How do we process information? Daniel Kahneman uses the concept of System 1 and System 2 to describe our fast, automatic, intuitive approach (System 1) and our slower, deliberate, reasoning approach (System 2) [1]. Is that the sum total of how we process information? In this article, I suggest that there might be an additional two or even three modes of information processing that are worth considering.
Firstly, let’s consider System 1. If we see a snake we recoil and if we see a red traffic light we stop. In both cases, we exhibit lightning fast information processing. However, it is important to recognise that in the case of the snake, we would recoil even if it were the first time we had even seen a snake. Why? Although it might be our first time seeing a snake, we are the descendants of thousands of generations of ancestors who have seen many snakes. And those who did not recoil did not live to reproduce. So, we are the ancestors of those who did recoil and now the instinct to recoil is our genes. Traffic lights have been around for less than a century or about four generations. Our ingrained response to stop at red is not the product of our ancestors’ experience, but our own life experience. We have had enough life experience of red lights for what was at first a conscious and deliberate act (stopping on red) to have developed into a pattern (like riding a bike) such that this behaviour has become effortless and fast. So, is there merit in splitting System 1 into a System 1a (inherited) and a System 1b (learned)?
Now, let’s consider System 2. When the nature of the problem we face is sufficiently novel, we have not experienced it often enough for it to have become an established pattern. As such, our information processing for this problem requires our deliberate consideration. In contrast to 2 + 2 which is dealt with by System 1, Kahneman asks us to compute 17 x 24 as an illustration of System 2 processing [1]. Consider the information processing done by a business executive who is considering expanding into a new market. The executive, assuming they are being diligent, will be assessing the prospects of success, the risk factors faced, the investment required, and so on. They will in some fashion compute how this initiative will perform and hence whether it is worth pursuing. In other words, they are assessing the value of going ahead versus not going ahead versus doing something else. Probably using some risk-adjusted return on investment value metric, either implicitly or explicitly. In this case the information processing is being done with some preferences in mind. In other words, how the alternatives score against the preferences or value metrics will inform the executive’s choice. This is unlike computing 17 x 24 where preferences simply don’t feature and there is no choice to be made. So, is there merit in splitting System 2 into a System 2a (without preferences and choice) and a System 2b (with preferences and choice)?
Lastly, let’s consider how information processing occurs, especially as our world becomes increasingly augmented by technology. When we face System 2 challenges, we often outsource some or all of the information processing, i.e. we often call on external support – in the form of both man and machine. The “man” side includes using experts, facilitators, and social constructs like teams. The “machine” side started centuries ago with tools like the abacus, but has bounded along to include tools like log tables (I am giving away my age), calculators, computers, and increasingly now artificial intelligence. The nature of information processing by outsourcing at least part of it might be considered sufficiently different to warrant being labelled System 3 (augmented), at least that is the opinion of The Strategic Decisions Group [2]. See also [Strategic Decision Making in the Augmented Age].
In summary:
System 1a is information processing by virtue of the experience of our ancestors, i.e. it is in our genes.
System 1b is information processing by virtue of System 1a plus our own life experience.
System 2a occurs when the context is sufficiently novel that our information processing cannot rely on established patterns and requires deliberate consideration.
System 2b is System 2a, but requires a choice that is arrived with (explicit or implicit) consideration of our preferences.
System 3 is when, as part of our deliberate consideration, we augment System 2a or 2b with external tools and/or experts.
Please challenge!
Acknowledgements to Dr Christoff Prinsloo [4] for triggering the initial debate that led to this article.
References:
[1] “Thinking, Fast and Slow”, Daniel Kahneman
[2] https://www.mccombs.utexas.edu/execed/take-a-class/strategic-decision-and-risk-management/biases-in-decision-making
[3] https://www.linkedin.com/pulse/strategic-decision-making-augmented-age-wayne-borchardt/
[4] https://www.linkedin.com/in/dr-christoff-prinsloo-66789016/