 |
| Uwe Peters |
Uwe is currently writing up his dissertation on the accessibility and self-knowledge of beliefs at King's College London. He develops an account of self-knowledge on which our privileged, non-inferential self-knowledge of beliefs is grounded in and dependent on our inferential knowledge of other people’s mental states.
In my research, I often appeal to cases of atypical cognition. In a recent paper, I argue that findings on autism pose a challenge to the following view of conscious thinking. On the basis of empirical studies and theoretical considerations, it has been argued that conscious thinking, i.e., thinking that involves the deployment of the components of working memory, involves only sensory-imagistic events such as episodes of inner speech or visual imagery but no attitudes, e.g., judgments or decisions.
The claim is that in conscious theoretical or practical reasoning, we only have indirect access to our own attitudes via their sensory-imagistic expressions which don’t themselves qualify as attitudes, for they don’t play the right causal role. They can attain attitude-like roles (e.g., settle reasoning and justify inferences and report) only after they have been unconsciously interpreted by the mindreading system and their underlying attitudes have been self-ascribed (see Carruthers 2011, 2013; Fletcher and Carruthers 2012; Frankish 2009, 2012; Frankish agrees that conscious thinking is based on sensory-imagistic events but claims that some of them can be “virtual” attitudes. He too, however, maintains that they depend on unconscious mindreading and metacognitive attitudes to become “effective”).
This view of conscious thinking is challenged by data on autism. It is well known that many subjects with autism have an impaired mindreading capacity (Baron-Cohen 1995; Frith and Happé 1999). If the events and states involved in conscious thinking are sensory-imagistic in nature and have to be unconsciously interpreted by the mindreading system, then subjects with autism who have a defunct mindreading faculty should perform poorly in tasks that require conscious thinking.
However, this is not generally the case. For instance, Scott and Baron-Cohen (1996) found that children with autism and an impaired mindreading capacity performed comparably to neurotypical controls on, e.g., analogical reasoning tasks which required conscious thinking, for subjects had to access the contents of the beliefs that their conclusions were based on in order to justify their conclusions. Whatever events or states they accessed at the conscious level, for them, they settled reasoning and justified their responses.
There are other studies that indicate that in subjects with autism first-order reasoning implicating working memory remains often intact (e.g., Dawson et al. 2007; Morsanyi and Holyoak 2010) while their theory-of-mind capacity is impaired (e.g., Scott et al. 1999; Baron-Cohen 1997, Baron-Cohen et al. 2001). The evidence of such a dissociation is intriguing and suggests that, unlike some philosophers claim (e.g., Fletcher and Carruthers, 2012; Carruthers 2011, 2013 Frankish 2009, 2012 , it is not necessary for the events or states involved in conscious thinking to be effective and play attitude roles that they are accompanied by theory of mind processing and metacognitive attitudes. For it seems unlikely that comparable performance in the same kind of reasoning requires in neurotypical subjects more sophisticated cognitive processing than it does in people with autism.
It would be nice if more experiments on autism were conducted that combine conscious reasoning tasks (e.g., System 2, working-memory dependent judgment- and decision-making) with a theory-of-mind test. The results of these studies might provide useful constraints on philosophical theorizing about conscious thinking and the nature of the states involved in it.
In my research, I often appeal to cases of atypical cognition. In a recent paper, I argue that findings on autism pose a challenge to the following view of conscious thinking. On the basis of empirical studies and theoretical considerations, it has been argued that conscious thinking, i.e., thinking that involves the deployment of the components of working memory, involves only sensory-imagistic events such as episodes of inner speech or visual imagery but no attitudes, e.g., judgments or decisions.
The claim is that in conscious theoretical or practical reasoning, we only have indirect access to our own attitudes via their sensory-imagistic expressions which don’t themselves qualify as attitudes, for they don’t play the right causal role. They can attain attitude-like roles (e.g., settle reasoning and justify inferences and report) only after they have been unconsciously interpreted by the mindreading system and their underlying attitudes have been self-ascribed (see Carruthers 2011, 2013; Fletcher and Carruthers 2012; Frankish 2009, 2012; Frankish agrees that conscious thinking is based on sensory-imagistic events but claims that some of them can be “virtual” attitudes. He too, however, maintains that they depend on unconscious mindreading and metacognitive attitudes to become “effective”).
This view of conscious thinking is challenged by data on autism. It is well known that many subjects with autism have an impaired mindreading capacity (Baron-Cohen 1995; Frith and Happé 1999). If the events and states involved in conscious thinking are sensory-imagistic in nature and have to be unconsciously interpreted by the mindreading system, then subjects with autism who have a defunct mindreading faculty should perform poorly in tasks that require conscious thinking.
However, this is not generally the case. For instance, Scott and Baron-Cohen (1996) found that children with autism and an impaired mindreading capacity performed comparably to neurotypical controls on, e.g., analogical reasoning tasks which required conscious thinking, for subjects had to access the contents of the beliefs that their conclusions were based on in order to justify their conclusions. Whatever events or states they accessed at the conscious level, for them, they settled reasoning and justified their responses.
There are other studies that indicate that in subjects with autism first-order reasoning implicating working memory remains often intact (e.g., Dawson et al. 2007; Morsanyi and Holyoak 2010) while their theory-of-mind capacity is impaired (e.g., Scott et al. 1999; Baron-Cohen 1997, Baron-Cohen et al. 2001). The evidence of such a dissociation is intriguing and suggests that, unlike some philosophers claim (e.g., Fletcher and Carruthers, 2012; Carruthers 2011, 2013 Frankish 2009, 2012 , it is not necessary for the events or states involved in conscious thinking to be effective and play attitude roles that they are accompanied by theory of mind processing and metacognitive attitudes. For it seems unlikely that comparable performance in the same kind of reasoning requires in neurotypical subjects more sophisticated cognitive processing than it does in people with autism.
It would be nice if more experiments on autism were conducted that combine conscious reasoning tasks (e.g., System 2, working-memory dependent judgment- and decision-making) with a theory-of-mind test. The results of these studies might provide useful constraints on philosophical theorizing about conscious thinking and the nature of the states involved in it.