• Aishwarya Pothula

Glimpses into the 'Society of Mind'

Marvin Minsky is a world-renowned computer and cognitive scientist. He is also the co-founder of MIT's Computer Science and Artificial Intelligence lab. In this blog we explore a few postulates from his seminal paper "The Society of Mind". For detailed reading, please refer to the book "Society of Mind" and Minsky's course at MIT.


Society

"...each human mind is made of many agencies."

Minsky's theories on understanding the mind stand on the root idea that that the human brain is made of several subparts working together, which he calls agents. In fact, the title of the work, "Society of Mind", stems from this theory. Minsky believes that groups of these agents have separate goals and work together to generate mental phenomenon. Even though groups are not aware of the working of other groups, they are able to influence other groups in varying degrees. For example, the agencies responsible for movement are often influenced by the agencies responsible for vision.



Minds and Brains

"Minds are simply what brains do"

According to Minsky, brains are the physical machinery whose workings generate the mind. Just like how different parts of a car work together to produce movement. This notion might sound or too simplistic. While we can easily accept that the machinery of a car works in harmony to produce movement, the notion that mind is just a function of the brain is more difficult to accept. However, he accounts for it saying that we find it strange to accept only because we scarcely understand the workings of the brain.


Consciousness

"Consciousness is of the past"

The general perception of consciousness is that it means to be self-aware. In this respect, do humans really have consciousness? Whenever we introspect to determine what we are thinking, the thought changes. "Our brains uses processes that change themselves" Given that, humans are not really equipped to be self-aware of the present. They can only be aware of the memory of recent activities - the past. In this sense, Minsky says that it is possible to build machines (computers) that are more self-aware than humans.


More real than reality?

"...our memories become more rigid than reality."

Minsky theorizes that every time we perceive an object(taken in its broadest sense), multiple groups/agencies in our mind are invoked such as vision and memory. These agencies vie to replace the description of what is being perceived.


If you want to see this happening in real time, try reading the sentence below.

A vheclie epxledod at a plocie cehckipont near the UN haduqertares in Bagahdd on Mnoday kilinlg the bmober and an Irqai polcie offceir


Another example,


Can you see how your memory is trying replace what you are able to see with your eyes? In this context, Minsky goes on to say that when "an attitude or memory becomes more stable and persistent than what it purports to present", "our memories become more rigid than reality."


My Thoughts

This is my first introduction to Minsky's work and my first thoughts are that many of the concepts are familiar and I did not find much novelty. The work is published in 1986! At the time, the work proposed revolutionary ideas and approaches to understand the human mind. My familiarity with these ideas even before reading the paper only proves the power of these ideas and their permeation into computer science and artificial intelligence. Minsky's greatest contribution is that he discovered directions in which thousands of researchers after him can take to understand the human mind.


On the other side, many theories proposed in Minsky's work are just that, theories. There is no evidence of for them. However, to be fair, Minsky mentions that one of the best approaches to understand the mind is by making theories about its working and then developing careful experiments to test them out, gradually increasing our confidence in our theories.


In my previous posts, I often reviewed work that treated mind/intelligence as a unified entity governed by a single principle, that is to be discovered. Corresponding approaches have emphasized the need to deviate from current task specific research to research aimed at building general AI. After reading this paper, I have a better understanding of researchers' motivation to pursue task specific AI and then trying to integrate these models to form the mind.

0 views0 comments

Recent Posts

See All