Damn you, Theory of Mind!
Arguably one of humanity’s greatest powers is our ability to attribute states of mind to other people, and to understand that these may be different from our own. But it is a double-edged sword…
US president Joe Biden’s announcement, on Sunday 21st July, that he would no longer pursue his candidacy for the presidency, was certainly headline-grabbing. It was also and, for many at least, as predictable as the sunrise the next morning – it had increasingly become a matter of when, rather than of whether he would withdraw. What was intriguing was how many of his copartisans in the Democratic leaning population considered his decision to take a step back as a selfless act of heroism. Many of them had not long ago joined a growing group of Democratic representatives, senators and supporters, calling for his withdrawal (and no doubt castigating him for his stubbornness). On the opposing side, Biden was seen as weak and cowardly, and some saw the hand of the “backroom elites”, the “men in grey suits” who “plunged the knife in”. How on earth could all these people know so clearly what Joe Biden was thinking?
Mind readers
The concept of the Theory of Mind (ToM) finds its origin in research by two primatologists, David Premack and Guy Woodruff, in which they explored whether non-human primates can understand others’ mental states (e.g., intent or purpose, knowledge, doubt, belief, liking, etc). Nearly fifty years later, the question whether non-human animals have a ToM remains controversial, but for humans, the concept has become central in developmental psychology, cognitive science and the philosophy of mind.
The theory refers to the capability of recognizing and understanding distinct mental states in oneself and in others. It is thought to develop in humans between the ages of 3 and 5, though some of the more complicated aspects, such as sarcasm and irony, mixed emotions or white lies, emerge later. Much like the fish in the parable the writer David Foster Wallace used in his commencement speech are (presumably!) unaware of what water is, because it is there all the time, we are not typically consciously aware of using our ToM because we use it continually. But let’s take just a small step back, and reflect on the guy who sat opposite us on the train, incessantly looking at his phone and at his watch – how do we know what we think we know about his mental state? Or, thinking of that time when, being a European in America, we were waiting for our client on the wrong floor, ignorant as we were about the difference in floor counting – how did we realize that two people could think differently about the same thing? Just briefly engaging in this kind of metacognition, thinking about how we, and others, think, we can see the power of ToM. While we cannot literally read other people’s minds, we can infer, from their actions, what might be their mental state, and explain their present or past behaviour. We can similarly try to predict what they might do based on what we think we know about their intentions, and about their beliefs, their knowledge, their goals and so on.
Imagine you are at a party, and a friend is sitting on her own, staring out of the window and not engaging with the other guests. Without ToM, you would be unable to notice her emotional state, or understand why she might be acting differently from everyone else. If you continued acting cheerfully, she might feel ignored or misunderstood. But thanks to your ToM you can recognize the signs of emotional upset, and act accordingly: ask whether she is OK and offer to go somewhere more private to talk. Or say you have to explain a complex topic to a colleague. Without ToM you would be unaware of what he already knows, and what he doesn’t – potentially even succumb to the curse of knowledge and use technical jargon and advanced concepts that goes over his head, leading to frustration all round. Our ToM allows us to establish our colleague’s background and prior knowledge, adjust our explanation accordingly, use sensible analogies and regularly check his understanding. ToM is clearly indispensable across a wide range of social interaction.
Misreading minds
Unfortunately, this is not the whole story. We cannot actually read other people’s minds, and so any conclusion we draw will be imperfect. That doesn’t stop us being overconfident, though, and being more certain about our inferences of other people’s mental state than we really should. Yet, the potential for imperfections is substantial, thanks to a variety of biases that we exhibit. Egocentric bias, for example, our tendency to use our own thoughts and feelings as a reference point, may lead us to overestimate how much others share our beliefs, attitudes and behaviour, and assume consensus where there is none (false consensus effect). It may also cause us to project our thoughts, feelings and motivations onto others, thus misinterpreting their behaviour. Our propensity to prefer simple explanations and quick judgments can set us up to make the fundamental attribution error, ascribing other people’s behaviour to their personality rather than to circumstantial factors, or to rely on thinking in cultural stereotypes when inferring their mental states.
But perhaps the most pernicious influences on how we apply our ToM involve these two tendencies: confirmation bias and motivated reasoning. The former describes how, rather than using the observed behaviour (and any other knowledge we think we have) to establish someone’s mental state, we start from an assumed mental state and only take into account any evidence in the observed behaviour that corroborates our assumptions. The latter starts from the emotions the observed behaviour evokes, and based on whether these are positive or negative, we reason towards a corresponding mental state.
A combination of these might explain some of the interpretations of Joe Biden’s decision. For the democrats who favoured Biden’s withdrawal, the decision was positive, and since he is part of their tribe, his mental state is praiseworthy. This naturally makes him a hero for making a huge sacrifice. (The fact that he could actually have chosen not to run in the first place and allow younger, capable candidates to throw their hat in the ring is quickly forgotten, as is the fact that the ‘heroism’ required a good deal of pressure.) A good outcome in a precarious situation makes Biden a hero, and therefore his decision was selfless, which, of course, confirms his heroism.
For the Republican supporters of Donald Trump, the decision was not so positive. Biden was an opponent easy to attack, a bumbling, clumsy old man, especially after the disastrous TV debate. To them, he is therefore not wise and heroic, but a coward running away from certain defeat given his lagging behind in the polls – or better still, a weakling whose decision has been made for him by a shady backroom elite. A bad choice cannot be made by someone of good character, and the fact itself that he pulled out confirms that there is a lot amiss with him.
Unlike the potential flaws mentioned earlier, which can be tackled through deliberate practice and metacognitive reflection, these two are a lot harder to address. Sometimes we simply are not after accuracy and nuance, but after feeling good and signalling allegiance, and that makes confirmation bias and motivated reasoning mighty allies.
Theory of Mind is an incredibly powerful ability, but we should beware of believing we can actually know what is in someone else’s mind. Unfortunately, polarization motivates us to do just that: we are convinced we know exactly what outgroup members are thinking and feeling.
It is tempting to wish for it not to exist – without ToM, we would simply observe facts and act accordingly, without ascribing intentions, desires, beliefs, preferences and aversions to others – and to think, “Damn you, Theory of Mind!”. But we cannot turn the evolutionary clock back. We are stuck with this double-edged sword, and we have to use it responsibly if we don’t want it to get the better of us. Good luck to us!
(PS: I leave it as an exercise to the reader to spot the instances in this article where my use of ToM is at least debatable.)