Site Logo

STPrism

Herb Simon

6 min read

Based on the works of Herbert A. Simon provided, the concepts of hierarchy, non-ergodicity (path dependence), speciation, causality, and constraint are deeply intertwined. They form a cohesive framework for understanding how complex systems—whether biological, social, or artificial—come into existence, survive, and function.

Here is the “untangling” of these relationships:

1. Hierarchy (Near-Decomposability) is the Architecture

The central connecting principle is Hierarchy, specifically defined by Simon as “Near-Decomposability” (ND).

• Definition: Complex systems are organized as “boxes-within-boxes”[1]. Interactions within a subsystem (e.g., a molecule, a department, an organ) are strong and rapid, while interactions between subsystems are weak and slow[2],[3].

• Relationship to Causality: Hierarchy creates the structure necessary for causal analysis. In a dynamic system, the “slow” variables (the higher levels of the hierarchy) act as constants or “exogenous” variables relative to the “fast” variables (the lower levels)[4]. This separation of time scales allows us to identify causal mechanisms because we can treat the upper level as the “unmoved mover” affecting the lower level in the short run[5].

2. Speciation and Evolution driven by Hierarchy

Hierarchy explains how complex systems (speciation) evolve within feasible timeframes.

• The Watchmaker Parable: Complex systems that are hierarchic evolve much faster than non-hierarchic ones. If a system is built from stable sub-assemblies (hierarchy), it does not collapse completely when disturbed; it falls back only to the previous stable level[6],[7].

• Relationship to Speciation: Speciation is facilitated by ND. Because the subsystems are “loosely coupled” (weakly connected), a specific organ or component can be improved or specialized (speciated) without requiring simultaneous changes in all other parts of the organism[8],[9]. This allows for “niche elaboration”—as new species evolve, they create new niches for other species (e.g., dogs create a niche for fleas), leading to a proliferation of diversity rather than a static equilibrium[10],[11].

3. Non-Ergodicity (Path Dependence) and History

Simon argues that these systems are non-ergodic; they do not explore all possible states and do not settle into a single, pre-determined equilibrium independent of their starting point.

• Local vs. Global Maxima: Evolution is a “myopic” hill-climbing process. An organism climbs the nearest hill of fitness (local maximum) but cannot cross a valley to reach a higher hill[12]. Therefore, the final form of the system depends heavily on where it started and the specific history of variations presented to it[13].

• Process over State: Because the system is path-dependent (non-ergodic), we cannot describe it simply by its final state (as in classical thermodynamics). We must use a “process description”—a recipe or algorithm of how it developed[14],[15].

• Relationship to Causality: In a non-ergodic world, we cannot assume that correlations represent a static, universal law. We must identify the specific “structural equations” or mechanisms that remain invariant under intervention[16],[17].

4. Constraints as the Mechanism of Definition

In Simon’s framework, “goals” and “causes” are operationally defined by constraints.

• Goals are Constraints: In complex design (or biological survival), there is rarely a single unitary goal. Instead, there is a set of constraints (requirements) that must be satisfied[18],[19],[20].

• Generators and Tests: Constraints function in two ways: as generators (limiting the search to a specific subspace) and as tests (verifying if a solution is acceptable)[21],[22].

• Relationship to Hierarchy: Constraints are applied hierarchically. High-level decisions (e.g., “produce 1 million gallons of paint”) become constraints for lower-level subsystems (e.g., “produce X amount of red pigment”)[23],[24]. This loose coupling allows the lower levels to work out the details independently, provided they satisfy the aggregate constraints imposed from above[25].

Summary of the Relationships

ConceptRole in the SystemRelationship to Others
HierarchyThe StructureEnables Causality by separating high/low frequency interactions; accelerates Speciation by allowing local adaptation.
CausalityThe LogicDefined by the asymmetry between “slow” (exogenous) and “fast” (endogenous) levels of the Hierarchy; identified by invariance under Constraint modification.
SpeciationThe DynamicDriven by the modular stability of Hierarchy; leads to Non-Ergodicity because new niches change the fitness landscape.
Non-ErgodicityThe TrajectoryResult of myopic evolutionary search; implies that History determines which Constraints and Hierarchies currently exist.
ConstraintThe DefinitionDefines the “inner environment” of a system; operates hierarchically, with top-level Constraints bounding lower-level freedom[26],[27].

**In short:**Hierarchy (Near-Decomposability) isolates subsystems, allowing them to Speciate (evolve/adapt) locally and rapidly. This adaptation is governed by Constraints passed down from the upper levels of the hierarchy. Because this process is “myopic” and history-dependent, the system is Non-Ergodic—it evolves unique, path-dependent forms rather than converging to a universal optimum. Causality is the tool we use to analyze this structure, identified by finding which variables (the slow, upper-level constraints) determine the behavior of others (the fast, lower-level variables) without being immediately affected in return.

Graph View