represented as βS>βL (minority S gains more from successful out-group coordination and loses more when failing, reflecting lower disagreement point/bargaining power). Identity valuation adds payoff i≥0 to preferring a norm that is frequent in the in-group and rare in the out-group; this effect scales with perceived in-group frequency times perceived out-group rarity. Individuals cannot observe others’ preferred norms; they infer expected next-step coordination norms from current observed usage with in- and out-group members (community information/gossip). Using these perceptions, anticipated phenotype payoffs are computed (w̃) as functions of α, β, c, m, i, and perceived norm-use probabilities (Eqs. 3–6). Individuals who miscoordinate or pay c at time t update their phenotype for t+1 via a payoff-biased logistic decision with slope μ≥0 (norms cannot be forgotten).

Simplified dynamics: Under assumptions that all L are UCC with norm 2 and S has no remaining UCC preferring norm 1, dynamics reduce to a single difference equation (Eq. 7) describing transitions between S1X and S2X, with transition probability P (Eq. 8) depending on identity valuation i, dissonance cost c, and in-group interaction α, and F (Eq. 9) representing perceived norm-1 use in S. Structural power terms β drop out in this simplified system because all considered phenotypes can coordinate universally.

Prospects across generations: Although the main model is within-generation, a heuristic intergenerational criterion is derived: S1X parents prefer producing S1X over S22 when i > [m + c(1 − αF) − αF]/F (Eq. 10), highlighting sensitivity to learning cost m and identity valuation i. Parameter estimation plausibility checks compare model trajectories to cross-sectional phenotype frequencies; a more rigorous test requires longitudinal data.

• Field evidence shows substantial cross-cultural competence (CCC) on an Amazonian ethnic boundary, especially among individuals with greater inter-ethnic interaction. In the inheritance vignette, 75% of 77 Matsigenka preferred need-based division (norm 1) and 68% of 82 Mestizos preferred equal division (norm 2); CCC phenotypes (1X, 2X) occur at non-trivial frequencies in both groups.
• The model demonstrates that in a structurally unequal but integrative society, cultural diversity in a coordination domain can be sustained within a single generation when CCC coexists with sufficient identity-based valuation of the minority-typical norm. A disempowered minority can maintain its distinctive norm by learning the majority’s norm (developing CCC) while retaining preference for the minority norm.
• CCC in the minority effectively insulates minority norm preference dynamics from some group-level power differences; in simplified dynamics, structural inequality parameters (β) do not govern transitions between CCC phenotypes because universal coordination is possible.
• Sustaining diversity requires that identity valuation i offsets the cognitive dissonance cost c experienced by CCC individuals when coordinating on the non-preferred norm; higher learning costs m increase the threshold identity valuation needed to favor CCC across generations (Eq. 10).
• Model trajectories can reproduce cross-sectional differences in phenotype frequencies between subgroups with higher versus lower inter-ethnic interaction under ethnographically plausible parameters, suggesting the model’s plausibility.
• The framework avoids the classic tradeoff in prior models between cultural diversity and inter-group coordination by incorporating CCC, yielding stable mixed equilibria with universal coordination.

The research question—whether cultural diversity can persist under intense, unequal, yet mutually beneficial inter-group interaction—is addressed by combining empirical indications of CCC with a model that includes CCC and identity valuation. Findings show that a minority can strategically learn majority norms to improve coordination while maintaining a preference for its own norms if identity valuation is sufficiently strong. This dynamic reduces the influence of structural power asymmetries on norm preference change among the minority and supports stable mixed equilibria with full coordination. Practically, the model suggests interventions: develop CCC (particularly in the low-power group) and bolster identity-based valuation and norm distinctiveness (e.g., cultural institutions, celebrations) to sustain diversity without hindering integration. The results challenge assumptions that minority identity valuation threatens majority culture; instead, it can characterize CCC individuals who are highly engaged in integrative interaction. By replacing fears of inevitable cultural loss with conditions for durable diversity, the approach may create social space to tackle structural inequalities more directly.

The paper contributes an integrated empirical-theoretical account showing how cultural diversity can be sustained in unequal, integrative societies through cross-cultural competence and in-group identity valuation. Empirical data document prevalent CCC, and the model explains how minority groups can maintain distinctive norms by learning majority norms while retaining preference for their own, producing stable mixed equilibria with universal coordination and diminished sensitivity to power asymmetries. Future research should develop age-structured and network-informed models, incorporate migration and signaling, allow for non-random forgetting and preference falsification, and consider continuous norms and linkage across domains. Longitudinal empirical studies across multiple time points and contexts are needed to test dynamic predictions and refine parameter estimates.

• The model is within-generation and omits explicit demographic processes; intergenerational dynamics are only heuristically addressed and require age-structured modeling of marriage, socialization, and learning.
• Many simplifying assumptions may limit generality: binary discrete norms, additive payoffs, fixed interaction probabilities, inability to forget norms (or specific forgetting assumptions), and random assortment within groups.
• Social network structure, migration between groups, signaling of norms via overt/covert markers, linkage across multiple norms, and preference falsification are not fully modeled; these factors could materially affect dynamics.
• Empirical evaluation relies on cross-sectional data from one context and on an absolute CCC measure derived from a single vignette; a rigorous test requires longitudinal data, broader measures of CCC, and validation across populations.
• Estimated phenotype frequencies exclude several observed phenotypes that are hard to classify as CCC or UCC, which may reflect measurement artifacts or theoretical gaps.