Causal Models, Creativity, and Diversity

The paper addresses whether causal models can explain life’s creativity and diversity across domains from physics and biology to cities and the humanities. Motivated by historical reflections (e.g., Gottfried Semper’s attempt to explain stylistic diversity after the 1851 Great Exhibition and his use of mathematical thinking), the author proposes that causality, expressed mathematically as functions transforming past inputs into present outcomes, can unify explanations of surprising, creative, and diverse phenomena. The work seeks a general, testable framework that captures both the generation of novelty and the emergence of diversity across cultural and natural systems.

The article situates its framework within a broad tradition: Newton’s causal differential equations; Semper’s 19th-century ideas on style and causes; Hume’s notions of relatedness; perturbation theory and quasispecies evolution (Eigen, 1971) as additive creative processes; Simpson’s diversity index (Simpson, 1949) for quantifying diversity; logistic growth (Quetelet, Verhulst) and Lotka–Volterra dynamics for interplay and waves; Ross’s generalized Lotka–Volterra formulation; and the replicator–mutator equation (Hadeler, Bomze & Burger, Page & Nowak) unifying selection (play) and mutation (creativity). It also references applications and allied concepts in epidemiology, ecology, urbanism, AI (neural network layering as additive architecture), and constructal theory (Bejan) to underscore pan-disciplinary relevance.

The study develops a unified causal modeling framework using differential equations: (1) A universal causal statement: anything new is caused by the past, represented as continuous change (differential equations). (2) Creativity via addition: inserting additions yields linear combinations where causes add up and are weighted by coefficients, producing causal flows and relatedness among variables. This backbone encompasses perturbation theory, quasispecies, variation–selection, and architectures analogous to neural network layers. It leads to emergence of tightly linked creative groups (styles, genres, mutant swarms, clusters). (3) Diversity via multiplication (interplay): inserting multiplications models meetings and conjoint probabilities, inspired by Simpson’s index. Basic interplay corresponds to Lotka–Volterra dynamics producing waves (epidemics, fashions, trends) and can drive diversification through predation/consumption patterns and escape in physical/evolutionary space. (4) General interplay model: xi = xi f(x1, x2, …, xi), equivalent to generalized Lotka–Volterra and foundational for compartmental and game-theoretic models. (5) Creative play: coupling addition (creativity) and multiplication (interplay) yields the replicator–mutator framework, uniting mutation/innovation with selection/play. (6) Multilevel models: interactions among subcategories and overarching categories generate short waves and longer cycles (growth and reform), formalized with systems including variables for habituation/boredom at multiple levels. (7) Empirical demonstrations: diversity maps using Simpson’s index; clustering content into creative groups; analysis of term frequencies over centuries (Google Ngrams) for science and subfields; urban diversity mapping (Sassi di Matera) across time. Equations for basic models are provided, with coefficients representing cooperation, defection, or consumption, and annotated versions referenced in prior work.

• Additions model creativity: when causes add, causal flows link items, creating relatedness and forming creative groups (styles, genres, lineages, mutant swarms, clusters). - Multiplications model interplay/diversity: meetings and joint effects produce waves (slow–fast cycles) and can foster diversification via escape and coexistence. - Unified model (replicator–mutator) alternates between creativity and interplay, explaining transitions from stable building blocks to adaptive creative groups to unstable, system-level play. - Diversity radiates from centers of density: analysis of US science-news shows highest diversity in dense coastal metros (East Coast, Los Angeles–San Francisco corridor); nine exemplar groups among 300 creative groups illustrate radiation from dense zones. - Multilevel interplay yields temporal structure: short waves at subcategory level are overrun by long cycles of growth and reform when linked to overarching categories; in science, three century-scale cycles are observed over ~300 years. - Diverse empirical contexts align with the framework: epidemics, fashions, trends; predator–prey systems; HIV diversification under immune pressure; cultural boredom and renewal; urban gentrification cycles (Sassi di Matera) with Simpson’s index rising and falling over time.

The framework shows that causal models can indeed account for creativity and diversity. Additive structures capture how novelty emerges through recombination and small steps that accumulate within related lineages, forming creative groups. Multiplicative structures capture interplay among groups, producing waves, coexistence, and diversification. Unifying both in replicator–mutator dynamics explains observed empirical patterns across disciplines, from viral evolution to cultural trends and urban change, and clarifies why diversity concentrates where interactions are dense. The models connect measurable relatedness/diversity (e.g., Simpson’s index, clustering) to dynamical predictions about spread, resilience, and cycles, thereby addressing the research question with a mechanistic, pan-disciplinary account.

Causality provides a general, mathematical framework in which additions express creativity and multiplications express interplay and diversity. Life proceeds from stable building blocks, to adaptable creative groups, to higher-level play among groups, with diversity radiating from centers of density and with multilevel dynamics producing both short waves and long growth–reform cycles. The proposed unified framework, grounded in differential equations and encompassing perturbation theory, Lotka–Volterra, and replicator–mutator dynamics, applies across physics, biology, urban systems, and the humanities. Future work can deepen empirical calibration, extend multilevel coupling across domains, and develop predictive tools leveraging relatedness and diversity metrics for decision-making in culture, public health, and urban planning.