ToM (Theory of Mind)

Theory of Mind ToM is the ability to interpret the behaviour of the self and others in terms of mental states which are not directly observable. Most human behaviour is usually best understood in terms of desires and beliefs (i.e. going to the fridge. Why are you going to the fridge? To get a drink. One has the belief that there are drinks in the fridge).

ToM is present in all healthy, neurologically normal, adults. People and cultures may differ in their specific theories – for instance, some cultures believe that is it reality that a God or something entity (human or otherwise) may see thoughts. So though most believe thoughts are private, some ToM incorporate this influence.

Baron-Cohen argues that Autism is the extreme end of the Male spectrum (males are ‘things’ people – they like non-abstract toys as children, for instance, trucks and balls). Males are ‘systematizes’ and not ‘people’ people.

Schizophrenia is the extreme end of the female spectrum (females like people, i.e. play with dolls) – schizophrenia is characterized by hallucinating social interactions (loosely speaking). Premack and Woodrum (1978) first looked at ToM in chimps. By showing their chimp (Sarah) 4 videos of man attempting to unsuccessfully complete 4 separate tasks, then providing Sarah the opportunity to pair a tool or strategy to the task they concluded three things:

  1. Chimps understand photos and video
  2. They understand problem solving (means-end relations); most importantly they can
  3. attribute intention to an actor – thus they understand mental states.

There was much commentary and criticisms of the claims, the most salient being: In order to demonstrate a ToM one needs to demonstrate the capacity of an individual to act on a misrepresentation of the world. Sarah may not know the man is trying to complete task x, but she herself knows how to complete (or at least associate pairs) with that task.

Thus, if the shared understand of reality is the actual reality it need to be shared, as it is independent. However, understanding something in another which must be shared (for it is not otherwise self-evident) is much more convincing for evidence of ToM.

Thus, the false-belief task was developed. The false-belief task was first developed by Wimmer and Perno (1983) – and it’s various version include the Sally-Ann task, and the Smarties task. A meta-analysis in 2001 suggest that there is a conceptual shift between age 3 and 5 that results in development of above chance performance on ToM tasks.

A problem remains as to how to test ToM without language – as language and stories are the basic way to test competence in humans. ToM in action (football) – it was illustrated in class that a soccer player may run with the ball, fake a certain action to confuse a defender, then change their behaviour to run past successfully. Is this ToM? We would say it was as there was an intention to fool the defender, however, a behaviour such as described might simply be a successful strategy struck upon by luck and learning by conditioning.

Most evidence for ToM in non-humans may be explained in this manner. Behaviour indicator for mind-reading has proved frustratingly difficult to find (that is, trying to measure the extent/accuracy of behaviour in apes based on their ability to interpret our mental states / physical capacities, etc). So far apes have failed the false-belief task.

Chimps will beg to individuals with food. If a chimp has the choice of begging off two people with food (one of whom has a bucket on their head covering their eyes/face – the chimp is equally likely to try to beg off the ‘blind’ individual as the uncovered individual).

There is also the guesser/knower paradigm – whereby an experimenter hides a banana in one of two places (the chimp sees the two places, but cannot know where the banana is hidden). The two experimenters than point to who has the banana (based on the assumption only the seer can know. The chimp cannot extract this info. Chimps will, however, do better on a competitive task (not a cooperative task) more on this later. Povinelli suggests that ToM-like behaviour may not be caused by ToM itself.

Thus (according to Suddendorf and Whiten) ToM-like behaviour may not be directly caused by ToM reasoning – it may be because other more primitive mechanisms govern it (learning); or because it has become an automatic response.

Chimps will also select a box that has been marked for food over a box that has been unintentionally marked. Ownership of food is respected by chimps. If food is not owned then two chimps will scuffle for ownership – and may even engage in tactical deception and decision making. This is arguably closer to what chimps are capable of, as cooperation is not in their nature.

However, this may be more parsimoniously explained by ‘subdominant will not approach food if dominant has gaze at it’ because that is likely to lead to a beating. If they haven’t gazed at it, it’s not associated with a beating. Thus, it remains possible there is no display of ToM-like capacities at all. Though there is no shared intentionality, no sharing of info or attention, (but toddlers do and can), no evidence for metarep (i.e. false-belief, recursion, executive control); it is argued that language, MTT and ToM are requisites – which (currently) are beyond Apes. So, Suddendorf and Whiten reinterpret the reinterpretation hypothesis.

It is a good idea to check your scientific articles on the Robot Don Text checker to avoid mistakes and plagiarism.