Rational Thinking Is a Lot Less Rational Than You Thought

You and I both no doubt take some certain amount of pride in our abilities to think rationally. We can take a view of the world and apply objectivity to it, consider facts over emotions, and distrust promises made without proof. We are aware of biases in the world, including our own. If my rough demographic sketch is correct, you probably don’t believe in god either, or at least a knowable god.

This is a good way to exist in the world, but, according to a study in last week’s Science, it's missing something. There’s a force of perception known as “contrast effects” that applies to many animals, as well as humans. This is when our decisions/perceptions become influenced by how things have been in the past, good or bad or otherwise. There is some level of “normal” behavior, but in terms of “perception, cognition, and related performance,” that is heightened or decreased depending on, well, how things have been. Lift something heavy and then lift something light immediately afterward; the light thing will feel lighter than it actually is. That’s the very basic idea.

Contrast effects are what might be considered “irrational,” at least intuitively. After all, the second weight is not actually that light in the real world. Thus, our perception is inaccurate, and this must be bad, right? Not so fast. The new study demonstrates otherwise: contrast effects, the weight being lighter than in reality, are likely actually part of thinking rationally, or at least of thinking in ways likely to be beneficial in the future.

In a way, it means being tricked is good for you. Your sense of things is objectively inaccurate, but that is probably better for the future anyway. And the notion that looking at the world inaccurately is a good thing is a somewhat uncomfortable.

The Science study is based on optimality models—modeling the varying costs and effects of different behaviors—and comes via the Modeling Animal Decisions group at Bristol University. I reached out yesterday to the group's Tim Fawcett, one of the paper’s authors, with some questions.

In more everyday terms, how should we best look at the difference between strictly history based decision making and rational decision making?

In fact, our study shows that taking history into account is rational—so I don't see a mismatch between these things. In a changing environment where the animal is uncertain about the pattern of change, past experiences may give useful information because they indicate how things are likely to change in the future. So it is rational to pay attention to past experiences, and adjust your behaviour accordingly.

What this means in everyday terms is that humans and other animals have evolved to make comparisons with the past. We care about whether things are better or worse than before because this tells us something about how things may change in the future. Someone who is used to good living conditions will tend to take those conditions for granted, whereas someone who is used to much tougher conditions should really take advantage when things suddenly become easier, because they expect tough conditions to return soon.

Could you give a hypothetical for a human using only historical information in their decision making, rather than rational? I'm thinking, maybe incorrectly, about an election with differing outcomes depending on whether the past four years were "good" or "bad" rather than if the candidate is actually the best option to make the next four years good or bad …

In the circumstances we have studied, we expect that animals (including humans) will take into account both current information and past information when making decisions. Your election example is an interesting one—contrast effects could certainly occur in this scenario. The clearest case would be where the person who was in charge for the past four years is not up for re-election, such that only new candidates can be chosen. We would expect judgements of the new candidates to be affected not only by their own strengths and weaknesses, but also by how good or bad things have been in the past four years. If the past four years have been bad, all candidates are likely to be rated more favourably than if the past four years have been good. This is a type of contrast effect.

A nice example of a contrast effect in humans comes from a real study that was done in the 1980s, by Doug Kenrick and Sara Gutierres. They got male undergraduate students to rate the attractiveness of an unfamiliar woman either before or during an episode of the original Charlie's Angels TV series, which was renowned for having three attractive female leads. The students gave lower ratings to the unfamiliar woman while they were watching Charlie's Angels, whereas a control group who watched a different TV show gave similar ratings before and during that show. This is consistent with a negative contrast effect: people respond less to a test stimulus (the unfamiliar woman) if they have recently been exposed to stimuli of higher value (the Charlie's Angels actresses).

Do humans or other more advanced animals use contrast effects less in their decision making? Is this something we're evolving away from?

No, I don't think so. As mentioned above, our study shows that contrast effects can, in fact, be rational. We would expect animals to be insensitive to past conditions in highly stable environments, where past conditions tell them nothing useful about the future. But most natural environments are probably too variable for that. Contrast effects are seen in a wide variety of animals, from honeybees to humans, so it seems that sensitivity to past conditions may be a generally useful trait.

How different do you think things would look if human civilization were governed by objective, rational thinking?

One of the biggest problems facing human civilisation, it seems to me, is impulsiveness—a focus on short-term goals, at the expense of what might be best for us in the long run. This is another evolutionarily widespread trait—many animals appear to show impulsive behaviour. That's not necessarily 'irrational' from an individual's point of view, but I do think it can create problems for future generations. If we could somehow keep our impulsive tendencies in check, I think humankind would have a much rosier future.

In doing this kind of research, how much more aware are you of your own decision making vulnerabilities?

In the Modelling Animal Decisions group at Bristol, we are investigating various types of behaviour that appear to be 'irrational' in some way. And in fact, what we're finding is that there are often good evolutionary reasons for why animals (including humans) sometimes behave in these surprising ways. Our tendencies to be impulsive, optimistic, depressed, anxious, or preoccupied with the past may all be deeply rooted evolutionary traits that help us to cope with a complex and changing world.

In doing this research I wouldn't necessarily say that I have become more aware of my own decision-making 'vulnerabilities', but I am certainly more aware of why, from an evolutionary point of view, I have those vulnerabilities.

Reach this writer at michaelb@motherboard.tv.