Marvin Minsky (1986), The Society of Mind, New York: Simon and Schuster (hardback, ISBN 0-671-60740-5; Library of Congress 86020322, call number BF431 .M553 1986; 339 pp., illustrated, bibliography pp. [326]–332, index). A Touchstone paperback (ISBN 0-671-65713-5) followed. Google Books dates it 15 March 1988, and the Harvard Book Store listing agrees.
Key takeaways
Registered from the publisher’s description and the contents list, not a reading. What follows is what the book says about itself.
- The publisher’s description, verbatim: “Marvin Minsky — one of the fathers of computer science and cofounder of the Artificial Intelligence Laboratory at MIT — gives a revolutionary answer to the age-old question: ‘How does the mind work?’ Minsky brilliantly portrays the mind as a ‘society’ of tiny components that are themselves mindless. Mirroring his theory, Minsky boldly casts The Society of Mind as an intellectual puzzle whose pieces are assembled along the way. Each chapter — on a self-contained page — corresponds to a piece in the puzzle. As the pages turn, a unified theory of the mind emerges, like a mosaic.”
- The contents list Google Books gives for the Simon and Schuster paperback, with its page numbers: Prologue (13), Wholes and Parts (20), Conflict and Compromise (27), The Self (34), Individuality (43), Problems and Goals (66), A Theory of Memory (77), Summaries (89), Papert’s Principle (94), The Shape of Space (104), Learning Meaning (114), Seeing and Believing (128), Consciousness and Memory (146), Emotion (162), Development (169), Reasoning (181), Words and Ideas (191), Context and Ambiguity (202), Transframes (212), Expression (221), Frames (239), Frame-Arrays (249), Language-Frames (256), Censors and Jokes (269), The Mind and the World (278), The Realms of Thought (287), Mental Models (296), Appendix (305), Postscript and Acknowledgment (318). The list has gaps: it jumps from p. 43 to p. 66 and from p. 128 to p. 146, and it carries one stray heading (“Contents”, p. 152). Some chapter titles are therefore missing, and the list is not a complete table of contents.
Standing in the record
Filed on 2026-10-01 at the owner’s request for Marvin Minsky’s “Society of Mind” programme. The theory holds that a mind is a society of many simple, mindless agents. It was developed with Seymour Papert from the 1970s through this book to The Emotion Machine. No anthology topic holds the book. Its subject is the architecture of minds, human and artificial, and it carries no instruction for machine-learning practice. It is the programme’s central statement, and it goes on the record so that the programme’s other works have something to point at.
Deferred because it was not read. This record files books from their
publisher’s description and contents when no lawful text can be read, and
it does not rebuild a book from what others say about it. So the takeaways
above stop at the description and the contents list. What the record knows
of the book’s content comes from two readings filed with it:
- Push Singh’s Examining the Society of Mind (LIT-444), a sympathetic exposition by Minsky’s student. It quotes the book (§§2.5, 6.12, 10.4, 20.1, 30.8 and the Postscript) and lays out its agents, K-lines, nomes and nemes, frames and transframes, censors, and A- and B-brains. It is secondary, and it is a student’s.
- The 2005 draft of the sequel, The Emotion Machine (LIT-432), read in full. Its introduction says it replaces this book’s “agent” with “resource”. Many readers had taken an agent to be “a personlike thing (like a travel agent)”, and that was not the intent. Its §9-5 says plainly that many readers took this book to propose a “community view” of the mind, and “that was not my intention”. When the book is read, that reading should check whether the 1986 text invites the community reading the sequel disowns.
Minsky defended the book against its first scientific reviewers in Artificial Intelligence in 1991 (LIT-418). The earlier memos and papers that led to it (the frames memo of 1974, “Plain Talk About Neurodevelopmental Epistemology” of 1977, “K-lines” of 1980 and the 1980 jokes memo) are being filed in this record at the same time by another reader. Singh traces the agent idea to the procedure-bearing frames of the 1974 memo. He quotes the 1977 paper, which already names the theory “The Society of Minds”, and the 1980 K-lines paper, which pictures the mind as “Divisions” of agents. By its title, the book’s chapter “Censors and Jokes” takes up the jokes memo’s theme. Singh ties the book’s censors and humour to that memo.
The book bears on what this record holds in three places. These are my inferences from the description, the contents and the two readings above, not claims the book was read to make:
- The record’s work on wholes made of parts and on collective intelligence. Pilgrim et al. (LIT-046) analyse collectives of capable individuals. Minsky’s agents are too specialised to be individuals in that sense, and Singh says so.
- Group minds: Schwitzgebel’s argument that the United States may be conscious (LIT-159).
- The record’s emotion works. The book has a chapter called “Emotion”,
which is why
emotion-and-affectis among its tags. Its sequel’s account of emotion is set against Damasio’s (LIT-384, LIT-388) in the reading of that sequel (NOTE-348).
Access when seeded:
- Library of Congress catalogue record (lccn.loc.gov/86020322, through the loc.gov JSON API): publisher, date, extent, bibliography pages and call number.
- Open Library work OL4981641W: the edition list with both ISBNs.
- Google Books (volume bLDLllfRpdkC): the publisher’s description and the contents list above.
- Harvard Book Store listing: the paperback’s date.
- Simon & Schuster’s own page for the book refused the request (HTTP 403) on curl and on WebFetch alike.
- The Internet Archive holds four scans (societyofmind00mins, societymin00mins, societyofmind00marv, societyofmind0000marv). All are lending-only (access-restricted), so I did not open them. An unrestricted upload under another name is not the publisher’s or the author’s, so I did not open it either.
- HathiTrust holds two copies, search-only.
- Minsky’s MIT home page, as mirrored on MIT’s web server, links the book only to a bookseller.
- A full text circulating at aurellem.org is not from the publisher or the author, and I did not use it.