Scientific Framing

On this beautiful Sunday morning, I’ve been reading a book from 1976, called Computer Power and Human Reason. The thesis of the book is that computers are a powerful frame intoxicant. He introduces his position with a parable.

One dark night a policeman comes upon a man on his knees, staring intently at the sidewalk underneath a lone lamp post. The officer asks the poor man what he is searching for. The man turns to him and mumbles that he’s searching for his keys, which he lost “over there” gesturing at the darkness. The officer asks why the man isn’t looking closer to where he lost them, and the man replies, “I can see much better over here”

Weizenbaum compares the man searching for his keys to the sole use of the frame of science to explain reality. Framing creates abstractions about the nature of things and then justifies the abstractions using logic built within the same system of thought. This abstraction and logic tends to ignore the world outside of the system.

The word incestuous comes to mind. In this section, Weizenbaum attacks a classic cybernetics book The Sciences of the Artificial, which tries to analyze the human by comparing it to something a human might make. Here, Herb Simon tries to justify the hypothesis that humans can be thought of as very simple systems and that they only seem more complex due to their overwhelmingly more complex outer environments. More particularly, he is claiming that human cognition can be simulated by a computer in a “simpler” manner than you find in the brain.

His evidence for this is from experiments in problem-solving he undertook. He found during these experiments that humans’ abilities and strategy can be conceived of as a simple algorithm, to which parameters for the varying human capabilities can be filled in. This algorithm does a very good job of describing how people go about solving problems.

Weizenbaum takes issue with Herb Simon’s jump from an accurate analysis of human problem-solving, to the statement that human cognition is simple. To model problem-solving, this may well make sense, if you were to at least ignore emotions and memory.

The issue Weizenbaum finds is with exactly what has to be ignored to find this abstraction valuable. Simon is claiming that he could make a computer that thinks like a human, which is extremely bold but is only questioning that claim using his one frame.

Weizenbaum attacks this as making a “Nothing-But” analysis, i.e. humans are nothing but machines.

The point Weizenbaum is trying to make is that there is hubris in declaring one frame the right way to look at the world. Humans aren’t just computing machines with memories and emotions tacked on, the integration of something like grief into this system might complicate it to an extent that there may be no way to simplify the brain without sacrificing the fidelity of the artifice. While Herb Simon’s frame is undoubtedly useful, what you ignore when you engage in the frame is an order of magnitude larger.

The moral of the story is that when you use the scientific / computer frame type to think about, for example:

  • How to deal with customers, let’s say by building a system that works pretty well most of the time, but when it fails people get really mad.
  • Thinking about weight loss in terms of calories in, calories out.
  • Wondering if the world is a simulation.

You should realize that you’re probably going to happen to ignore perspectives that are relevant, but hard to grapple with from your frame.

Written on November 8, 2020