Thinking, Fast And Slow

This post is about Daniel Kahneman's book Thinking, Fast and Slow. This is a book that everyone should probably read at some point in their lives, or at least read some cliff notes on to get an understanding of the basic ideas. The reason is because it gets at some fundamental truths about how humans function that are both non-obvious and extremely hard to recognize, even after learning about them. These truths apply to virtually any context one can imagine and are not limited in their relevancy to any particular discipline or field of study.

The central thesis is a dichotomy between two modes of thought - "System 1" and "System 2". "System 1" is fast, instinctive, and effortless. It happens automatically without us even realizing it. "System 2" is slow, deliberate, and logical. It involves effort and focus. In reality there is no physical distinction between the two modes of thought, it's all happening inside our brains at the same time, but it's a useful abstraction because it captures all the ways that "System 1" thinking can mislead us. The way I think about it is "System 1" is a filter on the raw sensory input we're taking in every second. If we think about it in stream processing terms, "System 1" is reading from the raw input stream and applying successively higher-order functions to the stream so we can quickly make sense of the input. It's what allows us to recognize faces, to read a book, to drive a car, or to catch a ball. These things don't require pupil-dilating mental effort on our part - they just happen. We might be aware that they're happening, but it's a passive awareness. We don't have to expend mental energy to make it so.

"System 1" thinking is absolutely essential for us to be able to function on a basic level. But it can also lead us astray. The filters being applied automatically to our input stream lead to thoughts and actions that frequently deviate from the rational agent model of human behavior. They lead to all sorts of cognitive biases such as anchoring, availability, substitution, framing, and overconfidence. These effects can be overcome with deliberate thought and effort - in other words, engaging "System 2" thinking to come up with an objective, logical conclusion. The problem is, these cognitive biases kick in so often and with such ease that it's hard to even recognize when such an error in judgement has occurred.

It's jarring to see just how often we're unconsciously influenced by cognitive biases, and how dramatic of an effect they can have. In one example, Kahneman was illustrating our tendency to accept a default option if one is available. He cited organ donor statistics in two different countries. In one country, the rate of civilians that opted to be an organ donor was something like 90%. In another, similar country the rate was 4%. So what accounted for the difference? In the first country you had to opt OUT of being a donor, and in the second country you had to opt IN. That was it. This is a startling conclusion. It's possible that there were confounding factors that could account for some of this variation, but it's just one example of an effect that's been observed in many different contexts with a high degree of statistical significance.

The reason I think it would be valuable for everyone to read this book is because understanding these biases likely leads to clearer thinking overall. It may be impossible, or even undesirable, to recognize every time our "System 1" thinking is making judgments that objectively do not make logical sense. But with some effort it's certainly possible to weed out the worst offenses and recognize when we're making egregious errors. In my own mental models, I've incorporated what I learned from Kahneman into my generally Bayesian approach to rational thought. I think this is a fairly sensible way to approach solving problems and making decisions.