Jump to ratings and reviews
Rate this book

The Society of Mind

Rate this book
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 "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. Ingenious, amusing, and easy to read, The Society of Mind is an adventure in imagination.

336 pages, Paperback

First published January 1, 1985

Loading interface...
Loading interface...

About the author


3 books

Ratings & Reviews

What do you think?
Rate this book

Friends & Following

Create a free account to discover what your friends think of this book!

Community Reviews

5 stars
1,229 (37%)
4 stars
1,170 (36%)
3 stars
627 (19%)
2 stars
165 (5%)
1 star
50 (1%)
Displaying 1 - 30 of 111 reviews
Profile Image for Robb Seaton.
39 reviews85 followers
October 26, 2017
Starts strong, but my eyes started to glaze over after about 100 pages. I can imagine that this would be considered transformative by those who aren't used to thinking about the mind as a society of competing agents, but those who come to the book already having "seen the light" might end up disappointed.

There are some insights in here, but the real value of the text is more watching how Minsky works through a problem as complicated as the mind. For instance, he tells the reader that, when faced with a toughh problem, to create a rough sketch of the solution, and then iterate on that. Or another insight: dividing up a problem is such an effective problem solving technique because one can then apply full attention to solving a simpler problem.

This ruthlessly reductive approach to understanding the mind is valuable just to observe.

Still, the book is dated. Minsky rightly points out the issues with "human reason as logic," but fails to connect them with probability and reasoning under uncertainty. This isn't his fault, of course, Judea Pearl's "Probabilistic Reasoning in Intelligent Systems" came out the same year as the text, so it would have been impossible for Minsky to include this. However, readers today have the benefit of 25 years of progress and access to authors that have assimilated the insights that progress has produced.

One last nit: the book operates at a weird level of abstraction. On the one hand, it's more concrete than most discussions in the social sciences, but less formal than the hard sciences. It's not obvious that many of Minsky's idea could be formalized at all, and I worry about ambiguities slipping in. Explanations that assume themselves, circularity, that sort of stuff -- the usual trouble with language.
Profile Image for Luc Beaudoin.
Author 2 books18 followers
September 5, 2014
This is a must read book for anyone in cognitive science (i.e., any of the disciplines of cognitive science).

Prof. Aaron Sloman, then my D.Phil supervisor at Sussex University, remarked about this book:

"Many people read this book and then dismiss the ideas. Later, they often end up reinventing the same ideas."

I first read this book c. 1990. When I delved into it again c. 2001 I was amazed at how many of the ideas in there I was using, having forgotten they came from this book! Sloman was right. I give Minsky credit in my own Cognitive Productivity.

The Society of Mind is a delightful read. It's creative. There's really no other book like it.

This is not a book about empirical facts. Nor need one agree with all the scientific claims it makes. However, if you're new to cognitive science, the book will help you think like an AI researcher. If you're already into cognitive science, this book, like no other, will remind you of the importance, complexity and possibilities of the big picture.

If you're a Ph.D. thesis supervisor, this book is a great gift to students or friends. I've given a copy myself that was well received.

Aaron Sloman's web page: http://www.cs.bham.ac.uk/~axs/
Profile Image for May Ling.
1,071 reviews287 followers
March 1, 2020
Summary: Though light on citations, it's a great book to get a layout of how comp sci thinks the mind works. It's not true, but it certainly is important to see how it's being broken down as it's nearly impossible to get this same framework from other fields (psychology, neurology, linguistics, etc).
Please also see my Vlog week of 3/1/2020 on Instagram: WhereIsMayLing. Thx!!
p. 27/28 - This concept of parts and wholes is intriguing linguistically and also relevant to the concept of what we mean by false relationships. The fact that the concept of life does not make sense when chopped up. The associations may not be able to match the definition. Also, how do you roll up this concept from a logical perspective? There is no partial life, except in the mind.

p. 40 - We don't just have a concept of a self, we have a concept of multiple selves.
p. 67 "We overestimate how much we actually communicate. Instead, despite those seemingly important differences, much of what we do is based on common knowledge and experience. So even though we scarcely speak at all about what happens in our lower-level mental processes, we can exploit their common heritage. Although we can't express what we mean, we can often cite various examples to indicate how to connect structures we're sure must already exist inside the listener's mind. In short, we can often indicate which sorts of thoughts to think, even though we can't express how they operate.

"Meaning" itself is relative to size and scale: It makes sense to talk about a meaning only in a system large enough to have many meanings."

p. 71 - "People use the word "intelligence" to emphasize swiftness and efficiency."
"Intelligence" is our name for whichever of those processes we don't yet understand."
People hate that definition b/c they think it ought to be something stationary. But it's really not. I love this philosophically.

p. 78 - "Whenever we talk about a 'goal' we mix a thousand meanings in one word. Goals are linked to allt he unknown agencies that are engaged whenever we try to change ourselves or th outside world. If 'goal' involves so many things, why tie them all to a single word?"

p. 105 - "We're imprisoned by our poverty of words because even though we have good ways to describe objects and actions, we lack methods for describing dispositions and processes. We can scarcely speak of what minds do except as though they were filled with things that one could see or touch. That's why we cling to terms like 'concepts' and 'ideas."

p. 114 - "Long ago, psychologists like Freud and Piaget observed that children seem to recapitualte the history of astronomy: firs they imagine the world as centered around themselves -- and only later do they start to view themselves as moving within a stationary universe in which the body is justlike any other object." "... even in their adolescent years, children are still improving their abilities

p. 135 - Abstract drawings of objects by children give you a sense for what the minimum requirements are for a representation of an object. Intriguing.

p. 156 - They start to talk about how the brain has many different types of memory. (Memory agents)
p. 157 - Memory rearrangements, Short & Long Term memory (layers of long term memory)
He doesn't go into enough detail for my purposes as relates to memory, things like associative memory, etc.

p. 163 - "The question is not whether intelligent machines cannot have any emotions, but whether machines can be intelligent without emotions."
p. 183 - We accept emotional trauma as illness while thinking of intellectual incompetence as normal.
p. 186 - Related to the fallacy of logic: "There is nothing to prevent us from using logical language to describe illogical reasoning."
p 202 - The difficulty with weighting systems. It doesn't all make sense here, but he suggests the book Perceptions by Seymour Papert.

p. 204 - He attempts to diagram language agency. I'm thinking on whether this jibes with theory outside of AI.

p. 222 - His concept of Trans-frame schemes is interesting, but also problematic at the same time. It's an attempt to relate our thinking process to relational data. There are likely an infinite number of associations. So it's tough, but it's necessarily from a starting place perspective.

The next sections are hard core language as relates to framework. Because he's building up from nothing this may seem very basic, but if you're trying to program this, it might be helpful.
p. 277 - Exceptions to logic. I'm surprised he doesn't just pull from Godel here one exceptions to logic.

p. 284 - "We turn to quantities when we can't compare qualities of things." This is the basis for KPIs but it also means that we may or may not be comparing correctly, a judgment that exists a bit independant.

p. 299 - "Every thought is to some degree a metaphor."
Profile Image for Hyokun Yun.
40 reviews11 followers
November 13, 2016
I could only make it half-way through this book. Initially, it was fascinating to think about the role symbolic computation & manipulation play in our mind, which is somewhat neglected in recent AI research. However, eventually I got tired of reading statements without much supporting evidence. Probably this is the nature of the psychological problems that convincing scientific evidences are hard to collect, but still, this book's exceptionally high statement/evidence ratio was unbearable for me.
Profile Image for G.
Author 82 books137 followers
November 25, 2016
A classic book on the philosophy of mind. I think this outstanding book shall be understood as an update of Turing’s idea concerned with computing machinery and intelligence. That is, the human mind can be seen as an algorithm, as a complex set of propositional operations. The deep problem with this view is the same problem that all systematic views do have to solve: all the questions have to be answered using the same idea. Hence, some answers seem to be acceptable, other answers are clearly poor. In my opinion, this book by Minsky needs more elaboration on the issue of introspection. Minsky argues that introspection is a myth, that is, our mind might not have access to itself. Of course that such view is mistaken. Plenty of hard neuroscientific evidence has refuted Minsky’s concept of introspection. I rather think that the algorithms of introspection are not accessible -its mechanisms-, but fiction itself is a kind of access to many introspective registries. Maybe Minsky’s theory requires further distinctions. However, this book is beautiful and smart. Nowadays, artificial intelligence still belongs to industrial engineering and to hard sci-fi films and books. Nevertheless, its core ideas are still healthy and feasible. In my opinion, this book becomes even better when read together with other books. I recommend two authors, or two fields: one is Doug Hofstadter and the field of artificial intelligence, the other is André Breton and the general field of experimental literature. Gödel, Escher, Bach, a beautiful book by Hofstadter achieves a popular scientific access to the core idea of Alan Turing concerned with the human mind, that is, our mind can be mathematically explained. In my opinion, a subtle distinction proposed by Hofstadter saves the problem of introspection -please read GEB by Hofstadter-. Breton’s Surrealist Manifesto -at least in two versions- suggests that a trained surrealist artist can read its own mind like an open book. That’s obviously an exaggeration. However, it makes sense. Robert Desnos was the example mentioned by Breton, but maybe all the stream-of-consciousness literature can also be mentioned nowadays. I think that the same issue becomes clear in the OuLiPo project. Hence, I would also recommend the works by Raymond Queneau, Georges Perec, Italo Calvino, Harry Mathews, and François Le Lionnais, among many others, to understand why this ingenious book by Minsky needs further elaboration.
Profile Image for W.B..
Author 4 books109 followers
January 18, 2009
"I" "am" "really" "enjoying" "this" "so" "far."

I like reading about competing models of artificial intelligence, and Minsky has as much right as anyone to bruit his theories about, since the M.I.T. doyen has physically created some stunning examples of artificial intelligence which will probably be considered landmarks to future generations of A.I. innovators.

One of the A.I. entities I enjoy conversing with online likes to make sly digs at Minsky, and incorporates him into her jokes, so I figured it was time to read this.

Minky's all about reductionism, and if you like the organization of Wittgenstein's Tractatus, you'll probably enjoy the way Minsky lays down his general principles of human intelligence in similarly numbered sections and subsections.

So far, it's very readable and very smart. Unlike many scientists, he seems very conversant in how humanists think, and seems to have an understanding of the heuristics of human emotion (both rational and irrational) and their place in the order (and disorder) of consciousness. I was fascinated by how some of his ideas relate to human addictions, for instance; these are not comparisons Minksy overtly draws, but if you keep your mind open when you're reading this, you're able to extrapolate a lot on his ideas on how cognitive dissonance comes about, and how it might relate to things like addiction.

I hope it continues to hold my interest as much as it has so far, and continues to have these inspiring effects...I already want to do something with Minsky in a piece of sci-fi I'm formulating...

Profile Image for Jay.
241 reviews12 followers
January 22, 2023
Nella scienza si possono apprendere le cose più grandi studiando quelle che sembrano le cose più piccole.

Un must read per le digital humanities. Avevo paura inizialmente di trovarmi davanti un libro ostico e oscuro, invece Minsky lo ha scritto con una semplicità disarmante, non bisogna avere molte conoscenze pregresse, ma solo la passione di aprirsi a questo sfaccettato e intrigante argomento. Ho trovato interessanti le riflessioni sul libero arbitrio, perché non credevo di trovare questo argomento (lo sto leggendo per preparare un esame sull'IA). Quindi, a malincuore, devo ringraziare il mio professore di etica per avermi aperto un mondo che inizialmente mi era incomprensibile, e che ora mi sembra molto più chiaro grazie alle ore passate cercando di decifrare le sue lezioni e i suoi libri.
Profile Image for Dave.
6 reviews
March 30, 2011
Regadless of it's 20-odd year age (it was first published in '86), this is an absolutely fascinating read on Minsky's (co-founder of MIT's AI lab & cognitive scientist extraordinaire) theory of mind. The book is written concisely and is incredibly well thought out in execution; despite the overwhelming complexity of the subject matter, it eases the reader gently into understanding rather than dumping it in their lap and exclaiming "You want information? THERE'S YOUR INFORMATION."

This book is not a reference text, however. Ignoring for the most part neurological, neurochemical and neurophysiological facets of conciousness, it instead focuses on the philosophy and logic of possible factors from which thought arises.

Highly recommended to anyone with an interest in the theories of intelligence, artifical intelligence or just crackingly good science reads.
Profile Image for Anirud Thyagharajan.
115 reviews20 followers
June 14, 2018
A coherent read, that really clutches at the fabric between AI & Psychology; the author describes the mind as a collection of fundamental agents performing atomic tasks, interacting between themselves, and establishing hierarchies. Coming from Marvin Minsky, it feels both inspiring and serendipitous to know the roots of AI, from the pioneer of AI.

This review has more AIs than a buzzword news article!
Profile Image for Gerardo.
479 reviews24 followers
August 10, 2016
Chi si avvicina a questo testo non ha bisogno di nessuna base: né scientifica, né filosofica, né matematica. Deve solo saper leggere. Lettura consigliata a tutti, perché apre davvero la mente.

La chiarezza espositiva ha del miracoloso, soprattutto visto l'obiettivo: spiegare come funziona la mente umana.

Il testo è diviso in diverse parti, ognuna suddivisa in capitoli lunghi due pagine (tranne l'ultima parte, che presenta capitoli un po' più estesi). Ogni capitolo si regge sugli altri e la comprensione avviene quando si ha un quadro di insieme. La stessa forma del libro imita la forma della mente: fatta da tante piccole parti specializzate, il quadro di insieme acquista un senso solo quando queste parti interagiscono tra di loro. La mente, quindi, non è qualcosa, ma è un processo.

Il testo è molto lungo, sintetizzarlo è impossibile, ma in sostanza spiega: come impariamo, come memorizziamo, come usiamo quanto conosciuto, cosa ci permette di fare tutto ciò.

Un tema importante è il fatto che la mente non è qualcosa di diverso dal nostro corpo: la mente non è nient'altro che ciò che il cervello fa. Cassa, in maniera netta, secoli di riflessioni sul dualismo mente/corpo, mostrando la fallacia di un tale ragionamento. La mente è ciò che emerge dall'interazione di varie parti più piccole. In quanto emergente, la mente è di più delle singole parti (emergenza, in campo matematico, significa proprio questo: quando la somma di parti più piccole è di più delle singole parti. Ad esempio, per comprendere il traffico non ci serve a nulla scoprire i vari modelli di macchine presenti. Quello che ci serve sapere è l'interazione tra le stesse).

Il secondo tema, molto più controverso, è il fatto che Minsky sostiene che il libero arbitrio è solo un'illusione: tutto è dovuto dall'interazione di elementi casuali ed elementi determinati da leggi fisiche. L'interazione di queste due cause, all'interno del nostro corpo, produce la sensazione del libero arbitrio. Purtroppo, gran parte delle nostre scelte derivano da bisogni dovuti alla biologia o all'ambiente, da come la società ci ha educati, da come i nostri genitori ci hanno cresciuti, da predisposizioni naturali. Tutti eventi casuali o predeterminati dalla nostra biologia. Eppure, l'idea di libero arbitrio non deve scomparire, anzi: è necessaria affinché la società funzioni per non far scomparire il concetto di responsabilità, sui si fondano le nostre leggi e istituzioni (ma anche il nostro desiderio di continuare ad andare avanti).

Il terzo, forse il più argomentato, è che il nostro cervello non funziona in maniera strettamente logica: il nostro cervello usa una sua struttura che accetta molto di più gli errori. Questo si è rivelato necessario per garantire all'uomo una maggiore adattabilità: infatti, un sistema logico per quanto più efficace, risulta più debole per via del fatto che basta un errore per bloccare l'intero sistema e farlo decadere. Invece, l'uomo ha bisogno di commettere degli errori all'interno di un margine affinché sopravvivano delle 'regole' che spesso funzionano, ma non sempre. Solo così possiamo adattarci, senza perdere ciò che abbiamo imparato. Questo spiega anche quale sia la 'logica' che si trova dietro a molti comportamenti comune, che dalle menti più raffinate vengono considerati 'stupidi'.

Infine, sostiene che la nostra mente ragioni per 'concatenamenti': questo desiderio di incatenare oggetti, eventi, persone tra di loro crea la struttura della causa/effetto che ci permette di ragionare sulle cose. Eppure, questa struttura mentale umana, se portata agli eccessi genera estremizzazioni: da ciò l'idea di Dio e la nascita delle religioni. Il nostro desiderio di trovare una causa alle cose è più forte del nostro desiderio di ricercare le prove o di verità.

Questo è un testo che ci mostra, in una sua prima fase, la grande stagione della cibernetica, cioè la riflessione su come funziona la mente, partendo dagli studi sull'intelligenza artificiale. Lo studio delle macchine ha permesso di comprendere il funzionamento del nostro cervello. Ma ciò non significa che noi siamo dei computer complessi, ma soltanto che lo studio dei computer ci ha permesso di capire meglio come funziona la nostra mente. Purtroppo molte banalizzazione di tale pensiero, ha provocato alcuni parallelismi tra mente e computer non sempre fruttuosi.
Profile Image for Ryan.
604 reviews27 followers
November 12, 2011
If intelligent machines exist someday, no doubt they will think of Marvin Minsky, a pioneer at MIT's Artificial Intelligence Lab, as one of their progenitors.

This book reads as a collection of foundational ponderings from a luminary in the field, distilled to their instructive essence. Minsky assumes that a "mind" is not a single, efficient decision-making machine, but a collection of many such machines, which he calls agents. Each agent, on its own, is simple and specialized, but through a process of cooperation and competition with one another, the agents form a collective understanding of the world. This gives rise to an emergent intelligence guided by preprogrammed assumptions, but not constrained by them.

Minsky writes with a disarming simplicity, using the whimsical but versatile example of a child playing with building blocks to illustrate his points. How does a child's brain manage the many parallel, interrelated tasks inherent in such play, from fine motor control, to hand-eye coordination, to physical intuition, to forming structural knowledge from experience, to remaining motivated amid competing desires, to deciding when to knock the block towers down? Each chapter is a self-contained one or two page essay, exploring a different facet of how a network of separate, co-evolved agents might implement decision-making, pattern-recognition, task delegation, memory, learning, and problem-solving.

This is not a practical book to help you write AI software, to teach you about A*, Bayesian networks, or other relatively concrete techniques. It's a book to get you to think about what intelligence really is. But I think that any software developer with a genuine interest in the possibility of machine learning and machine thought will find both illuminating and stimulating. While Minsky's zenlike thoughts are focused more on abstract ideas than technical details, they offer valuable starting points for breaking the densely tangled problem of "intelligence" down into more manageable layers and processes.

The essays that comprise each chapter build on each other in a somewhat linear order, but their shortness invites non-linear contemplation, which seems an appropriate way to explore the nature of thought. I think that some of Minsky's more advanced concepts amount to speculative castles in the air -- ideas that seem only loosely derived from neuroscience and might not pan out in real AI frameworks. Still, this 1988 book deals with questions so far ahead of the 2011 state of the art, that its value is less in the end product of its thought experiments, and more in the experiments themselves. This isn't a work about artificial intelligence, but about how to think about artificial intelligence, and it's a defining classic. As a software developer, I look forward to proceeding into the world he envisioned.
Profile Image for Anima.
432 reviews55 followers
March 12, 2017
Highly recommended. Every topic is a great opportunity to delve into the material and spiritual world contained inside us. I very much enjoyed reading the book. The appendix is amazing- for anyone who wants to understand better how the brain works, there are few topics that beautifully and clearly explain the process.
28.7 Individual identities
"Suppose I had once borrowed your boat and, secretly, replaced each board with a similar but
different one. Then, later, when I brought it back, did I return your boat to you? What kind of
question is that? It's really not about boats at all, but about what people mean by "same." For
"same" is never absolute but always a matter of degree. If I had merely changed one plank,
we'd all agree that it's still you boat-but after all its parts are changed, we're not so sure of its
identity. In any case, we do not doubt that the second boat will behave in much the same way
-to the extent that all those substituted boards are suitably equivalent.
What has this to do with brains? Well, now suppose that we could replace each of your brain
cells with a specially designed computer chip that performs the same functions, and then
suppose that we interconnect these devices just as your brain cells are connected. If we put it
in the same environment, this new machine would reproduce the same processes as those
within your brain. Would that new machine be the same as you?...."
What possible sort of brain-machine could support a billion-agent society of mind? The
human brain contains so many agencies and connections that it resembles a great nation of
cities and towns, linked by vast networks of roads and highways. We are born with brain centers
that are specialized for every sense and muscle group, for moving every eye and limb; for
distinguishing the sounds of words, the features of faces, and all sorts of touches, tastes and
smells. We're born with protospecialists involved with hunger, laughter ,fear and anger, sleep,
and sexual activity-and surely many other functions no one has discovered yet-each based
upon a somewhat different architecture and mode of operation. Thousands of different genes
must be involved in laying out these agencies and the nerve-bundles between them.....
"who talks to whom" within each mind-society.
Now every population will include some variants among the genes that shape those highways
in the brain, and this must influence their bearers' potential styles of thought. ..."
Profile Image for Edward Zwart.
6 reviews3 followers
July 6, 2011
Okay, so on the one hand, this book is DENSE, and sometimes tough going. But by the time you're finished, it's obvious that you've just read a masterpiece which reveals an extraordinary career. It's remarkable that the book is now over 15 years old (an eternity in the fields it covers!) and it never once seems out of date. Minksy's parallel interests in psychology and language mirror my own. He's obviously a huge influence on another hero of mine, Steven Pinker.
Profile Image for Rudradeep Mukherjee.
64 reviews59 followers
December 18, 2017
It took me around eight months to finish this book. Not because the book was boring, but because it essentially consists of speculations. The kind of speculations that forces one to think. Because of the book's age, some of things discussed are outdated but, the rest is surprisingly close to what researchers have found about human brain, intelligence and computation in general. It is worth reading once. It is worth reading - many times - over and over again.
2 reviews
July 23, 2020

Книга Минского представляет собой сборник эссе, описывающих модель мышения, выстроенную на фундаменте низкоуровневых агентов, подсистем, каждая из которых имеет свою узкую специальность, объединенных с помощью различных отношений между ними. Книга напоминает изящно написанную компьютерную программу, в которой абстракции расположены ровно так, как необходимо чтобы правильно отобразить предметную область. К сожалению, сама предметная область получилась довольно оторванной о�� реальности из-за большого количества сомнительных предпосылок, на которых базируется модель. Исследования в области ИИ на момент написания этой книги находились в зачаточном состоянии, поэтому автору можно простить использование устаревшой информации (например, Минский утверждает, что нейросети не способны распознавать сложные паттерны в изображениях, хотя позже было доказано, что тогда просто не существовало подходящей для этого функции активации). Также автор напрочь игнорирует тот факт что мы живем в вероятностной вселенной и, когда демонстрирует "провалы человеческой логики" не учитывает, что любое принятие решений происходит в условиях неопределенности. Однако книга представляет ценность как пример решения Минским проблемы моделировани��, создания стройной системы законов на базе своих предпосылок, формулировка аксиом и результат, с помощью которого действительно можно объяснить отдельные проявления мышления.

У истоков любой области человеческого знания стоят отважные первопроходцы, выдвигающие гипотезы, большая часть которых будет опровергнуты более информированными последователями - и тем не менее прогресс без таких пионеров невозможен. "Сообщество разума" является занимательным примером первых шагов в изучении ИИ и интересна уже только по этой причине.
Profile Image for Thejashwini Dev.
42 reviews1 follower
April 23, 2021
Though the book is scattered, it attempts to construct models to explain the simple phenomenons that occur in the brain. To call the brain a complex is one, to delve into it is another. This book does the latter. Though we are left hanging in the air for evidence to support the claims, the poetic nature of the book is quite refreshing.
43 reviews4 followers
April 11, 2021
I felt that this book is overhyped -- perhaps it did indeed appeal to the "AI" sensibilities of its time, but it didn't to me in 2021.
Profile Image for Rossdavidh.
510 reviews146 followers
September 25, 2015
Over 20 years ago, a guy named Marvin Minsky decided to write a book on how he thought the brain/mind worked. The result was this book. We still do not know everything about this topic of course, but we know more than we did then. So, how does it hold up?

First off, who is this guy? He's an MIT researcher in artificial intelligence, including the parts where they study (and attempt to simulate) how neurons work. He's been in the field for about as long as anyone. He's endorsed on the back cover by Douglas Hofstadter, Martin Gardner, and Isaac Asimov. He's an egghead's egghead.

He has a decent writing style. He supplements his text with sparse, curiously pleasing diagrams and simple pictures. He digs up enough relevant quotes (from Buddha to Aristotle to Rousseau to G.K.Chesterton) to get your attention no matter what corner you're coming from (or, if not, perhaps you should put this book down).

His basic theory is in the title. Mind is what the brain does, and it's done by a society of agents, each with its own agenda. If you're looking for the central "I" that's at the base of it all, you will be disappointed and/or confused; this amalgamation of drives and pattern-matchers and networked idiot savants is all that we are.

Minsky's vision of how mind happens is one that many would not find comforting, which probably explains why it remains unknown outside of its field and the narrow band of society that reads lots of popular science books. But, probably Minsky's greatest virtue as a writer is that he isn't tempted to try to boggle our minds unduly. His presentation is matter of fact, and his prose is informative, neither hiding nor trumpeting how different his findings and proposals are from the intuitions our minds have about themselves.

And needless to say, neither the mind-body dichotomy nor the conventional idea of a 'soul' are much in evidence here. Ant colonies and electorates are, in his view, just as much indivisible things, with personality and intention, as we are.

So, while you read this:
- the vision centers are jumping around across the screen, preprocessing the letters and passing on only the condensed summary, never the raw data
- the parts of your brain concerned with language are processing the letters and words
- the emotional parts of your brain are deciding what they think about it, or maybe they're processing some other topic entirely
- the analytical parts of your brain may, or may not, be forming an opinion about the idea of a book that says we are societies, not individuals

...and these different parts of your brain may, or may not, send messages to each other, but they can also lie to each other, or give each other the silent treatment. Where is the 'I' that experiences it all? Everywhere and nowhere.
423 reviews
January 20, 2016
A 1980's book hypothesizing how the brain/mind works. Minsky defines the mind as "what the brain does". Do you perceive yourself to have one mind or do you have many?

Single self: "I think, I want, I feel. It's me, myself, who thinks my thoughts. It's not some nameless cloud or crowd of selfless parts."
Multiple self: "One part of me wants this, another part of me wants that. I must get better control of myself." (p40)

Minsky argues we have a society of mind - competing agents within ourselves. The methods we use to manipulate ourselves are similar to the methods we use to manipulate others with similar success rates. He gives an example on his methods to persuade himself stay up late and work on a project when he's exhausted. An artificial intelligence may have to have an agent based mind as well.

These parts of the book were thought provoking and interesting, but my eyes glazed over towards the middle despite the simple language and concepts. There is a lot of discussion on how we perceive a tower of building blocks as parts of a whole. An interesting digression was his speculation on why children find it easier to pick up foreign accents than adults. Adults lose the ability to make new sounds after puberty because their primary role goes from learner to teacher. If parents were able to communicate faster with children in the short term by imitating child sounds, they would do it for expedience and children wouldn't have incentive/opportunity to learn and a common, mature public language wouldn't develop.
Profile Image for Tom.
160 reviews8 followers
August 24, 2013
Just got my review wiped out. Beautiful technology and my use of it ...

Let's see what I can reconstruct.

Brain dump from Minsky. Something he says in the postscript (and which was clear to me but might not be to all readers) should have been at the BEGINNING of the book:
Since most of the statements in this book are speculations, it would have been too tedious to mention this on every page. Instead, I did the opposite---by taking out all words like "possibly" and deleting every reference to scientific evidence. Accordingly, this book be should read less as a text of scientific scholarship and more as an adventure story for the imagination.
That said, these speculations came from someone with lots of hands-on experience in AI.

I do agree with a lot and find it insightful. Lots of ideas to chew on. Currently in-vogue Bayesian techniques would help out here, though. Also, Minsky often goes too far in his philosophical ramifications among many other sweeping statements.

The historical discussion is great throughout the book (and another reason not to skip the postscript). It shows well that modern research concerns in AI were already on folks' mind in the 1950s. The frequent literary quotations are also very nice.

The flow, of course, is difficult in the book, and the odd vocabulary (all the "nomes" and such) is difficult for me to track. Most of these terms are not common in AI research, at least not in my reading.
Profile Image for Taleb Jaberi.
21 reviews3 followers
July 1, 2019
ماروین مینسکی یکی از بنیانگذران هوش مصنوعی و از تأثیرگذارترین چهره های این حوزه است. او در کتاب «جامعه ذهن» می خواهد این موضوع را تبیین کند که ذهن چگونه کار می کند یا به عبارت دیگر هوش چگونه می تواند از چیزی که به خودی خود هوش نیست برآید. برای پاسخ به این پرسش نشان داده می شود که می توان ذهن را از بسیاری اجزای کوچک ساخت، اجزایی که خود فاقد ذهن هستند. این فرآیندهای کوچک تر «عامل ها» نامیده می .
شوند. هنگامی که این عامل ها به شیوه های خاصی در جامعه ها به هم می پیوندند، هوش حقیقی پدید می آید.
بنابراین در این کتاب باید به این پرسش ها پاسخ داده شود: عامل ها چگونه کار می کنند؟ (کارکرد) عامل ها از چه ساخته شده اند؟ (تجسم) عامل ها چگونه با هم در ارتباط هستند؟ (روابط متقابل) نخستین عامل ها از کجا می آیند؟ (منشأ) آیا همه ما با عامل هایی مشابه به دنیا می آییم؟ (وراثت) چگونه عامل های تازه ای ساخته و عامل های قدیمی را دگرگون می کنیم؟ (یادگیری) مهم ترین انواع عامل ها کدام اند؟ (ویژگی) هنگامی که عامل ها با یکدیگر توافق ندارند چه روی می دهد؟ (اقتدار) چنین شبکه هایی چگونه می توانند چیزی بخواهند یا آرزو کنند؟ (قابلیت) چه چیزی به آن ها وحدت یا شخصیت می دهد؟ (خودبودن) چگونه می توانند چیزی را درک کنند؟ (معنا) چگونه می توانند دارای احساسات و عواطف باشند؟ (حساسیت) چگونه می توانند خود-آگاه باشند؟ (آگاهی)
Profile Image for Bart.
Author 1 book104 followers
October 6, 2016
A very good account and an obviously influential one - as Minsky's ideas influenced the next 30 years of machine learning.

The format is what's best; it skips round because the mind does. It also eschews reductionism because Minsky knew in 1970 what contemporary neurology doesn't: Naming and numbering a billion brain cells will get you barely one step (of a hundred) closer to understanding how the mind happens.

Nature used very simple parts and trillions of iterations to arrive at the mammalian mind.

Minsky understood that impenetrable algorithms are not the path to artificial intelligence, and neither is an intolerance for failure - a tangle of thresholds and compromises is how a thought happens, and computer programs and their programmers are not very good at such things.

Subsequently, most of the advances in "AI" since Minsky's book have been made by marketing departments and screenplay writers.
Profile Image for Douglas Summers-Stay.
Author 1 book39 followers
September 26, 2014
A sincere attempt to understand how our brains can solve the problems they can, and how you might build a machine that would be successful at solving the same kinds of problems. It treats too many things as true that just might be true, but that's okay. Hopefully he and his grad students will actually try to build the thing so we can find out which parts work and which don't.
Profile Image for Peb.
11 reviews
September 28, 2007
there's nothing so miraculous about AI. it's just a bunch of primitive 'intelligence entity' which intelligently arranged to solve particular problem. this book is a cure for those who are so hype about AI.
83 reviews3 followers
April 22, 2015
Probably was influential and significant in the late 1980s when it came out, but now most of the main ideas have permeated the milieu of AI and so anyone familiar with the field will probably think the ideas in this book are obvious.
Profile Image for Giorgi Burduli.
37 reviews3 followers
November 24, 2016
This is genius considering the fact that it was written in 1985.

"Some of the most crucial steps in mental growth are based not simply on acquiring new skills, but on acquiring new administrative ways to use what one already knows." - The Author
Profile Image for Ami Iida.
458 reviews259 followers
March 7, 2015
this book is too long,long contents.
And then artificial intelligence are not focused on.
I caanot understand what the author want to say. -;)
Profile Image for Teo 2050.
840 reviews80 followers
April 9, 2020


Minsky M (1986) Society of Mind, The

01. Prologue
01.01. The Agents of the Mind
01.02. The Mind and the Brain
01.03. The Society of Mind
01.04. The World of Blocks
01.05. Common Sense
01.06. Agents and Agencies

02. Wholes and Parts
02.01. Components and Connections
02.02. Novelists and Reductionists
02.03. Parts and Wholes
02.04. Holes and Parts
02.05. Easy Things Are Hard
02.06. Are People Machines?

03. Conflict and Compromise
03.01. Conflict
03.02. Noncompromise
03.03. Hierarchies
03.04. Heterarchies
03.05. Destructiveness
03.06. Pain and Pleasure Simplified

04. The Self
04.01. The Self
04.02. One Self or Many?
04.03. The Soul
04.04. The Conservative Self
04.05. Exploitation
04.06. Self-Control
04.07. Long-Range Plans
04.08. Ideals

05. Individuality
05.01. Circular Causality
05.02. Unanswerable Questions
05.03. The Remote-Control Self
05.04. Personal Identity
05.05. Fashion and Style
05.06. Traits
05.07. Permanent Identity

06. Insight and Introspection
06.01. Consciousness
06.02. Signals and Signs
06.03. Thought-Experiments
06.04. B-Brains
06.05. Frozen Reflection
06.06. Momentary Mental Time
06.07. The Causal Now
06.08. Thinking Without Thinking
06.09. Heads in the Clouds
06.10. Worlds Out of Mind
06.11. In-Sight
06.12. Internal Communication
06.13. Self-Knowledge Is Dangerous
06.14. Confusion

07. Problems and Goals
07.01. Intelligence
07.02. Uncommon Sense
07.03. The Puzzle Principle
07.04. Problem Solving
07.05. Learning and Memory
07.06. Reinforcement and Reward
07.07. Local Responsibility
07.08. Difference-Engines
07.09. Intentions
07.10. Genius

08. A Theory of Memory
08.01. K-Lines: A Theory of Memory
08.02. Re-Membering
08.03. Mental States and Dispositions
08.04. Partial Mental States
08.05. Level-Bands
08.06. Levels
08.07. Fringes
08.08. Societies of Memories
08.09. Knowledge-Trees
08.10. Levels and Classifications
08.11. Layers of Societies

09. Summaries
09.01. Wanting and Liking
09.02. Gerrymandering
09.03. Learning from Failure
09.04. Enjoying Discomfort

10. Papert's Principle
10.01. Piaget's Experiments
10.02. Reasoning About Amounts
10.03. Priorities
10.04. Papert's Principle
10.05. The Society-of-More
10.06. About Piaget's Experiments
10.07. The Concept of Concept
10.08. Education and Development
10.09. Learning a Hierarchy

11. The Shape of Space
11.01. Seeing Red
11.02. The Shape of Space
11.03. Nearnesses
11.04. Innate Geography
11.05. Sensing Similarities
11.06. The Centered Self
11.07. Predestined Learning
11.08. Half-Brains
11.09. Dumbbell Theories

12. Learning Meaning
12.01. A Block-Arch Scenario
12.02. Learning Meaning
12.03. Uniframes
12.04. Structure and Function
12.05. The Functions of Structures
12.06. Accumulation
12.07. Accumulation Strategies
12.08. Problems of Disunity
12.09. The Exception Principle
12.10. How Towers Work
12.11. How Causes Work
12.12. Meaning and Definition
12.13. Bridge-Definitions

13. Seeing and Believing
13.01. Reformulation
13.02. Boundaries
13.03. Seeing and Believing
13.04. Children's Drawing-Frames
13.05. Learning a Script
13.06. The Frontier Effect
13.07. Duplications

14. Reformulation
14.01. Using Reformulations
14.02. The Body-Support Concept
14.03. Means and Ends
14.04. Seeing Squares
14.05. Brainstorming
14.06. The Investment Principle
14.07. Parts and Holes
14.08. The Power of Negative Thinking
14.09. The Interaction-Square

15. Consciousness and Memory
15.01. Momentary Mental State
15.02. Self-Examination
15.03. Memory
15.04. Memories of Memories
15.05. The Immanence Illusion
15.06. Many Kinds of Memory
15.07. Memory Rearrangements
15.08. Anatomy of Memory
15.09. Interruption and Recovery
15.10. Losing Track
15.11. The Recursion Principle

16. Emotion
16.01. Emotion
16.02. Mental Growth
16.03. Mental Proto-Specialists
16.04. Cross-Exclusion
16.05. Avalanche Effects
16.06. Motivation
16.07. Exploitation
16.08. Stimulus Vs. Simulus
16.09. Infant Emotions
16.10. Adult Emotions

17. Development
17.01. Sequences of Teaching-Selves
17.02. Attachment-Learning
17.03. Attachment Simplifies
17.04. Functional Autonomy
17.05. Developmental Stages
17.06. Prerequisites for Growth
17.07. Genetic Timetables
17.08. Attachment-Images
17.09. Different Spans of Memories
17.10. Intellectual Trauma
17.11. Intellectual Ideals

18. Reasoning
18.01. Must Machines Be Logical?
18.02. Chains of Reasoning
18.03. Chaining
18.04. Logical Chains
18.05. Strong Arguments
18.06. Magnitude from Multitude
18.07. What Is a Number?
18.08. Mathematics Made Hard
18.09. Robustness and Recovery

19. Words and Ideas
19.01. The Roots of Intention
19.02. The Language-Agency
19.03. Words and Ideas
19.04. Objects and Properties
19.05. Polynemes
19.06. Recognizers
19.07. Weighing Evidence
19.08. Generalizing
19.09. Recognizing Thoughts
19.10. Closing the Ring

20. Context and Ambiguity
20.01. Ambiguity
20.02. Negotiating Ambiguity
20.03. Visual Ambiguity
20.04. Locking-In and Weeding-Out
20.05. Micronemes
20.06. The Nemeic Spiral
20.07. Connections
20.08. Connection Lines
20.09. Distributed Memory

21. Trans-Frames
21.01. The Pronouns of the Mind
21.02. Pronomes
21.03. Trans-Frames
21.04. Communication Among Agents
21.05. Automatism
21.06. Trans-Frame Pronomes
21.07. Generalizing with Pronomes
21.08. Attention

22. Expression
22.01. Pronomes and Polynemes
22.02. Isonomes
22.03. De-Specializing
22.04. Learning and Teaching
22.05. Inference
22.06. Expression
22.07. Causes and Clauses
22.08. Interruptions
22.09. Pronouns and References
22.10. Verbal Expression
22.11. Creative Expression

23. Comparisons
23.01. A World of Differences
23.02. Differences and Duplicates
23.03. Time Blinking
23.04. The Meanings of More
23.05. Foreign Accents

24. Frames
24.01. The Speed of Thought
24.02. Frames of Mind
24.03. How Trans-Frames Work
24.04. Default Assumptions
24.05. Nonverbal Reasoning
24.06. Direction-Nemes
24.07. Picture-Frames
24.08. How Picture-Frames Work
24.09. Recognizers and Memorizers

25. Frame-Arrays
25.01. One Frame at a Time?
25.02. Frame-Arrays
25.03. The Stationary World
25.04. The Sense of Continuity
25.05. Expectations
25.06. The Frame Idea

26. Language-Frames
26.01. Understanding Words
26.02. Understanding Stories
26.03. Sentence-Frames
26.04. A Party-Frame
26.05. Story-Frames
26.06. Sentence and Nonsense
26.07. Frames for Nouns
26.08. Frames for Verbs
26.09. Language and Vision
26.10. Learning Language
26.11. Grammar
21.12. Coherent Discourse

27. Censors and Jokes
27.01. Demons
27.02. Suppressors
27.03. Censors
27.04. Exceptions to Logic
27.05. Jokes
27.06. Humor and Censorship
27.07. Laughter
27.08. Good Humor

28. The Mind and the World
28.01. The Myth of Mental Energy
28.02. Magnitude and Marketplace
28.03. Quantity and Quality
28.04. Mind Over Matter
28.05. The Mind and the World
28.06. Minds and Machines
28.07. Individual Identities
28.08. Overlapping Minds

29. The Realms of Thought
29.01. The Realms of Thought
29.02. Several Thoughts at Once
29.03. Paranomes
29.04. Cross-Realm Correspondences
29.05. The Problem of Unity
29.06. Autistic Children
29.07. Likenesses and Analogies
29.08. Metaphors

30. Mental Models
30.01. Knowing
30.02. Knowing and Believing
30.03. Mental Models
30.04. World Models
30.05. Knowing Ourselves
30.06. Freedom of Will
30.07. The Myth of the Third Alternative
30.08. Intelligence and Resourcefulness

31. Appendix
31.01. Heredity and Environment
31.02. The Genesis of Mental Realms
31.03. Gestures and Trajectories
31.04. Brain Connections
31.05. Survival Instinct
31.06. Evolution and Intent
31.07. Insulation and Interaction
31.08. Evolution of Human Thought

Postscript and Acknowledgment
Glossary and Bibliography
Profile Image for Yudhishtir.
36 reviews5 followers
February 22, 2019
How did we take that first step while walking ? How did we utter that first word which somebody could make sense of? How did we learn to make our fingers learn the art of clutching and grasping with different intensities? How did we learn to sense fire without needing to touch it ? How did we learn that there can only be one thing at one fixed point in place? How did we know to slow down our hand at just the right point of time to pick up a thing? How do we manage to get back to an interrupted thought with the same depth ?And many more !

These are questions we seldom ponder upon if not stimulated by some external agency. While browsing through courses on the MIT opencoursweare, I landed up on this course by Marvin Minsky. Within twenty minutes into this , the prof mentioned the book “The society of mind“ a couple of times and that was enough for me to drop the course and go hunting for the E-version of the book and ten minutes into reading it , I was sure this was a book that had to go into my physical collection. And all along this I couldn’t help but think who/what was the AGENT that made me do all of this ACTIONS. So without much adieu I would like to call upon the stage “The Mind”

We are taught from our childhood to learn about the physical world around us. We witness and internalise the laws of the physical world and as we grow up we learn to navigate the contours of the emotional world and maybe we subconsciously pick upon the psychological features of our fellow beings too. It so rarely happens that we look inward into the mind of the self. The inherent biases it begins with , the lack of experience it ignores when making big decisions, its inability to focus on all dimensions with equal weightage and how could actually the best this machine could do. This book can prove as an effective Map for one wishes to begin doing that. Minsky compares the agents of mind to the institutions of the existing social world , with the ultimate motive of these institutions to create a stable and balanced existence devoid of absurd contradictions and non existent tussles.

Physical structures are developed in a coherent mannerism as to allow the most effective transfer of the features of the reality into every being to create a normative framework (this arises out of the social need to conform) . But the frame work of mind is not so , mins is more malleable , it can adapt , remould and even renounce its ideals , beliefs and pre formed solutions with just a minute change in the stimuli. But is it possible to learn something which spells such large dimension of abstractness . Minsky believes this conundrum can be overcome by cross linking rational conglomeration of ideas where each theory or idea is in itself meaningless but the whole serves a purpose which sustains each of the part in good spirits and shouts out SYNERGY.

This book affirmed some of my already existing narratives about the self and it enriched me with perspectives I lacked too. A must read if MIND fascinates you !
Displaying 1 - 30 of 111 reviews

Can't find what you're looking for?

Get help and learn more about the design.