The theory of mind module seems to fit Fodor’s criteria reasonably well. It does appear to be domain-specific and informationally encapsulated; Segal [6] offers the following example as evidence for information encapsulation; ‘Watching a good actor can generate a sort of theory-of-mind illusion: even though one knows that he is not really in pain, or in love, or trying hard to solve a problem, it still seems to one that he is.’ So despite having information about the actor’s true mental state, the workings our theory of mind module are not influenced. The module also seems to operate quickly and obligatorily, and to have a characteristic ontogeny [7]; generally, the developmental sequence is as follows; one year olds are able to monitor eye gaze and adjust their gaze to fix another’s attention on an object, two year olds report their desires, attribute desires to others and explain behaviour by means of desires, three year olds use propositional attitude locutions, and four year olds are generally able to pass false belief tests. It is unclear whether or not there is a characteristic pattern of breakdown; sufferers of autism are unable to attribute false beliefs, but they appear to completely lack the faculty, and so do not provide us with any evidence of breakdowns within it. It is also not yet known whether the theory of mind module has dedicated neural architecture (although it is currently thought that the orbito-frontal cortex plays a vital role [8]).
Modularity theorists assume that concepts are innate and do not develop over time; the child has the adult concepts from the beginning. What does develop is the child’s ability to employ the concepts, which may partly depend on some linguistic triggering from the social environment, and partly on the maturation of the child’s cognitive capacities. Fodor proposes, for example [9], that the three-year-old already has adult concepts of belief but that the child’s ability to integrate information about the beliefs of others is restricted by processing limitations (such as current linguistic skills), and therefore adversely affects his success in false-belief tasks.
The Theory Formation Approach
Several authors argue against the modularist view, making the case that theory of mind is better explained by appeal to developing theories. The theory-formation view, argued by those such as Gopnik and Wellman [10], proposes that at early stages of development, the child possesses concepts quite unlike the adult ones, but arrive at the adult concepts by a process of scientific reasoning similar to that used by adult scientists; they also gather data, formulate and test hypotheses and alter these concepts when faced by contradictory data thus: -
- The theorist holds an initial theory which is confronted by counterevidence. When first presented with the counterevidence, the theorist may ignore it, treating it as noise
- The next stage is to bring in new theoretical apparatus, but use it only in auxiliary hypotheses, allowing the retention of the core of the original theory.
- The third stage is to use the new theoretical apparatus in other parts of the theory, but only apply it in limited contexts, still keeping the original theory on centre stage.
- Finally the new apparatus becomes central, and a new theory is organised around it.
It is said between the ages of two and five, the development of theory of mind follows this pattern. By the age of five, children have fully acquired the adult conception of beliefs and other representational states and attitudes. Theory of mind development mirrors the standard theory change lifecycle. The first step is that of denial; three year olds do sometimes deny counterevidence: if a cup is blue and an adult says ‘I think it is white’, three year olds tend to insist that the adult thinks it is blue [6]. Children are notably impervious to any evidence that contradicts their current theories in other domains as well. We can attribute this persistent unwillingness to be corrected to the fact that the child has not acquired the concept of false belief. Equally, this seems to be the downfall of the theory theory view; how can a child not yet equipped with a theory of mind acknowledge this counterevidence if it is unable to recognise any representation of the world other than that which it currently holds? When the adult scientist is confronted by counterevidence, it occurs to him that his theory may be false, and therefore considers alternatives, but infants have not yet acquired the complex concepts necessary to entertain this distinction. So it seems that at least one of the processes responsible for theory change in science cannot be applied to the acquisition of a theory of mind. The rest of the story says that the child then begins to conceive new theoretical concepts in order to account for the counterevidence. These appear first in auxiliary hypotheses: three year olds do begin to show understanding of genuinely representational states, but this first applies only to desires and perception and not to other concepts of their theory. They finally arrive at an alternative to their original theory, the adult theory of mind.
Another objection to the theory theory view is illustrated by sufferers of Williams syndrome; children suffering from Williams syndrome are able to acquire theories of mind, but are severely impaired when it comes to acquiring theories in all other areas. This suggests that the theoretical acquisition processes involved in developing a theory of mind are specialised, a prediction made by the domain specificity of the modularist view. There is a clear parallel in the language faculty, where acquisition is also task-specific; children do not acquire language using general learning processes that can apply in various domains. Instead the faculty deploys very specific principles which are only suitable for learning language.
The theory theorists use the different worlds argument in defence of their view. As noted above, all children develop the same, adult end-state theory of mind. The theory theory predicts that were we to place the child in a radically different environment, this end-state theory, based on the data available to the child, would be accordingly different, while the modularity hypothesis predicts that the resulting theory would be the same regardless of the environment. Gopnik and Meltzoff [11] suggest,
There is, in principle, a simple experiment that could always discriminate modularity theory and theory theory. Place some children in a universe that is radically different from our own, keep them healthy and sane for a reasonably long period of time, and see what they come up with. If they come up with representations that are an accurate account of our universe, modularity is right. If they come up with representations that are an accurate account of their universe, the theory theory is right.
They go on to describe hypothetical situations of significantly different environments, predicting that human children subject to such altered input would develop accordingly accurate theories, theories different to those acquired in our normal environment. But according to the suggestion above, if modularity theory is correct, then all sane and healthy children will develop the same end-state theory regardless of their environment. And this would be true if the modularity theorist did actually claim that children were innately programmed to develop one specific theory in a given cognitive domain. But they do not. Again Chomsky’s account of language acquisition comes in handy as a modularity theory paradigm. Since there are currently estimated to be approximately 6000 languages, it is not entirely accurate to say that modularity theories require that all sane and healthy children eventually develop the same theory regardless of their environment. If we recognise that modular accounts allow for a massive range of options available for the child’s end-state theory, the shortcoming of Gopnik and Meltzoff’s experiment is brought to light; there are possible results which would disprove the theory theory, but none which would disprove the modularity theory.
Does the existence of early performance deficits favour either theory? The child’s representations change and become more accurate over time. The fact that the child is born with erroneous representations which are later replaced by more accurate versions seems to count in favour of the theory theory and against the modularity theory. The modularity theory is antidevelopmental; modules do not lead to predictions through ongoing observation and hypothesis formation, but because they are designed by evolution to so do. But why would evolution have designed an incorrect module, only to have it eventually replaced? Developmental biology shows us that ontogeny recapitulates phylogeny [12]; early in their development, primate foetuses have structures that resemble gills; as development proceeds, these structures disappear and are replaced by embryonic lungs; it appears that in the course of development, organisms pass through stages that resemble the adult stage of their evolutionary forebears. Since this pattern is common in ontogeny, it is reasonable for the modularity theorist to propose that human children develop a sequence of incorrect modular representational systems, each of which is replaced by a more accurate system as the child matures. These inaccurate systems may simply be the developmental vestiges of our evolutionary predecessors’ adult systems. It would be interesting to see if there is indeed a correlation between the adult chimpanzee’s metarepresentational abilities and those of the human infant; at what stage is recapitulation evident?
Conclusion
The main difference between the modularity theory and the theory theory is to do with the relation between environmental input and the resulting representational systems, or theories. On the theoretical view, input is evidence that is crucial for determining the nature of the child’s current theory and the sequence in which future theories appear, while on the modular view, relevant experience may or may not trigger the use of a new theory, but the nature of this theory is genetically predetermined, as is the relation between future inputs and theories.
So far I have assumed that modules and theories are antithetical to each other. But we have only considered one form of modularity, that where modules are genetically determined and static. We could also consider Karmiloff-Smith’s proposal of modularisation [13], that modules may be entirely acquired; a result of development rather than a means. In addition, there exists the possibility that both kinds of development could exist in parallel; the theory of mind might be grounded in a cognitive module and yet still afford internal development; auxiliary modules may come on-line later in development. Gopnik [14] also suggests the possibility that information from the modular systems might feed into the central theory-formation system,
Once in the theory-formation system the information guides action and language in a productive way and is subject to the same kinds of revision and restructuring as other kinds of information. … they [modules] could serve as the source of a particular class of innate theories. They could be an important source of information for theory formation.
The status of the language faculty
Here we consider two notions; that language is fundamentally involved in the development of a theory of mind, as well as the inverse, that a theory of mind is causally necessary for the acquisition of the child’s first language. There are good grounds for thinking that the language faculty and our theory of mind abilities are closely related. A theory of mind is clearly needed for the production and understanding of speech, especially for pragmatic elements such as irony and metaphor, and it appears that language is a necessary input for the normal acquisition of theory of mind in young children.
Language is Necessary for the Emergence of Theory of Mind
Chomsky [15] suggests that there is a level of linguistic representation, ‘logical form’ (LF), which is where the language faculty interfaces with the central cognitive systems, and that the language faculty has access to the outputs of the central system, and can therefore build LF representations which combine information across domains. This enables us to explain why the theory of mind needs to access the resources of the language faculty in order for us to describe the contents of the thoughts being attributed to ourselves and others, if language does indeed play a part in linking together different modules. Carruthers [16] uses the following example,
For if geometry and colour (say) can’t be combined in a single thought without language, then one could hardly expect the mind-reading faculty to be able to attribute a thought to another (or to oneself) which conjoins geometry with colour without deploying language! This would be to give that faculty almost-magical super-properties possessed by no other module. No, if entertaining the thought that the object is to the left of the blue wall requires tokening the LF representation, ‘The object is to the left of the blue wall’, then ascribing to someone the belief that the object is to the left of the blue wall would similarly require the use of that LF sentence.
One way of assessing whether or not language is involved in children’s theory of mind developing is looking at whether children with language impairments have comparable difficulties in acquiring the theory. Baron-Cohen [17] finds that although theory of mind concepts are difficult for sufferers of autism, all other concepts develop normally. Happé [18] uses verbal mental age to predict success of the theory of mind tasks; a verbal mental age greater than 11.5 years guarantees that the subject has a theory of mind. This finding provides evidence that linguistic skills support the development of theory of mind. Autistic children suffer from severe language impairments, so it is not surprising that they are impaired in the development of theory of mind. If theory of mind was an autonomous module specifically impaired in autistic children and independent of linguistic capacities, then it would never be acquired, but the fact that high-functioning autistics with adequate language skills are able to acquire a theory of mind illustrates that language is implicated in its development. However, Christopher, the polyglot savant [19] is a counterexample; despite having a verbal mental age which, according to Happé, should result in success at theory of mind tasks, he displays a marked theory of mind deficit.
Looking at cross-linguistic evidence should also indicate whether or not the development of theory of mind is parasitic on the language faculty. If children who are speakers of languages with a heavy dependence on evidentials such as Korean, can perform false-belief tasks earlier, as suggested by Papafragou [20], then this is surely evidence for the importance of language in the acquisition of theory of mind? It appears that this is the case, but at present there is insufficient evidence in support of the claim.
Theory of Mind is Necessary for Language Acquisition
As shown by pragmatists such as Sperber and Wilson, it is recognised that communication demands that both the speaker and hearer recognise mutual metarepresentations of each others’ mental states, ‘Full-fledged communicative competence involves, for the speaker, being capable of having at least third-order meta-representational communicative intentions, and for the hearer, being capable of making at least fourth-order meta-representational intentions.’ Sperber and Wilson’s view of the mechanics of the communicative process highlights the role that a theory of mind plays in normal communication, for all utterance interpretation involves an inference to a hypothesis about the speaker’s intentions. For instance, in order for the library user to interpret the statement ‘Is anyone sitting here?’ as a request to remove her belongings from the neighbouring seat clearly involves the construction of a number of complex inferences about the speaker’s intentions. When we consider more sophisticated communicative phenomena such as irony and reported speech, the required inferences become even more complex. However, Papafragou entertains the suggestion that the success of early communication may simply be due to the child’s increasing familiarity with conversational routines; for instance the child knows that an expression of his inability to perform a given task is likely to result in a parent coming to his aid.
Papafragou also suggests that a theory of mind is heavily involved in particular elements of vocabulary acquisition, especially for morphemes which encode metarepresentational concepts, such as evidentials (e.g. allegedly and probably), ‘the ability to reason evidentially about the origins, strength and reliability of our beliefs is part of our ability to reason about mental states in general’ [20] and epistemic modals (e.g. must and may), ‘successful use of epistemic modals requires the speaker to perform deductive operations on abstract propositions (i.e. on the content of her beliefs)’ [21].
Conclusion
Neither the modular nor the theoretical accounts of theory of mind development seem to have all the answers; it is possible that in reality a combination of both paradigms exist, where an innate core architecture is revised internally based on the child’s experiences. However it does seem that the purely theoretical view relies on the circular reasoning that the child must alter its theory when confronted by contradictory data, which cannot be recognised unless the child has already acquired a theory of mind.
It appears that the development of the child’s linguistic abilities does not proceed in isolation, but feeds into and is influenced by the development of other modules of the mind, such as that of theory of mind, which is in turn influenced by the workings of the language faculty; this mutual dependence implies that the development timetables of the two faculties are also intertwined at each stage of the child’s development.
References
-
Wimmer, H and J. Perner (1983) ‘Beliefs about beliefs: Representations and constraining function of wrong beliefs in young children’s understanding of deception’ Cognition 13:103-128
-
Baron-Cohen, S. (1994) ‘How to build a baby that can read minds: Cognitive Mechanisms in Mindreading’. Current Psychology of Cognition 13:513-552
-
Leslie, A. M. (1987) ‘Pretence and representation, the origins of theory of mind’. Psychological Review 94:412-426
-
Harris, P. (1990) ‘The child’s theory of mind and its cultural context’. In G. Butterworth and P. Bryant (eds) Causes of Development. New York: Harvester Wheatsheaf 215-237
-
Fodor, J. (1983) The Modularity of Mind Cambridge MA: MIT Press
-
Segal, G. (1996) ‘The modularity of theory of mind’ In P. Carruthers and P. Smith (eds) Theories of Theories of Mind. Cambridge, CUP. pp.141-157
-
Garfield, J, C. Peterson & T. Perry (2001) ‘Social cognition, language acquisition and the development of theory of mind’. Mind & Language 16:494-541
-
Segal, G. (1996) ‘The modularity of theory of mind’ In P. Carruthers and P. Smith (eds) Theories of Theories of Mind. Cambridge, CUP. pp.141-157
-
Fodor, J. (1992) ‘A theory of the child’s theory of mind’ Cognition 44:283-296
-
Gopnik, A. and Wellman, H.M (1992) ‘Why the child’s theory of mind really is a theory’. Mind & Language 7:145-171
-
Gopnik, A. and Meltzoff, A (1997) Words, Thoughts and Theories. Cambridge MA: MIT Press
-
Povinelli, D. (1996) ‘Chimpanzee theory of mind?: the long road to strong inference’ In P. Carruthers and P. Smith (eds) Theories of Theories of Mind. Cambridge, CUP. pp.293-329
-
Karmiloff-Smith, A. (1992) Beyond Modularity: A Developmental Perspective on Cognitive Science. Cambridge MA: MIT Press
-
Gopnik, A. (1996) ‘Theories and modules: creation myths, developmental realities and Neurath’s boat’ In P. Carruthers and P. Smith (eds) Theories of Theories of Mind. Cambridge, CUP. pp.293-329
-
Chomsky, N. (1995) The Minimalist Program. Cambridge MA: MIT Press
- Carruthers – cognitive functions of the language faculty
-
Baron-Cohen, S. (1995) Mindblindness. Cambridge MA: MIT Press
-
Happe, F (1995) ‘The role of age and verbal ability in the ToM performance of subjects with autism’. Child Development, 66:843-855
- Smith, N.V and Tsimpli, I.-M. (1995) The Mind of a Savant: Language Learning and Modularity. Oxford: Basil Blackwell
-
Papafragou, A (2002) ‘Mindreading and verbal communication’. Mind & Language 17:55-67
-
Papafragou, A. (1998) ‘The acquisition of modality: Implications for theories of semantic representation’. Mind & Language 13:370-399