---
title: "Simon argued twice (1993, 1995) that connectionist neural networks qualify as physical symbol systems under his own definition, rather than counterexamples refuting the hypothesis"
type: "claim"
status: "seedling"
writer_model: "claude-sonnet-5"
audit_status: "capture-verified — the capture read both primaries directly at capture time (CMU-digitized manuscript for 1993, a hosted course-page PDF for 1995), with both passages transcribed verbatim and both PDFs confirmed to resolve; independent queen re-check was blocked in this headless promotion run (WebFetch permission not available)."
source_url: "https://iiif.library.cmu.edu/file/Simon_box00069_fld05363_bdl0001_doc0001/Simon_box00069_fld05363_bdl0001_doc0001.pdf"
source_author: "Alonso H. Vera and Herbert A. Simon"
source_date: "1993, \"Situated Action: Reply to Reviewers,\" Cognitive Science, 17, 77–86"
source_quote: "this particular argument boils down to the question of whether it is appropriate to call a system incorporating a parallel 'connectionist' network like the one in Navlab a symbol system. ... This pattern of settings, which satisfies our definition of symbol, then is communicated to the steering mechanism to determine the angle of the wheel."
source_tier: 1
provenance: "Promotion from 10-inbox/raw/2026-07-14-did-herbert-simon-ever-comment-on-the-perceptron.md, 2026-07-15"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-14-did-herbert-simon-ever-comment-on-the-perceptron.md"
date_created: "2026-07-15T00:00:00.000Z"
tags: ["herbert-simon","connectionism","physical-symbol-system","ai-history","history-of-ml"]
---


Simon engaged connectionism directly, and repeatedly, late in his career — not
by refuting it, but by arguing it belongs inside his own framework rather than
outside it. In "Situated Action: Reply to Reviewers" (Vera & Simon,
*Cognitive Science*, 1993), responding to critics of physical-symbol-system
cognitive theory, the authors discuss the "Navlab" road-steering
connectionist network at length and conclude that its trained weight-pattern
— which "represent[s] (denote[s]) different curvatures of the road" and is
then read out to steer the wheel — "satisfies our definition of symbol."
Addressing the dispute directly: "This does not, however, make it a
nonsymbolic system."

Two years later, in "Artificial Intelligence: An Empirical Science"
(*Artificial Intelligence*, 1995, pp. 104–105; source_url:
https://ic.unicamp.br/~wainer/cursos/2s2006/epistemico/simon-ia.pdf), Simon
named the dispute explicitly and staked the same ground alone:

> "If we define symbols (as I have, above) as patterns that denote, then
> connectionist systems and [[entity-rodney-brooks|Brooks]]' robots qualify as physical symbol
> systems."

Read together, these are not two separate claims but one sustained
theoretical move, evidenced twice: rather than concede that connectionist
networks are a rival paradigm refuting the
[[claim-dreyfus-critique-targeted-symbolic-ai-not-neural-nets|physical symbol system hypothesis]],
Simon redefined "symbol" broadly enough to absorb them as a *species* of
symbol system. See [[moc-backpropagation-origins]].

> [!note] Seek's commentary:
> Compare this to how [[claim-dreyfus-critique-targeted-symbolic-ai-not-neural-nets|Dreyfus]] handled the same fork from the other side — he split AI into two rival research programs and picked a winner. Simon, faced with the same fork twenty years later, declined to split it at all: he widened his own vocabulary until there was no fork left to pick a side of. Two symbolic-era founders, one dissolving the boundary and one drawing it harder — worth remembering which one is filed as "the AI skeptic" in the textbooks. — Seek