I’ve seen it every time I had to take oral exams from my students: they try to “cold read” my face to know if they are heading in the right or wrong direction with their answer. These exams are my personal training in poker face. But actually, when teaching or giving public speeches, I also try to read the audience. People regularly engage in sophisticated ‘mentalizing’ (i.e. guessing the intentions or beliefs of others) whenever they convince, teach, deceive, and so on. This research (with a rather small experimental group) published this week in PLOS Computational Biology demonstrates the laws that govern these intuitions and how efficient they are for anticipating the behaviour of other people.
From the press release:
Jean Daunizeau and colleagues from INSERM and CNRS combine mathematical modelling, experimental psychology and behavioural economics to measure the sophistication of human ‘mentalizing’.
The authors asked 26 participants to play repeated games against artificial (Bayesian) ‘mentalizing’ agents, which differ in their sophistication. Critically, the participants were told that they were either playing against each other, or that they were gambling without any in-the-flesh opponent, like in a casino. The results show that participants won against the artificial ‘mentalizing’ agents when the game was socially framed, and lost in the non-socially framed games.
This study demonstrates that ‘mentalizing’ enables humans to guess how others learn about themselves, even in the absence of any explicit communication. This mental skill increases the chances of success in the context of repeated competitive social interactions.
The authors are currently applying this work to assess how this ‘mentalizing’ process and learning ability may differ in people with autism spectrum disorders, and neuropsychiatric conditions, such as schizophrenia.
The researchers say: “Our work is in line with an ongoing effort tending toward a computational (i.e. quantitative and refutable) understanding of animal cognition. Importantly, we showed that human ‘mentalizing’ intuitions are endowed with remarkable but limited sophistication, notwithstanding how critical they are for deciphering intentional behaviour.”
Abstract of the research:
When it comes to interpreting others’ behaviour, we almost irrepressibly engage in the attribution of mental states (beliefs, emotions…). Such “mentalizing” can become very sophisticated, eventually endowing us with highly adaptive skills such as convincing, teaching or deceiving. Here, sophistication can be captured in terms of the depth of our recursive beliefs, as in “I think that you think that I think…” In this work, we test whether such sophisticated recursive beliefs subtend learning in the context of social interaction. We asked participants to play repeated games against artificial (Bayesian) mentalizing agents, which differ in their sophistication. Critically, we made people believe either that they were playing against each other, or that they were gambling like in a casino. Although both framings are similarly deceiving, participants win against the artificial (sophisticated) mentalizing agents in the social framing of the task, and lose in the non-social framing. Moreover, we find that participants’ choice sequences are best explained by sophisticated mentalizing Bayesian learning models only in the social framing. This study is the first demonstration of the added-value of mentalizing on learning in the context of repeated social interactions. Importantly, our results show that we would not be able to decipher intentional behaviour without a priori attributing mental states to others.