Near-decomposability and hierarchy (Simon)

Introduction. Simon’s 1962 paper argues that complex systems in nature and society are almost always hierarchic — systems of subsystems of subsystems — and explains why: hierarchic assemblies evolve faster and survive interruption better, illustrated by the parable of the watchmakers Hora, who builds in stable sub-assemblies, and Tempus, who does not. Such systems are “nearly decomposable”: interactions within a subsystem are strong and fast, interactions between subsystems weak and slow. The mathematical result, from Simon and Ando (1961), is that in a nearly decomposable system the short-run behaviour of each subsystem is approximately independent of the others, and the long-run behaviour of the whole depends only on aggregate variables of the subsystems. Decomposability is therefore a claim about timescale separation, and it is an empirical property of the system — the interaction matrix has a block structure or it does not — rather than a convenience the analyst chooses. Simon adds that hierarchic systems can be described economically because redundancy is high, and that a description in terms of state versus process is a further decomposition. The paper is the standard account of hierarchy in complex systems and is reprinted in The Sciences of the Artificial.

Important authors. Herbert A. Simon (1916–2001), Carnegie Mellon University, Nobel Memorial Prize in Economics (1978), ACM Turing Award (1975); the paper bridges his organisational work and his later work on design and complexity. Albert Ando (University of Pennsylvania) co-authored the aggregation theorem. Pierre-Jacques Courtois developed the queueing applications. Howard Pattee edited the 1973 hierarchy-theory volume that took the idea into theoretical biology; Carliss Baldwin and Kim Clark (Harvard Business School) and Daniel Levinthal (Wharton) carried it into the modularity and organisational-design literature.

Importance for cybernetics and the VSM. Simon worked alongside but not within cybernetics. Beer’s recursion principle is a special case of the hierarchic thesis, and the VSM literature cites Simon for that. What it has not done is apply the empirical half: whether an organisation is nearly decomposable at the levels the analyst draws is a testable property, measurable from the timescales of its variables, and the VSM’s recursion levels have never been tested that way. Simon also bears on the model’s own claim-structure. Beer held that the five systems are identifiable but not separable; if that is right the VSM is not nearly decomposable, and claim-by-claim audit may test something that is no longer the VSM.

Importance for the article. Simon carries three loads in v02. §2.5 names the decomposability bet: treating the model as an assembly of claims is “a substantive bet that its claim-structure is near-decomposable,” and decomposability “is an empirical property of a system, not a choice by an analyst (Simon, 1962).” §12.7 repeats the bet as a limitation, notes one adverse data point (§5: Schwaninger and Scheef’s factor analysis could not separate System Three from Three-star) and points to Demonstration IV as its instrument. §8.1 reformulates recursion as timescale stratification (IIIc): “Simon’s near-decomposability is fundamentally a claim about timescale separation,” so System Three is not above System One but on a longer time constant. The imports are Input from Cognitive Science (C2: “cited in §3.2 and then not applied”) and Variety and Channels Now (C4 §7 and closing note). A reviewer will press on whether a failed decomposability test refutes the protocol or the model (§12.7: either result is recordable), and may object that near-decomposability and “identifiable but not separable” are compatible, since Simon’s subsystems interact weakly, not zero.

Sources in the reading list.

  • the sections on near-decomposability and on the evolution of hierarchic systems; use the 1962 pagination.

Other important sources and authors.

  • Simon, H. A., & Ando, A. (1961). Aggregation of variables in dynamic systems. Econometrica, 29(2), 111–138 — the theorem behind near-decomposability; the formal basis for IIIc.
  • Simon, H. A. (1996). The Sciences of the Artificial (3rd ed.). MIT Press — the reprint with Simon’s later reflections, and the chapter on design that bears on the VSM as design method.
  • Courtois, P.-J. (1977). Decomposability: Queueing and Computer System Applications. Academic Press — the engineering treatment, with error bounds for the aggregation approximation.
  • Pattee, H. H. (Ed.) (1973). Hierarchy Theory: The Challenge of Complex Systems. Braziller — the theoretical-biology response, including Simon’s own chapter.
  • Baldwin, C. Y., & Clark, K. B. (2000). Design Rules, Volume 1: The Power of Modularity. MIT Press — decomposability as a design variable, with the design-structure-matrix method that could be applied to a VSM specification.
  • Ethiraj, S. K., & Levinthal, D. (2004). Modularity and innovation in complex systems. Management Science, 50(2), 159–173 — simulation of what happens when the designer’s decomposition does not match the system’s; the organisational form of the decomposability bet.
Scroll to Top