Conway, J., Long, E., Sevi, L., Catmur, C., & Bird, G. (2024). Theoretical Limitations on Mindreading Measures: Commentary on Wendt et al. Psychological Assessment.
A Theory of Mind (ToM) ability is made up of a wide range of skills that allow us to guess the mental state of others. This may include predicting what someone else is thinking, feeling, their attitudes or intentions and more. It is an ability that is thought to rely on a whole bunch of other cognitive skills.
This paper sees potential problems with traditional ToM tests by identifying the types of skills that are necessary for ToM judgements. Often, as psychologists, we are attempting to design experiments provide measurement for one skill without influence from other processes. With ToM doing this is difficult (maybe even impossible!) because there might not be a specific ‘theory of mind module’. Instead, our general cognitive ability allows us to make guesses based on previous information and experience. For example, we may use our knowledge that introverts often don’t like large groups to predict that they may not be as happy at a party compared to an extrovert. And this uses the same cognitive skills to predict match day outcomes based on football player statistics. So, what is special about theory of mind?
Geoff and colleagues believe that it is possible to separate ToM into two parts; general cognitive ability and relevant social knowledge, this is the Mindspace framework. This theory predicts that you will be better at inferring the mental state of a friend compared to that of a stranger because you have more information about your friend that can be used to predict their behaviour. The problem is that most of our traditional ToM tasks do not provide information about the imaginary characters involved and only measure the ToM ability that would exist towards a stranger. This neglects the skills required to use social knowledge to make accurate inferences. This means that the traditional ToM tasks measure a too simple model of ToM and so may underestimate overall ToM ability. In separating these two components, Geoff and colleagues highlight the importance of using psychological theories to figure out exactly what an experiment is measuring.
Traditional ToM tasks also use ‘correct answers’ based upon general opinion. For example, many people would say that if an actor is crying then they are portraying sadness (if the actor is good enough) but the chances that the actor is actually feeling sadness is perhaps quite low. It might be just as accurate to predict the actor is simply wanting to get paid! Some more recent studies have corrected for this by using videos of real social interactions where a true mental state is experienced but these tasks still rely on behaviour being similar to those that the participant has experienced or seen. This suggests that our performance in traditional ToM tasks may be culturally bound to our level of experience with neurotypical behaviour.
This paper shows that we – as psychologists – need to adapt ToM tests… or at least acknowledge that they might not tell the whole story!