On `Computer Simulation of Human Thinking' (Newell & Simon 1961)
On: `Computer Simulation of Human Thinking' (Newell & Simon 1961)
What a great article:
Newell, A., & Simon, H. A. (1961). Computer Simulation of Human Thinking. Science, 134(3495), pp. 2011-2017.
It has section titles like:
Current Problems In Research.
Those are always good, if you're looking to do something new & useful & surprising (read: creative. That's the 3-part standard definition of a creative artifact: new & useful & surprising).
Anyway let's go for a spin through this article, noticing how great it is whenever possible.
They open strong: they mention (on p. 2011) you have 2 types of problems, or, discovered thing:
1) Advances in theory (models) or application;
2) Good new techniques (or, methods, or algorithms.)
When stuff goes pear-shaped, is: when scientists `schizm', and work on the opposite/inverse stuff:
1) important problems - but, with lousy/inadequate methods/techniques...
or, even worse -
2) unimportant problems - that are easily handled by the current tools/methods/models (i.e. big deal, & so what, & who even cares what you did, there? It doesn't even matter.)
Then they go to town on Psychology Science, in 1961.
They say that:
"Gestaltism" = question-oriented Psychological research
"Behaviourism" = method-oriented Psych research.
(& I guess, it was/is true?)
Then they say: Cybernetics and the Communication Sciences have changed things, since WW2. (Good point, and well made.)
Then, they (basically) say:
Let's use computers, to figure out how thinking works. (e.g. Goal-seeking, and learning, etc.)
Then they go through getting some college sophomores (as test subjects) to do a problem solving exercise (transforming one symbolic logic expression into another, given: a set of transformation rules.) And noting how the sophomores do: their problem-solving.
So then: How to make a model of what the student's doing there, when they work on / solve the problem?
And, they note, computers are not just number-manipulating devices, but also symbol-manipulating devices (words, letters, whatever symbols you want).
Then they bust out this awesome diagram / model / thinking tool:
And they say: the student's behavior is governed by a program... (An algorithm.)
(And - it is!)
And - if they can make a computer program with subroutines, that pretty much emulates the student's behaviour, then the program is a theory of the student's behavior.
(Nice point.)
They also remind us that - all theories are prized on their parsimony and generality... "on how wide the range of phenomena it explains, and and on how economical of expression it is." (p. 2013)
(On that - see also, my HOLON/parton theory of the meme, the unit of culture, 2016.)
(and i=P, and all that.)
Anyways - they say, who cares about the hardware (brains or computers), it's all just info-processing.
Then they introduce their program, called a General Problem Solver. Or GPS.
It has objects, goals, and operators...
So you have: ends (goals) and means (operators).
It (this article) even has funky flowcharts, like this:
(Above is part of their `Methods for Means-End Analysis' diagram). You ask me, this is how people think too...?
At least, I do, anyway.
Then they even have this awesome comparison - of the "Computer Trace" (what the GPS computer program does / did) and what the person, thinking aloud did (i.e., "Protocol of Subject"), and you can see how they're: basically, the same thing... More or less.
e.g.:
(omg - so awesome!)
Anyway so - the GPS is a pretty good approximation (and, theory) of human thinking and problem solving, from an information processing point of view.
Well done, you guys!!! (Newell & Simon).
Also - hey, Why didn't anyone think of - or do, this, before.
Anyway they did, so they're awesome. Or were, in 1961.
I like this statement, in the Conclusion part:
"A digital computer is a general purpose symbol-manipulating device." (p. 2016)
Darn tootin'!
They say, the common element, in lots of computer programs (e.g.: chess players, music-writers, etc) is "selective search for possible solutions based on rules of thumb, or heuristics." (p. 2016)
In summary - this stuff reduces the problems in Psychology (& in figuring out how human thinking works, and how to do it better) of, the: "bad problems" and "bad techniques" problem.
Great work Newell and Simon!
I love this 1961 paper. So easy to read, and has lots of fun stuff like diagrams, flowcharts, models, and comparisons of human and computer thinking.
(But, I think humans are computers anyway. We're probably in a Sim. See Bostrom 2003 & whatnot)
Anyway read the article sometime, it's great.
You have been reading / viewing a blog-post by:
Dr J T Velikovsky Ph.D
(aka: Velikovsky of Newcastle)
AI Researcher & Enthusiast & Evolutionary Culturologist & Filmmaker & Writer & Artist & Actor & Muso & Random Guy
(and, also The StoryAlity Guy)
aka Humanimal
More stuff:
Transmedia Blog: On Writering
IMDb (Movies, Videogames):
Music: Texas Radio & Zen Stupidity
Youtube channel: https://www.youtube.com/joeteevee (over 100 videos, some are even: good)
Academia page: https://newcastle-au.academia.edu/JTVelikovsky
Researchgate page: https://www.researchgate.net/profile/Jt_Velikovsky
My ouvre...etc etc.
For more, see On Writering and StoryAlity News
-------------//--------------
