|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 LANG UAGE AND THEORY OF MIND IN 36 MONTH OLD CHILDREN By MARY ELLEN MCKAY EASTERS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER O F SCIENCE UNIVERSITY OF FLORIDA 2009
2 2009 Mary Ellen McKay Easters
3 To Jane
4 ACKNOWLEDGMENTS I would like to thank M. Jeffrey Farrar for his continued support and invaluable input I would also like to t hank Scott Miller and Bonnie Johnson for their important contributions to my education. In addition, I thank Virginia Holloway and Sarah Lynne Landsman who helped in innumerable ways, and I thank my husband Michael and my parents for everything
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 7 LIST OF FIGURES ................................ ................................ ................................ ......................... 8 LIST OF ABBREVIATIONS ................................ ................................ ................................ .......... 9 ABSTRACT ................................ ................................ ................................ ................................ ... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 11 Theory of Mind ................................ ................................ ................................ ....................... 11 Theory of Mind and a Focus on Tasks ................................ ................................ ................... 12 The Question of t he Role of Language ................................ ................................ ................... 15 The Question of the Role of Executive Function and Working Memory ............................... 25 Study Aims ................................ ................................ ................................ ............................. 28 2 M ETHOD ................................ ................................ ................................ ............................... 31 Participants ................................ ................................ ................................ ............................. 31 Measu res ................................ ................................ ................................ ................................ 31 Theory of Mind ................................ ................................ ................................ ................ 32 Language Ability ................................ ................................ ................................ ............. 36 Executive Functioning ................................ ................................ ................................ ..... 37 Working Memory ................................ ................................ ................................ ............ 38 3 R ESULTS ................................ ................................ ................................ ............................... 39 Demog raphic Information ................................ ................................ ................................ ...... 39 ................................ ................................ ......................... 39 Study Aim 1 ................................ ................................ ................................ ..................... 39 Study Aim 2 ................................ ................................ ................................ ..................... 40 Study Aim 3 ................................ ................................ ................................ ..................... 42 4 D ISCUSSION AND CONCLUSION ................................ ................................ .................... 48 Discussion ................................ ................................ ................................ ............................... 48 Addressing Study Aims ................................ ................................ ................................ ... 48 Study aim 1 ................................ ................................ ................................ ............... 48 Study aim 2 ................................ ................................ ................................ ............... 49 Study aim 3 ................................ ................................ ................................ ............... 52
6 Strengths and Limitations ................................ ................................ ................................ 55 Study strengths ................................ ................................ ................................ ......... 55 Study limitations and future directions ................................ ................................ .... 55 Conclusion ................................ ................................ ................................ .............................. 57 LIST OF REFERENCES ................................ ................................ ................................ ............... 58 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ......... 62
7 LIST OF T ABLES Table page 3 1 ................................ ................................ ............... 43 3 2 Paired samples t tests for latency and uncertainty ................................ ............................. 44 3 3 working memory (WM) ................................ ................................ ................................ ..... 45 3 4 Correlation s for false belief, CELF P 2 Core Language Scores and individual subtests ................................ ................................ ................................ ............................... 47
8 LIST OF FIGURES Figure page 3 1 Group comparisons (t tests): Children passing vs. failing false belief .............................. 46
9 LIST OF ABBREVIATIONS CELF P 2 Clinical Evaluation of Language Fundamentals Preschool Second Edition E1 Experimenter 1 E2 Experimenter 2 EF Executive functioning or executive function FB False Belief IC Inhibitory control M Mean PPVT 3 Peabody Picture Vocabulary Test Version 3 SD Standard deviation WM Working memory
10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science LANG UAGE AND THEORY OF MIND IN 36 MONTH OLD CHILDREN By Mary Ellen McKay Easters August 2009 Chair: M. Jeffrey Farrar Major: Psychology Language has been found to be a predict or of false belief performance in a wide range of studies and across normally developing and special populations such as SLI and William s Syndrome children. Serious questions remain concerning the precise role of language in false belief competence. Is la nguage a necessary and fundamental precursor to false belief s or are the demands of traditional false belief tasks driving this association? Also, is general language or are specific aspects of language (i.e., complementation syntax) differentially importa nt for false belief ? The current study used a test of false belief that is low in language demands designed for use with 36 month old children in order to address these questions, and to examine the roles of executive functioning and working memory in fals e belief performance. Results indicate that general language is the best predictor of false belief performance using this task. Results are consistent with previous work supporting the role of general language for false belief and are interpreted using Kat herine Nelson as the mode of development of Theory of M ind in young children.
11 CHAPTER 1 INTRODUCTION Theory of Mind Theory of ind can be described as the understanding that individuals are well as emotion. Theory of M ind includes an understanding that individuals are intentional age nts, that is, that mental state s lead to action. In addition, Theory of M ind includes the understanding that the mind represents the world in a way that may or may not accurately reflect the state of reality (i.e., that false beliefs are possible), and tha t mental states can differ among of human interaction. In fact, while comparative research has revealed many commonalities among the cognitive capabilities of hu mans and other primates, particularly in the social realm, it has been proposed that the essential social cognitive difference between humans and other primates is that only human s have a Theory of M ind (T omasello, 1998). The fact that Theory of M ind is so research and debate in child social cognitive development. Garfield, Peterson, and Perry (2001) outlined some fundamental questions that have been the su bject of ongoing d ebate in the Theory of M ind literature. First, is the nature of the child s natur ive capacitie s, evidenced by performance on Theory of M ind (as well as othe r) cognitive tasks? Second, is Theory of M ind a modular cognitive capability? Evidence that individuals with a utism have particular difficulty with Theory of M ind tasks despite perf orming normally on other cognitive
12 measures (e.g. Baron Cohen, Leslie, & Frith, 1985; Happe, 1995) has caused some to assert that Theory of M ind is a cognitive ability separate and distinct from other cognitive capabiliti es. Third, is the emergence of Theo ry of M ind due to an innate capacity, or are experience and is: What is the specific role of language in the emergence of Theory of M ind? Examining the role o f individual differences, including individual differences in language but also in other cognitive factors, has been vital to researchers attempts to address some of these, as well as other questions surrounding the nature of Theory of M ind. Theory of Min d and a Focus on Tasks Traditionally, Theory of M ind has been assessed in children using a standard set of tasks which tap the child s ability to hold conflicting representations in mind, as well as predict the behavior or thoughts of others based on these conflicting representations and a given set of premises. For example, to determine whether children are capable of understanding the distinction between an object s appearance and its true identity, researchers often use the 986). In this task children are shown a piece of sponge realistically painted to look like a rock. When asked what they see, children typically report seeing a rock, but after being allowed to touch the sponge rock, children are surprised to find that the really soft and spongy. However, children without advanced mental representational abilities cannot coordinate the two conflicting representations of the sponge rock simultaneously. Very young children, when asked again, may report that the objec t really is a sponge and looks like a sponge, or that the object looks like a rock and really is a rock, despite evidence to the contrary. Older children show the ability to distinguish the sponge rock s appearance (that it looks like a rock) from its real ity (that it really is a sponge). Taking this task a step further, a child s ability to hold the conflicting representations of one s own previously held incorrect belief and a currently
13 known true state of reality can be determined. Thus, very young child ren show an inability to reason about this representational change and report that they always thought that the object was a sponge. More advanced children can recognize that, despite their new knowledge, they previously thought that the sponge was a rock (Flavell, 1986). representations is the false belief task. A standard false belief task is the mistaken locations task, otherwise known as the Sally Anne task (Baron Co hen, 1985). In this task the child is presented with a puppet that hides a toy in one of two containers, and then leaves the scene. A second puppet then enters the scene, switches the location of the toy from the original location to the other container, a nd then leaves the scene. When the first puppet returns, the child is asked to predict where the first puppet will look for the toy. A child who answers correctly will indicate that the puppet will look in the location where the toy was first hidden, under standing that the puppet was not present for the location switch, therefore unaware of the true location of the toy. A child who is not yet capable of understanding false belief s will predict that the puppet will search the container in which the toy is re ally hidden. In their meta analysis of research on Theory of Mind Wellman, Cross, and Watson results of a massive body of research on children s false belief p erformance, they found that, in general, over the period from age 3 to age 4 children progress from performing significantly below chance to significantly above chance on false belief measures. In addition, any attempts performance of younger children, by re wording the test questions or otherwise changing the traditional tasks, merely succeeded in making the task easier
14 for all children without revealing any deviation from the general developmental trend, and did not ra ise the performance of younger children to above chance levels (Wellman, et al., 2001). Researchers interest in adapting traditional false belief measures in order to uncover a rmance on false belief tests is an accurate indicator of the child s actual Theory of M ind competence. Skeptical researchers (e.g., Chandler, Fritz, & Hala, 1989) question whether the above chance performance of children around age 4 reflects a truly eme rgent competence in false belief understanding or rather a heightened ability to perform well on the specific tasks presented to them due to advances in language and other cognitive abilities. Further, is the below chance performance of 3 year olds due to the fact that they do not have the cognitive competence necessary to understand false belief s, or does it simply reflect that the tasks they are presented with are simply too difficult for a very young child to understand? It is possible that the tradit ion al set of tasks used to tap T heory of M ind is masking the true capabilities of younger children. There are several aspects of traditional false belief tasks that are of special concern for researchers (Carpenter, Call, & Tomasello, 2002). First, tradition al tasks have a test like format: information is presented that the child must attend to and process, followed by a test like question which the child must answer correctly. This format alone may pose difficulties for younger children simply because it is unfamiliar and demanding. Second, young children are false belief tasks. That is, they may have difficulty producing a verbal response that is in conflict with the current state of r eality (that the puppet will look in a container that does not now hold the toy) despite having some knowledge of the correct answer. Some researchers (Clements & Perner, 1994) have noted that children around 3 years of age will look in the direction of th e location corresponding to the
15 correct answer (where the puppet originally hid the toy) in mistaken locations tasks, yet verbally respond incorrectly, indicating that latent knowledge exists which is not expressed verbally. Finally, an important issue ha s been raised surrounding the language demands of traditional false belief tasks (Jenkins & Astington, 1995). While seemingly simple, the linguistic capabilities required to process information presented during these tasks, and to answer the tests question s, may be too advanced for younger children, or for children with specific difficulties with language (se e Miller, 2001, for a study of Theory of M ind in children with SLI). As reviewed below, language ability has been shown to be an important predictor of children s false belief performance, but because traditional tests use complex language structures it has been difficult to precisely identify what role language plays. Is language a necess ary precursor to Theory of M ind competence, or is language merely necessary for a child to understand the test that is being given? In addition to language, the role of executive functioning in false belief reasoning has been of interest (e.g., Carlson & Moses, 2001; Carlson, Moses, & Breton, 2002; Carlson, Wong, & Lemke 2005). The Question of the Role of Language Considerable research has examined the role o f language in the emergence of Theory of M ind, and in particular, false belief understanding. Many researchers have found language to be related to 4 to 5 year olds performance on Theory of M ind assessments (e.g., Jenkins & Astington, 1995; Astington & Jenkins, 1999; Ruffman, Slade, Rowlandson, Rumsey, & Garnham, 2003). Most of these researchers argue that language developme nt is a necessary precursor to Theory of M ind understanding although there is debate regarding the nature of its influence. It may be that language is a tool that allows for successful on line processing of the information associat ed with Theory of M ind tasks, and advanced language is required to think about beliefs, false beliefs, and the intricate nature of the world of mental states. This position of
16 linguistic determinism for false beliefs is contrasted with claims of a weaker relationship between language and Theory of M ind. It may be that lan guage is important for understanding Theory of M ind in the sense that the child must be able to comprehend a nd produce language to perform Theory of M ind tasks, in the same way that language would be needed to perform other cognitive tasks. In even starker contrast to the position of linguistic determinism, it is conceivable that language and T heory of M ind are merely two cognitive developments occurring during the same developmental period. This position, however, is not generally accepted, and most resear chers agree that language does play a specific role in the development of Theory of M ind. There is an ongoing debate regarding why and what aspects of language are necessary. Longitudinal studies of typically developing children have provided primary suppo rt of the position that language is a necessary and fundamental precursor to T heory of M ind (Astington & Jenkins, 1999; de Villiers & Pyers, 2002, Farrar & Maag, 2002). These studies have shown that earlier la nguage abilities predict later Theory of M ind p erformance. For example, Farrar and Maag (2002) found that language ability at age 2 pre dicted performance on tests of Theory of M ind at age 4. Similarly, Astington and Jenkins (1999) found that language ability at age three predicts Theory of M ind perfor mance at age four, but that the reverse is not true (see also de Villiers & Pyers, 2002). Some studies have identified a more bi directional relationship between language and false belief performance (Slade & Ruffman, 2005). Evidence from special populat ions provides additional support for the necessary and fundamental role of language skill in Theory of M ind. Studies of children with specific language impairment (SLI) and children with William s Syndrome provide one complementary picture of this position Children with SLI show deficits in linguistic abilities while other cognitive abilities, particularly non verbal intelligence, remain intact (unlike other forms of language delay
17 which occur in conjunction with other cognitive impairments; Conti Ramsden, 1994). Studies have shown that children with SLI perform poorly on tests of Theory of M ind. Ziatas, Durkin and Pratt (2003) found that at 8 years of age, 17% of their sample of SLI childre n failed Theory of M ind tasks. Only 8% of their normally developing language matched peers of age 5 years 8 months failed the same tasks. Alternatively, children with Williams Syndrome show high language skills when compared to their impaired performance on tests of non verbal intelligence. These children perform simil arly t o normally developing peers on Theory of M ind tas ks (for a review of studies of Theory of M ind in children with William s Syndrome, as well as a utism and deafness, see Garfield, et al., 2001). Intact lang uage tends to accompany intact Theory of M ind, and impaired langua ge tends to accompany impaired Theory of M ind. The case of a utism illustrates Theory of M ind and language deficiencies occurring simultaneously. Individuals with a utism experience an overarching difficulty with communication, with a ra nge of abilities in language, and impaired functioning in the social realm (Tager Flusberg, 1994). Specifically, individuals with a utism show particular difficulty passing tests of false belief (Fisher, Happe, & Dunn, 2005). Research has provided evidence that higher levels of language ability in individuals with autism coincide with better Theory of M ind performance. Thus, the more impaired an individual is in one, the more impaired they are likely to be in the other. Katherine Nelson and her colleagues h ave offered an explanation of how language developme nt leads to the development of Theory of M ind (Nelson, Skwerer, Goldman, Henseler, Presler, & Walkenfeld, 2003). This viewpoint emphasizes the social nature of language and the role of socia lization in th e development of Theory of M ind understanding. Thus, the role of
18 language in Theory of M ind development is not limited to providing a c ognitive tool for representing Theory of M ind tasks and situations, but is a key for entering what Nelson refers to as a s interaction with a social partner who is wise to the intricacies of the social world, through language, enables the child to enter the world of understanding the inner workings of the minds of others. Language is a crucial w indow into the minds of others. In essence, we speak to each other because our minds differ. When we speak to a child the child comes to understand this difference of minds. At the same time we cultivate the child s language skill. This is the basis for th e link between language skill and Theory of M ind. This viewpoint emphasizes how cognitive and social skills interact and lead to a mature understanding of the social world. Further, this is a process that occ urs throughout childhood, s not seen as a sudden acquisition that results from a qualitative change in social or linguistic competence. Research has supported the essential role of social int eraction in the development of T he ory of M ind understanding. Ruffman, Slade, and Crowe (200 2) found that the amount of mental state talk that mothers used when interacting with their children was predictive of their children s later Theory of M ind. Some have suggested that the pragmatic role of language may be the key to the rel ationship between language and Theory of M ind (Harris, de Rosnay, & Pons, 2005). There is an undeniable link between language and Theory of M ind. However, the specific aspect of language responsible for this connection has been questioned. Semantics, the aspect of langu age concerned with the meanings of words, has been considered to be crucial to an understanding of Theory of M ind. In this view, a child must have an understanding of the nd to
19 understand false beliefs (Olson, 1988). Others claim that it is syntax, or the aspect of language concerned with the structure of language (grammar) that is most closely tied to Theory of M ind. In this view, having grammatical structures that dictat e the sequencing of words enables the child to mentally orchestrate the sequence of events in false belief tasks (Astington & Jenkins, 1999). false belief s. One specific syntactic structure t hat has been scrutinized in the literature as being particularly relevant to Theory of M ind is mental state complementation syntax (de Villiers & de Villiers, 2000; de Villiers & Pyers, 2002). This grammatical structure involves a propositional phrase that is embedded within a sentence following a mental state verb. An example of this would be the following: She thought that there was a cookie in the jar Whether or not it is true that there is a cookie in the jar does not determine whether the sentence its elf is true, based on the fact that it is possible for the subject to have a thought that is not true. de Villiers and de Villiers (2000) position of linguistic determinism states that when a child does not have the capacity to understand mental state comp lementation, the child also does not have the ability to mentally represent the content of the sentence in memory. Thus, even if information is presented explicitly, for example, if the child is told, She thought that there was a cookie in the jar, but the re really was not, this information cannot be remembered, and if asked the child may not be able to repeat the information that was just presented. The ability to understand mental state complementation is thought to be crucial to the ability to understand false beliefs for this reason. The child who cannot have a memory based on embedded complements cannot mentally represent false beliefs. In large part, the data which provide support for the causal role of complementation syntax have shown relationships among false belief and complementation tasks beyond the role
20 of general language (de Villiers & Pyers, 2002; Schick, de Villiers, de Villiers, & Hoffmeister, 2007) However, these researchers agree that this type of data do not allow for clear causal expl anations. Training studies have provided additional support for the role of complement syntax for false belief performance. Hale and Tager Flusberg (2003) found improved false belief performance in groups of children trained in either mental state compleme ntation or a series of Theory of M ind tasks, as compared to a control group that did not receive training. Using an experimental design, Lohmann and Tomasello (2002) revealed that training children in complementation syntax resulted in improvements in fals e belief posttest scores, suggesting that an understanding of this type of syntax is essential to false belief understanding. However, in another experimental condition, children who simply engaged in discourse with an experimenter about alternate identiti complementation syntax or mental state terms, showed improved false belief scores as well. The only group in this study not showing improved false belief scores was a group that did not receive in any language based training. Thus, language interaction, whether involving complementation syntax or not, was effective in improving children s false belief performance, and the combination of discourse using mental state terms, complementation syntax, and reference to deception (deceptive objects) was the best condition for improving false belief scores (Lohmann & Tomasello, 2002). Despite both correlational and training studies suggesting a role of complement language, the claim that sententi al com plement syntax is essential to Theory of M ind has been questioned on the grounds that understanding mental state complementation requires having a concept of false beliefs. That is, the two measures are confounded (Perner, 1991). In answer to this t ype of
21 claim, Schick, et al. (2007) offered evidence that complementation syntax is correlated with false belief even when using low verbal false belief measures. Cross linguistic studies have revealed that the relationship between false belief and complem entation syntax is certainly not straightforward. Perner, Sprung, Zauner, and Haider (2003) examined the relationship between false belief performance and complementation syntax in German l state of desire despite the fact that a person may want something that may or may not be possible in reality, o follow the verb (e.g. Mother wants the child to clean her room ). The mental state of desire is understood earlier in English speaking children than mental states of thought and belief. It may be the case that because children are not required to understa nd complex syntax, they are able to understand desires sooner. higher level mental state verbs (e.g. Mother wants that the child should clean her room ). Following from the perspective of linguistic determinism, one might expect that German speaking children would begin to understand desire at the same time that they begin to understand other mental state verbs due to their shared syntactic structure. Perner et al. (2003) fo und that German false belief tasks. They suggest that the link between complementation and false belief understanding is not cl early established. Importantly, no study has shown above chance performance on a test of false belief for children who fail de Villiers and Pyers (2002) complementation comprehension task.
22 Several researchers have argued that general language skills are as sociated with false belief performance rather than semantics or syntax per se. Ruffman and colleagues (2003) argue for the importance of general language ability, or the combination of semantics and syntax. Often abilities in syntax and semantics overlap a nd the two abilities interact in development. Semantic development influences syntactic development. To understand how the words in a sentence are related syntactically, the child must first have semantic knowledge of the words in the sentence. Syntactic d refers to the idea that children use word order as a frame of reference to determine meaning for unknown words (Ruffman, et.al, 2003). Further, any test that examines syntax necessarily requires some understanding of semantics. In several longitudinal studies, Ruffman and colleagues found that syntax and semantics abilities were eq ually strongly associated with Theory of M ind performance (e.g., Ruffman, et al., 2003; Slade & Ruffman, 2005). They suggest that language enables the child to mentally reflect upon mental states, beliefs, and false belief situations, leading to the ability to express an implicit understanding of social situations. It is clearly established th at l anguage is associated with Theory of M ind development, and in particular with false belief reasoning. Questions remain regarding the nature of this association. Is a certain level of language ability necessary for false belief reasoning, or are the languag e rich tasks used to assess false belief s responsible for the association? Are these linguistic task demands masking younger children s true false belief competence? Most prior research has adopted the perspective that language is an important development required for false belief reasoning (e.g., Astington & Jenkins, 1999; de Villiers & Pyers, 2002). Unfortunately, these studies have all employed language rich false belief tasks. This is understandable, in that purely non verbal false belief tasks are few and far between.
23 Call and Tomasello (1999) developed a non verbal false belief task for use with chimpanzees, but found that this task was even more difficult for 4 year olds than traditional tasks. Langdon and colleagues (1997) developed a non verbal p icture sequencing task for use with schizophrenic patients which includes a false belief trial based on the mistaken locations task. Porter, Coltheart, and Langdon (2008) utilized this picture sequencing task in a study of William s Syndrome children. They found that William s Syndrome children had more difficulty with false belief s when they were not able to rely on their keen language skills, suggesting that non verbal tasks have much to reveal about the language false belief relationship. Schick, et. al. (2007) used two low verbal measures of false belief in examining the relationship between complementation syntax and false belief in deaf children. Complementation syntax was found to be a predictor of false belief using these low verbal false belief mea sures. However, no study has addressed the relationship between individual differences in language and performance on a non verbal or low verbal false belief task in normally developing children. We have yet to determine the influence of linguistic task d emands on the language false belief association and on younger children s false belief competence. A false belief task that can be used effectively with younger children and does not rely on complex linguistic structures in the presentation and solution of the task itself is needed. Carpenter, Call, and Tomasello (2002) have developed a test of false belief that does just that. The language used in this task is minimal, ) even very young children will be familiar with. Carpenter and her colleagues (2002) developed this task for use with children around the age of 36 months. This task will be referred to as the 36 month old false belief task. This test has several other a dvantages as well. It does not involve a
24 requiring the child to p This task involves the use of novel toys and a container to hide the toys in across several training, control, and false belief trials. An experim enter interacts with the child using a novel toy, and then hides the toy inside a container. For false belief trials, one experimenter leaves the room, and is not present as the toy is moved from the container by another experimenter. Upon returning to the in the container where it was last hidden, but is not able to open the container. Then, seeing the order to choose correctly from two options, the child is required to understand which toy that the experimenter has falsely believed to be in the container. A complete description of task procedure is provided later. To address the inevitable question of whether this new task truly measures false belief Carpenter et al. (2002) included some ingenious measures to determine whether children truly understand the false belief s involved in the task or are simply guessing. The researchers determined that in fa lse belief trials, children chose the object that the adult requested quickly, equally as quickly as when choosing an object that the adult requested directly (that is, not based on a false belief ). This was the case regardless of whether or not the child chose correctly on the false belief trials. Carpenter et al. (2002) hypothesized that children felt that they knew which object the adult was requesting. In addition, on false belief trials, children showed signs of uncertainty, checking for adult guidance when making their choices. Their uncertainty was similar for trials during which the child was required to simply guess which object the adult was requesting. This is viewed as an indication that the children recognized that the false belief trials
25 were m ore challenging than trials during which the adult asked for an object directly. Based on understood that the task was more complex than a direct request, Carpen ter et al. (2002) assert that their task is an assessment of false belief understanding. Carpenter et al. (2002) found that over half of three year old children in their sample performed at above chance levels on false belief trials. Their results raise e xciting possibilities for the st udy of the role of language in Theory of M ind development. This task may enable researchers to determine what specific role langua ge plays in the development of Theory of M ind. Because the language demands of this task are l ow, researchers can be assured that performance on this task is not merely an artifact of the child s advanced linguistic capa bilities, but is the result of Theory of M ind reasoning itself. In addition, because variability in performance among children rem ains for this task, it is possible to determine if individual differences in language explain this variability. Thus, if language predicts performance on the 36 month old false belief task despite the minimal language demands of the task itself, we can hav e a deeper understanding that language does not simply aid children in understanding the langua ge structures used in tests of Theory of M ind, but plays a more fundamental role. Alternately, if language is not associated with the variation in performance on these tasks, it would suggest that the rel ationship between language and Theory of M ind found previously may have been an artifact of th e type of tasks used to assess Theory of M ind. The Question of the Role of Executive Function and Working Memory In a ddition to language, an important cognitive ability thought to be intricately related to Theory of M ind development is the child s executive functioning ability, specifically, inhibitory control (Carlson & Moses, 2001; Carlson, Moses, & Breton, 2002). Fals e belief tasks inherently require that the child inhibit the prepotent response of responding based on what they know to be
26 the true state of reality in favor of the correct response which conflicts with reality. Carlson and colleagues have found that perf ormance on tasks which tap the child s inhibitory control ab ilities predict performance on Theory of M ind tasks above and beyond the role of language. In general, 3 year olds have difficulty with traditional executive functioning tasks, such as the Dimensi onal Change Card Sorting Task, which requires the inhibition of a previously learned rule in order to sort cards along a different dimension (Zelazo, Frye, & Rapus, 1996; Perner & Lang, 2002). Four year olds are better able to perform this task, indicating that this type of executive functioning skill develops during this time of life. The 36 month old false belief task allows for the exploration of how strongly inhibitory control is related to performance on a false belief task that is low in language dema nds, yet similar in inhibitory control requirement. Indeed, it may be that executive func tioning is more fundamental to Theory of M ind than is language. Perhaps young children with advanced inhibitory control are best able to pass tests of false belief an d are best able to coordinate conflicting representations of a false belief and reality in mind simultaneously. Indeed, it may be possible that inhibitory control ability is a prerequisite for both advanced Theory of M ind and advanced language. As with la nguage, the nature of the relationship between exe cutive functioning ability and Theory of M ind understanding has been the subject of debate. It is possible that inhibitory control is a requirement for giving the correct response to questions about false b elief s. However, researchers such as Perner, Lang, & Kloo (2002) have suggested that this explanation is not sufficient in explaining the link bet ween executive functioning and T heory of mind performance. Their research has shown that executive functioning ability is highly correlated not only with responses to false belief questions, but also with versions of false belief tasks that require children to give an explanation along with their responses. It is not simply that executive
27 functioning is necessary for inhibition of prepotent responses, but appears to be related to a deeper understanding of false beliefs. As with the rel ationship between language and Theory of M ind, the relationship bet ween executive functioning and Theory of M ind is not clear. New v ersions of false belief tasks that reduce language demands and provide evidence of above chance performance in younger children may prove to be the most useful tools in explaining these associations. In addition, working memory has been conceptualized as a possible third factor explaining the rel ationship between language and Theory of M ind (Slade & Ruffman, 2005). This possible underlying factor has been found to correlate both with language ability (e.g., Baddeley, Gathercole, & Papagno, 1998) and Theory of M ind ability (e.g., Davis & Pratt, 1995; Gordon & Olson, 1998). It is possible that younger children who are not capable of passing false belief tasks are not capable of remembering all of the information presented over the course of a task. With regar ds to the 36 month old false belief task, it may be that the children who are most capable of remembering a sequence of events will be most capable of passing the task. Further, working memory may allow young children who do not have advanced language to p ass the 36 month old false belief task, as this skill may develop earlier. Because the new task does not use advanced language it is possible to disentangle the roles of language, Theory of M ind, inhibitory control and working memory. The role of language and other factors in performance on the 36 month old false belief task The proposed study is intended to address two primary questions: 1) What is the role of language in performance on the 36 month old false belief task and, 2) What is the role of execu tive function and memory in performance on this task? To answer these questions, children
28 will be assessed using the new false belief task as well as measures of language ability, executive functioning (specifically, inhibitory control), and working memory Study Aims Study aim 1 The first aim of this study is to attempt to replicate the procedure and results of Carpenter and her colleagues (2002) in the original report of the 36 month old False belief task. Study aim 2. The second aim of this study is to determine if language ability is associated with performance on the 36 month false belief task despite the low language demands of the task, and to determine the nature of this association. It is hypothesized that that 3 year olds performance on the 36 month old False belief task will be found to be associated with language ability because it is thought that langua ge plays a fundamental role in Theory of M ind development, and not that language is merely required to understand Theory of M ind tasks. If la nguage is found to be associated with false belief performance, this study will examine what aspect of language in particular is involved in this relationship. That is, it will be determined if general language ability is the best predictor of Theory of M i nd performance, or if semantics, syntax, and mental state complementation are differentially important to the child s ability to pass the new task. It is hypothesized that general language ability will be more strongly associated with false belief performa nce than specific aspects of language ability (syntax or semantics). Thus, Core Language Scores from the CELF P 2 and PPVT 3 scores will be more highly correlated with performance on the false belief test than any one individual subtest of the CELF P 2 or the test of mental state complementation syntax. This prediction is based on Slade and Ruffman s (2005) finding that syntax and semantics are equa lly important as predictors of Theory of M ind.
29 Further, this study will examine whether children will be able to pass this test of false belief without having the ability to pass de Villiers and Pyers (2002) complementation task. If so, this would be compelling evidence against their position of linguistic determinism (that complementation syntax determines false belief understanding). It is hypothesized that performance on the false belief task will not be dependent upon complementation syntax competence, as children of this age group are not generally expected to pass the complementation task (de Villiers & Pyer s, 2002). Study aim 3 Finally, the third aim of this study is to explore the roles of working memory and inhibitory control in performance on this test of false belief Working memory and executive function m ay be strongly associated with Theory of M ind p erformance in 36 month olds. By removing language demands of traditional false belief tests, it may be possible to reveal that working memory and inhibitory control (rather than language) are the key factors supporting an understanding of false beliefs In hibitory control may be particularly significant in the child s ability to pass the 36 month old false belief task in that the child will be required to inhibit a prepotent behavioral response in favor of the correct response in order to pass tests of inhi bitory control as well as the 36 month old false belief task. This ability to inhibit prepotent responses may be fundamental to the devel opment of a more sophisticated Theory of M ind. Thus, it is hypothesized that false belief performance will be associate d with performance on measures of inhibitory control and working memory. Next, in the case that language, inhibitory control and working memory are all found to be correlated with false belief the possibility will be explored that language and false belie f are two advancements that develop as a result of more basic cognitive advancements at this age. Inhibitory control and working memory may be basic cognitive advancements that support the
30 de velopment of both language and Theory of M ind in general. It is h ypothesized that the relationship between language and false belief will remain after controlling for the effects of inhibitory control and working memory. That is, these cognitive competencies are expected to be relevant to false belief competence, but ar underlying the relationship between language and false belief
31 CHAPTER 2 METHOD Participants Participants were recruited through classroom teachers and research assistants who provided parents with a flie r explaining the study. Compensation included a $5.00 per child incentive paid to the preschool center to be used for a classroom activity or supplies of the teacher s choice. Thirty seven children from local preschools were tested. Eleven children were dr opped from the study. Four began testing but either left the area or simply did not return to preschool within the testing period. Six were found to be non English speakers, thus not eligible for our language assessments. One was dropped due to the inabili ty to pass training items for the false belief black/white task, cat/dog task, and the CELF P 2. The final sample included 19 male and 7 female participants, and was largely white (21 white, 3 Hispanic, 2 Asian). Participants ranged in age from 30 to 45 months ( M =38.62, SD=3.837). The mean age of the sample was s lightly older than the sample used by Carpenter, Call, and Tomasello (2002, M == 36.2 months ) in the original study of the 36 month old task. The age range was also somewhat wider. This was the re sult of the wide age range recruited, and the lag between flier distribution and date of testing. Mother s highest level of education was elevated in this sample, with all reporting mothers having at least some college, and 14 mothers having a degree highe r than a bachelor s degree. Measures Participants were tested using a battery of measures assessing Theory of M ind, language ability, inhibitory control, and working memory. Tasks were presented in a fixed order (Carlson & Moses, 2001). The 36 month old Fa lse belief task was always presented last. This was done in
32 large part to allow the children to establish a rapport with the testers before engaging in the false belief task, in an attempt to reduce apprehension or anxiety on the part of the children. The false belief measure was videotaped, with parental permission. Theory of Mind 36 month old false belief task. In this study the experimenters used the 36 month old False belief task (Carpenter, et al., 2002). Order of presentation of the four false belief trials was randomized across children, and one baseline condition followed each false belief trial. Thus, each child experienced 8 trials. Procedure for false belief trials is as follows: 1.) Child is seated on floor with two experimenters (E1 and E2), w ith blanket placed about 6 feet behind the child. E1 places 1 of 2 familiar toys into a box and removes it, then places it next to the other toy on the blanket. E1 t hen asks the child to get only the toy that was in the box. This is repeated 3 times. 3.) E1 then begins the false belief unfamiliar objects (for which the child is not likely to have already learned a ve rbal label) in her lap, covers them with a cloth, and brings out a container. While the child and E2 watch, E1 places 1 of the 2 unfamiliar objects (object A) in the container, shows the child and E2 what is inside, then closes the container. Different obj ects and containers are used for each of the 4 false belief trials. 4.) False belief B: While both the child and E2 watch, E1 switches locations of objects A and B, so that object B was the last object seen in the container by E2 before E2 excuses herself from the room and shuts the door.
33 5.) False belief A: E2 excuses herself from the room before E1 switches the locations of objects A and B, so that the last object E2 saw in the container was object A, though the last object actually inside the container was object B. 6.) All c onditions : While E2 is gone, E1 takes object B out of the container, closes the container, and places both objects on the blanket. E1 then calls E2 back to the room. When E2 returns, she announces that she wants to get an object wh ile looking directly at the container and ignoring the objects on the blanket. 7.) Novel w ord c onditions: E2 announces that she wants to get the object by referring to the object using a novel word (dax, toma, or gazzer). She announces her intention to ge t the object 8 times while walking over to the container, holding onto the container, and trying to open the container before declaring it stuck. 8.) Generic w ord c onditions 9.) At test, E2 turns the child in the direction of the blanket, and with the child s back to s the (toy/dax)! Could you get just the (toy/dax) for me? Thus, t he four false belief trials are False belief A Novel Word, False belief A Generic Word, False belief B Novel Word, and False belief B Generic Word. A correct response to False belief A conditions is retrieval of object A, indicating that the child understa nds that E2 would have thought that object A was the last object in the container, based on the fact that this was the last object seen in the container by E2, even though the last object actually in the container was object B. A correct response to False belief B conditions is retrieval of object B, indicating that the child understands that E2 believes that object B was in the container even though there actually is NO object in the container. Both sets of conditions are included to indicate not only whet her children choose the
34 correct object that is requested, but also to indicate that children who have an understanding of false belief will switch strategies based on the condition. Scores on the 36 month old false belief task range from 0 to 4 correct res ponses. The baseline conditions are as follows: Knowing Baseline Conditions: Each child experienced two knowing baseline conditions. In the familiar word condition, the experimenter places two toys with familiar labels on a blanket (such as a toy car and a teddy bear), and asks the child to retrieve one of the objects using the generic word condition, the experimenter places one et the one on toy the child should be retrieving. Guessing Baseline Conditions: Each child experienced two guessing baseline conditions. In the novel word condi tion, the experimenter places two novel objects on the blanket and asks the child to retrieve an item by using an unfamiliar, novel label. For example, the experimenter In the generic word condition, the experimenter asks the child to retrieve one of two novel objects by child makes will be based only on guessing, as the chi ld will have no other indicator of which object is really being requested. In sum, the 36 month old False belief task consists of 8 separate trials, four trials assessing false belief and four baseline trials. Coding and reliability. All conditions were scored live for correct or incorrect choices on false belief trials, and each child received a score of 0 4 correct. In addition, independent
35 raters viewing the videotape of each session coded for latency to choose and uncertainty. Latency to choose was re corded in seconds. Indicators of uncertainty included questioning of the indicators of uncertainty was recorded. (In Carpenter et al, 2002, observation of any or all of Reliability for latency and uncertainty coding was assessed by having a different coder recode the videotaped false belief sessi on for eight randomly selected children (32.5%). Pearson s correlations between coders were as follows: for latency r = .961, p < .001, and for uncertainty r = .974, p < .001. Mean times to decide (latency) for guessing and knowing baseline conditions and false belief trials were compared using paired samples t tests. Mean levels of uncertainty were compared for guessing and knowing baseline conditions and false belief trials as well. These analyses were intended to support the claim by Carpenter et al. (2 002) that children s performance on false belief trials is not based on random guessing, but is indeed based on their understanding of false belief s. Specifically, Carpenter et al. (2002) found that latency to choose an object was similar for knowing and f alse belief trials (latency for guessing trials was longer). choose in false belief conditions. In addition, they found that indications of uncertainty were simil ar for guessing and false belief conditions (showing less uncertainty in knowing conditions). The researchers reason that this indicates children behaved as if they felt the choice was a more s. Replication of these findings
36 in the present study will bolster their claims that children s performance is not based on random guessing. Language A bility Three language measures were used 1. Clinical Evaluation of Language Fundamentals Preschool, S econd Edition (CELF P 2, Semel, Wiig, & Secord, 2004). The CELF P 2 was used as a measure of various aspects of children s language development: morphology, syntax, expressive vocabulary, and, when the three are combined, general language. Children comple ted the three subtests Structure, testing syntax, and Expressive Vocabulary). The CELF P 2 can be used with children as young as age 3. The measure involves showing ch ildren simple pictures and probing children to describe the pictures. Probes are directed at eliciting specific linguistic items. Standardized scores are reported. 2. Peabody Picture Vocabulary Test Version 3 (PPVT 3, Dunn & Dunn, 1997). The PPVT 3 was us ed as a measure of receptive vocabulary, and is also often considered a proxy for general language ability. Children are asked to point to the picture representing the word uttered by the experimenter. Standardized scores are reported. 3. The Complementati on Comprehension Test. de Villiers & Pyers, 2002, developed this test for use in the investigation of the role of the specific ability to comprehend complementation syntax as it relates to Theory of Mind ability. This measure involves the experimenter read ing a sentence containing complementation syntax while pointing to photos of actors and the relevant pieces of information shown in the photos. The child is then asked a question which requires the child to have understood complementation syntax. For examp le, the
37 the man something contrary to reality, that she saw a ghos t. The test consists of 12 sentences and sets of photos. de Villiers and Pyers (2002) use a criterion of 10 or more correct responses for passing the task. All scores below 10 are considered to be failing scores. Executive F unctioning Executive functioni ng was assessed using two tasks used in previous studies of the relationship between executive functioning and Theory of Mind The types of tasks chosen have been shown to be significant predictors of performance on traditional false belief tasks (Carlson & Moses, 2001; Carlson, Moses, & Breton, 2002). These tasks are intended to assess inhibitory c ontrol or the aspect of executive functioning concerned with inhibition of a prepotent response. 1. Dimensional Change Card Sorting Task In this task (Frye, Ze lazo, & Palfai, 1995), children are required to inhibit a prepotent response and suppress a rule already learned in a simple card sorting task. The child is instructed to sort cards according to one dimension, either color (red and blue) or shape (rabbits and boats). Five cards, in a fixed order, are presented, and children are asked to place the cards into one of two containers labeled with a red rabbit or a blue boat (if sorting according to shape, rabbits go in the rabbit container, and so on). After com pleting this, the experimenter instructs the child that the rule for sorting has changed, and that for the next cards, the child should sort according to the opposite dimension (red cards go in the red container, and so on). Five cards, in a fixed order, a re presented and the experimenter repeats the new rule before presenting each card. Scores are the number of cards sorted correctly out of the five presented after the rule switch. 2. Black/ w hite t ask. A second measure of executive functioning a simplifie d adaptation of the day/night stroop task (Gerstadt, Hong, & Diamond, 1994) requires the child to inhibit a prepotent urge to correctly label the colors black and white (Simpson & Riggs, 2005). Children
38 are instructed to say the word white when shown a bla ck card, and to say the word black when shown a white card. Children will be given the rules, and then given up to three practice trials for each rule. When the child is able to respond correctly for each rule, test trials are started, during which no feed back is given. Scores are number correct out of 16 trials. If a child is unable to pass 3 practice trials, with the rule repeated between each trial, the child is given a score of 0 and not forced to complete the 16 test trials. Working M emory Cat/ d og t ask Working memory was assessed using an arbitrary rule task (Simpson & shown cards with unrelated line drawings (purple horizontal stripes or green vertical stripes). Children are given the rules, and then given up to three practice trials for each rule. When the child is able to respond correctly for each rule, test trials are started, during which no feedback is given. Scores are number correct out of 16 te st trials. If a child is unable to pass 3 practice trials, with the rule repeated between each trial, the child is given a score of 0 and not forced to complete the 16 test trials. The order of presentation was as follows: PPVT 3, dimensional change card sorting task, CELF P 2, complementation comprehension, cat/dog task, black/white task, and 36 month old false belief task.
39 CHAPTER 3 RESULTS Demographic Information Preliminary analyses indicated no significant differences between males ( n = 19) and fem ales ( n = 7) on any measure, and no significant differences between white ( n = 21) and non white (Hispanic and Asian, n = 5) participants on any measure. Mother s highest level of education (on a scale of 1 4, 1=some college, 2=bachelor s degree, 3=master s degree, 4=doctorate) did not correlate significantly with any measure. Subsequent analyses did not take these factors into account. 1. Co re Language Scores (standardized) for the CELF P 2 ranged from 63 137, M = 100.5 SD = 17.14 Standard scor es for the PPVT 3 ranged from 83 128, M = 106.88, SD = 13.26 In addition, children s pe rformance on inhibitory control and working memory tasks cove red the expected range and means for these tasks fell near the median. This suggests that this sample is representative of a wide range of abilities, averaging about the same as the population being characterized in this study. For false belief M = 2.19 correct responses ( SD = 1.1 ). Ten children passed false belief using the criteria used by Carpenter et al, (2002) of 3 or 4 correct out of 4 false belief trials, and of this 10, three children chose correctly on 4 out of 4 trials. Study A im 1 The first a im of the study was to replicate the procedure used by Carpenter et al., (2002) in the original report of the 36 month old false belief task, as well as to replicate the results of ndeed a true test of false
40 belief Paired samples t tests were performed for latency (time to choose an object, measured in seconds) and uncertainty (total number of behavioral indicators of uncertainty). Results indicate that children in this sample behav ed similarly to children in the original sample ( see Table 3 2). Time to decide was significantly different for false belief and guessing conditions ( t (39) = 2.1 4 p = .038), and not significantly different for false belief and knowing conditions. Specifi cally, children took 3.55 seconds longer to choose when guessing ( M = 11.76 seconds, SD = 8.94 seconds) than in false belief trials ( M = 8.32 seconds, SD = 5.91 seconds), and chose only 1.03 seconds sooner in false belief trials than in knowing conditions ( M = 9.11 seconds, SD = 5.93 seconds). Number of indications of uncertainty was significantly different for false belief and knowing conditions ( t (41) = 2.07 p = .045), and not significantly different for false belief and guessing conditions. Specifically children displayed more uncertainty in false belief conditions ( M = .67 indicators, SD = 1 indicator) than in knowing conditions ( M = .26 indicators, SD = .7 indicators). For guessing conditions ( M = 1.17 indicators, SD = 1.5 indicators) number of indica tors was similar to false belief conditions and significantly greater than in knowing conditions ( t (41) = 3.49 p = .001). Study A im 2 The second aim of this study was to determine if language ability is associated with false belief performance despite th e low language demands of the task, as well as to determine if working memory and inhibitory control are associated with false belief Pearson Product Moment correlations were used to address this aim ( see Table 3 3 ). Before examining relationships among f alse belief and other cognitive factors, it is important to consider the role played by the age of the child. First, age was not found to be correlated with performance on the 36 month old false belief task (Because age was not a factor for perfor mance on the false belief test, analyses employ standardized scores for CELF P 2 and PPVT 3) As expected, a correlation
41 was found between age and the complementation comprehension task ( r ( 25) = .40, p = .047). Interestingly, age was not correlated with inhibitory control as measured by the dimensional change card sorting task in this sample. Correlations were found between age and working memory ( r ( 24) = .58 p = .003) and age and executive functioning as measured by the black/white task ( r ( 24) = .53 p = .008). Th ere is a clear positive relationship between language and false belief Performance on the false belief task was significantly correlated with standardized scores on the CELF P 2 ( r (26) = .39 p = .047) and marginally correlated with standardize d scores on the PPVT 3 ( r (25) = .38 p = .060 ). However, this was not the case for the relationship between false belief and complementation syntax. A second set of analyses was performed to address the first aim as well. Independent samples t tests were used to re veal whether the children who passed the false belief trials differ from those children who did not pass the false belief trials in language, executive functioni ng, working memory, and age ( see Figure 3 1). Children were grouped as passing (answering corre ctly on 3 or 4 out of 4 false belief trials, n = 10) or failing (answering correctly on less than 3 false belief trials, n = 16) false belief then compared using independent samples t tests for the dependent variables. This comparison may be more meaningf ul in the examination of false belief reasoning, as the four false belief test trials in this task do not measure qualitatively different aspects of false belief Findings from the first and sec ond sets of analyses are consistent. False belief passers and failers differed significantly on the CELF P 2, t (24) = 2.26 p = .033, as well as on the PPVT 3, t (23) = 2.2 p = .039. Specifically, for passers, CELF P 2 M = 109.4, SD = 15.46 and PPVT 3 M = 113.50, SD = 9.48 For failers, CELF P 2 M = 94.94, SD = 16.14 and PPVT 3 M =
42 102.47, SD = 13.83 False belief passers and failers did not differ significantly on any other measure, nor did they differ in age. Given that language was found to be positive ly related to false belief performance, the next analyses explored what aspect of language in particular is involved in this relationship. Correlations among false belief performance and the three subtests of the CELF P 2 elucidate the relationships ( see T able 3 4). Performance on the false belief test correlated similarly, but only marginally, with each of the subtests of the CELF P 2 (specifically, for Word Structure, the measure of morphology, r ( 26) = .31 p = .124, Sentence Structure, the measure of syn tax, r ( 26) = .34 p = .093, and Expressive Vocabulary r ( 26) = .36 p = .070). The Core Language Score, combining these various aspects of language skill, proved to be the most highly predictive of false belief perfor mance ( r ( 26) = .39 p = .047). Next, th e relationship between false belief performance and performance on the complementation task is of particular interest. Not only did complementation syntax ability not correlate with performance on this false belief test, but false belief passers and failer s did not differ significantly on scores for the complementation task. Finally, and perhaps most clearly, whereas 10 children passed the false belief task, only one child met de Villiers and Pyers (2002) criteria of 10 12 correct responses out of 12 on the complementation task, and this one child failed false belief Study A im 3 Finally, the third aim of this study was to explore the role of working memory and inhibitory control in false belief performance. None of the measures of inhibitory control and working memory were significantly correlated with false belief (though all were in the expected direction see Table 3 3 ). In addition, independent samples t tests comparing false belief passers and failers indicate that the two groups did not perform diff er ently on the Inhibitory Control and
43 WM measures Because no significant relationships were found to exist between false belief and the measures of inhibitory control and working memory separate analyses were not performed to answer the question of wheth er inhibitory control and working memory are more basic cognitive tools that underlie the false belief la nguage association. Table 3 Range M SD Age in months 30 45 38.62 3.84 False Belief 0 4 2.19 1.10 CELF P 2 63 137 100.5 0 17.14 PPVT 3 83 128 106.88 13.26 Complementation 0 10 3.52 2.62 Card Sort 0 5 2.72 2.25 Black/White 0 16 6.33 6.04 Cat/Dog 0 16 12.88 3.76
44 Table 3 2. Paired samples t t ests for latency and u ncertainty Latency (seconds) Uncertain ty (number of indicators) M SD M SD False Belief ( FB ) 8.32 5.91 .67 1.00 Guessing 11.76 8.94 1.17 1.5 0 Knowing 9.11 5.93 .26 .7 0 Mean Diff SD t (39) p Mean Diff SD t (41) p FB/ Guessing 3.55 10.48 2.14 .038 .5 00 1.74 1.86 .0 70 FB/ Knowing 1.03 6.92 .94 .351 .405 1.27 2.0 7 .045 Guessing/ Knowing 2.62 9.12 1.8 2 .077 .905 1.68 3.49 .001
45 Table 3 3. orrelations for age, false belief language, inhibitory control (IC) and working memory (WM) F alse B elief Ag e in months CELF P 2 PPVT 3 Comp.^ Card Sort (IC) Black/ white (IC) Age in months .0 3 CELF P 2 *.39 .2 7 PPVT 3 .38 .10 **.6 9 Comp. .15 *.40 10 .2 5 Card Sort (IC) .2 6 .15 *.46 .15 .1 9 Black/white (IC) .10 **.5 3 .3 6 .22 .1 4 17 WM .0 6 **.5 8 .08 .0 1 .3 8 .0 5 **.5 4 ^Complementation task p < .05, two tailed ** p < .01, two tailed
46 Figure 3 1 Group comparisons (t tests): Children passing vs. failing false belief 30 35 40 45 Age in months Pass Fail 63 83 103 123 CELF P 2 Pass Fail 83 93 103 113 123 PPVT 3 Pass Fail 0 2 4 6 8 10 12 Complementation Pass Fail 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Card Sort Pass Fail 0 5 10 15 Black/White Pass Fail 0 5 10 15 Working Memory Pass Fail t(24) = .4 p = .695 t(24) = 2.26 p = .033 t(23) = 2.2 p = .039 t(23) = .34 p = .739 t(23) = 1.66 p = .111 t(22) = .65 p = .520 t(22) = .14 p = .894
47 Table 3 4 Correlations for false belief CELF P 2 Core Language Scores and individual s ubtests Core Language Score Sentence Structure Word Structure Expressive Vocabulary False belief *.39 .3 4 .31 .36 Core Language Score **.8 2 **.8 6 **.8 7 Sentence Structure **. 50 **.5 8 Word Str ucture **.65 p < .05 (two tailed) ** p < .01 (two tailed)
48 CHAPTER 4 DISCUSSION AND CONCLUSION Discussion The current study addressed one of the most debated questions in Theory of M ind research: What is the nature of the relationship between langu age and false belief in young children? By using a task that does not involve advanced language in its administration and by testing younger children who have previously not been able to pass the ling uistically complex traditional Theory of M ind tasks, a b etter understanding of this relationship is possible. Addressing Study Aims Study a im 1 The first, and possibly most critical, aim of the present study was to attempt to replicate the procedure and results of Carpenter and her colleagues (2002) in the ori ginal report of the 36 month old False belief task. Accomplishing this was supremely critical as all of the results and implications presented here rely on the assumption that this false belief task is a valid test of false belief The results suggest that this aim was realized. Children from this sample behaved similarly to the children in Carpenter et al s (2002) sample when choosing objects at test. When examining latency to choose, children chose objects in the false belief trails quickly, as if they kn ew immediately (right or wrong, based on false belief ) which object they felt the experimenter was looking for. When examining uncertainty, on the other hand, children displayed feelings of being uncertain similar to a guessing situation. They behaved as i f they understood that something was amiss, so to speak, and that they knew the experimenter who had left the room had missed out on something important.
49 Study a im 2 The second aim of this study was to determine if language ability is associated with perfo rmance on the 36 false belief task despite the low language demands of the task Of particular interest was to determine whether general language ability or specific aspects of language are responsible for this association. The results of this study clearl y indicate that performance on the 36 month old False belief task is related to language ability. Thus, we have compelling evidence for the position that language plays a fundamental role in the development of Theory of M ind, and against the position that language and false belief are related due to task demands inherent in traditional false belief tasks. Further, we have shown that performance on a low verbal false belief task is related to general language ability and not to any specific language skill No individual measure of syntax or semantics was more strongly associated with false belief than our measure of language which took into account syntax, semantics, and morphology (CELF P 2, core language score). In addition, ten children were able to pas s this test of false belief without passing de Villiers (2002) test of complementation syntax. Thus, we have provided the first evidence of this kind against the strong position of linguistic determinism proffered by de Villiers and de Villiers (2000). If one accepts that Carpenter, et al (2002) have developed a true test of false belief that 3 year olds are capable of passing at above chance levels, one must consider that the relationship between false belief and complementation syntax previously found ma y not be evidence for the strong position of linguistic determinism Perhaps complementation syntax is a reflection of the quality of the linguistic environment, or plays some other role for Theory of Mind development. These findings are especially compe lling when combined with the findings of Slade and Ruffman (2005) that general language, and not specific syntax, predicted false belief They conclude that both semantics and syntax are equally responsible fo r aiding in the development of
50 Theory of M ind. In this view, language allows the child to reflect upon implicit knowledge of social situations and false belief s, and leads to the ability to make these understandings explicit. Indeed, th ind know ory of M ind (Ruffman, 2000), and refers to the knowledge that children display through eye gaze but cannot express verbally early in development. In the present case of the relationship between general language and the low verbal 36 month old false bel ief task, it is theorized that having more developed language skill enables the child to mentally reflect upon the situation presented in the task. The child who is able to use language to think about the world can use this skill to reflect upon what the a ctors in the scenario see, know, and believe. The child is then able to act upon a sophisticated understanding of the situation and choose the correct object based on the actor s false belief Instead of the strong position of linguistic determinism whic h states that specific syntactic structures are responsible for emergence of false belief competence, these results are seen as support for the more pragmatic, experiential view of development of language and Theory of M ind. Ruffman and colleagues (e.g., T aumoepeau & Ruffman, 2008) support the role of social interaction and experience with language (particularly mother s mental state talk with infants) as a mechanism by which social understanding develops. Additionally, though this study did not examine the process of social construction or soci al experience as it relates to Theory of M ind specifically, the results presented here can be viewed as support for the experien tial view (Milligan, Astington, & Dack, 2007) as defined by Nelson According to this vie w (Nelson, et al, Rather than experie ncing an abrupt acquisition of Theory of M ind competence, children undergo a gradual transition towards social competence aided b y physical developments, cognitive
51 developments and importantly, interaction with more knowledgeable others. First, physical developments allow the child to experience the environment more fully. Innate drives for socialization, such as the preference for faces and infants social smiles, develop into social interactions such as joint attention and, later, early language. Language becomes the critical tool used between individuals to allow insights into the unseen mind, starting with an understanding of des ires and feelings, culminating in an adult like understanding of beliefs and intentions. This study provides an additional component to this depiction of development. The children in this sample showed above chance performance on this test of false belief and perhaps more compellingly, their performance was not related to chronological age. This test allows us to see continuity in development of false belief reasoning based on language development and not merely chronological age. Indeed, to successfully navigate the social world it is critical to understand the intentional behavior of others. We communicate with language to inform each other that our minds differ. This is the essence of false belief and the adult like social compe tence of Theory of M ind A child who is exposed to language is exposed to the contents of the minds of others. More experience with language, and better quality experience of language, is what a child requires to gain entry into the community of minds. This high quality experien ce with language leads to the child s own high quality language skill as well. The results presented here, that general language comprised of semantics, morphology and syntax is the best predictor of 3 year olds false belief performance using a task that requires very little language, support the view that it is not what aspect of language, but the quality of a child s language environment that determines a child s understanding of the minds of others.
52 Children whose language skills have been cultivated ea rlier have been escorted into the community of minds earlier as well. children into the community of minds is directly related to Vygotsky s (1986) concept of the zone achievement of attainable next steps in development. This is seen as the process at play for developing language, a representational mind, and ultimately, false be lief competence (Nelson, et al, 2003). As suggested by Slade and Ruffman (2 005), the relationship between Theory of M ind and language is bidirectional. Language and Theory of M ind interact in development in a manner similar to the interaction of syntax an d semantics in development. The child experiences language in social situations, enabling the child to gain insights into the minds of others, as well as gaining experience with language as a social tool. Social skills (or more specifically, Theory of M ind ) developed through this language interaction lead to a more skillful and savvy use and understanding of language itself. Study a im 3 The third aim of this study was to explore the roles of working memory and executive functioning, specifically, inhibitory control, in performance on this test of false belief These results would suggest that working memory and inhibitory control are not underlying factors in children s ability to understand false belief Here, our results are inconsistent with Nelson s (200 2) account of how the child comes to enter the community of minds. Nelson (2002) s Theory of M ind competence. False belie f tasks, including the 36 month old false belief task, require that the child suppress a prepotent
53 response, which coincides with the true state of reality, in favor of a correct response that coincides with a false belief, a function of inhibitory control These tasks also require that a child remember a string of related events that result in a false belief, a function of working memory. It is unclear why these factors were not relevant to false belief performance in the present study, as previous work h as shown the importance of executive functioning for false belief (e.g., Carlson & Moses, 2001). It is possible that the measures employed were unsuccessful at tapping these deeper mental competencies. In particular, the black/white task (for inhibitory co ntrol) has not been widely used. The more widely accepted test of inhibitory control that was used, the dimensional change card sorting task, showed some association with false belief although this association was not statistically significant. Thus, it i s possible that inhibitory control was indeed not an important factor for false belief in this case. The cat/dog task (working memory) also has not been widely used. More widely accepted tests of working memory, such as the backward digit span task, would not have been appropriate for children of this young age, as basic number knowledge is not yet expected of this group. A reliable test of working memory is needed for use with very young children. Given the case that inhibitory control was indeed operation alized successfully by the tasks used in this study, an alternate explanation is offered. Because the specific role of executive functioning and inhibitory control in Theory of Mind development has been debated, and empirical findings have been mixed, some researchers (Bull, Phillips, & Conway, 2008) have recognized the need to use multiple executive functioning tasks which tap separate aspects of executive functioning. This allows the researcher to view executive functioning as a fractionalized, rather tha n a unitary, construct. As a result, it is possible to determine precisely what aspects of executive functioning are involved in false belief performance. It is this
54 perspective that may explain why the measures of inhibitory control employed here did not correlate with performance on the 36 month old false belief task. The black/white inhibitory control task employed here is verbal in nature. The child was required to inhibit prepotent verbal responses in favor of a novel rule (say black when you see white false belief tasks (in which the child must verbally respond contrary to the state of reality for the false belief test question), the 36 month old false belief task reduced the need to inhibit a prepotent verbal response. Instead, responses were behavioral (choosing an object). If the child was indeed required to inhibit a prepotent response, this response would be a behavioral one (always choosing the last object actually in the container). It i s possible that the dimensional change card sorting task was slightly more predictive of false belief in this case because the child s responses here are a physical behavior (placing a card in a box), not a verbal response. However, the fact that the dimen sional change card sorting task tests for rule switching may have made this task a non significant factor for false belief in this case as well. An inhibitory control task that is more strictly behavioral/physical, rather than overlapping physical behavior and rule switching within one task, would elucidate more clearly the role of inhibitory control in the 36 month old false belief task. A possible example of this would be the version of Luria In this task the child must do the opposite hand gesture (fist), rather than imitating the same hand gesture (open hand), displayed by the experimenter. In this case physical imitation is the prepotent response to be inhibited in favor of the correct resp onse, the opposite gesture. As a result, it is not suggested that these results are in conflict with Nelson s Theory of Mind in
55 children. Instead, more sensitive measures must be employ ed to truly determine the role of executive functioning in false belief Strengths and Limitations Study s trengths This study is the first of its kind in two important ways. First, we examine d individual differences in multiple aspects of language as they relate to performance on a low verbal test of false belief s. This type of research is critical to gaining answers to the debate concerning the role of language in the development of Theory of M ind. By controlling the need for language skill to meet task de mands for false belief this study has made clearer the necessary and fundamental role played by language in the development of T heory of M ind. Second, this is the first research which has shown above chance performance on a test of false belief s by young children who are unable to pass de Villiers and Pyers (2002) test of complementation syntax. This type of evidence is the strongest argument against their position of the specific, and critical, role of complementation syntax in the development of false be lief competence, Additionally, this study utilized multiple measures of language and executive functioning (focusing on inhibitory control), allowing for a highly sensitive examination of the roles of multiple aspects of cognitive development for false bel ief s. Study l imitations and f uture d irections The present study is limited in that only cross sectional, individual difference data are presented. To best understand any aspect of development, longitudinal data are needed to track changes over time. Next, the data presented here correlate scores on standardized language measures with a test of false belief s. To provide clear support for the exp eriential view of language for Theory of M ind development, research examining the role of social interactions for the language false belief association is necessary.
56 In addition, the 36 month old false belief task may be questioned as a true test of false belief competence. The checks and balances included as a part of the task (latency and uncertainty, as discussed above) indicate that children are behaving as a result of their level of understanding of false belief s, when interpreted similarly to Carpenter et al (2002). However, these checks and balances are based on inferences drawn on children s behaviors and are open to alternate interpretations. It is possible that this task is an indicator of young children s false belief ls his or her belief during task administration. That is, when the experimenter is looking for the toy in false belief trials, he or explicitly stating a belief, th ese actions may be clear enough to a young child that the false belief of the false belief on his or her own. The child must understand that what the experimenter last saw indicates which object is sought (explaining latency and uncertainty patterns), but the false belief aspect of the task may be weakened or eliminated, keeping the task from tr uly matching traditional false belief tasks in its ability to tap this critical achievement. performance on false belief trials. In each and every case, the child is req uired to choose one of two objects, allowing for a 50/50 chance to choose correctly randomly. The authors of the task (Carpenter, et al., 2002) include two types of false belief s ( False belief A, and False belief B, as described above) to require that the child switch strategies to show adherence to false belief understanding to address this issue. A task with more false belief trials, or more objects to choose
57 false b elief s based on random guessing. Finally, inclusion of a wider variety of executive functioning tasks would allow for closer power of the data presented. Conclusio n 36 month old false belief task have been replicated, 2.) That general language is the best predictor of performance on a low verbal test of false belief s for young children, and 3.) That inhibitory control and working memory, as measured, are not important predictors of performance on this test of false belief s.
58 LIST OF REFERENCES Astington, J. W., & Jenkins, J. M. (1999). A longitudinal study of the relation betwe en language and theory of mind development. Developmental Psychology, 35 1311 1320. Baddeley, A., Gathercole, S., & Papagno, C. (1998). The phonological loop as a language learning device. Psychological Review, 105, 158 173. Baron Cohen, S., Leslie, A., & Frith, U. (1985). Does the autistic child have a "theory of mind"? Cognition, 21 37 46. Bull, R., Phillips, L. H., & Conway, C. A. (2008). The role of control functions in mentalizing: Dual task studies of Theory of Mind and executive function. Cognitio n, 107 663 672. Carlson, S. M., & Moses, L. J. (2001). Individual differences in inhibitory control and children s theory of mind. Child Development, 72, 1032 1053. Carlson, S. M., Moses, L. J., & Breton, C. (2002). How specific is the relation between ex ecutive function and theory of mind? Contributions of inhibitory control and working memory. Infant and Child Development, 11 73 92. eginning understanding of false belief. C hild Development, 76, 73 86 Carpenter, M., Call, J., & Tomasello, M. (2002). A new false belief task for 36 month olds. British Journal of Developmental Psychology 20 393 420. Chandler, M., Fritz, A. S, & Hala, S. (1989). Small scale deceit: Deception a s a marker of two three and four year olds' early theories of mind. Child Development, 60 1263 1277. Clements, W.A., & Perner, J. (1994). Implicit understanding of false belief. Cognitive Development, 9, 377 395. Davis, H. L., & Pratt, C. (1996). The development of children s theory of mind: The working memory explanation. Australian Journal of Psychology, 47 25 31. de Villiers, J., & de Villiers, P. (2000). Linguistic determinism and the understanding of false beliefs. In P. Mitchell & K. Riggs (Eds. ), Children s reasoning and the mind (pp. 189 226). Hove, UK: Psychology Press. d e Villiers, J. G., & Pyers, J. E. (2002). Complements to cognition: A longitudinal study of the relationship between complex syntax and false belief understanding. Cognitive D evelopment, 17 1037 1060. Dunn, L., & Dunn, P. (1997). Peabody Picture Vocabulary Test 3 rd Edition. Circle Pines, MN: American Guidance Services.
59 Farrar, M.J., & Maag, L. (2002). Early language development and the emergence of a theory of mind. First Lan guage 22 197 213. Fisher, N., Happe, F., & Dunn, J. (2005). The relationship between vocabulary, grammar, and false belief task performance in children with autistic spectrum disorders and children with moderate learning difficulties. Journal of Child Ps ychology and Psychiatry, 46 409 419. Flavell, J. H. (1986). The development of children s knowledge about the appearance reality distinction. American Psychologist, 41 418 425. Frye, D., Zelazo, P. D., & Palfai, T. (1995). Theory of mind and rule based r easoning. Cognitive Development, 10, 483 527 Garfield, J.L., Peterson, C.C., & Perry, T. (2001). Social cognition, language acquisition, and the development of the theory of mind. Mind & Language 16 494 541. Gerstadt, C. L., Hong, Y. J., & Diamond, A. ( 1994). The relation between cognition and action: Performance of children 3.5 7 years old on a Stroop like day night test. Cognition, 53 129 153 Gordon, A. C., & Olson, D. R. (1998). The relation between acquisition of a theory of mind and the capacity t o hold in mind. Journal of Experimental Child Psychology, 68, 70 83. Hala, S., Hug, S., & Henderson, A. (2003). Executive function and false belief understanding in preschool children: Two tasks are harder than one. Journal of Cognition and Development, 4 275 298. Hale, M.C., & Tager Flusberg, H. (2003). The influence of language on theory of mind: a training study. Developmental Science, 6 346 359. Happe, F. (1995). The role of age and verbal ability in the ToM performance of subjects with autism. Child Development, 66, 843 855. Harris, P. L., de Rosnay, M., & Pons, F. (2005). Language and children s understanding of mental states. Current Directions in Psychological Science, 14, 69 73 Hughes, C. (1998). Finding your marbles: Does preschoolers' strategic behavior predict later understanding of mind? Developmental Psychology, 34, 1326 1339. Jenkins, J. M., & Astington, J. W. (1995). Cognitive factors and family structure associated with theory of mind development in young children. Developmental Psychology 32 70 78. Langdon, R., Michie, P. T., Ward, P. B., McConaghy, N., & Catts, S. V. (1997). Defective self and/or other mentalising in schizophrenia: A cognitive neuropsychological approach. Cognitive Neuropsychiatry, 2, 167 193. Lohmann, H., & Tomasello, M. (2002). The role of language in the development of false belief understanding: a training study. Child Development 74 1130 1144.
60 Luria, A. R., Pribram, K. H., & Homskaya, E. D. (1964). An experimental analysis of the behavioral disturbance produced by a left frontal arachnoidal endothelioma (meningioma). Neuropsychologia, 2, 257 280. Miller, C. (2001). False belief understanding in children with specific language impairment. Journal of Communication Disorders, 34 73 86. Milligan, K., Astington, J. W., & Dack, L. A. (2007). Language and Theory of Mind: Meta analysis of the relation between language ability and False belief understanding Child Development, 78 622 646. Nelson, K., Skwerer, D.P., Goldman, S., Henseler, S., Presler, N., & Walkenfeld, F.F. (2003). Entering a community of minds: an experiential approach to 'theory of mind'. Human Development, 46, 24 46. Olson, D. R. (1988). On the origins of beliefs and other intentional states in children. In J.W. Astington, P. L. Harris, and D. R. Olson, ( Eds.), Developing theories of mind (pp. 414 426). Cambridge: Cambridge University Press. Perner, J. (1991). Understanding the representational mind. Cambridge: Bradford Books/MIT Press. Perner, J., & Lang, B. (2002). What causes 3 year olds difficulty on the dimensional change card sorting task? Infant and Child Development, 11 93 105. Perner, J., Lang, B., & Kloo, D. (2002). Theory of mind and self control: More than a common problem of inhibition. Child Development, 73 752 767. Perner, J., Sprung, M., Zauner, P., & Haider, H. (2003). Want that is understood well before say that, think that, and false belief: A test of de Villiers s linguistic determinism on German speaking children. Child Development, 74 179 188. Porter, M. A, Coltheart, M., & Langdon R. (2008). Theory of mind in William s Syndrome assessed using a nonverbal task. Journal of Autism Developmental Disorders, 38, 806 814. mothers mental state language and th eory of mind understanding. Child Development, 73, 734 751 Ruffman, T., Slade, L., Rowlandson, K., Rumsey, C., & Garnham, A. (2003). How language relates to belief, desire, and emotion understanding. Cognitive Development, 18, 139 158. Schick, B., de Vill iers, P., de Villiers, J., & Hoffmeister, R. (2007). Language and Theory of Mind: A Study of Deaf Children. Child Development, 78 376 396. Semel, E, Wiig, E. H., & Secord, W. A. (2004). Clinical Evaluation of Language Fundamentals Preschool, Second E dition. San Antonio, TX: Pearson.
61 Simpson, A., & Riggs, K. J. (2005). Inhibitory and working memory demands of the day night task in children. British Journal of Developmental Psychology, 23, 471 486. Slade, L., & Ruffman, T. (2005). How language does (an d does not) relate to theory of mind: A longitudinal study of syntax, semantics, working memory and false belief. British Journal of Developmental Psychology, 23, 117 141. Tager Flusberg, H. (1994). The relationship between language and social cognition: L essons from autism. In Y. Levy (Ed.) Other child ren, other languages: Issues in the theory of language acquisition (359 381). Hillsdale, NJ: Erlbaum. Taumoepeau, M., & Ruffman T (2008). Stepping stones to others' minds: Maternal talk relates to child men tal state language and emotion understanding at 15, 24, and 33 months. Child Development, 79, 284 302. Tomasello, M. (1998). Uniquely primate, uniquely human. Developmental Science, 1 1 30. Vygotsky, L. (1986). Thought and Language. Cambridge: MIT Press. Wellman, H., Cross, D., & Watson, J. (2001). Meta analysis of theory of mind: the truth about false belief. Child Development, 72 655 684. Zelazo, P. D., Frye, D., & Rapus, T. (1996). An ag e related dissociation between knowing rules and using them. Cogni tive Development, 11, 37 63. Ziatas, K., Durkin, K., & Pratt, C. (2003). Differences in assertive speech acts produced by children with autism, Asperger syndrome, specific language impairment, and normal development. Development and Psychopathology, 15 73 94.
BIOGRAPHICAL SKETCH Mary Ellen McKay Easters (called Molly) was born in 1978 in Ridley Park, Pennsylvania the daughter of Daniel and Beverly McKay and younger sister of Andrew Molly and her family moved to Jacksonville, Florida in 1990. Molly atte nded Mandarin High School, graduating in 1996. Molly graduated with honors from the University of Florida with a Psychology major in May 2000. After graduation, Molly worked in the mental health care field with a focus on children and adolescents. After li ving and working in Lancaster, Pennsylvania, Denver, Colorado, and Jacksonville Florida Molly began graduate school in the Developmental Psychology program at the University of Florida in 2003 While in graduate school, Molly mar ried Michael Easters in 2 004, and the couple welcomed d aughter Jane in 2008.