Second-order False Belief Tasks Research Articles

Towards Epistemic-Doxastic Planning with Observation and Revision

Epistemic planning is useful in situations where multiple agents have different knowledge and beliefs about the world, such as in robot-human interaction. One aspect that has been largely neglected in the literature is planning with observations in the presence of false beliefs. This is a particularly challenging problem because it requires belief revision. We introduce a simple specification language for reasoning about actions with knowledge and belief. We demonstrate our approach on well-known false-belief tasks such as the Sally-Anne Task and compare it to other action languages. Our logic leads to an epistemic planning formalism that is expressive enough to model second-order false-belief tasks, yet has the same computational complexity as classical planning.

Longitudinal relations between theory of mind and academic achievement among deaf and hard-of-hearing school-aged children

This 2-year longitudinal study investigated the bidirectional relations between the development of theory of mind (ToM) and academic competences in a sample of 270 deaf and hard-of-hearing (DHH) children (Mage at Wave 1 = 7.52 years, SD = 0.99; 58.5% boys and 41.5% girls). Across three waves (10 months apart), children were assessed for their ToM abilities, using the ToM scale and a second-order false belief task, as well as for their language and mathematics skills. Cross-lagged correlational analysis revealed significant bidirectional associations between ToM and academic achievement (language and mathematics). That is, ToM predicted academic achievement with similar strength as ToM development itself was predicted by academic achievement. Our results highlight the bidirectional nature of the association between ToM and academic achievement, and they show that ToM development plays a crucial role in DHH children’s school functioning.

Development and validation of a task battery for verbal and non-verbal first- and second-order theory of mind

This paper presents a new toolkit for assessing Theory of Mind (ToM) via performance in first and second-order false belief (FB) tasks. The toolkit includes verbal and non-verbal versions of first and second-order FB tasks; the verbal version is currently available in Greek and German. Scenarios in the toolkit are balanced for factors that may influence performance, like the reason for the FB (deception, change-of-location, unexpected content). To validate our toolkit, we tested the performance of neurotypical adults in the non-verbal and verbal versions in two studies: Study 1 with 50 native speakers of German and Study 2 with 50 native speakers of Greek. The data from both studies yield similar results. Participants performed well in all conditions, showing slightly more difficulties in the second- than first-order FB conditions, and in the non-verbal than the verbal version of the task. This suggests that the task is at the high end of the sensitive range for neurotypical adults, and is expected to be well inside the sensitive range for children and populations that have difficulties in ToM. Factors like deception and type of outcome in the video-scenarios did not influence the behavior of neurotypical adults, suggesting that the task does not have any confounds related to these factors. The order of presentation of the verbal and non-verbal version has an influence on performance; participants beginning with the verbal version performed slightly better than participants beginning with the non-verbal version. This suggests that neurotypical adults used language to mediate ToM performance and learn from a language-mediated task when performing a non-verbal ToM task. To conclude, our results show that the scenarios in the toolkit are of comparable difficulty and can be combined freely to match demands in future research with neurotypical children and autistic individuals, as well as other populations that have been shown to have difficulties in ToM. Differences between baseline and critical conditions can be assumed to reflect ToM abilities, rather than language and task-based confounding factors.

Preschoolers’ ingroup bias in predicting others’ sharing: The role of contexts and theory of mind

The current study investigated preschoolers’ ingroup bias in predicting people’s sharing across contexts and its relation to second-order theory of mind (ToM) ability. In Experiment 1, 96 5- and 6-year-old children were assigned to one of two groups in a minimal group paradigm. They heard a story about fictional ingroup and outgroup peers sharing in a public or private condition and were asked to predict and evaluate their sharing behavior. Children predicted that ingroup peers would share more than outgroup peers and also showed ingroup bias in evaluation regardless of the equal actual sharing of ingroup and outgroup peers. Moreover, 6-year-olds displayed a flexible ingroup bias in predicting others’ sharing across conditions because they held such a bias only in public conditions and did not expect ingroup and outgroup peers to share differently in private conditions. Experiment 2 tested a new sample of 80 6-year-olds with the same sharing story and a second-order false belief task. Results showed that only 6-year-olds who fully passed the false belief task showed a flexible bias in predicting sharing across conditions. Results indicate that children’s ingroup bias in predicting others’ sharing is becoming flexible across contexts as they grow up and ToM skills contribute to the development of their increasingly sophisticated prosocial reasoning.

Advanced theory of mind in children with mild intellectual disability and deaf or hard of hearing children: A two-year longitudinal study in middle childhood.

The present study investigates the development of advanced theory of mind (AToM) among typically developing (TD) children, children with mild intellectual disability (MID), and deaf or hard of hearing (DHH) children. The 2-year longitudinal study comprised three waves and included a large sample of children from Poland in middle childhood aged around 7.5-9.5years (N=779; M=7.7, SD=0.92 at wave 1). The analysis of children's understanding of second-order false belief and the Faux-Pas Recognition Test showed that TD children outperformed children with MID and DHH children on both measures. At 7.5years, almost 60% of the TD children correctly solved the second-order false belief task; correct performance at 7.5years in children with MID and DHH children was 27 and 38% respectively. Two years later, correct performance rose to 80% (TD children), 45% (children with MID), and 63% (DHH children). Despite these differences, the speed of AToM development did not differ across the groups. The development of faux-pas recognition followed a non-linear pattern, with TD children showing no further significant development after mid-elementary school. Our findings show differences in AToM development between TD children, children with MID, and DHH children, and they suggest that children's development of AToM may follow different developmental pathways, depending on the aspect of AToM under study.

Conscience and theory of mind in children aged 4 to 7 years

Associations between three dimensions of early conscience—moral reasoning, the capacity to experience guilt, and the moral self—and theory of mind (ToM) were examined in children aged 4–7 years (N = 80). Participants were administered a task assessing their understanding of the intentions and actions of a transgressor in situations entailing intentional and accidental wrongdoing, a moral self scale, and a battery of first-order and second-order false belief tasks. Children’s capacity to experience guilt was measured via parent report. Expressive vocabulary was also measured. Repeated-measures analysis of covariance with ToM, age, and their interaction as covariates revealed that children who had higher ToM scores attributed more positive intentions to the accidental transgressor than to the intentional transgressor and judged the intentional transgressor’s action as more wrongful than children who scored lower on these tasks. Ηierarchical regression analyses also indicated that a more advanced ToM performance predicted higher levels of guilt and the moral self after accounting for age and expressive vocabulary.

Children’s understanding of unintended irony and unsuccessful irony

ABSTRACT This study examined children’s understanding of irony when a difference existed between a speaker’s intended meaning and a listener’s interpretation of the meaning. Three irony contexts were presented to 87 7/8-year-olds (second graders), 90 11/12-year-olds (sixth graders), and 103 adults. In the normal irony context, the speaker intended to convey irony, and the listener felt that he/she was being criticized by the speaker. In the unintended irony context, the speaker intended to convey praise, but the listener felt that he/she was being criticized by the speaker. In the unsuccessful irony context, the speaker intended to convey irony, but the listener felt that he/she was being praised by the speaker. The results showed that 11/12-year-olds and adults understood all three types of irony and appropriately judged whether or not the speaker’s intention was ironic and whether or not the listener thought that he/she was being criticized by the speaker. However, 7/8-year-olds did not understand the context differences, and across all types of irony they consistently judged that the speaker intended to praise the listener and that the listener thought that he/she was being praised. Furthermore, performance was generally correlated with passing second-order false belief tasks, especially with unintended and unsuccessful irony.

Gaining or losing wisdom: Developmental trends in theory of mind in old age

Theory of mind (ToM) is a tool in the field of social cognition that is employed to judge and predict others’ intentions, motives, and psychological states. At present, studies on the development of ToM in older adults have offered some meaningful insights. However, researchers still lack an overall understanding of the differences of ToM abilities based on different age brackets. Accordingly, the present study utilized a variety of tasks to investigate the development of ToM in the elderly at different ages. The results demonstrated that: (1) The ToM abilities showed a declined tendency with the increasing age. (2) While overall, different types of ToM tasks varied in accordance with each age brackets. In the double emotion task, the development of ToM abilities showed no difference among three groups of people aged 65–74, 75–79, 80–84, while in comparison, the people aged 85–89 had a notable decline in ToM abilities. In the double bluff task, the ToM abilities of people aged 65–74 were notably stronger than that of the people aged 85–89, while the other groups did not varied much. In the false belief task, ToM abilities of people aged 65–74 were significantly stronger than that of people aged 65–74 and 85–89, while the other groups did not varied much. In the second-order false belief task, the ToM abilities of people aged 65–74 and 75–79 were greatly stronger than that of other two groups that had similar results. In the faux pas task, the ToM abilities of people aged 65–74 were notably stronger than that of other three groups, people aged 75–79 than that of 85–89, while other groups showed no difference. In general, the ToM abilities of the elderly displayed the inner consistency of development, but manifested particularity in different tasks.

Being Deceived: Information Asymmetry in Second-Order False Belief Tasks.

We analyze four well-known second-order false belief tasks. Superficially, all four tasks share a common logical structure: All are based around a principle of inertia, which says that an agent's beliefs are preserved over time, unless the agent receives information to the contrary. However, a deeper analysis reveals details that are both suggestive and puzzling. First, the four tasks exemplify all four possibilities inherent in the two dimensions of being-deceived versus not-being-deceived and change-in-world versus change-in-belief-only. Second, there is a feature common to all four tasks: All come with a "built in" first-order false belief. We call these inner first-order false beliefs. They introduce an informational asymmetry that has the same logical form in all four tasks, but whose role is unclear. We do two things in this paper. First, we show that inner first-order false beliefs play an important (though seemingly unremarked) role in the experimental design of the tasks. Second, we present some empirical results (for both typically developing children and children with Autism Spectrum Disorder) on the effects of being-deceived versus not-being-deceived and change-in-world versus change-in-belief-only on second-order reasoning ability.

Predictors of Theory of Mind performance in bilingual and monolingual children

Aims and objectives/purpose/research questions: The goal of the current study was to examine whether language and Executive Function (EF) skills predict Theory of Mind (ToM) performance in bilingual and monolingual children. Design/methodology/approach: Participants included 44 monolingual English-speaking children ( MAge = 7.03, SD = 1.23), 44 simultaneous English-Spanish bilingual children ( MAge = 7.36, SD = 1.18), and 27 English first language (L1) bilinguals (native English speakers learning Spanish through dual immersion programs; MAge = 7.58, SD = 1.24). In the second-order false belief task, children were asked to reason about one character’s thoughts about another character’s thoughts. Three components of EF were measured: inhibition (via a flanker task); updating of working memory (via a Corsi blocks task); and shifting (via a card sort task). The Clinical Evaluation of Language Fundamentals 4th Edition indexed omnibus English language skills. Data and analysis: Separate logistic regression models were constructed to analyze the predictors of ToM performance for each group controlling for age, socioeconomic status, and non-verbal IQ. Findings/conclusions: Language ability, but not EF skills, predicted ToM performance in simultaneous bilinguals. In contrast, EF skills but not language predicted ToM performance in monolinguals and in English L1 bilinguals. Originality: This is the first study to show that children with distinct language acquisition histories approach verbal ToM problems differently. Significance/implications: It appears that the linguistic demands of the ToM task take priority for children with lower levels of target-language ability, while the cognitive demands of the task take priority for children with higher levels of target-language ability.

The Relationship of Clinical Symptoms with Social Cognition in Children Diagnosed with Attention Deficit Hyperactivity Disorder, Specific Learning Disorder or Autism Spectrum Disorder.

Objective One of the areas of social cognition is Theory of Mind (ToM) is defined as the capacity to interpret, infer and explain mental states underlying the behavior of others. When social cognition studies on neurodevelopmental disorders are examined, it can be seen that this skill has not been studied sufficiently in children with Specific Learning Disorder (SLD). Methods In this study, social cognition skills in children diagnosed with attention deficit hyperactivity disorder (ADHD), SLD or Autism Spectrum Disorder (ASD) evaluated before puberty and compared with controls. To evaluate the ToM skills, the first and secondorder false belief tasks, the Hinting Task, the Faux Pas Test and the Reading the Mind in the Eyes Task were used. Results We found that children with neurodevelopmental disorders as ADHD, ASD, and SLD had ToM deficits independent of intelligence and language development. There was a significant correlation between social cognition deficits and problems experienced in many areas such as social communication and interaction, attention, behavior, and learning. Conclusion Social cognition is an important area of impairment in SLD and there is a strong relationship between clinical symptoms and impaired functionality.

Accelerating the Development of Second-Order False Belief Reasoning: A Training Study With Different Feedback Methods.

One-hundred-six 5-year-olds' (Mage =5;6; SD=0.40) were trained with second-order false belief tasks in one of the following conditions: (a) feedback with explanation; (b) feedback without explanation; (c) no feedback; (d) active control. The results showed that there were significant improvements in children's scores from pretest to posttest in the three experimental conditions even when children's age, verbal abilities, or working memory scores were controlled for. The training effect was stable at a follow-up session 4months after the pretest. Overall, our results suggest that 5-year-olds' failures in second-order false belief tasks are due to lack of experience and that they can be helped over the threshold by exposure to many stories involving second-order false belief reasoning, including why questions.

Irony comprehension and mentalizing ability in children with and without Autism Spectrum Disorder

BackgroundDespite evidence suggesting that individuals with Autism Spectrum Disorder (ASD) may have difficulties in comprehension of figurative language, and irony in particular, previous studies examining this ability among individuals with (ASD) have reported inconsistent findings, resulting in different suggested etiologies of this difficulty. In view of the inconclusive findings, the current study assesses the contribution of various factors to irony comprehension, with specific focus on the association between mentalizing ability and irony comprehension. MethodIrony comprehension was examined in 20 individuals with ASD (age range 10–15) as compared to 20 typically developing (TD) peers (age range 10–15) through a task of reading ten ironic comics. Participants were matched by age, gender, vocabulary, executive function abilities, and their results on a second-order false-belief task. Their mentalizing abilities were examined by the Hinting Test for the ability to understand intentions. ResultsA significant difference in irony comprehension was observed between the groups. Re-adjusting the group matching according to the Hinting Test scores eliminated the group difference in irony comprehension. Multilevel regression with logit link analysis showed that hinting and mental flexibility contributed to irony comprehension. ConclusionsThe study's findings demonstrated that individuals with ASD showed adequate ability to comprehend irony but nonetheless were outperformed by TD peers (matched on age, language, ToM, and executive functioning abilities). A comparison of the two groups by their mentalizing abilities (through understanding intentions) revealed similar comprehension abilities. These findings highlight the importance of using several tools that each focus on different aspects of mentalizing when assessing this skill in studies of figurative language in ASD.

Theory of Mind Development in School-Aged Left-Behind Children in Rural China.

The current study aimed to investigate differences in theory of mind between left-behind children and non-left-behind children in rural China and to examine the potential protective role of general reasoning ability in left-behind children’s theory of mind. Participants included 213 children aged 7.10–13.67 years (111 boys and 102 girls, M = 10.51 years, SD = 1.33), 101 of whom were left behind in rural areas by one or both migrating parents for at least 6 months. The Strange Stories task, a second-order false belief task, and a faux pas task were used to measure children’s theory of mind, and Sessions B and C in Raven’s Standard Progressive Matrices were used to test children’s general reasoning ability. The results showed that left-behind children scored lower on both the faux pas task and Strange Stories task. Additionally, on second-order false belief understanding, left-behind boys performed worse than non-left-behind boys, while left-behind girls scored higher than non-left behind girls. Moreover, children’s general reasoning ability moderated the relationship between parental migrant status and children’s faux pas understanding: For children with high levels of general reasoning ability, left-behind children performed similarly to non-left-behind children, while for children with low levels of general reasoning ability, left-behind children scored lower than non-left-behind children, indicating that general reasoning ability buffered the negative effect of being left behind on children’s theory of mind development. The implications of these findings for training directed at left-behind children are discussed.

Prior Knowledge, Episodic Control and Theory of Mind in Autism: Toward an Integrative Account of Social Cognition.

Over the last 30 years, research has explored theory of mind (ToM), the ability to attribute mental states to oneself and to others. Work on ToM in typical and atypical populations has shed light on the neurocognitive mechanisms underlying social understanding and interaction. The ToM hypothesis has long been regarded as one comprehensive explanation of the severe cognitive and behavioral impairments encountered by individuals with autism. However, high-functioning individuals can often pass both first-order and second-order false belief tasks using cognitive compensation strategies. To provide more sensitive measures of mental state attribution abilities, researchers have introduced more difficult, “advanced” theory of mind tasks. In this article, we argue that in attempting to bypass compensation strategies, these new advanced ToM tasks, such as the Faux Pas and the Strange Stories tasks, impose cognitive demands beyond those specific to the domain of ToM. We then provide an integrative account of social deficits in autism that takes into account several distinct components of mental state understanding, including both general cognitive capacities and processes specific to ToM. We argue that a number of related cognitive abilities, including episodic cognitive control and inferencing from prior knowledge, are necessary to understand how both people with autism and typical development navigate challenging, real-life social situations.

Does a good argument make a good answer? Argumentative reconstruction of children's justifications in a second order false belief task

Abstract This paper proposes a novel approach to interpret the results of a classical second-order false belief task (the ice cream man task) administered to children in order to investigate their Theory of Mind. We adopted a dialogical perspective to study the adult-child discussion in this research setting. In particular, we see the adult-child conversation as an argumentative discussion in which children are asked to justify their answers to the questions asked by the researcher. We analysed the specificities of the research setting as an argumentative activity type; we reconstructed and analysed the children's answers on the basis of two models taken from Argumentation theory (the pragma-dialectical model and the Argumentum Model of Topics). Our findings show that some of the children's partially “incorrect” answers depend on the pragmatics of the conversation, the relation between explicit and implicit content, and a misunderstanding of the discussion issue. Other “incorrect” answers are actually based on correct inferences but they do not meet the researchers' expectations, because the children do not share the same material premises as the researchers. These findings invite further research on children's reasoning and on the characteristics of argumentation within a research task.

Impaired theory of mind in Chinese children and adolescents with idiopathic generalized epilepsy: Association with behavioral manifestations of executive dysfunction

BackgroundEpilepsy is a common neurological disorder with a core feature of cognitive impairments. Previous studies showed that patients with focal epilepsy have deficits in both theory of mind (ToM) and executive function (EF). However, there are few studies of ToM in patients with idiopathic generalized epilepsy (IGE), especially in populations with pediatric epilepsy. The aim of this study was to examine the characteristics of ToM and EF, including some of their subcomponents, and explore the relationship between them in Chinese children with IGE. Materials and methodsWe recruited 54 children and adolescents with IGE as the experimental subjects and 37 typically developing children and adolescents as control subjects. Both groups completed ToM tests, namely, second-order false belief tasks (FBTs) and faux pas tasks (FPTs). Their caregivers completed the Behavior Rating Inventory of Executive Function (BRIEF) at the same time. ResultsChildren and adolescents with IGE displayed worse performance on some of the FBTs and FPTs than healthy controls (p<0.01). They also exhibited widespread EF deficits, comprising eight subcomponents (p<0.05). Pearson's correlation analysis revealed that several subcomponents of EF (inhibition, emotional control, initiation, working memory, and monitoring) were unequally correlated with FBT and FPT. Regression analysis showed that ToM had associations with inhibition, working memory, and duration of seizures. Analysis of variance (ANOVA) indicated that children with newly diagnosed epilepsy displayed significant deficits in FBT, FPT, and distinct subscales of EF. ConclusionsOur results revealed significant impairments in ToM and EF in children and adolescents with IGE compared with healthy controls. We found significant correlations between ToM and two subcomponents of EF (inhibition and working memory) in children with IGE. Additionally, the duration of seizures affected ToM in patients but was a less powerful predictor than the two subcomponents of EF. Even for children with new-onset seizures and without medication, the deficits in ToM and some distinct subscales of EF were apparent. This result has clinical implications for both nonpharmaceutical therapies and cognitive rehabilitation.

Reliability and validity of advanced theory-of-mind measures in middle childhood and adolescence.

Although theory-of-mind (ToM) development is well documented for early childhood, there is increasing research investigating changes in ToM reasoning in middle childhood and adolescence. However, the psychometric properties of most advanced ToM measures for use with older children and adolescents have not been firmly established. We report on the reliability and validity of widely used, conventional measures of advanced ToM with this age group. Notable issues with both reliability and validity of several of the measures were evident in the findings. With regard to construct validity, results do not reveal a clear empirical commonality between tasks, and, after accounting for comprehension, developmental trends were evident in only one of the tasks investigated. Statement of contribution What is already known on this subject? Second-order false belief tasks have acceptable internal consistency. The Eyes Test has poor internal consistency. Validity of advanced theory-of-mind tasks is often based on the ability to distinguish clinical from typical groups. What does this study add? This study examines internal consistency across six widely used advanced theory-of-mind tasks. It investigates validity of tasks based on comprehension of items by typically developing individuals. It further assesses construct validity, or commonality between tasks.

Testing the validity of a continuous false belief task in 3- to 7-year-old children

In two studies, we examined young children’s performance on the paper-and-pencil version of the Sandbox task, a continuous measure of false belief, and its relations with other false belief and inhibition tasks. In Study 1, 96 children aged 3 to 7years completed three false belief tasks (Sandbox, Unexpected Contents, and Appearance/Reality) and two inhibition tasks (Head–Shoulders–Knees–Toes and Grass/Snow). Results revealed that false belief bias—a measure of egocentrism—on the Sandbox task correlated with age but not with the Unexpected Contents or Appearance/Reality task or with measures of inhibition after controlling for age. In Study 2, 90 3- to 7-year-olds completed five false belief tasks (Sandbox, Unexpected Contents, Appearance/Reality, Change of Location, and a second-order false belief task), two inhibition tasks (Simon Says and Grass/Snow), and a receptive vocabulary task (Peabody Picture Vocabulary Test). Results showed that false belief bias on the Sandbox task correlated negatively with age and with the Change of Location task but not with the other false belief or inhibition tasks after controlling for age and receptive vocabulary. The Sandbox task shows promise as an age-sensitive measure of false belief performance during early childhood and shows convergent and discriminant validity.

Theory of Mind in Euthymic Bipolar Patients and First-Degree Relatives.

The present study analyzed the capacity for mentalization of patients with bipolar disorder (BD) and their first-degree relatives (FDR) and examined the implications of clinical variables and cognitive deficits. The study recruited 31 patients with type I BD, 18 FDR, and 31 paired healthy controls. Their capacity for mentalization was explored by means of first- and second-order false-belief tasks, the hinting task, and the Movie for the Assessment of Social Cognition (MASC). Patients and FDR were found to have a theory of mind (ToM) deficit when they were evaluated with the MASC, which was also related to a worse neurocognitive performance and to being a patient or FDR. The evidence of ToM deficits in FDRs supports the hypothesis that these deficits could be an independent trait marker for cognitive deficit. Further research is needed on FDR of patients with BD, using sensitive ToM assessment instruments such as the MASC.