Complex language, complex thought?

(1)

Complex language, complex thought?

Spontaneous and elicited production of complex syntax

by children with different levels of false-belief reasoning.

Suzanne Theresia Maria Hazenberg (1885847) Research Master Linguistics Language and cognition

(2)

Abstract

The fact that Theory of Mind (i.e. the ability to attribute mental states to oneself and others) develops during childhood, has inspired many studies on what factor(s) drive this development. One of the factors that is thought to influence Theory of Mind development is language. It is however unclear what aspects of language (general abilities, semantics or syntax) are a prerequisite for the development of first-order and second-order false-belief. The current study tested what parts of language correlated significantly with Theory of Mind development.

Two experiments were carried out: the first analysed the use of complex syntax and mental state verbs by 103 children (ASD, ADHD, TD) aged 6 to 12 in a spontaneous narratives production task, the other tested 55 children aged 7 to 10 on a double embedding sentence elicitation task. The spontaneous production task revealed that children with a better working memory and false-belief performance produced longer clauses and fewer morphosyntactic errors. Also, both first- and second-order false-belief reasoning were related to the use of main clauses, MLU and clause coordination. The strongest correlational pattern was found for the hypothesis on the link between general language abilities and Theory of Mind performance: the better the overall quality of the child’s narratives (longer MLU, lower rate of morphosyntactic errors, etc.), the better the child’s performance on the false-belief task.

Given that children are not likely to produce double embedded sentences in spontaneous production, a sentence elicitation experiment was designed to test a complex variant of De Villiers’ syntactic hypothesis: the foundation for Theory of Mind is the child’s understanding of tensed complements embedded in another clause. In the task, children had to produce single and double embeddings for pictures that showed cartoons with single or double embedded thought-bubbles and thinking clouds. There was a significant correlation between the production of double embeddings and the performance on second-order belief questions (r=0.49), but not between the double embeddings and the first-second-order false-belief questions. Children who performed well on both FB1 and FB2 questions were more likely to produce double embeddings than children who only mastered first-order false-beliefs; they were most likely to produce single embeddings.

(3)

(4)

(5)

1 Introduction

The French philosopher Simone Weil (1901-1943) once wrote that a mind enclosed in language is in prison. Her controversial claim points to the fact that language and mind are supposed to be intertwined: what people think may be limited to the relations they can think about in words. “So the mind moves in a closed space of partial truth, which may be larger or smaller, without ever being able to glance at what is outside” (Weil, 1986, p. 69). Although Weil’s philosophical analysis of language is suffused with a spirit of dreariness – emphasizing how language restricts human thought – it shows an interesting premise: language has an influence on the mind.

This premise has fueled philosophical, psychological and linguistic debates for many ages. One aspect of the mind that is studied in relation to language, is the ability to read the minds of other persons, or the fact that we possess a so-called Theory of Mind. It is fascinating that we, humans, not only think about ourselves and our beliefs, but also about others and their beliefs. When we do, the linguistic perspective is changing from speaking about observable facts to claiming things about the beliefs of other persons on these facts, whether these beliefs are in accordance with the reality around us (and our own point of view) or not.

Reasoning about the beliefs of another person is called first-order Theory of Mind, an ability that is cross-culturally acquired around the age of 4 (Wellman, Cross & Watson, 2001). It is also possible to add an additional level of complexity and claim things about the beliefs someone has about the beliefs of a third person, as in “Suzanne thinks that David thinks that his grandma will cry on their wedding day”. This so-called second-order Theory of Mind (double embedded mindreading) is a key competence for full-fledged social interactions and develops between the age of 7 and 9 (Hollebrandse, Van Hout & Hendriks, 2014).

(9)

Three hypotheses have been proposed to clarify the linguistic requirements for the development of first-order false-belief reasoning. These linguistic requirements vary from more semantic to more syntactic. The first hypothesis from the language-first proponents claims that mature language abilities are necessary for mindreading. A more specific hypothesis is the claim that children must have acquired the semantics of mental state verbs like to know or to think. A third prominent hypothesis was proposed by Peter and Jill de Villiers and colleagues: the driving force for first-order false-belief reasoning is the understanding of tensed complements.

De Villiers’ syntactic hypothesis claims that children must have acquired not only the lexical meaning of mental state verbs, but also the syntactic structure that is associated with them (i.e. tensed complements) and they must know that the embedded proposition can be false while the entire sentence is true (De Villiers, 2007). The child needs an advanced or complex syntax: utterances consisting of more than one clause, either through coordination, subordination or embedding (Vasilyeva, Waterfall & Huttenlocher, 2008). For false-belief reasoning, the child must understand that in the sentence ‘Orpheus thinks that Eurydice likes his music’, it is not certain whether Eurydice likes Orpheus’ music or not; we focus only on Orpheus’ beliefs. So only if a child masters this sentence structure and meaning, (s)he can pass first-order false-belief tasks. So, in De Villiers’ syntactic hypothesis, the complex syntax (and specifically the construction in which a tensed complement clause is embedded in another clause) is considered to be the driving force behind understanding different points of view: In other words, language provides an important foundation for understanding other peoples’ thoughts and beliefs.

Research into the relationship between complex syntax and second-order Theory of Mind is rather scarce, possibly due to the difficulties of eliciting double embedded syntactic structures. Yet, more insight in the linguistic mechanisms underlying Theory of Mind development is valuable in order to see how language and thought interact. A new study could contribute considerably to the debate when it could distinguish the way first-order false-belief relates to language from the way second-order false-belief relates to language. That is, it is not yet clear whether the development of second-order false-belief depends on the same linguistic requirements as first-order false-belief reasoning (De Villiers, 2007).

Therefore, this thesis focuses on the relationship between both first and second-order false-belief reasoning and children’s ability of producing complex syntax in Dutch. This ability is tested in both a spontaneous production task (exploring many aspects of complex syntax) and in an

elicited production experiment focused at eliciting double embedded sentential complements only

(10)

heavily on children’s pragmatic abilities, a visual sentence elicitation task will be developed, so that syntactic ability can be measured without interference of pragmatic knowledge. The main advantage of this approach is that it keeps linguistic abilities apart from false-belief an sich. So the central question of this thesis is: What elements of complex syntax are related to first- and second-order false

belief reasoning?

(11)

2 Theory of Mind

Before proceeding to a chapter on the relationship between language and Theory of Mind, it is important to clarify what the latter term refers to. Over the past decades, the term ‘Theory of Mind’ has become a catch-all term: its meaning nowadays varies from mindreading to intersubjectivity (Astington & Baird, 2005). Therefore, it is crucial to clarify how Theory of Mind is defined in the current study. The goal of this introductory chapter is (2.1) to present briefly how the concept of a Theory of Mind has entered into the scientific arena, (2.2) how Theory of Mind can be defined and (2.3) what is globally known about the development of this cognitive skill in both typically developing children and clinical populations (ASD and ADHD).

2.1 A brief history

The concept of a Theory of Mind was first defined in an article on primate psychology entitled

Does the chimpanzee have a Theory of Mind? (1978). Premack and Woodruff came up with the term to

explain observations where chimpanzees demonstrated behavior that required them to make inferences about what another chimpanzee might be doing. Although nowadays the term describes an aspect of human cognition, its original use was in the description of chimpanzee’s cognitive abilities. The authors state that “an individual has a Theory of Mind if he imputes mental states to himself and others” (Premack & Woodruff, 1978:515).

The idea of describing this ability as a ‘Theory’ was inspired by two reasons: firstly, mental states are not directly observable and secondly, the Theory of Mind is as a whole system that makes specific predictions about the behavior of others (e.g. their purpose, intention, knowledge, belief, thinking, doubt, guessing, pretending and liking). The concept of a Theory of Mind soon came to the notice of developmental psychologists and the focus on young children quickly became the main research area. Only a few years after the publication on chimpanzee Sarah, the concept was applied to young children to study a variety of cognitive skills like communicative ability and false-belief performance (Astington & Baird, 2005). Theory of Mind, being a metarepresentational ability, was seen as the finally discovered missing link for other important developmental milestones such as moral responsibility, self-consciousness, and social interaction. The child needs to discover his own mind through introspection and must assume by analogy that other people also have minds, to which they have no direct access.

(12)

wrong belief requires explicit representation of the wrongness of this person’s belief and the ability to compare it to one’s own different knowledge or belief. The task format was as follows:

“A story character, Maxi, puts chocolate into a cupboard x. In his absence his mother displaces the chocolate from x into cupboard ‘y. Subjects have to indicate the box where Maxi will look for the chocolate when he returns. Only when they are able to represent Maxi’s wrong belief #(‘Chocolate is in x’) apart from what they themselves know to be the case (‘Chocolate is in y’) will they be able to point correctly to box x” (Wimmer & Perner, 1983:109).

The results of this very first Theory of Mind study with children already showed that the ability is developing around the ages of 4 to 6 years. From this moment on, children are able to represent the relationship between the ‘epistemic states’ of other people. Although a myriad of experiments was carried out after this first discovery, with many different adaptations (other tests, other groups of children), a meta-analysis of 178 separate studies showed that children’s performance on false-belief tasks are strikingly consistent across various countries and test manipulations (Wellman, Cross & Watson, 2001). Studies in the early 90s showed that the development of Theory of Mind is culturally invariant and relatively independent of intelligence: there is little variation between children in their acquisition of a Theory of Mind (Fletcher et al., 1995). One important exception were people with an autism spectrum disorder: they do not or only partially acquire a Theory of Mind (Baron-Cohen, Leslie & Frith, 1985; Baron-Cohen, 2001). Since the study of Baron-Cohen et al. (1985), clinical psychologists also started to study Theory of Mind in a wide variety of clinical populations

(13)

2.2 The multifaceted Theory of Mind

After the first chimpanzee study, Theory of Mind has become an important theoretical construct in many research areas. Theory of Mind allows people to see themselves and others in terms of mental states and opens up the way to considering every person as a psychological being, with their own motives, thoughts and emotions (Wellman et al., 2001). With a Theory of Mind we understand other people as if we were all clairvoyant: mental-state attribution helps us to predict future behavior of other people, because we understand that their behavior is driven by the intentions and emotions that live in their mind. This insight is crucial for communication and social cognition: “Theory of Mind underpins our ability to deceive, cooperate and empathize, and to read others’ body language. It also enables us to accurately anticipate other people’s behavior, almost as if we had read their minds” (Gallagher & Frith, 2003:77).

Theory of Mind is not only the bare reasoning about the behavior of others. According to Baron-Cohen (2001), it is a much broader ability that also underlies skills like teaching or persuading others, intentionally deceiving and communicating, pretending and sharing a focus of attention. Many scholars have given lists with abilities that do or do not fall under the term Theory of Mind, but unfortunately, quot homines, tot sententiae: it seems as if every scholar has a slightly different definition of Theory of Mind. Astington and Baird (2005) state that the fundamental problem with defining Theory of Mind is that it refers to at least three different phenomena: (1) the area of research into cognitive abilities that are related to mindreading, (2) the cognitive abilities themselves and (3) the theoretical explanations for these abilities (Astington et al., 2005). So the cognitive abilities clustered under the term Theory of Mind are very diverse and range from false-belief understanding to social intelligence and mentalizing, which is why they state that Theory of Mind has a ‘multi-faceted nature’.

The common denominator in all definitions is the idea that Theory of Mind allows us to represent other minds in one brain (Fletcher et al., 1995). It is the necessary step to distinguish appearance from reality, and the thoughts of others from the thoughts of ourselves, which is in turn necessary for social and emotional awareness, but also for understanding and predicting the behavior of other people (Perner, 1991). Despite all variation in definitions, there is general agreement that Theory of Mind involves recognizing that other individuals have sensations, desires, thoughts, knowledge and beliefs, as a product of their own unique experiences – and that their point of view may therefore differ from ours (Wilson & Fox, 2013).

(14)

understanding that other individuals also have a mind. Although this definition was originally applied to the mindreading capacities of chimpanzees, the definition has not drastically changed, although it is nowadays claimed to be human-specific (Astington & Jenkins, 1999; Call & Tomasello, 2003). In developmental psychology, Theory of Mind is seen as a milestone in child development: it is the point at which children begin to take a cognitive, representational view of the mind, which has consequences for the child’s cognitive abilities: they have a better episodic memory, more self-control and social understanding (Perner, 1991). Once a Theory of Mind is acquired, it is impossible to lose it: “Just as we perceive certain patterns of movements in terms of physical causation, we cannot help but perceive certain behavioural interaction patterns in terms of intention” (Fletcher et al., 1995:110).

Some scholars have tried to restrict the use of the term, because of its multifaceted applications in different studies. In the more restricted sense, Theory of Mind is “a domain-specific, psychologically real structure, composed of an integrated set of mental-state concepts employed to explain and predict people's actions and interactions that is reorganized over time

when faced with counterevidence to its predictions” (in: Astington & Baird, 2005:4). Human beings make

predictions about other people’s behavior based on assumptions about their internal world of thought. When Suzanne is preparing for a walk and gets her umbrella, one might assume that she thinks that it is raining outside. When, after a phone call, Suzanne leaves the umbrella, but picks up her sunglasses instead, one might assume that Suzanne has changed her mind: it is likely that she has now other thoughts about the weather outside. So the reference to counterevidence that is mentioned in this definition acknowledges that the way we deal with false-belief can be seen as the quintessence of Theory of Mind. Comprehending false-belief means that a child understands that the mind is characterized by two things: subjectivity and its susceptibility to manipulation by information (Perner & Ruffman, 2005).

(15)

Shaked & Solomonica-Levi, 1998). Given that false-belief is a core concept in recent studies on Theory of Mind, we will use the term Theory of Mind especially for referring to this restricted definition, namely as the ability that enables people to attribute mental-states and make inferences about these beliefs in relation to the physical world.

2.2.1 First-order Theory of Mind

Soon after the first study on Theory of Mind it was theorized that a mindreading task could have varying degrees of complexity for which a different level of Theory of Mind is required. Baron-Cohen (2001) makes a distinction between Theory of Mind of the first and second order, but we should also indicate the zero level, which is the absence of a Theory of Mind (Pijl, 2011). A child with a zero-level Theory of Mind has perceptions, but is unaware of the fact that other people also possess a mind with their own beliefs and perceptions. (S)he has no theory about the mental states of others. For instance, when a little boy is playing hide and seek with a parent, he may just cover his eyes and think that his mother will not know that he is there, because he cannot see his mommy. So the child makes no mental-state attribution concerning his mother.

What follows zero-order is first-order Theory of Mind. This refers to the ability to infer the thoughts of one other person. For instance, Mary understands that John is happy, because his mother just bought his favorite toy. There is only one step of Theory of Mind: X attributes a mental state to Y. In the more specific context of false-belief reasoning, this would mean that X knows a false-belief that Y has, because this false-belief is in contrast with reality. So, in a first-order false-belief task, one needs to make an attribution about a false-belief of one other person with regard to real events.

(16)

“Where is the marble really?” (reality question) and “Where was the marble in the beginning?” (memory question). The Sally-Anne task also checks for memory effects by asking the child not only belief questions, but also a memory question (so that it can be excluded that a child gives the wrong answer, because (s)he has simply forgotten the plot). If a child answers the belief question with reference to the false belief of the doll, it means that (s)he understands the distinction between mind and world: mental states may reflect reality, but they are internal and fundamentally distinct from real-world events or behaviors.

Figure 1 Sally-Anne task (Baron-Cohen et al., 1985)

First-order Theory of Mind cannot only be measured through an unexpected displacement task, but also by the so-called unexpected contents false-belief task (Wilson & Fox, 2013). The most renown version of this task is the Smarties task (Gopnik & Astington, 1988), where children are shown a familiar container (i.e. the Smarties tube). Children are asked what they think is in the box (“Smarties!”). The tube is then shown to contain pencils only. Another character is introduced and the children are asked what (s)he will think is in the box. If the children correctly answer with “Smarties”, they successfully attribute a false-belief.

(17)

Namely, at the time t0 Sally sees that the marble is in the box. At the time t1, after Sally has

returned, she deduces that the marble is still in the box as she has no belief to the contrary (the principle of inertia). This is not the case for the Smarties task of Gopnik and Astington (1988). So one could expect that children perform a little worse on the Sally-Anne task than on the Smarties task, because of the extra complexity (i.e. they need to assume the principle of inertia). However, the Smarties task comes with other complexity: the child must suppress the initial belief that a Smarties tube contains candy. So, in practice the two false-belief tasks yield the same results.

2.2.2 Second order Theory of Mind

Second-order Theory of Mind entails another ‘embedded’ step in reasoning and typically involves a third person. The term refers to reasoning about the thoughts that someone else may have about the mental state of a third person: it is embedded mental-state attribution (Baron-Cohen, 2001). In other words, person X reasons about the mental-state attribution that person Y may have about the thoughts of person Z. Second-order false-belief tasks involve reasoning about what people think about other people’s thoughts. So, in a second-order false-belief task, one must attribute the false-belief of one person based on the thoughts of another (and not on reality, as in the first-order false-belief tasks). In this set-up, X is often an observer who reasons about the ‘first-order Theory of Mind’ of two other persons. However, it may also involve only two persons in cases where X thinks that Y thinks that X thinks something, as in “I think that you think that I have written a complex sentence”. This type of social reasoning is quite complex in everyday situations, although people are

able to attain this complexity as long as there is a sufficient amount of scaffolding (Meijering, Van Rijn, Taatgen & Verbrugge, 2011). It is another form of second-order Theory of Mind that allows reasoning about what other people think about us. Figure 2

shows the three different levels of ToM. Figure 2 Representation of 0th_{to 2}nd_{Theory of Mind}

(18)

(1985) there were overlapping beliefs at the second-order. In the ice-truck story, John and Mary think that the ice cream truck will stay in the park all day. Then, Mary and John go their own way. John learns that the truck is moving to the church. He thinks that Mary does not know that the truck is moving, but Mary has heard the same news. This is where the second-order false belief comes in: John doesn’t know that Mary already knows that the van has moved. Both protagonists know the same location of the truck and so there are overlapping beliefs on the second-order level.

A more recently developed task tests the ability to understand one person’s belief (the first layer) about a belief attributed to another person (the second layer), without overlapping beliefs. The eight different stories of this task do not contain second-order syntactic embedding, but they do involve second-order reasoning (Hollebrandse et al., 2014). The set-up of the stories is different from the task of Wimmer and Perner, because in this new task, the second protagonist Y does not change his belief in the same belief as that of the first protagonist X (John and Mary believe the same), but in another belief. Therefore, X has a first-order belief: his belief does not correspond to the initial belief and not to reality, whereas Y has a second-order belief about X: he thinks that X still has the initial belief (which is different from reality). The prototypical story in this Bake Sales task is as follows:

Sam and Maria are playing together and look out the window and see that the church is having a bake sale. Sam watches Maria grab her money and run out of the house and over to the church. “Ah ha,” Sam says. “She’s going to buy chocolate chip cookies, her favorite.” And then Sam goes back to playing with his toys. On her way to the bake sale, Maria runs into the mailman and tells him, “I’m going to get a nice pumpkin pie for Grandma.” When Maria gets to the bake sale, she finds out that all they have are brownies. So she couldn’t buy a pumpkin pie for her grandma, but buys a bunch of brownies to bring back to her family instead. What did Maria think she was going to buy at the bake sale? Why? (1st_{order FB) Back at the house, Mom comes in and says to}

Sam “I noticed that the church is having a bake sale.” “Oh yes,” Sam says. “Maria went there.” Then Mom asks, “oh, what does Maria think they’re selling at the bake sale?” What does Sam tell his mom? Why does he tell her that? (2nd_{order FB)}

(19)

2.3 Theory of Mind development in children

In this section, a brief overview of empirical results is given for the false-belief tests that were discussed above and for some interesting related tasks. The results are discussed for typically developing children (2.3.1), children with an autism spectrum disorder (2.3.2) and children with an attention-deficit hyperactivity disorder (2.3.3).

2.3.1 Typically developing children (TD)

Understanding false-beliefs can be seen as the fruit of a long development that begins in early infancy and has its culmination when the child can pass a second-order false-belief task. The first signs of Theory of Mind may be apparent by 18 months when children engage in symbolic play and by the age of 3, children start to acquire some understanding of mental states (De Villiers, 2007). However, they do not yet grasp the colourful palette of mental states: although they understand that certain perceptions influence thoughts and that someone may have different desires from themselves, they don’t grasp the idea of pretence, potentially because pretence does not stand in a direct referential or causal relation with reality (Yirmiya et al., 1998).

(20)

Theory of Mind in the preschool years. As the graph shows, there is a substantial development in false-belief performance: at 44 months children are at

50% correct on a first-order false-belief question and the children continue improving in the early school years (Figure 3; Wellman et al., 2001:25). Additional analyses showed that children can do better on false-belief tasks dependent on task salience, motivation, motive and real presence and that this is cross-culturally valid.

There are also scholars who believe that children can pass a Theory of Mind task much earlier, as long as the task is less verbal or non-verbal. For

instance, on a hide-and-seek board game, children as young as 2½ years old were able to use deceptive strategies “that both trade upon an awareness of the possibility of false beliefs and presuppose some already operative Theory of Mind” (Chandler et al., 1989:1263). A series of studies that used young children’s eye-gaze as a measure for Theory of Mind also suggested that children might have some implicit false-belief understanding around 2½ year old (Perner et al., 2005) and that they understand the false-belief in a violation of expectation task (in which a toy slice of water melon is hidden in one of two different boxes and where an actor acts out true and false belief) when they are only 15 months old (Onishi & Baillargeon, 2005), but other scholars criticize these methods and point to the fact that some nonverbal tasks may even be more difficult than verbal false-belief tasks, especially when second-order false-belief is tested (Hollebrandse & Van Hout, 2013; see Chapter 3). Still, it does no harm to consider false-belief reasoning as the culmination of a long developmental pathway that started in infancy (De Villiers, 2007).

2.3.2 Children with an autism spectrum disorder (ASD)

Both renown psychiatric handbooks containing the criteria for the classification of psychiatric and mental disorders, the DSM-5 and the ICD-10, present problems in social communication and interaction as critical for the diagnostic of an autism spectrum disorder (Durand, 2014). Recent studies show a prevalence of autism spectrum disorders (ASD) of 6-9 in 10.000 people (Verhulst, Verheij & Ferdinand, 2007). Children with ASD are characterized by communication impairments, social impairments and restricted, stereotypical patterns of behavior and interests and a decreased imagination (APA, 2000; Volkmar, State & Klin, 2009). For most forms of

(21)

autism, the language phenotype is conspicuous: children with ASD often confuse the first and second person and also display direct and mitigated echolalia in their language use (Belkadi, 2003; Kanner, 1943).

Since autism was described, independently and almost simultaneously, by Leo Kanner (1943) and Hans Asperger (1944), a wide variety of theories have been proposed to explain this enigmatic condition (Rajendran & Mitchell, 2007). The discredited ‘refrigerator mother’ is an infamous example of one of these theories that suggested that autism was caused by the mother’s attitude toward her child (Bettleheim, 1967). A scientifically more fruitful explanation was developed in the late eighties by Baron-Cohen, Leslie and Frith (1988). They suggested that many aspects of the impaired functioning in autism could be explained by a single cognitive deficit: a lacking Theory of Mind.

For instance, such a deficit would make pragmatic language use very difficult and lead to problems in understanding communicative intention, presupposition, indirect requests or discourse planning (Geurts & Embrechts, 2008). It would also hinder all kinds of social activities in which one has to make mental-state attributions and predictions based on the mental-state of others, and it would also limit one’s ability to engage in imaginative play. Indeed, a large number of studies reveal Theory of Mind deficits or degrees of mindblindness in people who are diagnosed with an autism spectrum disorder (Cohen, 1995; Frith, 2003; Colle, Baron-Cohen & Van der Lely, 2008).

(22)

Another study by Baron-Cohen (1989) showed that success on a second-order false-belief task seems to be non-existent for children with autism even when they are almost 15 years old (having a mental age of around 12 years), while 60% of children with Down syndrome are successful in the same task. So children with autism seem to display a severe delay in their ability to mental-state attribution. But, even when they develop a Theory of Mind– albeit delayed, they will lack a spontaneous ‘Intuition of Mind’ (Vermeulen, 2008).

2.3.3 Children with ADHD

Problems in false-belief judgments are not only reported for autism spectrum disorders, but also for ADHD (Perner, Kain & Barchfeld, 2002). The cardinal features of the attention-deficit hyperactivity disorder (ADHD) are an inappropriate level of inattentiveness, hyperactivity and impulsivity (APA, 2000). In recent years, the prevalence of the diagnosis has drastically risen and nowadays affects 3-7% of school-aged children, especially boys (APA, 2000). Language impairments are not officially stated in the diagnostic criteria for ADHD, but language-related problems are observed in many studies (Geurts & Embrechts, 2008) and show that children with the diagnosis have pragmatic difficulties, such as more stereotyped conversation patterns, in which they hardly differed from children with ASD, although the degree of these difficulties was more profound for children with ASD. Comorbid language impairments are reported for 35-50% of children with ADHD (Baker & Cantwell, 1992).

ADHD is basically a dysfunction of executive functioning. Since executive functioning and Theory of Mind development are thought to be related, it is reasonable to assume that children with ADHD also have difficulties with false-belief reasoning (Korkmaz, 2011), although others refuse to assume that problems in executive functioning influence Theory of Mind development (Perner et al., 2002). Children with ADHD also display impairments involving emotion, face and reduced empathy, which is closely related to mindreading abilities (Uekermann et al., 2010). Explanations that have been proposed for Theory of Mind difficulties in ADHD are twofold. Firstly, children with ADHD are impulsive and lack the ability to focus attention (needed for good false-belief judgements), and secondly, their behavioral problems may hinder Theory of Mind development (Perner et al., 2002).

(23)

deviant pattern from typically developing children (Buitelaar, Van der Wees, Swaab-Barneveld and Van der Gaag, 1999; Kuijper et al., 2013).

2.4 Summary

In this chapter it was shown how the concept Theory of Mind entered the scientific arena through a study with chimpanzees. The basic definition of it as imputing mental states to oneself and others has not fundamentally changed and still inspires studies in developmental and clinical psychology, linguistics, logic and many other fields. False-belief performance, as measured through unexpected content and unseen displacement tasks, is nowadays seen as the key to measuring Theory of Mind, and the degree of complexity of these tasks allows us to make additional distinctions between first-order and higher-order Theory of Mind.

(24)

3 Language and Theory of Mind

Although the first reference to the term Theory of Mind emerged in a psychological article on the mindreading capacities of chimpanzees (Premack & Woodruff, 1978), it is nowadays claimed to be specific to humans. Although great apes are capable of some kind of mental-state attribution (i.e. in deceiving other apes), this capacity is rather limited and not even developed in all primates. For instance, a study by Call and Tomasello (1999) showed that seven chimpanzees and orangutans were incapable of passing a nonverbal false belief task. So the answer on the question whether primates have a Theory of Mind does not sound as unanimously anymore as it did thirty years ago. The primates’ behavior that was explained as a Theory of Mind in the 70s may have been nothing more than an acquired behavioral pattern (Byrne & Whiten, 1991). Therefore, mental-state attribution is nowadays considered to be a defining aspect of the human mind (Astington & Jenkins, 1999; Call & Tomasello 1999) or even a quintessential ability that makes us human (Baron-Cohen, 2001). The idea of Theory of Mind as being human-specific has inspired the thought that it may be relying on the other mysterious skill that only human beings possess: language (Gauker, 2011).

Philosophical considerations on the relationship between conceptual skills, or in a more general sense, between thought and language is as old as the hills. Its scientific starting point brings us to the 18th_{century with the writings of the philosopher and linguist Humboldt (De}

Mulder, 2011). In his essay on The heterogeneity of language and its influence on the intellectual development

of mankind (1838), Humboldt stated that the inner life and thought of people is expressed in the

structure and character of the language they use. Still, the relationship between language and thought has been provoking controversy for many ages and is as much under debate as the question on the relationship between language and a specific aspect of thought: Theory of Mind. The debate on language and Theory of Mind can be summarized in the question whether the acquisition of language has an effect on non-linguistic areas of cognition (cf. the Sapir-Whorf hypothesis), or whether advances in cognition are a prerequisite for language acquisition. The dominant view on this topic has been shifting back and forth as a ‘pendulum’ over the ages, and it is still doing so (De Mulder, 2011; Slade & Ruffman, 2005).

(25)

and a wide range of false belief tasks do not reveal the direction of the relationship, if there is any. It is even possible that a third factor is causally implied which influences the variation in both factors. Therefore: ‘More precision is needed about the nature of the relation and the aspects of

language that are involved’ (Astington & Jenkins, 1999:1311). This chapter discusses the three

different possibilities for the relationship between language and Theory of Mind in more detail and the empirical support that has been found for each of these possibilities.

3.1 Language depends on Theory of Mind

It is beyond any doubt that an infant must have at least some idea that the utterances of his caretaker mean something and are said with a certain purpose. In the very beginning of language acquisition, a basic attribution of mental states is needed. For instance, in the process of word learning children need to understand the intention of the speaker, which is often transferred by pointing gestures or eye gaze. The resulting triangulation between speaker, listener and object focuses the attention of the child on an object to connect to a word, which helps the child to fix the referent for the word they hear (De Villiers, 2007). The understanding of intention is a crucial concept which explains how children learn to map events onto language (Golinkoff & Hirsh-Pasek, 2008).

Without the basic meta-representational skill of considering a caregiver as an intentional agent with mental states, the child would ignore his parents’ utterances “instead of trying their hardest to assign a representation to them” (De Mulder, 2011:5). So this suggests that mindreading mechanisms have a profound impact on verbal interaction between children and their caregivers (Papafragou, 2002). In this respect, the eye gaze studies with autistic infants are interesting: autistic children look less at the eyes of their caregivers and seem to lack the gift to grasp spontaneously their caregiver’s intentions (Hutt & Onsted, 1966; Baron-Cohen, 1988). However, the chicken-and-egg debate remains unsolved: is it the lack of eye contact that results in a troubled understanding of other people’s intentions, or does some innate mindblindness make that autistic infants do not feel the need to follow someone’s eyegaze? Whatever the relationship, it is known that eye gaze and intention reading are crucial for word learning and that impairments in these abilities will harm children’s word learning ability. Indeed, many people with a severe disorder in the autistic spectrum show delayed language abilities or do not acquire functional language at all (Diehl, Bennetto & Young, 2006; Lord & Paul, 1997).

(26)

language use. This concept-first opinion is nourished by the work of the cognitivist psychologist Jean Piaget who claimed that children first acquire a conceptual understanding of the distinction between appearance and reality (i.e. false belief). Only after this milestone, child language becomes more elaborated, as a reflection of this cognitive growth spurt (Piaget, 1980). Likewise, modern researchers like Papafragou (2002) suggest that a child needs to understand other people’s minds before being able to understand and properly use mental state language. Mindreading precedes communicative sophistication and is a fundamental requirement for successful communication.

In this view, it is language that follows from a growth spurt in cognitive abilities such as the unfolding Theory of Mind. The consequence of this reasoning is the idea that a child has to use non-linguistic mental models to represent false beliefs, so that Theory of Mind does not primarily depend on language. It is the claim that implicit knowledge precedes explicit understanding, which is also supported by Josef Perner (1991). His research focused on the role of implicit knowledge in false belief tasks. He showed that children show implicit false belief reasoning, although they cannot verbalize it in an explicit task.

In one classic study, Clements and Perner (1996) videotaped children’s eye movements while they were watching a location switch false-belief story. A few seconds before the return of the story character, the children heard the following prompt: “I wonder where he’s going to look?” This was not a direct question to the child regarding where the protagonist would look, but rather a means of seeing whether the child would look at the initial location (i.e. the ‘belief’ of the story character) or at the real location (i.e. the ‘belief’ of the child). Almost 90% of the children between 2;11 and 4;5 months old looked at the initial location (showing false-belief knowledge), but in contrast, only 45% of the children mentioned that location in answer to an explicit false belief question (Clements & Perner, 1994). It suggests that the young children had non-linguistic resources that enabled them to pass the non-verbal aspect of the task. However, De Bruijn and Newen (2012) suggests this type of responses in false-belief tasks with very young children do not necessarily imply Theory of Mind reasoning, but are the reflection of some kind of associative memory.

(27)

understanding of false belief, starting with implicit knowledge that develops towards explicit understanding with low confidence and finally, to a stage with full confidence (Ruffman, Garnham, Impart & Connolly, 2001). But the basics of a Theory of Mind remain non-verbal and implicit. In my opinion, it is difficult to determine whether this unconscious eye-gaze response can be explained as implicit false-belief reasoning. The test of Clements and Perner (1994) measures at least an expectation bias, but it may not be a valid test for actual false-belief reasoning. More work should be done to clarify what implicit false-belief understanding entails and what it does not. Under the interpretation of Onishi and Baillargeon (2005), this means at least that the infants realize that others act on the basis of their beliefs and that these beliefs are representations that may or may not mirror reality.

In the discussion on language and thought, some scholars point to the work of Varley (1998). He showed that when an adult loses language through brain damage after puberty (i.e. in aphasia), he is still capable of false belief reasoning. This would not be the case if false belief reasoning was primarily dependent on linguistic command. However, direct inferences to the situation of the child acquiring language and Theory of Mind are a delicate business: an adult losing language possesses already mature cognitive abilities and is no longer in a process of developing mindreading abilities, like children (De Villiers, 2007).

A more bidirectional relationship between language and Theory of Mind was found in a longitudinal experiment by Slade and Ruffman (2005). They tested 44 three- and four-year-olds and came to the conclusion that the understanding of mental states has a facilitative role in semantic development. At the first moment of testing, it seemed that language predicted first-order Theory of Mind, but at the second moment of testing, Theory of Mind was a strong predictor for language. This was more recently confirmed in another longitudinal study (De Mulder, 2011). So the mindreading seems to influence a child’s further language development, although the relationship for higher order false-belief remains unclear.

(28)

the child has to take into account the perspective of the listener or speaker. This is where mindreading abilities come into play.

However, the majority of studies focus on lexical aspects of language, and particularly the acquisition of mental state verbs as think or guess. The production and comprehension of these verbs seems to be related to performance on a false-belief task. Ziatas, Durkin and Pratt (1998). claim that the correct use of these verbs follows from Theory of Mind. Similarly, children with an impaired Theory of Mind used less advanced mental state language in a narratives task: they do use mental state expressions as often as typically developing peers, but do not “locate characters’ internal states within a causal framework” (Capps, Losh & Thurber, 2000:201). This means that even though autistic children use mental state verbs that refer to thoughts, desires or beliefs, they do not explain and interpret events with reference to these thoughts, desires and beliefs. When children with and without an autism spectrum disorder are compared, it is shown in several studies that a better performance on false-belief tasks is significantly correlated with a higher score on a task on mental state verbs. According to Ziatas and colleagues, this provides “further support to a growing body of studies indicating a relationship between impaired Theory of Mind and the specific profiles of communicative impairment observed in children with autism” (Ziatas et al., 1998:760).

Yet, we must note that both this conclusion and this observation concerning the use of mental state verbs is not shared by everyone, especially in studies that analyzed spontaneous speech samples. Several studies did not find differences in the spontaneous production of mental state verbs between autistic children (with an impaired Theory of Mind) and typically developing children (Mäkinen, Loukusa, Leinonen, Moilanen, Ebeling & Kunnari, 2014; Norbury & Bishop 2003; Colle, Baron-Cohen, Wheelwright & Van der Lely, 2008). And except for the work that was done on implicit false-belief reasoning task, the direction of the relationship is not convincingly and unequivocally proved.

3.2 Theory of Mind depends on language

Other than the concept-first camp (i.e. cognitive growth is the driving force behind development in linguistic abilities), the scholars of the language-first camp claim that the child’s language is the vehicle for acquiring a Theory of Mind. It may be inspired by the classic idea of Humboldt (1838) that language drives intellectual development in general. In its most radical formulation, the idea communicates a version of linguistic determinism: if different languages have different ways to describe events, then these differences also influence human thought. In the early 20th

(29)

their grammars toward different types of observations and different evaluations of externally similar acts of observation, and hence are not equivalent as observers but must arrive at somewhat different views of the world” (Whorf, 1940). This so-called Sapir-Whorf hypothesis entails that cognition is profoundly affected by the language that someone possesses, which may also apply to the acquisition of a Theory of Mind. However, if Theory of Mind is language-driven, then the question remains what part of language is crucial to this development. Three positions will be discussed in the following sections.

3.2.1 The general proficiency hypothesis

In the weak version of the language-first hypothesis, it is claimed that a great deal of language processing is a prerequisite for successful performance on most Theory of Mind tasks, due to the nature of these tasks: many assessment methods heavily rely on language. In an ordinary false-belief task, children hear a story or an explanation and must also process quite complex questions (Astington & Jenkins, 1999). The interplay between language and Theory of Mind may be a problem in the experimental assessment procedure: 3-year-olds may be unable to show their understanding of false-belief, because the linguistic demands of the assessment procedure are too complex (Chandler et al., 1989). Theory of Mind may not be dependent on language, but the false-belief tasks are. Consequently, non-verbal tasks may show fundamentally different results than verbal tasks. In an assessment procedure that did not rely heavily on linguistic commands, Chandler et al. (1989) showed that even 2 ½-year-olds were capable of deceiving others and had some rudimentary Theory-of-Mind. It sheds another light on the eye-gaze studies that were discussed in the previous section that provided some proof for implicit false-belief understanding (Perner et al., 2005; Clements et al., 1996). Still, if the bad performance of young children is due to the linguistic complexity of the assessment procedure, then a completely nonverbal false-belief task should be easier for (young) children. Interestingly, it is not true that a completely nonverbal task is easier for children: their performance on a verbal and nonverbal false belief task (which was also suited for chimpanzees an orangutans) seems to be highly correlated (Call & Tomasello, 1999).

Also, a study by Hollebrandse, Van Hout and Hendriks (2011) showed that 7-year-olds performed successfully on a verbal false-belief reasoning task with an extra layer (2nd_{order false}

(30)

‘From this point of view, it is not that linguistic immaturity masks children’s underlying competence, but rather that Theory of Mind development and language development are themselves fundamentally related and interdependent. Language development may itself provide children with resources, such as syntactic ability or semantic understanding, that promote or allow false-belief understanding’ (Astington & Jenkins, 1999:1312; emphasis added).

Several longitudinal studies did not only find an effect of Theory of Mind on early language development, but also a reverse effect from language on later false belief performance in the opposite direction (Slade & Ruffman., 2005; De Mulder, 2011). Although this might serve as an argument for both the language-first and the concept-first camp, it was found that the effect of language on later false belief performance was considerably larger in size than the effect of early false-belief performance on language. These findings suggest that the direction of causality might shift in the course of development and that the concept-first position is especially true in the early years of development, but that the linguistic hypothesis becomes increasingly important over the years. De Mulder (2011) suggests that for a further developed Theory of Mind, certain advances in linguistic ability are a necessary condition. But then it becomes relevant to state what kind of advances in language are needed for Theory of Mind development and how specific or general these advances should be – which constitutes a further point of contention.

(31)

3.2.2 The semantic hypothesis

Other scholars believe that the relationship between language and Theory of Mind is more specific. For instance, several researchers propose a semantic hypothesis, emphasizing that the driving force behind Theory of Mind development are the semantic aspects of language. More specifically, the child’s lexicon should contain mental state terms like think or guess. The specific semantic aspects of these terms would foster the understanding of other people’s minds. Mental state verbs are the conceptual underpinning that is needed for further Theory of Mind development: they refer to unobservable phenomena. When they are referred to by means of language, this unobservable realm of thought is brought to the child’s attention (Astington & Jenkins, 1999). When the child encounters, in conversation, verbs that refer to cognition or other mental states, (s)he will actively look for the meaning of these words.

When children are 2 or 3 years old, they start to acquire specific lexical terms that are related to mental states. After the acquisition of words that indicate perception, desire and emotion (e.g. see, happy, love, want), children start to use cognition terms (e.g. know, think) (Astington & Baird, 2005). However, most children use these terms in conversation before they understand their semantic meaning. For instance, they may use I think as a pause-filler, without making the association with his or her own mental state (Shatz, Wellman & Silber, 1983). The concept of desire is mastered before the concept of belief, because desire is more directly linked to visible behavior like reaching or stopping these activities once the desired object is achieved (De Villiers, 2007). The discovery that cognition related verbs have something to do with the realm of thought will lead children into the understanding that there is something as a realm of thought in their own, and by extension, in other’s minds.

(32)

Moore, Pure and Furrow (1990). Apparently, children need linguistic means to express their understanding.

This version of the language-first hypothesis would predict that children and adults with (severe) problems in false-belief reasoning would make less extensive use of mental state terms. Narratives studies on children with autism do not quite fit into this theory and have repeatedly shown that the spontaneous production of mental state terms by children with ASD is not significantly different from that of typically developing peers (Mäkinen et al., 2014; Norbury & Bishop, 2003; Colle et al., 2008). However, it is possible that differences emerge in elicited production tasks: spontaneous production tasks may not challenge children to use their full language competence, whereas the fixed format of an elicitation experiment could provide insights into the limits of a child’s language.

3.2.3 The syntactic hypothesis

Other scholars do not seek the solution of the problem in semantics, but rather in syntactic ability. That is, in order to represent states different from current reality, more complex syntactic constructions are needed. For Theory of Mind computations, the child needs symbolic representations that override perceptual information. In the acquisition process of false-belief reasoning, the child would elaborate on these representations (Plaut & Karmiloff-Smith, 1993). The most prominent version of the hypothesis was proposed by Jill De Villers. She and her colleagues believe that a specific linguistic structure is needed for mindreading (De Villiers, 2005; 2007; De Villiers & Pyers, 2002): for successful performance on a false-belief task, children need to master the syntactic structure of tensed complements. In other words, the key to a well-developed Theory of Mind is the understanding of sentential complementation constructions, in addition to the understanding of mental state verbs.

The underlying idea is that the representation of other’s mental states is not only encoded in the semantics of verbs that refer to these mental states (like to think, to believe), but that they are also encoded in the special syntax of complementation in which these mental state verbs can be used (De Villiers & Pyers, 2002). Without the understanding of complex propositional structures, children will fail to understand false-belief. Consider the following sentence:

(1) Suzanne thinks (that) hot chocolate milk is healthy.

(33)

mind and therefore, the truth value of the sentence is not evaluated with respect to this world, but with respect to another possible world: Suzanne’s mental world (even if she is mistaken in her beliefs). For the verb to say it is easy to verify whether the proposition is true or false, because it involves overt evidence; however, for to think the truth or falsity is covert. De Villiers (1999) argues that there is something in the syntactic structure of the embedded proposition that indicates for mental state verbs the potential falsity of the proposition that follows. It must be an abstract feature, because in many sentences with a comparable truth structure, the complementizer ‘that’ is optional.

This syntactic analysis then opens the door to false-belief reasoning, where a child does not only have to represent his own mental world or reality, but also another possible world: the mind of someone else (De Villiers & Pyers, 2002). The embedded complement makes it possible to predict what will happen: if Suzanne thinks that hot chocolate is healthy, she is likely to drink it a lot. The propositional context and particularly the mental state verb give another type of information: it signals the likelihood of this prediction (De Villiers, 2005). If the verb were to

doubt, then Suzanne would be less likely to drink a lot of chocolate milk. So although the lexical

meaning of mental verbs is needed to draw attention to mental states, it is only in the combination with their typical syntactic structures that they scaffold Theory of Mind: the embedding structure organizes meaning.

Peter de Villiers states that the child must put three things together to understand mental state verbs: (1) lexical meaning, (2) syntactic structure (in a wide range of occurrences) and (3) the knowledge that embedded propositions can be false. The acquisition of the language of ‘propositional attitudes’ would then open up a “classification into worlds, or different points of view of reality” (De Villiers, 2005:188). More importantly, children do not only need this full syntax of mental verbs and their complements for reporting about false-belief, but they need it “in order to represent in his own mind the belief states of other people” (De Villiers & Pyers, 2002:1056). So, a sufficient mastery of this grammatical phenomenon is claimed to be crucial to a successful (first order) false-belief performance: without sentential complementation, a child has no means to represent and compare their own and other’s beliefs.

(34)

contained complex questions in which the question word at the beginning of the sentence was coreferential with the unexpressed object in the embedded clause (sentence 2):

(2) What did the girl say she bought __ ?

In the underlying structure of the sentence, the Wh-element originates in the embedded clause. Suppose that the girl bought a fluffy rabbit, but that she said she bought a kitten with quite big ears. The correct answer on the question would be ‘a kitten with big ears’, which was what the girl said. However, children under the age of 4 years old would answer with ‘a fluffy rabbit’. De Villiers (1999) concluded that young children do not understand that the question concerned the joint effect of both ‘saying’ and ‘buying’ and that they only concentrate on the embedded ‘buying’. This would be due to a lack of understanding of both the syntax and meaning of sentential complements (Diessel, 2004). So in other words, young children cannot deal with embedded false complements.

In the following years, a series of studies were designed to test the aforementioned claims. One prediction was that deaf children would acquire false belief later on, after the acquisition of complex syntax (of complementizers). Indeed, language-delayed deaf children who had difficulty with complex syntax had an impoverished way of speaking about beliefs, but not with speaking about emotions. They also experienced problems in both verbal and non-verbal false-belief tasks. In addition, they were delayed in their understanding of tensed complements in combination with mental state verbs. But the more advanced their language skills, the better their understanding of false beliefs (De Villiers, 2005).

(35)

Lohman and Tomassello (2003) took the hypothesis even further and started training children in order to make them better in passing false-belief tasks. Their interesting finding was that direct false-belief training had less effect than training children explicitly on complementizer structures. Children who received this syntactic training improved considerably in their false-belief performance than children who were directly trained on false-false-belief reasoning. Also, a regression analysis on the performance of 28 typically developing children on a wide variety of language tasks and false belief understanding (e.g. unexpected contents, unseen displacement, explanation of action) showed that the mastery of tensed complements ‘is a precursor and possible prerequisite of successful false-belief performance’ (De Villiers & Pyers, 2002:1037). Based on these ideas, Mary Sweig-Wilson and others recently developed a series of Special Needs software under the name Language for Theory of MindTM where children learn words like want and

desire, and are shown what others can and cannot perceive. The program states that it “develops a

language foundation for a Theory of Mind” (Sweig-Wilson, 2013).

Despite the overwhelming evidence and these promising results, the hypothesis has been challenged as well. For instance, Ruffman et al. (2003) showed that the predictive value of tensed complements on false-belief reasoning was not always significant, whereas general language ability was a good predictor. The same was found by De Mulder (2011) who claims that the theory on sentential complementation structures as prerequisite for Theory of Mind may be wrong. She states that general language ability still is the best predictor for later Theory of Mind development. Additional criticism came from a study by Lely, Hennessey and Battell (2002). They showed that children with SLI did well on false-belief tasks, despite their poor performance on several sentential complement tasks.

More recent studies elaborate on the work of De Villiers and test her claims in the field of double embedded sentences in relation to second-order false-belief. For instance, Hollebrandse and Van Hout (2013) state that for single belief ascription, one can avoid syntactic complementation structures:

(3) Hot chocolate is healthy. Suzanne thinks that.

However, there is no such alternative for second-order belief ascription, where a third person thinks something about the belief ascription of someone else. For instance, if the dentist knows that Suzanne thinks that hot chocolate is healthy, there is an extra layer of beliefs. This cannot be represented with semantic embedding as in the following sentence:

(4) # Hot chocolate is healthy. Suzanne thinks that. The dentist knows that.

(36)

the syntactic dependencies between propositions (Hollebrandse, Hobbs, De Villiers & Roeper, 2008). Over the past few years, the focus of linguistic research has shifted towards the study of double embedded syntactic structures and second-order false belief tasks. In studies that showed a better performance on verbal false-belief tests than on non-verbal false-belief tasks, it was suggested that this is due to a process of scaffolding: “The recursive linguistic representations involved in syntactic embedding may provide the scaffolding to perform the recursive step of a second-order FB reasoning” (Hollebrandse & Van Hout, 2013:104). It provides a whole new kind of support for the language-first hypothesis, by extending the domain of its predictions.

3.3 Language and Theory of Mind depend on a common underlying mechanism

So far, we have seen the position that a growth spurt in cognitive abilities are reflected in the increased complexity of the child’s language (the concept-first position) and the opposite idea that the acquisition of more complex language is the crucial step towards more advanced cognitive abilities (the language-first position). It is also possible that there is an underlying factor or mechanism that causes development in both language and Theory of Mind. In that case, the relation between language and false-belief reasoning is only symptomatic of underlying cognitive development.

3.3.1 Executive functions and language

Several mechanisms have been proposed as common underlying factor, such as executive functions that seem to influence both preschooler’s narrative competence and their Theory of Mind development (Friend & Bates, 2014). For instance, there is a marked development in executive attention and behavioral inhibition around 4 years of age, simultaneously with the development of explicit Theory of Mind abilities. Children learn to focus attention, to shift quickly between contrasting rules and become more resistant when they are confronted with distraction. So, maybe it is the development of the skill to inhibit reality-based responses that makes children succeed at false-belief tasks.

(37)

functioning: the fact that one is able to inhibit prepotent responses, to stop ongoing responses or to be able to deal with interference during processes is a fundamental ability for the functioning of working memory, self-regulation motivation and arousal.

Despite the jumble of vaguely defined terms, most researchers believe that language is crucial to the development of executive functions and inhibitory control (Astington & Baird, 2005). Namely, at the same moment, children’s narrative abilities also show a tremendous development. This does not automatically imply causality: it might still be the case that these abilities just happen to emerge and develop concomitantly, but one could imagine that developing a meaningful and causally coherent narrative implies planning and inhibition, and besides that also the ability to remember all important bits of information (Friend et al., 2014).

A recent study compared children’s narratives with their performance on Luria’s Tapping Task, which measures response inhibition. In this task the child is asked to tap twice on a wooden dowel when the experimenter taps once, and vice versa. The results of the experiment revealed that advanced narratives at 4.5 years of age predicted a faster performance on the Tapping Task at 5 years of age, but in turn, this faster response rate did not predict later narrative skill (Friend et al., 2014). The exact role of executive functions seems to be rather complex, but a wide variety of empirical studies show that there exists some kind of complex relationship between EF, language proficiency, narratives and Theory of Mind.

Complex language, complex thought?