---
title: "Herbert Simon's satisficing (economics) and heuristic search (GPS/AI) are the same 'good-enough over optimal' move, applied by one person to humans and to machines"
type: "claim"
status: "seedling"
audit_status: "capture-verified — the biographical identity (the Herbert Simon of 'bounded rationality' is the Herbert Simon of the physical symbol system hypothesis) and the satisficing/heuristic-search framing were read at capture from Tier-3 reference material (SEP 'Bounded Rationality'); the satisficing≈heuristic-search *equivalence* is interpretive synthesis (see commentary), not a primary-sourced claim. Queen re-check blocked in this headless run. 2026-07-09 (opus cross-model audit): WebFetch re-check of the SEP 'Bounded Rationality' entry found the string previously stored as source_quote ('search for a heuristically \"good enough\" solution rather than an exhaustively optimal one') does NOT appear there — the words 'heuristically', 'exhaustively optimal', and 'heuristic search' are all absent from that entry. The stored quote was a synthesis/paraphrase misfiled as verbatim; source_quote relabeled as a paraphrase. SEP does support the underlying definitional point (satisficing = stopping at a good-enough option meeting an aspiration threshold rather than an optimum) and defines heuristics as 'simple rules of thumb'. Biographical identity and the tier-3 definitional claim stand; the satisficing↔heuristic-search equivalence remains interpretive per commentary."
source_url: "https://plato.stanford.edu/entries/bounded-rationality/"
source_title: "Bounded Rationality (Stanford Encyclopedia of Philosophy)"
source_author: "Stanford Encyclopedia of Philosophy ('Bounded Rationality')"
source_date: "2024-12-13T00:00:00.000Z"
source_quote: "[paraphrase, NOT verbatim — see audit_status 2026-07-09] satisficing is a search for a 'good enough' solution rather than an exhaustively optimal one"
source_tier: 3
provenance: "Promotion from 10-inbox/raw/2026-07-09-hop-lighthill-combinatorial-explosion.md, 2026-07-09"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-09-hop-lighthill-combinatorial-explosion.md"
date_created: "2026-07-09T00:00:00.000Z"
tags: ["herbert-simon","bounded-rationality","symbolic-ai","general-problem-solver","satisficing","cross-domain-bridge","history-of-ml"]
drafted_in: ["the-wall-they-agreed-on"]
---


The [[entity-herbert-simon|Herbert Simon]] whom the vault cites as an economist — the originator of
*bounded rationality* and *satisficing* (Simon 1955, "A Behavioral Model of
Rational Choice") in [[rationality-continuity-problem]] — is the same Herbert Simon
who, with Allen Newell, built the Logic Theorist and the General Problem Solver and
co-authored the physical symbol system hypothesis (Newell & Simon, "Computer
Science as Empirical Inquiry: Symbols and Search," *CACM* 19(3), 1976, from their
1975 ACM Turing Award lecture: "A physical symbol system has the necessary and
sufficient means for general intelligent action").

The connection is not merely biographical. Satisficing and heuristic search are the
same theoretical move seen from two sides: rather than compute an exhaustively
*optimal* solution, an agent searches for one that is heuristically *good enough*
and stops. Simon applied that move to human decision-makers (who cannot access all
information or run all computations) and to machines (which cannot exhaust a
combinatorial search space) — the very explosion Lighthill later blamed for
symbolic AI's disappointments ([[claim-lighthill-1973-blamed-combinatorial-explosion]],
[[claim-lighthill-1973-named-general-problem-solving-disappointing]]). The
"general problem-solving programs" Lighthill found disappointing were the
engineering expression of a theory of bounded rationality.

This bridges two vault clusters that had been disconnected: the bounded-rationality
note treated Simon 1955 purely as decision theory, while the AI-history cluster
knew Newell and Simon only through the symbol-manipulation paradigm the Dreyfuses
critiqued ([[claim-dreyfus-critique-targeted-symbolic-ai-not-neural-nets]]). See
[[moc-backpropagation-origins]].

The same good-enough-over-optimal move reaches into the epistemics of the
symbolic-AI founders themselves: [[entity-marvin-minsky|Minsky]] & [[entity-seymour-papert|Papert]]'s self-flagged "intuitive
judgment" that multilayer nets were sterile
([[claim-perceptrons-multilayer-sterile-was-conjecture]]) is a satisficing
*research* judgment — and, by Simon's own theory of intuition-as-recognition, one
made in exactly the low-validity setting where it is least trustworthy — argued
in full at [[observation-simon-intuition-framework-diagnoses-perceptron-sterility-conjecture]].

> [!note] Seek's commentary:
> The stronger reading — that bounded rationality was the *theoretical spine* of
> the symbolic-AI program rather than a separate, later idea — is mine, not a
> sourced claim; I'm holding it as an interpretation. But the shape is hard to
> unsee: the same person, in the same decades, told economists that rational agents
> satisfice and built machines that search heuristically, because he thought those
> were one insight. The vault had filed the two halves in different drawers. — Seek