These two items are intended to tap the understanding of false belief with specific regard to beliefs about an unexpected location (item 8) and unexpected contents (item 12). As the term implies, false belief (FB) understanding refers to the understanding that a person can hold a belief that contradicts reality. For example, in the case of an unexpected location, Patty believes that her keys are on the table when they are really in a drawer. That is, Patty has a mistaken or false belief about the location of an object. In the case of unexpected contents, Patty might be shown a cereal box and (quite reasonably but falsely) believe that the cereal box contains cereal when in fact, the cereal has been removed and replaced by cookies without her knowledge. Although there are myriad ways in which people can hold a mistaken belief, FBs about object locations and the contents of deceptive containers have dominated research on theory of mind.

The most well-known version of the FB task (alternately referred to as the ‘standard’ FB task, the ‘classic’ FB task, the ‘change location’ task, and the ‘Sally-Anne’ FB task) is the one originally used by Baron-Cohen, Leslie, and Frith (1985)[1] in their influential article “Does the autistic child have a theory of mind”? In this task, children were shown two doll protagonists and told a story in which an object is moved from an old location to a new location without the knowledge of the main protagonist. Specifically, Sally puts a marble in a basket and leaves the room. In her absence, Anne enters and moves the marble from the basket to a box and then Anne leaves. Children are asked, “When Sally returns, where will Sally look for her marble?”[2] Children who answer with the new (incorrect) location fail the question whereas children who answer with the old (correct) location pass the question by presumably demonstrating their knowledge that Sally has a false belief. Credit for the correct response is only given if children also pass two additional control questions. The first is “Where is the marble really?” (to ensure that the child understands the current actual location of the marble). The second question is “Where was the marble in the beginning?” (to ensure the child’s correct memory of the original location).

Over the last three decades, performance on the FB task has become a general marker or ‘litmus test’ for theory of mind (Carpendale & Lewis, 2006; Wellman, 1988). Yet, the importance of FB understanding has long been debated (e.g., Bloom & German, 2000; Russell, 2005; Wellman, Cross, & Watson, 2001) and “many questions still remain about what exactly children are doing when they pass false belief tests” (Lohmannn, Carpenter, & Call, 2005, p. 451). Indeed “we still know relatively little about which skills feed into false belief and which later skills it feeds into” (Carpendale & Lewis, 2006, p. 78; also see Perner, Leekam, Wimmer, 1987; Wellman et al., 2001). Consider the Sally-Anne task described above. To pass this task, the child needs to understand more than the concept of FB. With regard to theory of mind, the child also needs to appreciate the concept of ‘seeing-leads-to-knowing’ (item 9). That is, the child needs to understand that because Sally did not seeAnne move the marble, Sally will not know the marble’s new location. Another example involves agency or ‘intentionality’ (item 24): the child must understand that people act in accordance with their goals and that Sally wants her marble. Still another example involves counterfactual reasoning (item 29). “In order to acknowledge false belief, children must suppress or ignore what they know to be true of the world” (Riggs, Peterson, Robinson, & Mitchell, 1998, p. 74). In light of these considerations, it seems fair to conclude that the FB task is not ‘pure’ but rather confounded in that it typically taps multiple theory of mind domains.

With regard to the development of FB-understanding as measured by standard FB tasks, the overwhelming consensus across hundreds of studies is that typically developing 3-year-olds fail the task whereas 4-year-olds tend to pass the task (e.g., Perner et al., 1987; Wiessman, Friederici, Singer, & Steinbein). So one important question is: Why do 3-year-olds fail? As Bloom and German (2000) explained:

“Several investigators have modified the FB task to make it simpler – for instance, by making the questions simpler, more specific, and more pragmatically natural, making the change of location less salient, giving the children a memory aid for false belief content, and so on. Such modified FB tasks are often passed by 3-year-olds, a finding that has been used to support the argument that younger children have sophisticated conceptual competence when it comes to understanding that beliefs can be false, but that this competence is filtered through inefficient processing capacities” (p. B27).

Hence, it could be that younger children understand FBs but that their knowledge is masked by memory and language task demands. To further address this possibility, researchers have begun to explore the performance of infants on ‘implicit’ or ‘spontaneous’ FB tasks (i.e., nonverbal tasks that do not rely on eliciting any specific response such as pointing or answering questions[3]). Although spirited debate on the matter continues (with some arguing that infants show early conceptual competence in FB-understanding; e.g., Baillargeon, Scott, & He, 2010; Roth & Leslie, 1998; Scott, Richman, & Baillargeon, 2015), the majority of researchers have concluded that implicit and explicit FB-understanding are qualitatively distinct, independent, and under the control of different cognitive mechanisms (e.g., ApperlyButterfill, 2009; Low & Wang, 2011; Perner et al., 1987; Wellman et al., 2001; Weissmannet al., 2016)[4].

When it comes to FB-understanding as it has traditionally been assessed, recall that the consensus is that typically developing children pass the task between 4- and 5-years of age which provides good rationale for FB-understanding as a Basic theory of mind capacity. As alluded to above, FB performance is affected by the processing demands of the task (e.g., Lohmann et al., 2005). It is also positively correlated with level of executive functioning (e.g., Carlson, Claxton, & Moses), pragmatic language comprehension (Frank, Baron-Cohen, & Ganzel, 2015),and language ability. Regarding language, some researchers have proposed that the acquisition of specific linguistic elements is necessary for FB-understanding. For example, Bartsch and Wellman (1995) argued that theory of mind relies on acquiring a mental state vocabulary (e.g., words like ‘think’, ‘know’) which help children reflect on unobservable mental processes. In another line of research, de Villiers and colleagues (e.g., de Villiers and Pyers, 2002) have argued that the mastery of sentential complements is responsible for the acquisition of FB-understanding because this kind of syntactic structure (e.g., “He said he went to the store but he went to the beach”) provides a cognitive framework for thinking about conflicting mental representations. Contrariwise, other researchers have argued that it is not any specific language skill, but rather general language ability that is the primary driver of FB-understanding and the bulk of the evidence contrasting specific vs. general language accounts tends to favor the general language hypothesis (e.g., Cheung et al., 2004; Lohmann et al., 2005; Milligan, Astington, & Dack, 2007). Still other researchers take a different view and conclude that it is not any specific language skill (or general language for that matter) that is causally responsible for FB-understanding (and a theory of mind), but rather the simple fact that language allows interaction to happen among people (essentially a Vygotskian view; e.g., Astington, 1996; Tomasello, 2009). In light of the evidence for the various ways that language may contribute to FB-understanding, most researchers tend to take an ‘all of the above’ position and adopt the view that multiple aspects of language are causally and bi-directionally implicated.

Given the links between language and FB-understanding (and in line with the Vygotskian view of language socialization), it is not surprising that FB-understanding is also strongly associated with social experience and believed to be reliant upon social experience for its full expression (as are most aspects of theory of mind; for review, see Garfield, Peterson, & Perry, 2001). Indeed, FB-understanding is predicted by a variety of social-environmental factors: the age at which children demonstrate FB-understanding is negatively correlated with family income and parental education (Cole & Mitchell, 1998; Cutting & Dunn, 1999; Weimer & Guajardo, 2005) and positively correlated withsecure attachment in infancy (e.g., Symons & Clark, 2000), the child’s number of older siblings (Ruffman, Perner, Naito, Parkin, & Clements, 1998),and the amount and quality of early conversations about the mind (especially talk that is rich in mental state terms and focuses on the nature, causes, and consequences of mental states; e.g., “Which one do you like?”, “She thought that chickens could fly but now she knows they really can’t”, “How would you feel if someone said that to you?”; de Rosnay,Fink, Begeer, Slaughter, & Peterson,2014; RuffmanParkin, 2001; Ruffman, Slade, & Crowe, 2002). Finally, FB-understanding is correlated with a range of child cognitive and academic outcomes. For instance, it is associated with the frequency of children’s own mental state term productions (e.g., Hughes & Dunn, 1998), has been implicated as an important factor in children’s emerging math and literacy ability (Blair & Razza, 2007; Mason & Just, 2009),and has been argued to be causally-related to children’s social emotional maturity and social competence (de Rosnay et al., 2014; Lalonde & Chandler, 1995; Peterson, Slaughter, Moore, & Wellman, 2016).

In typical development, a summary of FB-understanding vis-à-vis the development of other theory of mind domains shows that:

  • The development of desire-understanding (item 59), pretense (item 26), the mental-physical distinction (item 16), seeing-leads-to-knowing (and not seeing leads to not knowing; item 9), and the understanding of true beliefs precede the understanding of FBs (e.g., Hogrefe, Wimmer, & Perner, 1986; Perner et al., 1987; Peterson, O’Reilly, & Wellman, 2016; Peterson, Wellman, & Slaughter, 2012).
  • FB-understanding emerges around the same time as a wide range of other basic level theory of mind competencies including the understanding of the appearance-reality distinction (item 11), level 2 visual-perspective-taking (item 21). and counter-factual thinking (item 29) (Gopnik & Astington, 1988; German & Nichols, 2003). This has been interpreted as reflecting the development of a more foundational understanding that the mind itself is representational (e.g., Gopnik & Astington, 1988; Perner, 1991).
  • Theory of mind competencies that emerge after the mastery of FBs include (but are not limited to), the understanding of biased cognition (item 40), the mind as an active interpreter (item 41), emotional display rules (item 17), mixed emotions (item 36), sarcasm (item 2), pragmatics and the metalinguistic aspects of language (see pragmatics subscale), and second-order understanding of beliefs (items 22 & 23) (e.g., Carpendale & Chandler, 1996; Carpenter & Lewis, 2006; Peterson et al., 2012, 2016). Indeed, there is much more to a mature theory of mind than the mastery of FBs.

FB understanding in ASD

Since Baron-Cohen et al.’s (1985) seminal article, hundreds of studies have documented deficits in FB-understanding in ASD. At the same time, the individual differences within studies are interesting and not all individuals with ASD fail standard tests of FB. As in typical development and ADHD, the ability of children with ASD to pass FB tasks is positively correlated with level of executive functioning (e.g., Pellicano, 2010; Pellicano, Mayberry, Durkin, & Maley, 2006),the production of cognitive mental state terms (e.g., “think”, “know”, “remember”) in children’s spontaneous speech (Tager-Flusberg, 1992; 2003), andlanguage ability. In fact,Happé(1995) found that children with ASD required a far higher verbal mental age to pass FB tasks than did other participants. While typically developing children had a 50% probability of passing the tasks at the verbal mental age of 4 years, those with ASD took more than twice as long to reach that probability of success (at the advanced verbal age of 9;2). This and other similar findings have led several researchers to propose that children with ASD with high intellectual and verbal skills may be using a compensatory (essentially logical and nonmentalistic) strategy to ‘hack out’ the correct solution to FB tasks (Bowler, 1992; Durrleman & Franck, 2015; Eisenmajer & Prior, 1991; Happé, 1995; Senju, 2012). This argument has recently received additional support from a study examining both standard FB task performance and implicit or ‘spontaneous’ FB-task performance. Senju, Southgate, White, and Frith (2009) found that although individuals with Asperger syndrome can pass the standard FB tasks, they nevertheless failed tasks designed to assess the spontaneous attribution of mental states. This notion gains importance when we consider that being able to compute the logic of mental states is no guarantee that one can or will apply the principles more broadly: that is, to real-world theory of mind dilemmas. As Senju (2012) argued:

“Unlike experiments, the real social world is fluid and rapidly changing. We have to process socially relevant information rapidly, spontaneously, and online in order to achieve day-to-day social interaction…[the notion of a compensatory process is] consistent with…the findings that training on FB tests does not necessarily improve social adaptation in ASD: the capacity for FB- attribution may not be sufficient to deal with its spontaneous use in a fluid and rapidly changing ‘real’ social world” (p.111).

In fact, while research has demonstrated that children with ASD can be taught to pass tests of FBs, generalization of skills to the real world has proven elusive (Hadwin,Baron-Cohen, Howlin, & Hill, 1996; Swettenham, 1996). One notable exception involves studies implementing the use of ‘thought bubbles’ to train FB-understanding which have demonstrated that training can transfer reliably to a variety of untrained contexts (e.g., Wellman et al., 2002). Thus, for many children with ASD, thought bubbles are a useful prosthesis for teaching not only FBs but a variety of mental processes. As Wellman et al. (2002) explained:

“Thought-bubbles arguably provide a particularly natural or effective way of depicting

thoughts pictorially, one that could come to aid autistic individuals’ reasoning about people, behavior, and mental states. In particular, thoughts are representational mental states and thought-bubbles depict a person’s thoughts in a straightforward representational fashion…Teaching an explicit compensatory strategy could help children with autism bypass their deficit in understanding mental states” (p. 346).

FB understanding in ADHD

Several studies have examined FB-understanding in individuals with ADHD. Most of these studies compare children with (or at risk for) ADHD with typically developing controls and conclude that there are no differences between groups when FB-understanding is assessed using the traditional (laboratory-type) Sally-Anne task (Charman, Carroll, & Sturge, 2001;HappéFrith,1996; Hutchins, Prelock, Morris, Benner, Lavigne, & Hoza, 2015; Perner, Kain, & Barchfeld,2002;Sodia,Hulsken,Thoermer, 2003; Yang, Zhou, Yao, Su, & McWhinnie, 2009).Bycontrast, meta-analyses and studies using applied measures of FB-understanding (as would be needed to solve FB problems in the real world) report that children with ADHD are impaired relative to their typically developing counterparts in FB-understanding and lie intermediate between ASD and typically developing samples (Bora & Pantelis, 2016; Hutchins et al., 2015).These patterns of performance, combined with the fact that executive function deficits occur in both conditions, may be responsible for the high degree of clinical overlap that is observed between ADHD and ASD (e.g., Gonzalez-Gadea et al., 2013); nevertheless, social incompetence in ASD and ADHD likely results from very different processes. First, while children with ADHD tend to pass standard FB-tasks but fail to apply this understanding in the world, children with ASD tend to fail in both contexts (Hutchins et al., 2015). This suggests a core conceptual deficit in ASD but a theory of mind performance-related deficit in ADHD. This is important in light of the robust finding that FB-understanding significantly correlates with executive functioning (Mary et al., 2015; Perner et al., 2002). What this means is that children with ADHD are likely to be unimpaired on tasks that do not place great demands on working memory, attention shifting, or inhibitory control (i.e., the Sally-Anne task) but significantly impaired in tasks/contexts that do require these cognitive resources (i.e., the real world) (Caillies, Bertot, Motte, Raynaud, & Abely, 2014; Hutchins et al., 2015; SodianHulsken, 2005)[5] and this conclusion fits with the finding that social cognitive deficits in ADHD improve with age (Bora & Pantelis, 2016). Although we are not aware of training studies involving FBs and children with ADHD, much research indirectly suggests that social cognition and social competence can be improved through training in theory of mind and executive function which has been documented as successful for other populations and which, in turn, has implications for improving academic achievement (for a review see Kloo & Perner, 2008).

FB understanding DoHH

Theory of mind “is one of the main research fields on cognitive development in deaf children today” (CourtinMelot, 2005, p. 16) and studies on the development of FB-understanding are quite numerous. The first study on this topic was conducted by Peterson and Siegal (1995) who evaluated the FB task performance of signing, prelingually-deaf children (ages 8 – 13 years) of hearing parents[6]. Results indicated that 65% of the children failed the FB task which typically developing children routinely pass at age 4 years and the performance of the DoHH children was no different from the performance of children with autism who had been tested in previous research. This study was important not only for demonstrating the potential for profound theory of mind deficits in oral and late-signing DoHH children but it was also theoretically influential for our understanding of theory of mind development. Once thought to be uniquely disrupted in autism, evidence of similar performance deficits in a sample of deaf children underscored the importance of language, social experience, and early family conversations about the mind for theory of mind development. Subsequent studies have largely confirmed and extended the findings of Peterson and Siegal (1995) and have concluded: