Language and theory of mind in preschool children with specific language impairment

Language and theory of mind in preschool children with specific language impairment

Available online at www.sciencedirect.com Journal of Communication Disorders 42 (2009) 428–441 Language and theory of mind in preschool children wit...

199KB Sizes 0 Downloads 27 Views

Available online at www.sciencedirect.com

Journal of Communication Disorders 42 (2009) 428–441

Language and theory of mind in preschool children with specific language impairment M. Jeffrey Farrar a,*, Bonnie Johnson b, Virginia Tompkins a, Molly Easters a, Andrea Zilisi-Medus b, Joann P. Benigno c b

a Department of Psychology, University of Florida, United States Department of Communication Sciences and Disorders, University of Florida, United States c School of Hearing, Speech, and Language Sciences, Ohio University, United States

Received 28 March 2008; received in revised form 10 April 2009; accepted 2 July 2009

Abstract Language plays a critical role in the development of theory of mind (ToM). There is limited research, however, examining the role of specific components of language in ToM development for typical and clinical populations. The purpose of the current study is to examine the relative contributions of general grammar, grammatical tense markers, syntactic complementation, and receptive vocabulary on understanding standard ToM tasks in preschool children with specific language impairment (SLI). Thirty-four children with language disorders, ages 42–65 months, were administered a series of language and ToM measures. Hierarchical regression analyses were conducted to examine the relative contributions of language subcomponents to ToM task performance. The results indicated that general grammatical development and vocabulary contributed uniquely to ToM reasoning. Sentential complementation abilities did not make an independent contribution. Theoretical and clinical implications of the findings for different accounts of the role of language in ToM reasoning will be discussed. Learning outcomes: Readers of the current study will be able to: (1) understand the relationship between language and ToM development in both typical and clinical populations; (2) understand the different ways in which various language components are related to false belief reasoning; and (3) consider the implications for intervening with children with language disorders. # 2009 Elsevier Inc. All rights reserved.

1. Introduction The overarching aim of the current study was to examine the role of language in the theory of mind (ToM) development of children with specific language impairments (SLIs). ToM, broadly defined, refers to the understanding of self and others as mental beings who have beliefs, desires, emotions, and intentions, as well as the understanding that behavior is motivated by these mental states (Flavell & Miller, 1997). A particularly important achievement of ToM development is the understanding of false beliefs, which refers to the knowledge that one may hold a belief that differs from reality and that people can have different beliefs about the same situation. False belief (FB) understanding

* Corresponding author at: Department of Psychology, University of Florida, P.O. Box 112250, Gainesville, FL 32611, United States. Tel.: +1 352 273 2140; fax: +1 352 392 7985. E-mail address: [email protected] (M.J. Farrar). 0021-9924/$ – see front matter # 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.jcomdis.2009.07.001

M.J. Farrar et al. / Journal of Communication Disorders 42 (2009) 428–441

429

is a well-documented milestone of social cognitive development that occurs around 4–5 years of age in typically developing children (Wellman, Cross, & Watson, 2001). Recent research has clearly established the role of language in the emergence of ToM development (see Astington & Baird, 2005). Despite the recognition of the critical role that language plays in this understanding, there is considerable debate regarding the nature of this relation. Of primary interest has been which component of language is necessary for understanding false beliefs and other conflicting representation tasks. Most of this research has explored these relations in children who are typically developing (e.g., Astington & Jenkins, 1999; de Villers & Pyers, 2002). In addition, the role of language in ToM development in clinical populations, such as individuals with deafness, autism spectrum disorder, and language impairment has been explored (e.g., Gale, de Villiers, de Villiers, & Pyers, 1996; Ziatas, Durkin, & Pratt, 1998). The purpose of the current study is to examine the relative contributions of different grammar and vocabulary measures to ToM in preschool children with SLI. 1.1. Typical populations Children’s ability to pass false belief (FB) tasks and other conflicting representation tasks (e.g., appearance–reality) emerges between 4 and 5 years of age (Astington & Baird, 2005). In the standard Sally–Anne mistaken location task, children are provided a story in which Sally puts her marble in a basket as Anne watches. Sally then goes to play. While she is gone, Anne moves the marble to a different basket. The child is then asked: ‘‘Where will Sally think her marble is when she returns to the room?’’ If the child understands the representational nature of beliefs, then she should predict that Sally thinks the marble is in the place Sally left it, even though the child knows that it is another location. Passing the task indicates that the child understands that different people can have conflicting beliefs about the same situation, i.e., that he or she understands false beliefs (Frith, 1989). Past theorists have proposed a variety of accounts regarding the emergence of ToM, including biological/ modularity (Baron-Cohen, 1995), theory construction (Gopnik & Wellman, 1994), and simulation of mental states (Harris, 1996). More recently, a considerable volume of research has focused on the role of language in FB reasoning (see Astington & Baird, 2005). Well-established findings, from both correlational (e.g., Astington & Jenkins, 1999) and training studies (e.g., Lohmann & Tomasello, 2003), demonstrate that language predicts FB performance. However, there is considerable debate concerning the nature of this relation. One of the central issues concerns whether semantic (i.e., vocabulary) development, general grammatical development, specific grammatical constructions, or more general language ability is central to FB understanding. Jenkins and Astington (1996) found associations between performance on ToM tasks and a wide range of language abilities as measured by the Test of Early Language Development (TELD; Hresko, Reid, & Hammill, 1999), a general measure of semantic and syntactic ability (see also Farrar & Maag, 2002; Ruffman, Slade, Rowlandson, Rumsey, & Garnham, 2003). Similarly, vocabulary size, as indexed by performance on the Peabody Picture Vocabulary Test (PPVT), predicts FB performance not only for children with typical development but also for children with autism and those with cognitive disability (Happe´, 1995). Measures of receptive vocabulary, the most frequently used measure of language development in these studies, reflect that general language abilities are thought to play a role in FB reasoning. Further, children’s understanding of mental verbs such as think and know also relates to the performance on ToM tasks (Moore, Bryant, & Furrow, 1989). The general explanation of the link between measures of mental state vocabulary and ToM reasoning is that in order to pass FB tasks, the child must be familiar with the language of the mind. Other researchers have argued that grammatical development plays a more fundamental role in FB reasoning (Astington & Jenkins, 1999; de Villiers & de Villiers, 1999). In the strongest version of this argument, de Villiers and her colleagues have argued that children’s comprehension of complement constructions is causally linked to FB understanding. In complement constructions, the use of mental or communication verbs involves the embedding of propositions in a recursive fashion, such as Sally thought that Anne was sad, but she was really happy. In these constructions, the truthfulness of the embedded proposition, e.g., Anne was sad, is irrelevant to determining the actual truthfulness of the entire sentence because of the presence of a mental verb, e.g., thought in the above example. Thus, such sentences directly distinguish between the way things are in the mind from the way things are in the world— which is at the core of a representational ToM understanding, i.e., FB. Complement understanding is typically assessed with a memory for complement task (de Villiers, 2000). Children are shown a picture of a person thinking or saying something that is false, which is then qualified by the true state of reality. For example, ‘‘She told the man she saw a ghost, but it was really a blanket. What did she tell the man?’’ To be credited with a correct response, children must

430

M.J. Farrar et al. / Journal of Communication Disorders 42 (2009) 428–441

have remembered the ‘‘mistake’’ (she saw a ghost) that occurred prior to hearing the complementation clause (it was really a blanket). Most importantly for this argument is the finding that children’s use of complements predicts their performance on FB tasks (e.g., de Villiers, 2000). For instance, in a longitudinal study examining the relationship between memory for complements and FB reasoning, de Villers and Pyers (2002) reported that performance on the memory for complements tasks was a precursor of later FB reasoning. Training studies, however, provide only partial support for this claim. Children were exposed to a deceptive object, such as a candle that resembled an apple. The deceptive object was noted in several different ways. Children who were exposed to the deception using a complement construction performed significantly better on FB reasoning than children in a control condition with no training. However, they did not perform any better than children exposed to discourse that highlighted the deception but did not involve using a complementation (Hale & Tager-Flusberg, 2003; Lohmann & Tomasello, 2003). Thus, training studies suggest that while complementation may facilitate FB reasoning it is not critical. Further, the unique and predictive value of complementation understanding has not always accounted for variations in performance on FB reasoning tasks. In a cross-linguistic comparison of children learning Cantonese and English, complementation did not play a unique role in predicting ToM after controlling for general language comprehension (Cheung et al., 2004; see also Perner, Sprung, Zauner, & Haider, 2003). Similarly, although Astington and Jenkins (1999) found that syntactic development is uniquely associated with ToM, they did not argue that complement understanding accounts for variations in this relation. In a cross-lagged correlational study, Astington and Jenkins (1999) found that children’s language development, and in particular syntax, at 3 years of age was related to later ToM at 4 years. They suggest that children’s syntactic skills allow them to keep track of linguistic information used to represent the changing locations or contents involved in ToM tasks. However, since a specific complementation measure was not used in their study the relative role of general syntactic skills versus complementation understanding remains unclear. Other researchers have proposed a more general language account and have argued that syntax does not play a unique role in FB reasoning. For example, Slade and Ruffman (2005) argued that both semantics and syntax are important in ToM development. They argue that understanding a proposition, including false ones, requires both semantic and syntactic understanding. As they observe, most grammatical measures of language ability require both semantic and syntactic skills. In fact, many standard linguistic assessments, such as the TELD, confound semantics and syntax, making it difficult to tease apart their relative contributions. Across several studies, Ruffman and colleagues report that general language plays a critical role in ToM development and found no unique role for syntax. While they suggest that complementation does not play a special role, they did not measure complementation in their sample. Thus, multiple studies have found convincing support for a role of language in ToM understanding. A recent metaanalysis of studies investigating the relationship between language and ToM in typically developing children confirmed that different measures of language played significant roles in FB reasoning (Milligan, Astington, & Dack, 2007). The largest effect size was found for the memory for complement measures (.66). However, the other language measures, including general language, receptive vocabulary, semantics, and syntax, all had significant effect sizes ranging from .32 to .54. However, a consensus still has not emerged regarding whether particular linguistic abilities are required for ToM reasoning. One of the reasons for this lack of agreement is that different studies use different sets of language measures, with very few using a range of measures that systematically tap into different language skills. Only a few studies have directly compared the explanatory power of different language measures (e.g., semantics, grammar, and complementation) in predicting FB performance. These studies have not always found that grammar plays a unique role beyond language comprehension (Astington & Jenkins, 1999; Slade & Ruffman, 2005) but they have often used more general measures of grammar or language that assess both semantics and syntax. Perhaps more critically, these studies have typically not directly assessed or compared the relative contributions of semantics, general grammar and complementation. For example, of the 104 studies included in the meta-analysis, only 4 studies used a complementation measure. Thus, it is difficult to fully assess the differential role of semantics, general grammar, and complementation in ToM development without examining the relative contribution of different language measures. 1.2. Clinical populations Studies of clinical populations of children support the relation between language and ToM. Both children with autism spectrum disorder and deaf children who experience limited language input show delays in ToM understanding. The extent of their ToM delay is related to their language abilities (e.g., Peterson & Siegal, 1999;

M.J. Farrar et al. / Journal of Communication Disorders 42 (2009) 428–441

431

Tager-Flusberg & Joseph, 2005). For instance, Ziatas et al. (1998) reported relationships between belief term development and ToM performance across groups of children with autism, Asperger syndrome, and specific language impairment (SLI) (see alsoTager-Flusberg, 2000). Studies of children with SLI have been used to explore and clarify the link between language and ToM. Children with SLI are characterized by normal intelligence, normal hearing abilities, typical emotional development, and typical behavioral development, but below normal performance on language measures (Leonard, 1998). If language is crucial to ToM performance, then it would be expected that children with SLI would perform poorly on ToM tasks. Further, children with SLI might show differential development of various language skills, perhaps making it easier to disentangle their relative roles. If children with SLI are delayed on memory for complements or general grammar than their performance on a FB task might be impacted. On the other hand, if grammar is not necessary for ToM, but rather only vocabulary development, then children with SLI may not show ToM performance deficits if their vocabulary is not impaired. Similarly, if children with language impairments have poor general grammar skills then again they might be delayed on the ToM tasks and the degree of current language impairment would be predictive of ToM performance. There have been a limited number of ToM studies of children with SLI. The results have been mixed regarding whether children with SLI have difficulty passing ToM FB tasks. Miller (2001) reported that the performance of children with SLI was significantly lower than typically developing controls when the linguistic demands of the ToM tasks were high, but that these differences tended to dissipate as the linguistic demands were reduced. Similarly, Farrant, Fletcher, and Mayberry (2006) reported that 5-year-old children with SLI were delayed in ToM reasoning and visual perspective taking relative to typical age-matched controls. Other studies have compared children with autism spectrum disorder to children with SLI. A few studies found that children with autism perform more poorly on ToM tasks than children with SLI (Shields, Varley, Broks, & Simpson, 1996; Ziatas et al., 1998); however, one study found few differences in ToM performance between children with high-functioning autism and children with SLI (Gillott, Furniss, & Walter, 2004). These studies suggest that children with SLI and other language disorders may be delayed in their ToM development because of their language impairment. However, compared to research on typically developing preschoolers, studies of children with language impairment have not typically explored the relation between FB and different language abilities (Farrant et al., 2006; Miller, 2001). Of primary interest would be which language skill, namely, semantic development, general grammar, memory for complements, or some combination of language abilities, accounts for the variance in this relation. Miller (2004) examined the relation between understanding sentential complements and passing FB tasks for children with SLI and their language- and age-matched controls. Overall, the children with SLI as a group performed similarly to age-matched controls on the FB tasks, but performed worse on the complementation task. Their performance on the complementation task was more similar to the performance of the language-matched control group. Of particular importance was the relation between performance on the complement and FB tasks. For the three groups combined together, the best predictor of performance on the FB tasks was the sentential complements task. Due to sample size, however, it was not possible to fully explore this relation separately for each group. Thus, the link between FB and complementation was generally supported in the group of children with SLI. The exact nature of this relation remains unclear since children’s general grammar and vocabulary ability was not assessed. Furthermore, the complement sentences used in Miller’s study differed from those used in the standard memory for complements task because the sentences did not contain a contrast with reality, making it difficult to compare across studies. In the largest assessment of language and FB in children with SLI, de Villiers, Burns, and Pearson (2003) tested nearly 300 children with SLI from 4 to 9 years of age on a variety of language and ToM tasks. Complement understanding was the best predictor of performance on ToM tasks followed by communicative role taking. In contrast, general measures of morphosyntax contributed less to children’s performance on ToM. The complementation measure was based on a single item from the Diagnostic evaluation of language variation (Seymour, Roeper, & de Villiers, 2003) which is different from the more extensive complementation measure used in prior studies with typically developing children. In addition, measures of semantic development were not taken either, and thus, the relative impact of semantic and grammar remains unclear. 1.3. Current study and predictions In summary, only a few studies have examined the role of language in ToM development for typical and clinical populations using a variety of language measures. The purpose of the current study is to examine the relative roles of grammar, including complementation, and semantics in understanding false beliefs in preschool children with SLI.

432

M.J. Farrar et al. / Journal of Communication Disorders 42 (2009) 428–441

Our study adds to the body of literature on children with atypical language development and theory of mind. The first research question addressed whether there was a relationship between language and ToM performance for children with SLI. It was expected that children’s language abilities would be related to the level of performance on the ToM tasks. The second, and primary, research question was which aspect of language was most predictive of ToM reasoning. Of particular interest were the relative contributions of vocabulary, grammar, and complementation on ToM performance. It was hypothesized that all measures of language would be related to TOM reasoning. 2. Methods 2.1. Participants Thirty-four children with specific language impairment participated in this study (11 females; 23 males). They ranged in age from 42 to 65 months (M = 56.12, SD = 5.24). Maternal education ranged from 9th grade to a bachelor’s degree, with an average of 1 year of education completed beyond high school (SD = 2). By parent report, 24 of the children were White/non-Hispanic, 3 were White/Hispanic, 2 were Asian, 2 were African American, and 3 did not provide race/ethnicity information. 2.1.1. School site All children who participated in this study were enrolled at the same preschool in an intensive speech and language intervention program. Teachers, who were certified speech–language pathologists, provided an intensive languagerich program 3 h per day, 5 days per week. In each classroom the student–teacher ratio was no greater than 5:1, with a maximum of 15 children per classroom. To qualify for this preschool program, children had to be evaluated by a certified speech–language pathologist. All children had to meet the state of Florida criteria for eligibility for services. During the period of our recruitment (May 2004–December 2005) the state criterion was a standard score of 77 or less on one or more language measures and/or use of clinical judgment. 2.1.2. Participant language inclusion criteria The use of a cutoff score on standardized tests has been justly questioned in terms of its diagnostic utility (Dollaghan & Campbell, 1998; Spaulding, Plante, & Farinella, 2006); hence, classification of children as having specific language impairment for this study was a multifaceted process. We obtained triangulation of a language delay status through (a) continued concern by the child’s lead teacher/speech–language pathologist at the time of the study; (b) reviewing the child’s language assessment and treatment history in their school file documenting initial language test scores done at the school; and/or (c) a Structured Photographic Expressive Language Test – Third Edition (SPELTIII; Dawson, Stout, & Eyer, 2003) score of <95 at the time of the study. There was an average of 13 months between initial school diagnosis and language and ToM testing for the current study. The SPELT was one component in our inclusionary criteria for participation. This test was chosen because the validity of the SPELT-III for identifying language impairment has been documented through both exploratory and confirmatory analyses indicating 90% sensitivity and 100% specificity when a cutoff standard score of 95 is applied (Perona, Plante, & Vance, 2005). Table 1 provides test score information reported in the children’s school files from the children’s initial diagnostic tests that were used in conjunction with clinical judgment to qualify these children for the speech–language program. As seen in Table 1, four of the children did not receive an initial diagnosis of language delay, however, at the time of current testing, they were delayed on the SPELT-III and were still in treatment for language delay. In addition, five participants who were referred by teachers were excluded because they received a score of 95 or greater on the SPELT-III and their original language diagnosis reflected only a phonological language disorder. Thus, to be included in the study, children had to still be in treatment for a receptive or expressive language disorder. In addition, children had to receive an initial diagnosis of a receptive or expressive language delay and/or have a current SPELT score of below 95. 2.1.3. Other participant criteria File review and/or caregiver interview confirmed that participants had no known neurological impairment, behavioral disorder, or emotional disorders. In addition children received a standard score of 75 or higher on the Leiter

M.J. Farrar et al. / Journal of Communication Disorders 42 (2009) 428–441

433

Table 1 Initial diagnostic tests and scores for school entry. Participant number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

Diagnostic test a

PLS-3 PLS-3a PLS-3a PLS-3a TELD-3b PLS-4c PLS-4c PLS-4c R/EOWPVTd PLS-4c PLS-4c PLS-4c PLS-4c OWLSe PLS-4c PLS-4c CELF-Pf TELD-3b PLS-3a TELD-3b PLS-4c PLS-4c PLS-4c PLS-4c REELS-2g PLS-4c PLS-4c PLS-4c PLS-4c PLS-3a PLS-3a TELD-3b PLS-3a PLS-3a

Composite standard score

Receptive standard score

Expressive standard score

81 74 80 81 119 52 104 69 n/a 92 73 96 93 77 100 61 76 84 87 74 94 84 85 104 76 79 95 81 87 77 83 82 73 73

83 76 77 89 119 53 114 81 84 109 82 102 95 81 102 67 93 86 92 80 92 90 109 106 92 90 96 92 85 74 66 86 70 70

82 77 76 77 112 61 93 62 68 77 75 91 92 76 99 62 73 88 84 77 86 80 63 101 61 72 96 73 91 84 84 84 81 81

a

PLS-3: Preschool Language Scale – Third Edition (Zimmerman, Steiner, & Pond, 1992). TELD-3: Test of Early Language Development – Third Edition (Hresko et al., 1999). c PLS-4: Preschool Language Scale – Fourth Edition (Zimmerman, Steiner, & Pond, 2002). d R/EOWPVT: ROWPVT: Receptive One Word Picture Vocabulary Test (Brownell, 2000a) and EOWPVT: Expressive One Word Picture Vocabulary Test (Brownell, 2000b). e OWLS: Oral and Written Language Scales (Carrow-Woolfolk, 1995). f CELF-P: Clinical Evaluation of Language Fundamentals – Preschool (Semel, Wiig, & Secord, 1992). g REELS-2: The Receptive–Expressive Emergent Language Scale – Second Edition (Bzoch, League, & Brown, 1991). b

International Performance Scale – Revised (Leiter-R; Roid & Miller, 1997) indicating nonverbal intelligence within the normal range. All participants passed a hearing screening administered by the on-site certified audiologist. 2.2. Materials and procedure Language measures. Four language measures were obtained from the children for the purposes of the current study: an overall expressive grammar measure, a morphosyntax measure, a memory for complement structure measure, and a receptive vocabulary measure. 1. Children’s score on the SPELT-III was used as a measure of overall expressive grammatical abilities. In this standardized test, children are shown photographs and asked questions about each photograph. Items initially require short phrases, such as ‘‘on the chair,’’ and build in syntactic complexity to items that require

434

M.J. Farrar et al. / Journal of Communication Disorders 42 (2009) 428–441

production of complex sentence structures. This test does not include a ceiling; children are administered all test items. 2. A measure of children’s expressive morphosyntax was derived using a subset of items from the SPELT-III yielding a tense composite score. There are 53 items on the SPELT-III but only 28 of these items require the child to produce a tense morpheme to be credited with a correct response. For example, in item #9 the child sees a picture of a sheep eating and is asked ‘‘Tell me what the sheep does everyday.’’ To be credited with a correct response, the child must produce a verb with the third person singular -s morpheme (i.e., he eats). An example of an item that was not included in the tense composite is item #8 where the probe question is ‘‘What’s happening.’’ The item is probing for progressive -ing. The child only has to say ‘‘writing’’ to be credited with a correct response and the -ing morpheme is not an example of a tense morpheme. Although the child may have responded with ‘‘he is writing’’ this item was excluded from the tense composite because the tense morpheme ‘‘is’’ was not required. 3. Children’s understanding of complement structures was assessed using the memory for complements task (see de Villers & Pyers, 2002). This task consists of 12 items in which the child is told a brief scenario using pictures as prompts. Each of the 12 items involves the actor thinking or saying something that is false, and is then qualified by the true state of reality. For example, ‘‘She told the man she saw a ghost, but it was really a blanket. What did she tell the man?’’ To be credited with a correct response, children must have remembered the ‘‘mistake’’ (she saw a ghost) that occurred prior to hearing the complementation clause (it was really a blanket). Children received 1 point for each mistake correctly remembered. The total score ranged from 0 to 12. In addition to a total score, separate subscores were calculated for communication verbs (range 0–6) (e.g., say) and mental verbs (range 0–6) (e.g., think). The use of mental verbs may be redundant with standard FB measures, and thus communication verbs may provide a purer measure of complement understanding that is not potentially confounded with ToM measures. 4. Children’s receptive single word vocabulary was assessed using the Peabody Picture Vocabulary Test – Third Edition (PPVT-III; Dunn & Dunn, 1997). This measure assesses general receptive vocabulary and is commonly used in ToM studies. Theory of mind. Children were administered several ToM tasks designed to assess their understanding of conflicting mental representations. These included the typical tasks used to assess false belief understanding and included unexpected content, unexpected location, and appearance–reality tasks. Order of presentation of these tasks was counterbalanced across children. 1. Unexpected contents task (Perner, Leekam, & Wimmer, 1987). In the mistaken contents task, children were shown a box of crayons and were asked, ‘‘What is in this box?’’ After the child answered that the box contained crayons, the box was emptied to reveal it contained candles, and the child identified the true contents. After returning the candles to the box, children were asked about their original belief, ‘‘When you first saw the box, all closed up like this, what did you think was in the box?’’ The child was awarded 1 point for answering this FB question correctly. Children were then told that the Cookie Monster puppet had never seen what was in the box and they were asked what Cookie Monster thought was in the box. The child was awarded 1 point for answering this FB question correctly. Children were also asked a memory control question regarding the original content of the box. In order to receive credit for the FB questions, children had to answer the memory control question correctly. Scores on the FB unexpected content task could range from 0 to 2. 2. Unexpected location task (Wimmer & Perner, 1983). FB was also assessed using the mistaken locations task. In this task, children were told a story using Ernie and Elmo puppets, in which Ernie plays with a toy, puts it away in one location in full view of the child, then ‘‘goes outside to play.’’ While Ernie is away, Elmo removes the toy from its hiding place, plays with it, then puts the toy away in a different location. Children are told that Ernie comes back to play with the toy again, and are asked the FB question: ‘‘Where will Ernie look for the toy?’’ The child was awarded 1 point for answering this FB question correctly. Memory control questions were asked to ensure whether the child remembered where the character had put the object initially and its later location. In order to receive credit for the FB question, children had to answer the control question correctly. Scores on the unexpected location task could range from 0 to 1. 3. Appearance–reality tasks. Children completed three appearance–reality tasks requiring them to distinguish between real and apparent object identities and properties, reflecting an understanding of conflicting representations. In the mistaken attribute task, children were presented with a white paper butterfly and were asked, ‘‘When you look at this butterfly right now, what color is it?’’ The researcher placed a blue filter paper over

M.J. Farrar et al. / Journal of Communication Disorders 42 (2009) 428–441

435

the butterfly and asked, ‘‘When you look at this butterfly right now what color is it?’’ This was then followed by the critical appearance–reality question, ‘‘What color is the butterfly really and truly? Is it really and truly white or is it really and truly blue?’’ Children were awarded 1 point for answering that it was white. The second mistaken attribute task was similar in procedure, using a white fish and red filter paper. Children were awarded 1 point for answering that it was white. Scores could range from 0 to 2. Children also completed the mistaken identity task, an appearance–reality task involving object identity. After first viewing a sponge painted to look like a rock, children were allowed to touch it, and then name it. The researcher then asked, ‘‘What does this look like to your eyes right now, does it look like a rock or a sponge?’’ They were then asked the appearance–reality question: ‘‘What is it really, is this really a rock or really a sponge?’’ The child was awarded 1 point for answering it was really a sponge. Following Gopnik and Astington (1988), the researcher then asked ‘‘When you first saw this, before you touched/squeezed it, what did you think it was?’’ The child was awarded 1 point for answering ‘‘rock.’’ This was followed by the FB question. The experimenter showed the child a cartoon cat puppet and said ‘‘Kitty has never touched this before. What does Kitty think this is, a sponge or a rock?’’ The child was awarded 1 point for responding rock. Total scores on this appearance–reality task ranged from 0 to 3. Thus, total scores of the four ToM tasks ranged from 0 to 8. Children had to correctly answer the control questions to be given credit for the unexpected location and unexpected content questions. Eight of the children failed one of the control questions for one of the ToM tasks and thus they were not credited with a correct answer on that specific task. 3. Results Overall performance. Table 2 provides the means, standard deviations, and ranges for the language and theory of mind measures. As reflected in the table there was wide variation in performance in the children on both the language and ToM measures. Language and theory of mind. Of primary interest was the relationship between language and ToM performance. As Table 3 indicates, Pearson bivariate correlations showed that all the language measures were significantly related to ToM performance. Age and performance on the Leiter-R nonverbal IQ test were not related to ToM. Table 2 Means, standard deviations, and range of measures.

Age in months Leiter-Ra SPELT-IIIa Morphosyntax b Complementc PPVT-IIIa Theory of mindd a b c d

Mean

SD

Range

56.12 105.71 83.18 22.07 6.47 95.94 2.94

5.24 10.89 10.82 21 4.27 13.48 2.37

48–65 88–128 64–113 0–68 0–12 63–122 0–8

The values represent standard scores (M = 100, SD = 15). The values represent percent accuracy on 28 items from the SPELT-III. Maximum score = 12. Maximum score = 8.

Table 3 Correlations of age, nonverbal IQ (Leiter-R), language, and theory of mind measures. Leiter-R Age Leiter-R SPELT-III Morphosyntax Complementation PPVT-III * **

p < .05. p < .001.

.37 – – – – –

*

SPELT-III .07 .10 – – – –

Morphosyntax .02 .13 .92** – – –

Complementation .01 .02 .59 ** .44 * – –

PPVT-III *

.38 .31 .57 ** .39 * .42 * –

ToM .17 .04 .55 ** .48 * .51 * .52 *

436

M.J. Farrar et al. / Journal of Communication Disorders 42 (2009) 428–441

Table 4 Results of hierarchical regression analyses for language and theory of mind. Variable

B

B (S.E.)

Beta

.09

.026

.52**

Step 2 PPVT-III SPELT-III

.07 .17

.03 .08

.31 .38*

Step3c PPVT-III SPELT-III Complementation

.05 .12 .15

.03 .09 .09

.28 .25 .25

a

Step 1 PPVT-III b

a b c * **

At Step 1, R2 = .27, p < .01. At Step 2, DR2 = .09, p < .05. At Step 3, DR2 = .04, p (n.s.). p < .05. p < .01.

To explore these relations more fully, we conducted two hierarchical regression analyses using the different language measures as predictors of ToM scores. Since the morphosyntax and SPELT measures were highly correlated, we ran separate regressions for these two grammatical measures using raw scores in the regression. In the first regression, the variables were entered in the order of PPVT-III, SPELT-III, and complementation scores. This order was used because it goes from most general to most specific language skill. As seen in Table 4, in Step 1 of the regression, PPVT was significantly related to ToM scores. However, once the SPELT-III score was added in Step 2, only it predicted ToM and PPVT was no longer associated with ToM performance. In Step 3, complementation scores did not make any unique contribution to ToM performance in the complete model. This pattern did not change if performance on the complementation task was broken down to items involving belief verbs and those involving communication verbs. Thus, the overall regression analysis suggests that general grammar was the most important predictor of ToM and that complementation did not uniquely contribute. In the second regression model we examined the role of morphosyntax measure more specifically. As before, the variables were entered in the order of PPVT-III, morphosyntax, and complementation scores. As seen in Table 5, in slight contrast to the first regression, in Step 2 of the regression both PPVT and the morphosyntax measure uniquely predicted ToM performance. As before, complementation scores were not uniquely related to ToM performance, using the overall or subscore measures, in the complete model. Table 5 Results of hierarchical regression analyses for language and theory of mind. Variable

B

B (S.E.)

Beta

Step 1 a PPVT-III

.09

.027

.52**

Step 2 b PPVT-III Morphosyntax

.07 .17

.03 .08

.40* .32*

Step3c PPVT-III Morphosyntax Complementation

.05 .12 .15

.03 .09 .09

.32 .23 .28

a b c * **

At Step 1, R2 = .27, p < .01. At Step 2, DR2 = .09, p < .05. At Step 3, DR2 = .06, p (n.s.). p < .05. p < .01.

M.J. Farrar et al. / Journal of Communication Disorders 42 (2009) 428–441

437

We performed one additional analysis, however, to explore whether differences in task performance on the complementation task related to ToM performance. de Villers and Pyers (2002) recommended distinguishing passing and failing the complementation task based on cutoff scores of 10 and above for passing and below 10 for failing the task. In our sample, we used the pass/fail cutoff for complement understanding and found significant differences in ToM scores between those who passed the complement task (M = 4.60), versus those who failed (M = 2.71), t(37) = 2.27, p < .05. These results are consistent with the correlational analyses. However, when the other language variables were considered in the regressions, complementation did not play a unique role. In summary, we found that current language ability in both vocabulary and grammar was correlated with ToM performance. To further explore the relationship between current language status and ToM performance we used the SPELT-III to form subgroups of children with mild and moderate language impairment. The moderate impairment group was defined as a SPELT-III score of 84 or below (n = 18) and the mild impairment as 85 and above (n = 16). Not surprisingly, on the morphosyntax tense composite measure, the mild impairment group performed significantly better (M = 34% correct) than the moderate impairment group (M = 12% correct), t(34) = 4.76, p < .001. There were no significant differences in performance on the Leiter-R IQ test. The two groups also differed on the PPVT-III with the mild group scoring higher (M = 102.63) than the moderate group (M = 90.00), t(34) = 3.65, p < .01. Similarly, the moderate group scored significantly worse on the complementation measure (M = 4.00) than the mild group (M = 9.29), t(34) = 4.50, p < .001. Of primary interest was whether there were differences in ToM performance for these two groups. There was a significant difference in the ToM performance, t(34) = 2.89, p < .01 between the moderate (M = 4.06) and mild language impairment (M = 1.94) groups. The children with moderate impairment performed significantly lower than the mild impairment group on the ToM tasks. Mean performance of these children was substantially below 50% correct. In contrast, mean performance of the mild impairment group on ToM was more than twice as high as the children with moderate impairment. There was also a wider range of performance with many of the children scoring above 50% correct. Thus, the language abilities of children with SLI were related to their performance on ToM tasks. Of particular importance was the finding that complementation did not have a unique role to play in once vocabulary and general grammar were accounted for in the regressions. 4. Discussion The purpose of the current study is to examine the relative role of grammar and semantics on ToM performance in children with language disorders. This has been a topic of considerable interest in studies of typically developing children. A variety of different proposals have been made for the role of language in ToM development. de Villiers argued that an understanding of complementation is necessary for the development of ToM reasoning because it allows children to represent false beliefs. Astington and Jenkins (1999) agree that syntax is important, but argue that it is not the acquisition of this structure that is responsible for ToM development since even some 3-year olds are capable of producing complements in their speech. They suggest, for example, that syntax may provide a format for representing the changing locations of objects. Other accounts argue for a more general role of language in theory of mind development (e.g., Slade & Ruffman, 2005). Many of these studies use different language measures with most not using a complementation measure. This makes it difficult to clearly identify the critical language abilities contributing to ToM development. Of particular interest in the current study was which aspect of language significantly predicted performance on the ToM measures in children with SLI. All language measures (SPELT-III, morphosyntax, complementation, and PPVT-III) were significantly related to ToM scores. The two regression models found, however, that complementation did not play a unique role in FB reasoning. The regressions did indicate that more general language abilities made unique contributions to ToM performance. In the regression analysis that included the morphosyntax and PPVT measures, each made independent contributions to ToM performance, whereas in the other regression analysis, the SPELT-III was related but vocabulary was not. This is not particularly surprising, since the SPELT-III reflects both grammatical and semantic knowledge, whereas the morphosyntax measure more directly reflects grammatical ability. Of particular importance was the finding that after accounting for general language ability, memory for complements was not related to ToM. The current results differ from de Villiers et al. (2003) who found that

438

M.J. Farrar et al. / Journal of Communication Disorders 42 (2009) 428–441

complementation was the most important predictor in children with SLI. There were differences, however, in how these studies measured both complementation and general grammar. It was true that even in the current study that using a cutoff criterion of passing 10 or more of 12 complement sentences showed that those children who passed the complementation task scored significantly higher on the ToM tasks than those who failed the complementation task. However, this is likely because complementation and grammar are correlated. It should be noted that if only a complementation measure had been used, then the present results also would have also suggested a special role of complementation. It is only when more general grammar and vocabulary measures were used does it become clear that more general language abilities may account for this relationship with ToM. As reflected in both the correlations and group comparisons between the moderate and mild groups, those children with more extensive language delay have more difficulty with the ToM tasks. Thus, the level of difficulty children with SLI have on ToM tasks, at least those tasks that have a language component, is related to their language abilities. These results have important implications for understanding the role of language in ToM development. The current findings indicated that for children with SLI, general grammar is particularly important for the development of FB understanding—more so than either vocabulary or complementation. A similar pattern has been reported for children with autism (Tager-Flusberg & Joseph, 2005). This finding may be particularly relevant for interventions designed to improve FB understanding in children with SLI. Understanding the best predictors of FB can help clinicians to target the specific aspect of language (i.e., general grammar) that is most likely to facilitate FB understanding. In a training study with children with typical language skills, Lohmann and Tomasello (2003) showed that children’s FB understanding improved significantly when children were exposed to perspective-shifting discourse and sentential complements compared to a group who did not receive language training. Thus, there is evidence that language training can be used to successfully improve children’s FB understanding. Our results suggest, however, that training on general grammar rather than sentential complements may be a more productive approach to training designed for children with SLI. What are the implications, if any, for understanding the role of language in ToM in typically developing children? Investigations of typically developing children have sometimes reported an important role of complementation understanding in FB reasoning (de Villiers & de Villiers, 1999), although not always (e.g., Cheung, 2006). In many of these studies, complementation measures have not always been systematically compared to more general grammatical measures. In fact in Milligan et al.’s (2007) meta-analysis of 104 studies, only 2 published studies included a complementation measure at all, and of these, only 1 included a general grammar measure (Cheung et al., 2004). In this latter study, it was also reported that complementation was not uniquely related to FB after the effects of general grammar were controlled. These and related findings have led to the development of a general language perspective for the emergence of ToM in which both semantic and grammatical abilities are believed to play important roles (e.g., Nelson, 2005; Slade & Ruffman, 2005). While caution is warranted in extending the current findings to typically developing children, the current results are consistent with this emerging perspective regarding the role of language in ToM development. Specifically, these findings support a more general language account of ToM development, rather than a syntax (i.e., complementation) specific version. Together, studies of both typically developing and atypically developing children can contribute to a comprehensive understanding of the role of language in ToM development. Further studies with both typical and atypical populations that assess children’s language abilities using a wide range of measures in order to clearly identify the critical language abilities are needed. There are several potential limitations to the current study. First, there might be concerns about the current status of the children’s language ability and its relation to performance on the ToM tasks. We used a triangulation approach to determining language status. The children had to currently be in treatment for an expressive or receptive language delay. In addition, they had to have an initial diagnosis of an expressive or receptive language delay and/or receive a SPELT score of 95 or below at the time of testing. Importantly, performances on the theory of mind tasks were predicted by level of current language ability as measured by the SPELT-III. At the time of the current study, all of the children had a normal level of intellectual functioning as measured by the Leiter. Further, as measured by the SPELT, about half the sample, the moderate impairment group, scored below a standard score of 85 on the SPELT. Group comparisons indicate that this moderate impairment group performed significantly worse on the FB task than the mild impairment group. This moderate impairment group also did significantly worse on the measures of morphosyntax, vocabulary, and complementation. Thus, these subsequent analyses support the claim that the level of language impairment is related to theory of mind performance.

M.J. Farrar et al. / Journal of Communication Disorders 42 (2009) 428–441

439

A second concern was the use of an expressive measure of syntax rather than a comprehension measure of syntax. This measure was selected because of its use in diagnosis of language impairment. It is interesting to note that this measure was the strongest predictor of ToM performance. Other studies of language and theory of mind have used either receptive (e.g., Astington & Jenkins, 1999) or expressive measures of syntax (de Villiers et al., 2003). Future studies should, however, include a measure of receptive syntactic ability as well. Similar effects are reported for both types of measures. Finally, there may be concerns raised regarding our semantic measures that used a general receptive vocabulary measure rather than a more specific measure of mental state vocabulary. We adopted this more general measure because it is the one, more typically employed and allows for more direct comparisons across studies. It may be, as Nelson (2005) suggests, that general measures of receptive language reflect children’s experiences conversing with adults about a range of topics, including topics involving the mind. Thus, measures of general vocabulary are markers of language development and may co-vary with more specific measures of mental state vocabulary. Future studies should also consider using more specific measures of mental state vocabulary. 4.1. Clinical implications From a clinical standpoint, the current results indicate that level of language delay in children with SLI (i.e., mild versus moderate impairment) is related to their ToM understanding. Based on these findings, as well as those of prior research, treatment goals which simultaneously target the enhancement of linguistic skills, particularly grammar, as well as social cognitive skills should be implemented (Miller, 2006). There are several strategies teachers and clinicians could use to targeting improvement in linguistic skills. For example, the ‘‘theory of mind’’ tasks could be converted into stories and embed the target words (e.g., mental state verbs), syntax (e.g., complement structures), and morphology into these stories. Many theory of mind tasks involve simultaneously managing multiple tenses such as present, past, and/or future verb tense depending on the task. Previous research has shown that interventions with preschoolers with SLI are more productive when clinicians embed training on grammar within a conversational context as opposed to direct training, in which grammatical targets are explicitly elicited (Bruce, Hansson, & Nettelbladt, 2007). Thus, using standard ToM tasks may improve not only language skills but also ToM abilities. In addition to using the standard ToM measures employed in the current investigation (e.g., false belief), other ToM assessments, such as joint attention, understanding and production of mental state terms, may also provide a window into identifying and developing language goals for treatment (see Miller, 2006 for an excellent review). Improving children’s language abilities may have a corollary influence on their social cognitive understanding. This is an important goal as well since ToM understanding affects various aspects of social development, including peer relations (Cassidy, Werner, Rourke, Zubernis, & Balaraman, 2003). A recent longitudinal study highlights the importance of improving children’s language abilities as they found that children who had a language disorder that persisted into adulthood had significant theory of mind deficits and significantly worse social adaptation compared to their non-language disordered siblings (Clegg, Hollis, Mawhood, & Rutter, 2005). Although this latter study did not directly examine correlations between language ability and theory of mind, it suggests that there is such a relationship. This study highlights the importance of understanding predictors of theory of mind understanding as there are negative effects of delays in acquiring theory of mind understanding for individuals with SLI. Further, it is possible that as children’s social cognitive development improves it may impact their subsequent language abilities as well. Several models of language development argue for the importance of social cognitive development for both semantic and grammatical development (Tomasello, 2003). In thinking past the current study, future work on language and theory of mind connections in children with language disorders should focus on enhancing the generalizability and translation of theory of mind understanding and its associated language skills to children’s everyday experiences and interactions in the classroom environment and other social settings (i.e., interactions with family members and peers). In addition, encouraging teachers to engineer interactions which place children with language disorders with typically developing children with more sophisticated linguistic and cognitive skills may also facilitate such development. Acknowledgements The authors thank the staff and teachers of Munroe Regional Medical Center’s Speech and Hearing Center and the parents and children who participated.

440

M.J. Farrar et al. / Journal of Communication Disorders 42 (2009) 428–441

Appendix A. Continuing education 1. Which of the following language skills has been shown to be related to false belief reasoning in prior research? a. vocabulary. b. grammar. c. complementation. d. language comprehension. e. all of the above. 2. Children with SLI typically do worse on theory of mind than typically developing children? True/False 3. In the current study, complementation did not make any unique contribution to theory of mind? True/False 4. False belief reasoning in children is typically achieved at what age? a. 2 years. b. 3 years. c. 4 years. d. 5 years. 5. Complementation is thought to be necessary for theory of mind because it permits the representation of false beliefs. True/False References Astington, J., & Baird, J. (2005). Why language matters for theory of mind. New York: Oxford University Press. Astington, J., & Jenkins, (1999). A longitudinal study of the relation between language and theory of mind development. Developmental Psychology, 35, 1311–1320. Baron-Cohen, S. (1995). The Eye Direction Detector (EDD) and the Shared Attention Mechanism (SAM): Two cases for evolutionary psychology. In C. Moore & P. Dunham (Eds.), Joint attention: Its origins and role in development (pp. 41–59). Hillsdale, NJ: Erlbaum. Brownell, R. (Ed.). (2000). Receptive One Word Picture Vocabulary Test (ROWPVT). Novato, CA: Academic Therapy Publications. Brownell, R. (Ed.). (2000). Expressive One Word Picture Vocabulary Test (EOWPVT). Novato, CA: Academic Therapy Publications. Bruce, B., Hansson, K., & Nettelbladt, U. (2007). Interactional style, elicitation strategies, and language production in professional language intervention. Child Language Teaching and Therapy, 23, 253–266. Bzoch, K., League, R., & Brown, V. (1991). The Receptive–Expressive Emergent Language Scale – Second Edition (REELS-2). Austin, TX: Pro-Ed. Carrow-Woolfolk, E. (1995). Oral and Written Language Scales (OWLS). Bloomington, MN: Pearson Assessments. Cassidy, K. W., Werner, R. S., Rourke, M., Zubernis, L. S., & Balaraman, G. (2003). The relationship between psychological understanding and positive social behaviors. Social Development, 12, 198–221. Cheung, H. (2006). False belief and language comprehension in Cantonese-speaking children. Journal of Experimental Child Psychology, 95, 79– 98. Cheung, H., Husan-Chih, C., Creed, N., Ng, L., Wang, S., & Mo, L. (2004). Relative roles of general and complementation language in theory of mind development: Evidence from Cantonese and English. Child Development, 75, 1155–1170. Clegg, J., Hollis, C., Mawhood, L., & Rutter, M. (2005). Developmental language disorders—A follow-up in later adult life. Cognitive, language and psychosocial outcomes. Journal of Child Psychology and Psychiatry, 46, 128–149. Dawson, J. I., Stout, C. E., & Eyer, J. A. (2003). Structured Photographic Expressive Language Test – Third Edition (SPELT-III). DeKalb, IL: Janelle Publications. de Villiers, J. (2000). Language and theory of mind: What are the developmental relationships? In S. Baron-Cohen, H. Tager-Flusberg, & D. Cohen (Eds.), Understanding other minds: Perspectives from autism and developmental cognitive neurosciences. Oxford: Oxford University. de Villiers, P. A., Burns, F., & Pearson, B. (2003). The role of language in theory of mind development of language-impaired children: Complementing theories. In Beachley, B., Brown, A., & Conlin, F. Eds. Proceedings of the 27th annual Boston University conference on language development (Vol. 1). Somerville, MA: Cascadilla Press. de Villiers, J., & de Villiers, P. (1999). Linguistic determinism and false belief. In P. Mitchell & K. Riggs (Eds.), Children’s reasoning and the mind. Hove, UK: Psychology Press. de Villers, J., & Pyers, J. (2002). Complements to cognition: A longitudinal study of the relationship between complex syntax and false-belief understanding. Cognitive Development, 17, 1037–1060. Dollaghan, C., & Campbell, T. F. (1998). Nonword repetition and child language impairment. Journal of Speech, Language, and Hearing Research, 41, 1136–1146. Dunn, L., & Dunn, P. (1997). Peabody Picture Vocabulary Test – Third Edition. Circle Pines, MN: American Guidance Services. Farrant, B., Fletcher, J., & Mayberry, M. (2006). Specific language impairment, theory of mind, and visual perspective taking: Evidence for simulation theory and the developmental role of language. Child Development, 77, 1842–1853. Farrar, M. J., & Maag, L. (2002). Early language development and the emergence of a theory of mind. First Language, 22, 197–213. Flavell, J., & Miller, R. H. (1997). Social cognition. In Kuhn, D., & Siegler, R. Eds. Handbook of child psychology (Vol. 2). NY: Wiley. Frith, U. (1989). Autism: Explaining the enigma. Oxford: Basis Blackwell.

M.J. Farrar et al. / Journal of Communication Disorders 42 (2009) 428–441

441

Gale, E., de Villiers, P., de Villiers, J., & Pyers, J. (1996). Language and theory of mind in oral deaf children. In Springfellow, A., Cahana-Amitay, D., Hughes, E., & Zukowski, A. Eds. Proceedings of the 20th annual Boston University conference on language development. Vol. 1 (pp.213– 224). Somerville, MA: Cascadilla Press . Gillott, A., Furniss, F., & Walter, A. (2004). Theory of mind ability in children with specific language impairment. Child Language Teaching and Therapy, 20, 1–11. Gopnik, A., & Astington, J. (1988). Children’s understanding of representational change and its relation to the understanding of false belief and the appearance reality distinction. Child Development, 59, 26–37. Gopnik, A., & Wellman, H. (1994). The theory theory. In L. A. Hirschfeld & S. A. Gelman (Eds.), Mapping the mind: Domain specificity in cognition and culture (pp. 257–293). New York: Cambridge University Press. Hale, M., & Tager-Flusberg, H. (2003). The influence of language on theory of mind: A training study. Developmental Science, 6, 346–359. Happe´, F. G. (1995). The role of age and verbal ability in the theory of mind task performance of subjects with autism. Child Development, 66, 843–855. Harris, P. L. (1996). Desires, beliefs, and language. In P. Carruthers & P. K. Smith (Eds.), Theories of theories of mind. Cambridge University Press. Hresko, W. P., Reid, D. K., & Hammill, D. D. (1999). Test of Early Language Development – Third Edition (TELD-3). Austin, TX: Pro-Ed. Jenkins, J. M., & Astington, J. (1996). Cognitive factors and family structure associated with theory of mind development in young children. Developmental Psychology, 32, 70–78. Leonard, L. (1998). Children with specific language impairment. MIT Press: Cambridge, MA. Lohmann, H., & Tomasello, M. (2003). The role of language in the development of false belief understanding: A training study. Child Development, 74, 1130–1144. Miller, C. A. (2001). False belief understanding in children with specific language impairment. Journal of Communication Development, 34, 73–86. Miller, C. (2004). False belief and sentence complement performance in children with specific language impairment. International Journal of Language and Communication Disorders, 39, 191–213. Miller, C. (2006). Developmental relations between language and theory of mind. American Journal of Speech Language Pathology, 15, 143–154. Milligan, K., Astington, J., & Dack, L. (2007). Language and theory of mind: Meta-analysis of the relation between language ability and false-belief understanding. Child Development, 78, 622–646. Moore, C., Bryant, D., & Furrow, D. (1989). Mental terms and the development of uncertainty. Child Development, 60, 167–171. Nelson, K. (2005). Language pathways into the community of minds. In J. Astington & J. Baird (Eds.), Why language matters for theory of mind (pp. 26–49). New York: Oxford University Press. Perner, J., Leekam, S., & Wimmer, H. (1987). 3-Year-olds difficulty with false belief: The case for a conceptual deficit. British Journal of Developmental Psychology, 5, 125–137. 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’ linguistic determinism on German-speaking children. Child Development, 74, 179–188. Perona, K., Plante, E., & Vance, R. (2005). Diagnostic accuracy of the Structured Photographic Expressive Language Test – Third Edition (SPELT3). Language, Speech, and Hearing Services in Schools, 36, 103–115. Peterson, C., & Siegal, M. (1999). Representing inner worlds: Theory of mind in autistic, deaf, and normal children. Psychological Science, 10, 126–129. Roid, G. H., & Miller, L. J. (1997). Leiter International Performance Scale – Revised. Woodale, IL: Steolting Co. Ruffman, T., Slade, L., Rowlandson, K., Rumsey, C., & Garnham, A. (2003). How language relates to theory of mind. Cognitive Development, 18, 139–158. Semel, E., Wiig, E. H., & Secord, W. A. (1992). Clinical Evaluation of Language Fundamentals – Preschool (CELF-P). San Antonio, TX: The Psychological Corporation. Seymour, H. N., Roeper, T. W., & de Villiers, J. (2003). Diagnostic evaluation of language variation. San Antonio, TX: The Psychological Corporation. Shields, J., Varley, R., Broks, P., & Simpson, A. (1996). Social cognition in developmental language disorders and high-level autism. Developmental Medicine and Child Neurology, 38, 487–495. Slade, L., & Ruffman, T. (2005). How language does (and does not) relate to theory of mind: A longitudinal study of syntax, semantics, working memory, and false belief. British Journal of Developmental Psychology, 23, 141. Spaulding, T. J., Plante, E., & Farinella, K. A. (2006). Eligibility criteria for language impairment: Is the low end of normal always appropriate? Language, Speech, and Hearing Services in Schools, 37, 61–72. Tager-Flusberg, H. (2000). Language and understanding minds: Connections in autism. In S. Baron-Cohen, H. Tager-Flusberg, & D. J. Cohen (Eds.), Understanding other minds: Perspectives from developmental cognitive neuroscience (2nd ed., pp. 124–149). Oxford: Oxford University Press. Tager-Flusberg, H., & Joseph, R. (2005). How language facilitates the acquisition of false belief understanding in children with autism. In J. W. Astington & J. A. Baird (Eds.), Why language matters for theory of mind (pp. 298–318). New York: Oxford University Press. Tomasello, M. (2003). Constructing a language. Cambridge, MA: Harvard University Press. Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theory of mind development: The truth about false belief. Child Development, 72, 655–684. Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children’s understanding of deception. Cognition, 13, 103–128. Ziatas, K., Durkin, K., & Pratt, C. (1998). Belief term development in children with autism, Asperger syndrome, specific language impairment, and normal development: Links to theory of mind development. Journal of Child Psychology and Psychiatry and Allied Disciplines, 39, 755–763. Zimmerman, I. L., Steiner, V. G., & Pond, R. E. (1992). Preschool Language Scale – Third Edition (PLS-3). San Antonio, TX: The Psychological Corporation. Zimmerman, I. L., Steiner, V. G., & Pond, R. E. (2002). Preschool Language Scales – Fourth Edition (PLS-4). San Antonio, TX: The Psychological Corporation.