Urbanized knowledge syndrome—erosion of diversity and systems thinking in urbanites’ mental models

The study investigates how urbanization along U.S. Northeast coastlines relates to residents’ ecological knowledge and systems thinking about human–environment interactions. Urban coastal development and shoreline armoring degrade habitats and ecosystem services, potentially weakening people’s connectedness to nature and ecological understanding. The authors hypothesize: (H1) residents in more urbanized areas exhibit more linear and less systems thinking in their mental models of coastal SESs; and (H2) increased urbanization is associated with homogenization of mental models (reduced cognitive diversity). The purpose is to empirically assess whether urbanization correlates with simplified, less diverse perceptions of SES interdependencies and whether such perceptions relate to lower self-reported pro-environmental behaviors, with implications for urban sustainability.

The introduction situates the work within literature on urbanization’s ecological impacts (biodiversity loss, pollution, impervious surfaces) and reduced ecosystem functioning and resilience in urban areas. Prior research suggests that diminished exposure to natural environments in cities can reduce environmental connectedness and ecological knowledge, potentially lowering pro-environmental behaviors. Cultural and physical homogenization in cities—driven by centralized governance, standardized media and education, and similar built environments—extends to ecological homogenization across urban areas. The authors argue this homogenization may also affect cognition, leading to more anthropocentric frames, human exceptionalism, and limited appreciation of reciprocal human–nature interdependencies. While many studies theorize these relationships, few empirically test how urbanization shapes mental models; this study addresses that gap using fuzzy cognitive mapping and graph/network analysis.

Design and sample: A cross-sectional survey of 1,397 adults residing in shoreline counties across eight U.S. Northeast states (ME, NH, MA, RI, CT, NY, NJ, DE) was conducted via Qualtrics Research Panels. After data quality filters (attention checks, speeders threshold at 50% of mean completion time, manual review) and excluding sparse FCMs (≤5 edges; bottom 10th percentile), 1,226 respondents’ fuzzy cognitive maps (FCMs) were analyzed. IRB approval obtained (Northeastern University); informed consent collected.

FCM elicitation: Participants selected relevant coastal system concepts from a provided list (Marine life; Water Quality; Marshes & Natural Habitats; Recreational & Cultural Activities; Beaches; Seawalls & Engineered Shorelines; Protection from Storms; Water Access; Commercial Fisheries & Livelihoods) and could add one concept. They then specified pairwise causal influences among selected concepts with signed, weighted edges (−1 to +1, magnitude 0–1), producing directed, weighted graphs (FCMs).

Context variables: County-level urbanization was appended using the NCHS Urban–Rural Classification Scheme (six categories; two nonmetro, four metro including large central metro), and county-level percent armored shoreline from NOAA ESI (Gittman et al., 2015).

Clustering mental models: Each FCM was transformed into a vector of general structural metrics (e.g., number of nodes/edges, sum of absolute edge weights, concept centralities, density, counts of drivers/receivers/ordinary nodes, receivers-to-drivers ratio). Hierarchical clustering using Ward’s minimum variance on Euclidean distances identified clusters of mental models.

Systems vs. linear thinking metrics: Systems thinking indicators included (1) Complexity score (receivers/drivers; proxy for trade-offs/synergies), (2) Cycle basis size (emergence from local relationships), (3) Reciprocal motifs (bidirectional dyads), and (4) Feedback motifs (triadic feedback loops). Linear thinking indicators adapted from Krackhardt (1994) were (1) Connectedness, (2) Hierarchy (unidirectional adjacency), (3) Efficiency, and (4) Least-upper boundedness (LUBedness). Independent-samples t-tests compared indicators across clusters with Bonferroni correction (α = 0.00625 for eight tests).

Cognitive distance (homogenization): Pairwise distances between FCMs combined (i) Jaccard distance on unweighted adjacency matrices and (ii) Euclidean squared distance between spectra (eigenvalues) of normalized Laplacians. Cognitive distance CD = d_E × d_J × ρ, normalized to [0,1]. Mean within-cluster pairwise distances were compared via t-tests.

Pro-environmental behavior: Participants self-reported adoption (yes/no) of six behaviors (e.g., donations to conservation, product boycotts for environmental reasons, attending public environmental meetings, contacting agencies, voting influenced by environmental positions, behavior change due to environmental concern). Odds ratios (Cluster-0 relative to Cluster-1) with 95% CIs were estimated; significance inferred when CI excluded OR = 1.

• Sample and clustering: Of 1,397 respondents, 1,226 FCMs were analyzed. Two distinct clusters of mental models emerged.
• Demographics: No significant differences across clusters for education, income, home ownership, political affiliation, gender; differences for race and age: race χ²(6, N=1226)=23.804, p<0.001 (Cluster-1 had smaller white proportion and larger Black/African American and Asian proportions); age t(1224)=7.211, p<0.001 with Cluster-1 younger (M=41.84, SD=13.83) than Cluster-0 (M=47.49, SD=13.61).
• Urbanization and shoreline armoring: Mental model clusters associated with NCHS urbanization, χ²(4, N=1226)=26.46, p<0.001; Cluster-1 overrepresented large central metros (level 6) and underrepresented small metros (level 3). Percent armored shorelines higher in Cluster-1 than Cluster-0, t(1224)=3.044, p=0.002.
• Systems vs. linear thinking: All four systems thinking indicators were significantly higher in Cluster-0 than Cluster-1 (Bonferroni-adjusted α=0.00625). Three of four linear thinking indicators were significantly higher in Cluster-1 than Cluster-0; the fourth (Hierarchy) was not significant. Interpretation: Urban-associated Cluster-1 shows less systems thinking and more linear thinking.
• Homogenization (cognitive diversity): Mean pairwise cognitive distance within Cluster-1 was significantly smaller than within Cluster-0 (p<0.001), indicating more homogeneous mental models among residents in more urbanized areas.
• Pro-environmental behaviors: Odds of self-reported adoption were higher in Cluster-0 relative to Cluster-1 across items; three of six behaviors showed statistically significant higher odds (95% CI excluding OR=1) for Cluster-0. Overall, stronger UKS symptoms in Cluster-1 correspond to lower reported pro-environmental actions.

The study provides empirical support that urbanization correlates with simplified and homogenized mental models of coastal social–ecological systems. Individuals in more urban settings (Cluster-1) demonstrated less systems thinking—fewer trade-offs/synergies, cycles, reciprocity, and feedbacks—and greater linearity, alongside more homogeneous cognitive structures. These cognitive patterns are associated with lower odds of several pro-environmental behaviors. The findings suggest that urbanization may foster an Urbanized Knowledge Syndrome (UKS): erosion of systems thinking and cognitive diversity that could limit residents’ ability to perceive complex human–nature interdependencies, potentially leading to decisions that inadvertently degrade ecosystems and reduce community resilience. The authors posit reinforcing feedbacks: diminished exposure to functioning natural systems weakens ecological knowledge and connectedness, influencing behaviors and preferences that contribute to further environmental degradation, which in turn may further erode knowledge and stewardship. Recognizing these dynamics is pivotal for urban sustainability strategies that cultivate richer, more diverse ecological mental models and support stewardship behaviors.

This work introduces and empirically supports the Urbanized Knowledge Syndrome concept, showing that greater urbanization is associated with more linear and homogenized mental models of coastal SESs and with lower self-reported pro-environmental behaviors. The study contributes methodological advances by combining fuzzy cognitive mapping with graph-theoretic and spectral measures to quantify systems thinking and cognitive diversity at scale. To counter UKS and enhance urban sustainability, the authors propose: (1) polycentric, nested governance that increases institutional diversity and local decision heterogeneity; (2) strengthening human–nature connections via equitable provision of high-quality green spaces and nature-based solutions grounded in biophilia; and (3) promoting adaptive learning and ecological knowledge (e.g., citizen science, civic ecology) to build systems thinking. Future research should test causal pathways, evaluate interventions that expand systems thinking and cognitive diversity, and investigate how changes in urban ecological infrastructure and governance affect mental models and behaviors over time.

The study is observational and does not identify causal directions between urbanization, mental models, and behaviors. Cross-sectional design and reliance on self-reported behaviors limit causal inference and may introduce reporting biases. The analysis focuses on county-level urbanization and armored shoreline metrics and does not incorporate individuals’ perceived urban–rural gradient or proximity to specific green/blue amenities. While demographic differences were mostly nonsignificant across clusters, unmeasured confounders may exist. The FCM approach, while powerful, is constrained by the provided concept list and respondent engagement, and clustering on structural metrics may overlook content-specific nuances.