You are on page 1of 6


Developing the Frith-Happe´ Animations: A Quick and Objective Test
of Theory of Mind for Adults with Autism
Sarah J. White, Devorah Coniston, Rosannagh Rogers, and Uta Frith
It is now widely accepted that individuals with autism have a Theory of Mind (ToM) or mentalizing deficit. This has
traditionally been assessed with false-belief tasks and, more recently, with silent geometric animations, an on-line ToM
task. In adults with milder forms of autism standard false-belief tests, originally devised for children, often prove
insensitive, while the Frith-Happe´ animations have had rather better success at capturing the on-line ToM deficit in this
population. However, analysis of participants’ verbal descriptions of these animations, which span scenarios from
‘‘Random’’ to ‘‘Goal-Directed’’ and ‘‘ToM,’’ is time consuming and subjective. In this study, we developed and established
the feasibility of an objective method of response through a series of multiple-choice questions. Sixteen adults with
autism and 15 typically developing adults took part, matched for age and intelligence. The adults with autism were less
accurate as a group at categorizing the Frith-Happe´ animations by the presence or absence of mental and physical
interactions. Furthermore, they were less able to select the correct emotions that are typically attributed to the triangles
in the mental state animations. This new objective method for assessing the understanding of the animations succeeded
in being as sensitive as the original subjective method in detecting the mentalizing difficulties in autism, as well as being
quicker and easier to administer and analyze. Autism Res 2011, 4: 149–154. & 2011 International Society for Autism
Research, Wiley Periodicals, Inc.
Keywords: autism; Theory of Mind; mentalizing; triangle animations; classification; test sensitivity

It is now widely accepted that a cognitive deficit in
Theory of Mind (ToM) or mentalizing can explain a
number of core social impairments in autism; these
individuals have difficulty in the intuitive understanding
of others’, as well as their own, mental states. This ToM
deficit was originally detected using false-belief tasks,
such as the Sally-Ann test [Baron-Cohen, Leslie, & Frith,
1985], but such tasks have proved insensitive with older
and more able individuals. A number of other methods
for assessing mentalizing ability have therefore been
devised, which are suitable for use with adults who have
only mild forms of autism. One such method uses silent
animations of geometric shapes moving around the
screen, based on a study by Heider and Simmel [1944]
in which it was established that people often attribute
human character traits to shapes in motion, and sometimes even attribute mental states.
Heider and Simmel’s original animations were never
intended to signify specific mental states and the
sequence of events could be sufficiently described in
terms of a goal-directed (GD) action. The Frith-Happe´
animations therefore consist of three novel sets of
geometric animations representing three different types
of motion: random, GD, and ToM. The ‘‘random’’ set
is intended to depict purposeless movement and no

interaction between two triangles, ‘‘GD’’ animations are
intended to depict an interaction consequent upon the
physical action or behavior of one of the triangles, while
the ‘‘ToM’’ animations are intended to depict an interaction involving one triangle reacting to the other’s
mental state. In the first experiment, participants
watched the animations and afterwards they were asked
for a verbal description of what they thought had
happened [Abell, Happe´, & Frith, 2000]. Both children
and adults were found to be predominantly correct in
their use of action terms for the random animations,
interaction terms for the GD animations, and mentalizing terms for the ToM animations. However, children
with autism gave less appropriate mentalizing descriptions and even children with autism who had successfully passed both first and second-order false-belief tasks
were unable to appropriately describe the ToM animations. These animations have since successfully been used
in a growing number of experiments to investigate ToM
and agency attribution in individuals with and without
autism [Bird, Castelli, Malik, Frith, & Husain, 2004;
Campbell et al., 2006; Castelli, Happe´, Frith, & Frith,
2000; Castelli, Frith, Happe´, & Frith, 2002; Corden,
Critchley, Skuse, & Dolan, 2006; Kana, Keller, Cherkassky,
Minshew, & Just, 2009; Klein, Zwickel, Prinz, & Frith,

From the Institute of Cognitive Neuroscience, University College London, London, United Kingdom (S.J.W., D.C., R.R., U.F.)
Received March 29, 2010; accepted for publication November 12, 2010
Address for correspondence and reprints: Sarah J. White, Institute of Cognitive Neuroscience, University College London, 17 Queen Square, London
WC1N 3AR, UK. E-mail:
Grant sponsors: Medical Research Council, and Economic and Social Research Council; British Academy; Grant numbers: PTA-037-27-0107; PDF/2009/213.
Published online 5 January 2011 in Wiley Online Library (
DOI: 10.1002/aur.174
& 2011 International Society for Autism Research, Wiley Periodicals, Inc.


Autism Research 4: 149–154, 2011


which was felt to be more obviously evident in their daily lives. Boulanger.11).86) 114. Leslie. divided into three sets of animations: random.’s [2000] study were the stimuli.13 (11. Stuss. 1985. & Berry. Each set contained four animations.65). Asperger syndrome. Po0.. Klin & Jones. Frith. 1995]. Procedure Testing took place individually in a quiet room at the Institute of Cognitive Neuroscience.. 2000. The aim was to develop an objective method that was as sensitive as the subjective method in detecting ToM deficits at a group level in autism.47) 14 AS 1 HFA 1 ASD 7 Autism 7 Autism Spectrum 2 None 15 (11:4) 36. Method Participants Ethical approval for the study was received from the Joint UCL/UCLH committee on the Ethics of Human Research and. HFA. WAIS III-UK./Objectively testing theory of mind in autism INSAR . as Skuse. 2010. 1993. matched for age (t(29) 5 0.. Autism Spectrum Disorder. Richir. Levine. one of these participants met criteria in the social domain but had a physical disability that made it impossible to score the two 150 gesture items in the communication domain.00 (10. The two remaining participants were not excluded as their ADOS scores revealed a degree of communication and social difficulty. Asterisks indicate when groups significantly differ from one another. Skinner. Seigal.2009. The participants were then instructed that. while being quicker and easier to administer and analyze by providing participants with multiple-choice questions (MCQ) about each animation. Autism-Spectrum Quotient. Materials The 12 animations originally used in Abell et al. A battery of standard ToM tasks was administered to assess each participant’s explicit mentalizing ability independently from the animations.84) – – – – – – Note. high-functioning autism. Lawrence. 1985. written informed consent was sought and received from all participants. while watching the animations. Matsuda. GD. Klin. & Leekham. verbal IQ (t(29) 5 0. Oatley & Yuill. they would be asked to categorize it as one of three types: ‘‘no interaction’’ (random). ASD. Po0. UCL. Luckett. Claxton. the responses to the geometric animations were analyzed with a subjective method of scoring: participants gave a verbal description after each animation.44 (12. Martin. & Schulz. Furthermore. Sixteen adults with Autism Spectrum Disorder (ASD) and 15 typically developing adults took part. AQ scores for the ASD group were significantly greater than for the comparison group (t(28) 5 5.74) 106. which was scored by several markers who were highly familiar with the animations and scoring system. & Corcoran. 1997]. 2004] and two secondorder false-belief tasks [Bowler. & Clubley. This is an extremely time-consuming process.64 (7. 1985.97). 2006. and ‘‘mental interaction’’ (ToM) (referred to hereon as MCQ-categorization).google.71) 19. As expected. 2007. AS. removal of these participants’ data made no difference to the results from the animations.53 (9. 1996]. 2006. this study focuses on devising an objective administration and scoring procedure for use in experimental studies involving the geometric animations. Montgomery & Montgomery.75) 37.82. indeed they both had an Autism-Spectrum Quotient score [AQ. See http://sites.. Salter. Rime´. 2000].001. Lawrence et al. Springer.. before the test session. Perner. Lord et al. Koelkebeck et al.97) 110. Zwickel & Mu have a number of other sets of similar animations [Berry & Springer. 1989. 2002. 2001] above the recommended cut-off of 32. Moriguchi. 2007.47 (13. 1992. Thornton.. Baron-Cohen.31) 111. they were to give a concurrent verbal description of what they thought was happening.25 (12. & Stroobants. research for examples of the stimuli. and ToM. see Table I). Values are given as mean (SD) unless otherwise stated. 1985] (total score of 7) as well as Happe´’s mental state Strange Stories [Fletcher et al. which is subjected to bias and variability between markers. Knickmeyer. 2000. Moriguchi et al. Bowler & Thommen. Herba. Wheelwright. Raggatt. Taylor. 2008. ‘‘physical interaction’’ (GD). and performance IQ [t(29) 5 0. Russell. & Hackett. 1999. Zwickel. Ohnishi.85. The following definitions of these three types were provided during practice.06 (8. At the end of each animation. The participants in the ASD group had all received previous diagnosis from a qualified clinician and all but two also met criteria for an ASD on the Autism Diagnostic Observation Schedule [ADOS-G. Perner & Wimmer. 2009. 2006. Mori. Wechsler Adult Intelligence Scale. Laubin. & Tulving. Table I. Motivated by the limitations of this subjective method of scoring. Reynaud. & ¨ ller. These included five first-order [Baron-Cohen et al. AQ. 2006. In all the above experiments. & Komaki. Rosenbaum. Meier. Participant Characteristics N (male:female) Age (yr) Verbal IQ Performance IQ AQ (cut-off 32) Clinical diagnosis ADOS ASD Comparison 16 (12:4) 33. 2007. Powell. Morris. Baron-Cohen. Messer.001. Wellman & Liu. gender (w2(1) 5 0. White et al. 2009]. In addition.

76% of responses rated as identical or 1 point apart on a six-point scale) and therefore the average of the two markers’ scores was taken.27 (1. For the subjective scoring.60) 4. Po0.036) and Happe´’s mental state Strange Stories (t(24) 5 3.30) 14.98) Analysis Background False-belief tasks (7) Strange stories (16) Objective MCQ-categorisation total (12) MCQ-feelings (8) Subjective Intentionality Random (5) Goal-directed (5) ToM (5) Appropriateness Random (5) Goal-directed (5) ToM (5) Given the low range of possible scores on all measures of the animations. An alpha level of 0. Figure 1 and Table II reveal that the ASD group’s performance was consistently worse for the MCQ-categorization compared to the comparison group across all 12 animations (U 5 198. Feedback was given on the practice trials.62 (0.45 (0. two independent markers (one blind to group. both on a six-point scale [see Zwickel.47 (1. Subjective scoring.5. no feedback was given.001.93) 5. Po0.00) 10. Participants could score a total of 12 for the MCQ-categorization.66) 2. However. participants could choose between one of five listed adjectives to best match the feelings of each of the triangles at the end of the animation. Scores for the ASD group were significantly worse on both the false-belief tasks (t(17.66 (0.48 (0. White. P 5 0. 2010 for further details].39) 3. Values are given as mean (SD) unless otherwise stated.71 (0. and Subjective Tasks Task (maximum score) ASD Comparison 6. Agreement between the markers was high for intentionality (Intraclass Correlation Coefficient (ICC) 5 0. P-values are also reported for results between 0. P 5 0.04 (0.91 (0. no additional questions were asked. background ToM tasks.05 and 0.85) 6. Po0.66) 2.57) 2.* No interaction: There is no obvious interaction between the triangles and movement appears random. P 5 0. this group difference was only significant for the ToM animations (U 5 59. P 5 0.30) 0.009). when looking at the three types of animation separately. If they correctly identified a ToM animation.5. In the experimental phase.26 (0.07 (1.51 (0. For the objective scoring. The verbal descriptions were analyzed for appropriateness and intentionality. using a similar scoring system to Castelli et al.70. In all other conditions. a maximum of 8 points could be scored.47 (0.00 (0. Test Scores for Background. corresponding to two possible correct answers for each of the four ToM animations. This was divided into a maximum of 4 for each of the three animation types. Po0.001).59) 3.3. * Mental interaction: An interaction between the triangles involving the manipulation of the emotions and thoughts of one triangle by the other. For the MCQ-feelings task.95) 2. This involved scoring the use of mental state terms (intentionality score) and the correct inference of the underlying scenarios (appropriateness score). whereas parametric tests were used for the Note. Po0.17 (1. * Physical interaction: An interaction between the triangles in which actions are directed toward each other in order to achieve specific goals.05 was selected. participants were shown one animation at a time in a pseudo-randomized order.05.54) 3. Objective.001.83. & Frith.24 (0. the other DC) rated the participants’ verbal descriptions of the animations for intentionality and appropriateness. Throughout the experimental phase. they were then presented with additional questions (MCQ-feelings) to test their understanding of the mental states depicted in the animation. There was no significant difference between the two groups on the intentionality scores for both the Table II.1) 5 2. 96% of responses rated as identical or 1 point apart on a 6 point scale) and for appropriateness (ICC 5 0.38 (2.01. see Table II).1.02 (0.10.27) 7.21 (0. The verbal descriptions that participants gave were recorded for later transcription and scoring. Asterisks indicate when groups significantly differ from one another. INSAR White et al.5.001. a separate list was provided for each triangle in each animation (see Appendix A). [2000].41) 3.005. Results and Discussion The groups significantly differed in their performance on the background ToM tasks. For the MCQ-feelings. After each animation. Scoring Objective scoring.94 (1./Objectively testing theory of mind in autism 151 . Participants were also informed that after some animations there would be further questions. They were then presented with two practice animations (GD and ToM) for familiarization to ensure that they understood the task. For the MCQfeelings. participants were presented with a list of MCQ-categorization types to choose from.03) 12.001). non-parametric tests were conducted on these data. the comparison group’s mean score was again significantly greater than that of the ASD group (U 5 28.49) 0.90) 1.70) 4. Po0.

the new method required only a fraction of the time necessary for the original subjective method to be scored. and ToM animations by ASD and comparison groups. Similarly the appropriateness scores for the ASD group were significantly lower than those of the comparison group for the ToM animations (U 5 186. see Table II). the MCQ-categorization total score was weakly related to the false-belief tasks (r 5 0. ASD. Nevertheless. random (U 5 143. (c) Anxious. despite this objective method providing more explicit information about the animations. even when they correctly categorized these animations. correctly identifying the ‘‘feelings’’ of the triangles in only about a quarter of the ToM animations. but did not differ for the GD (U 5 150. 2002]. 2000. (d) Annoyed. P 5 0. Performance on objective categorization of random./Objectively testing theory of mind in autism INSAR .5) or random animations (U 5 113. (e) Unsure How do you think the Big Triangle feels at the end of the clip? (a) Frustrated. and the MCQ-feelings measure was weakly related to the intentionality score on the ToM animations (r 5 0.0. Appendix A MCQ-categorization Which of these 3 categories best fits the film clip you have just seen? (a) No interaction (b) Physical interaction (c) Mental interaction MCQ-feelings (correct answers in bold) Coaxing: How do you think the Little Triangle feels at the end of the clip? (a) Proud. the ASD participants had more difficulty in understanding the mental states depicted in the animations. this novel objective method of administering and analyzing responses to the Frith-Happe´ animations was found to be as sensitive as the traditional subjective method in demonstrating the well-established mentalizing impairment in autism. (d) Satisfied. We hope that the ease and speed at which this method can be administered will make the animations an even more useful and objective research tool with high-functioning children and adults. all but four participants in the ASD group fell below the comparison group range of scores on this measure. (b) Puzzled. this difference was only significant for the ToM animations. The novel objective method presented in this study for administering and analyzing participants’ responses to the Frith-Happe´ animations therefore successfully detected mentalizing difficulties in this group of highfunctioning adults with autism.4 ASD Comparison Accuracy (4 max) 3 2 1 0 Random Goal-Directed ToM Figure 1. (c) Tense. Theory of Mind. (e) No feelings Mocking: How do you think the Little Triangle feels at the end of the clip? (a) No feelings. The ASD group’s performance was indicative of a general difficulty with identifying whether interactions were occurring and whether these involved mental states.007). (b) Loving.065). Further investigation revealed that. supporting the widely held belief of a mentalizing impairment in autism..478. (b) Fulfilled. the ASD 152 group did manage to correctly categorize approximately two-thirds of the 12 animations. (e) Devious How do you think the Big Triangle feels at the end of the clip? (a) Aggravated. goal-directed.5. Castelli et al. (b) No feelings. ToM. P 5 0. This is reminiscent of previous studies in which individuals with ASD did use some mental state language to describe these animations but often gave inappropriate mental states [Abell et al. These correlations were not significant in the comparison group however. in fact. most probably due to ceiling effects in this group’s performance.488. P 5 0. Interestingly. The mentalizing difficulties revealed by this novel method were in line with their impaired performance on standard ToM tests and their impaired verbal descriptions of the animations. (d) Mischievous. (c) Lonely.5) but the intentionality score for the ASD group was significantly lower than that of the comparison group for the ToM animations (U 5 188. (d) Frivolous. However. P 5 0. (e) No feelings White et al.. In conclusion.5) and GD animations (U 5 97. (c) Secure. Autism Spectrum Disorder.061).008). In the ASD group.0.

E.... 62. Development and Psychopathology. Knickmeyer. 205–223. Oreland. (c) No feelings. Cherkassky. & Ohrmann.. F.E. U. C. 5–17. White et al. S. G. 831–846. American Journal of Psychology. R.. T. Spektor.. C. K. Mandy.. W. Keller. Moriguchi.. R. & Komaki. B. Moriguchi. P. B.A.D.. (e) Lucky How do you think the Big Triangle feels at the end of the clip? (a) No feelings. 877–893. 125. Kueppers.. et al. 1472–1482. Skuse.. (2006). Autism. (2002). M.. F. H. The development of mental state attributions in women with X-monosomy. Journal of Cognitive Neuroscience../Objectively testing theory of mind in autism 153 . 889–897. & Clubley. Risi. Leslie. (2000). Arolt. W.D. (d) Pleased. Structure. & Husain. Impaired self-awareness and theory of mind: An fMRI study of mentalizing in alexithymia. (1999). Attributing social meaning to ambiguous visual stimuli in higher-functioning autism and Asperger syndrome: The social attribution task.. Cognitive Development. and the role of monoamine oxidase B in the sociocognitive phenotype. & Dolan. (2010). 243–259. (c) Bored..K. T. Journal of Child Psychology and Psychiatry.. Minshew. (c) Scared. Fear recognition ability predicts differences in social cognitive and INSAR neural functioning in men. (2007). Brain Cognition. Neuroimage. K. (1944).. D. Castelli. R..A. Leventhal.. M. F. The influence of movement and outcome on young children’s attribution of intention. Corden. G. Bowler. D. Montgomery. Frith. 17. scientists and mathematicians. Fetal testosterone and empathy. K. Mandy. Messer.. (1985).. O. K. Mori. Lawrence. J. & Springer. 18.. U. Luckett.J. P. Journal of Child Psychology Psychiatry. M. & Frith. 914–928. J. Frith. (b) Eager. D.. Critchley. Kana. (2004).. Brain. 33. Wheelwright. Lord. Klin.. (b) Cunning. S. males and females. Prinz. 355–363. The impact of extensive medial frontal lobe damage on ‘‘Theory of Mind’’ and cognition... Lawrence. Baron-Cohen. 115–123. Happe´.. D... et al. Journal of Autism and Developmental Disorders. D.. Psychiatry and Clinical Neurosciences.M. Atypical frontal-posterior synchronization of Theory of Mind regions in autism during mental state attribution.. Baron-Cohen. 147–171. T.... Pedersen.A. V. U. (2006)... motion and preschoolers’ perceptions of social causality. 5. Maeda. Brain.. 61.. 11. (2000). Animated triangles: An eye tracking investigation. 57. 127.C. Frith. & Frith. & Skuse.C. Hormones and Behavior. 15. 84–100. Ohnishi. L. (2006). Theory of Mind in first-episode schizophrenia patients: Correlations with cognition and personality traits. 1–16. (2009). Lambrecht. 49. Powell. & Simmel. 18. K. D. & Frith. Skinner. T. ‘‘Theory of mind’’ in Asperger’s syndrome. Malik. Fletcher. 41. 127–140. (2000). 1189–1197. (1992). C. U. & Hackett. (e) Uneasy References Abell.S. 237–244. A. Ohnishi. V. Asperger syndrome and brain mechanisms for the attribution of mental states to animated shapes. 37–46. 157–166. K. & Frith. C. A.. Thornton. NeuroImage. R. R. D. A. 4. Klin. 32. Berry. British Journal of Developmental Psychology.. An experimental study of apparent behaviour. Other minds in the brain: A functional imaging study of ‘‘theory of mind’’ in story comprehension. & Montgomery. 21. The autism-spectrum quotient (AQ): Evidence from Asperger syndrome/high-functioning autism. & Thommen. Journal of Autism and Developmental Disorders. Y. Attributing social and physical meaning to ambiguous visual displays in individuals with higher-functioning autism spectrum disorders. U.. Castelli. A. Quarterly Journal of Experimental Psychology. Klein. A. 109–128. Happe´. et al.. F... The Autism diagnostic observation schedule-generic: A standard measure of social and communication deficits associated with the spectrum of autism. Mitra.. Happe´..J. Movement and mind: A functional imaging study of perception and interpretation of complex intentional movement patterns. 119. Social Neuroscience. 30. W. Changes of brain activity in the neural substrates for theory of mind during childhood and adolescence. Jeyakuma.. (2000). et al. 1–11. Do children with autism who pass false belief tasks understand the mind as active interpreter? Journal of Autism and Developmental Disorders. Jones.. 102.M. (2000). S. Bird. E.. Taylor. (2002)... Mori. Zwickel. F. U. J.. F. P. Baron-Cohen. Castelli. Matsuda.E. Happe´. M.. (2001). (1995).J. R. S. Schizophrenia Research. Autism. E. (2006). 282–292. Koelkebeck.. 31. F. Do triangles play tricks? Attribution of mental states to animated shapes in normal and abnormal development.. 57. Meanings in motion and faces: Developmental associations between the processing of intention from geometrical animations and gaze detection accuracy.. Ecological Psychology. Does the autistic child have a ‘‘theory of mind’’? Cognition. (d) Composed. Dolan. 32.D. (2009).. (1993).. W. L.. C.. 40–53. (e) Foolish Surprising: How do you think the Little Triangle feels at the end of the clip? (a) Provoked. S. 245–261. Cognition. L. & Frith. D... 61.J..... Campbell.L. (d) Cheerful.A. F.. (2007). S.. (e) No feelings How do you think the Big Triangle feels at the end of the clip? (a) Excited. Frith. Martin. (b) No feelings. & Jones. Y. Cook. (d) Disappointed. Suslow.D. (2006). (c) Delighted.. F.. Heider.Seducing: How do you think the Little Triangle feels at the end of the clip? (a) Lost. & Schulz. T.L. 135–152. Cognition. A. T. (b) Terrified. M... Raggatt. H.M.. & Just. D. 4. Baker. Bowler.H. T.. 99–118. R.D. S. U. Attribution of mechanical and social causality to animated displays by children with autism. Lane. N.

Journal of Experimental Child Psychology.S. & Frith. D. ¨ ller. London: Psychological Corporation. K. 39. Reynaud.. Stuss. A. (1999).. Can autistic children read the mind of an animated triangle? Autism. & Stroobants. K. & Liu. 199–211. C. S. J.. Social Cognitive and Affective Neuroscience. 12. U. U. 20. 523–541. in press.. J. Perner. (1996).. 318. British Journal of Social Psychology. D... 60. White. H.M. R. Seigal. Scaling of Theory-of-Mind Tasks. Nonverbal bases of social perception: Developmental change in sensitivity to patterns of motion that reveal interpersonal events. Laubin. Exploring the building blocks of social cognition: Spontaneous agency perception and visual perspective taking in Asperger syndrome. T. M. Zwickel. Zwickel. Rime´. Meier. M. G. & Yuill. 349–371. (2004). 9.. Journal of Nonverbal Behaviour. 115–124. Claxton. J.. D. J. (2008).. K. 241–260. & Wimmer. & Berry. (1989). 24.. Morris. & Mu evaluate and improve robots. Levine.Oatley. Child Development.M. Wellman. K. Eye movements as a means to Zwickel. & Corcoran. 357–366.. Wechsler Adult Intelligence Scale (WAIS). Perner. H. Springer. (2010).A.. Psychonomic Bulletin and Review.. International Journal of Social Robotics. Perception of interpersonal action in a cartoon film. & Leekham. 75. N. P. Wechsler. H. Schizophrenia Research.. 101–111. (2009). (2007). 437–471.J. S. Frith. D. 154 Salter. E.. Russell. Science.T. Agency attribution and visuo-spatial perspective taking. 1. & Tulving.. 3UK ed. The perception of interpersonal emotions originated by patterns of movement.R./Objectively testing theory of mind in autism INSAR . Child Development. Motivation and Emotion.. 81. J. E. Leslie. D. R. Do you see what I see? Interpretations of intentional movement in schizophrenia. 688–700. Boulanger. A. 16. (2009).to 10-yearold children..A.. B.... Herba. J. 1089–1093. Exploration of the autistic child’s theory of mind: Knowledge. R. B... & Skuse. Theory of mind is independent of episodic memory.. (1985). Richir. (1985). 1257. B. belief and communication. Lawrence. Rosenbaum. John thinks that Mary thinks thaty: Attribution of second-order beliefs by 5. (1985). (2006). White et al.