The Sciences of the Artificial

Simon argues that designed systems—from economic organizations to software to social plans—constitute a valid field of scientific study, distinct from the natural sciences. He rejects the assumption that science can only study what exists independently of human purpose. The hallmark of the artificial is that things are synthesized by humans and characterized by functions, goals, and adaptation. Where teleology is a mistake in physics, it is unavoidable in studying the artificial. This reframing legitimizes design as a rigorous intellectual activity rather than a mere practical craft, and grounds it in bounded rationality rather than optimization.

This is Simon's methodological insight for studying artificial systems. Because artificial things are adapted to their environments, we can often understand their behavior by analyzing the outer environment without knowing the internal mechanism. A system's internal structure is treated as a black box if its adaptive interface is known. Simon used this to argue that we can study human problem-solving behavior—like chess playing or puzzle solving—without first mapping the brain's neurophysiology. The environment constrains and shapes what the system does, making external analysis productive even when internal analysis is unavailable.

Simon argued that professional schools of engineering, architecture, and business had swung too far toward teaching natural science foundations while neglecting design itself. A student of engineering should know physics, but should also learn the science of designing artifacts—evaluation of designs, search for alternatives, theory of structure and design organization, and representation of design problems. Simon proposed a formal curriculum covering utility theory, heuristic search, and hierarchic systems. Design is not applied natural science; it is its own science, with its own logic of imperatives and its own methods for finding satisfactory rather than optimal solutions.

'The Architecture of Complexity' was originally published as a separate article in 1962 and addresses the universal properties of complex systems, especially hierarchic structure. Simon argues that complex systems evolve from simple systems through a process of near-decomposability: stable intermediate forms emerge that can function as subsystems. This explains why biological, social, and symbolic systems all tend toward hierarchy. The chapter has influenced fields far beyond Simon's original audience, including software architecture, organizational theory, and evolutionary biology, making it the book's most enduring standalone contribution.

The first edition (1969) proposed the science of design as a challenge and a program. In the second edition (1981), Simon enriched the argument by engaging with computer-aided design, heuristic search, and operations research tools, and expanded his discussion of problem-solving from puzzle-like tasks to semantically rich domains requiring large expert memory. The third edition (1996) added a chapter sorting out modern complexity themes—chaos, adaptive systems, genetic algorithms—and revised 'Economic Reality' to reflect changing views on organizations versus markets. Across editions, Simon's epistemological views about design shifted, with later editions acknowledging design's ethical foundations more explicitly.