Bounded Rationality

Bounded Rationality is the condition that people make decisions with incomplete information, limited foresight, constrained capacity or willingness to calculate, and judgments affected by emotion. In The Wisdom of Crowds, it is not presented as a defect that rules out good decisions: when individual judgments are aggregated under the right conditions, their collective result can be remarkably strong. The concept also clarifies why gathering information is not enough; groups must allow distinct knowledge to be expressed and find ways to make it useful across the whole system.

Limits of Individual Judgment

Herbert Simon’s term, as the book uses it, names the gap between real human decision making and the ideal of a fully informed, consistently calculating chooser. People generally have less information than they would like and cannot see very far into the future. Most neither possess nor wish to perform sophisticated cost-benefit calculations. They may settle for an option that seems good enough rather than search indefinitely for the best possible one, and emotion can affect what they judge. Bounded rationality describes these ordinary constraints, not a claim that people are incapable of making useful decisions.

The distinction matters because the book’s account of collective intelligence does not depend on individual perfection. Imperfect judgments can produce an excellent aggregate when they are combined in the right way. A crowd need not be dominated by exceptional experts, nor must most of its members be especially informed or rational, for the group to make a wise decision. The argument is conditional: aggregation can overcome some limitations of individual judgment, but the conditions that make the information and judgments complementary still matter.

Local Knowledge and Aggregation

Decentralization helps people draw on knowledge tied to a particular place, job, or experience—knowledge that is difficult to summarize or transmit but can be highly valuable. People close to a problem may be better positioned to recognize a workable solution, while specialization allows different participants to develop distinct perspectives. This arrangement can increase the information available to a group without requiring each participant to know everything.

But distribution alone does not create collective intelligence. Information uncovered in one part of a system may never reach another part, and a group can remain fragmented even when its members hold relevant knowledge. The challenge is to preserve the specificity of local knowledge while making it collectively useful. The book’s account of Google offers an analogy: the knowledge embodied in millions of individual webpages can contribute to more effective searches when it is aggregated. More generally, a market, corporation, or intelligence agency needs both room for independent contributions and a means of bringing those contributions together.

When Organization Suppresses Judgment

The American intelligence community before September 11 illustrates the difference between decentralization and successful aggregation. Dividing work among specialized agencies could support expertise and diverse analysis, but the system lacked a real means of aggregating judgments as well as information. The result was decentralization without sufficient coordination: relevant knowledge did not become a shared assessment. Creating a central agency might make it easier for someone to oversee information, but would risk losing the diversity, independence, and local knowledge that specialized agencies offered. The proposed alternative in the book is not simply more central control, but better aggregation—including basic steps such as linking agency databases.

The Columbia accident investigation offers a different organizational failure. The Mission Management Team operated in a deeply hierarchical culture, and concerns raised by the Debris Assessment Team did not receive a serious hearing. The meeting process was both over-structured, with questions largely posed by the leader and answered by one person, and insufficiently structured, because other members were not systematically asked to contribute. Discussion began from a conclusion, so new information was interpreted to fit it. These practices restricted the available analysis and made it harder for the group to revise its judgment. The episode shows that having knowledgeable people in a room does not ensure that their knowledge will affect a decision.

The same problem appears in the book’s account of corporations with many management layers. At one General Motors plant in Van Nuys, a worker who had installed a bracket for two years did not know that it held the car’s sunshade. In a separate example of bureaucratic complexity, a chart from an unnamed company mapped 223 paths a new-product idea might have to travel before acceptance. Such distance between workers, information, and decision makers can keep practical knowledge from circulating. Hierarchy can therefore undermine the collective intelligence that makes a large organization useful in the first place.

Why the Concept Matters

Bounded rationality also helps frame coordination and cooperation. Some solutions in these domains do not arrive as a single definitive decision; they emerge over time through repeated interactions and shared practices. Such arrangements can work without being imposed from above, but they remain vulnerable to exploitation by free riders and tax evaders. Their success depends not on ideal rationality but on whether people’s limited judgments, expectations, and responses can combine into workable patterns.

The book extends this point to democracy without claiming that democratic decisions always reveal a single, objectively wise public answer. Democracy may be a poor instrument for solving some cognition problems, yet it can help people manage the problems of living together and coordinating their interests. Its importance lies partly in the experience of accepting an outcome one opposed, with the expectation that opponents will not destroy what one values and that another opportunity to prevail will come. Bounded rationality thus sets the terms for a modest but consequential claim: collective decision making need not depend on perfect individuals, but it does depend on institutions that let people contribute, hear one another, and continue to cooperate.