One hundred fifty years of sociology — from Durkheim to Bourdieu — described who people are, how they connect, and where they stand in the social structure. SocioGraph compiles those theories into executable populations.
Physics simulates. Climate science simulates. Biology simulates. Social science — one hundred fifty years of accumulated theory — still observes. The most sophisticated science of human behavior has no computational engine. Sociology has equations, typologies, field theories, and network models. It does not have a compiler.
Rigorous agent-based models stay small and slow. Large-scale data generation produces statistically plausible noise with no theory behind it. The field is trapped: formality or speed, depth or breadth, Bourdieu or big data. No tool combines both.
Historical periods without surveys. Vulnerable groups impossible to sample. Emerging markets with no longitudinal data. Cultural contexts with no validated instruments. Decision-makers need to understand populations they cannot observe — and the data does not exist. Not because it was lost. Because the science that describes these populations was never compiled.
SocioGraph reverses the arrow. Theory is not a lens for interpreting data. Theory is source code — and SocioGraph is the compiler. Not reduced to code — formalized as specification, preserving the mechanisms that make each theory explanatory.
Observe the world
Collect data
Fit theory to data
Theory explains the past
→ Start with theory
→ Formalize into specification
→ Compile theory into executable agents
→ Theory generates observable futures
Not GPT with a sociology prompt. Not random sampling with demographic labels. Formal specifications, compiled from peer-reviewed theory, producing populations you can audit at every layer.
SocioGraph doesn't invent theory. It compiles the theories that already exist — and have been waiting 150 years for a runtime.
Behaviorally diverse agents, not behavioral clones
Agent motivation frameworks: instrumental, value-rational, affective, traditional.
Populations with real social position, not just demographics
Agent capital vectors, dispositional patterns, social space coordinates.
Societies that cohere for different structural reasons
Emergent structural constraints, mechanical vs organic cohesion.
Information flow, social influence, structural holes
Network topology: bridges, clusters, tie-strength distribution.
Theory competition: run Coleman against Bourdieu on same data
Micro-macro link: individual actions → aggregate outcomes.
Populations that evolve, not static snapshots
Recursive agent-structure interaction: agents make structure, structure makes agents.
8 services. Each rooted in theory. Products consume these services in different combinations.
Generates N agents with demographics, values, capital vectors, and habitus — each parameter traceable to formal theory.
Builds social structure: ties, clusters, bridges, hierarchies. Not random graphs — theory-grounded topology.
Models four capitals — economic, cultural, social, symbolic — and their exchange dynamics within fields.
Assigns culturally coherent value profiles to agents. Values shape decisions, not the other way around.
Reconstructs populations of past periods from partial data and theory. When the survey does not exist, the theory still does.
Forward-projects populations under theoretical models. Not trend extrapolation — structural demographic modeling.
Ensures realistic heterogeneity within groups. No 'average representatives.' Every group contains the variation that makes it a group.
Adapts models between cultural contexts while preserving theoretical structure. A French Bourdieusian population and a Japanese one are different — but both are Bourdieusian.
Five levels of validation, applied specifically to this platform. Honest reporting of where we stand.
Reconstructive Validation
Reproduce known sociological phenomena from empirical data
Mechanistic Grounding
Every agent parameter traceable to formal theory via specification
Behavioral Plausibility
Domain experts — sociologists, demographers — recognize populations as realistic
Structural Validity
Network topology reproduces real-world social structure patterns
Statistical Fidelity
Generated distributions match known census and survey data
L4 for SocioGraph means every agent's capital portfolio traces back to Bourdieu, every network tie to Granovetter, every value profile to Schwartz. Not fitted to data — derived from theory.
SocioGraph is in development. We do not claim what we have not validated. What we do claim: every design decision is made with L5 as the destination, not L1 as the ceiling.
SocioGraph powers any product that needs to understand who people are.
Creative impact prediction. Test messaging, campaigns, and creative concepts on computed audiences grounded in social theory — before spending budget.
Universe in your pocket. Personal transformation through dialogue, rituals, and daily practices — powered by all three Quantumeon platforms: sociology, cognition, and emergence.
Social experiment environment on computed populations. Test competing theories on the same population.
Population reaction simulation. Test legislation on theory-grounded populations with real social structure.
Computed market twin for new geographies where no survey data exists.
Historical population reconstruction from partial data and formal theory.
