Elsevier

Journal of Communication Disorders

Volume 54, March–April 2015, Pages 2-14
Journal of Communication Disorders

Trajectories of pragmatic and nonliteral language development in children with autism spectrum disorders

https://doi.org/10.1016/j.jcomdis.2015.01.001Get rights and content

Highlights

  • Children with autism have slower rates of pragmatic and nonliteral language development with age.

  • Syntax and vocabulary abilities contribute to pragmatic and nonliteral language development.

  • Theory of mind also contributes to pragmatic and nonliteral language development.

Abstract

Children with autism spectrum disorder (ASD) often have difficulties with understanding pragmatic language and also nonliteral language. However, little is understood about the development of these two language domains. The current study examines pragmatic and nonliteral language development in 69 typically developing (TD) children and 27 children with ASD, ages 5–12 years. For both groups, performance on pragmatic language and nonliteral language scores on the Comprehensive Assessment of Spoken Language increased significantly with chronological age, vocabulary, syntax, and theory of mind abilities both for children with ASD and TD children. Based on a cross-sectional trajectory analysis, the children with ASD showed slower rates of development with chronological age relative to TD children for both the pragmatic language and nonliteral language subtests. However, the groups did not show significant differences in the rate of development for either pragmatic language or nonliteral language abilities with regard to their vocabulary abilities or TOM abilities. It appears that children with ASD may reach levels of pragmatic language that are in line with their current levels of basic language abilities. Both basic language abilities and theory of mind abilities may aid in the development of pragmatic language and nonliteral language abilities.

Learning outcomes: After reading this article, the reader will understand: (1) the relation between basic language abilities (vocabulary and syntax) and advanced language abilities (pragmatic and nonliteral language), (2) how the cross-sectional trajectory analysis differs from traditional group matching studies, and (3) how pragmatic and nonliteral language development for children with autism shows both similarities and differences compared to typically developing children.

Introduction

Individuals with autism spectrum disorders (ASDs) have difficulties with the development of social and communicative abilities. However, for domains of language development, some theories (such as the relevance theory; e.g., Happé, 1993) suggest that some areas of language (especially the domains of pragmatic language or nonliteral language) may be disproportionately impacted for those individuals with autism who do show considerable development in their vocabulary and/or syntax abilities. For typically developing (TD) children, skills in both nonliteral language and pragmatic language continue to improve throughout childhood (Rundblad and Annaz, 2010a, Ryder and Leinonen, 2003). Some research suggests that as a group, children with autism spectrum disorder (ASD) often have a difficult time understanding both pragmatic language (Lam and Yeung, 2012, Tager-Flusberg and Anderson, 1991) and nonliteral language (Rundblad & Annaz, 2010b). However, individuals with ASD show great variability in the rate of language acquisition across development (Tek, Mesite, Fein, & Naigles, 2014). Understanding the predictors of progress in pragmatic language and nonliteral language development for individuals with autism may be crucial for creating appropriate individualized targets for intervention (Klin et al., 2007).

Comprehension of pragmatic language and nonliteral language share a common feature in that they require children to not only understand the individual meaning of words embedded in a sentence structure, but also to understand and respond appropriately to the communicative intent of language embedded in social and linguistic contexts (McTear & Conti-Ramsden, 1992). The specific definition of pragmatic language versus nonliteral language (as well as the boundary between these domains) varies between research studies, as well as the tasks used to assess abilities in these domains. The current study examines predictors of the development of pragmatic language and nonliteral language abilities using two subscales from the Comprehensive Assessment of Spoken Language (CASL; Carrow-Woolfolk, 1999). Pragmatic language in this case refers to the understanding and use of the literal aspects of context during communication (e.g., understanding the socially appropriate use of language for relevant contexts, such as: greetings, expressions of gratitude, making direct requests and answering questions). Nonliteral language refers to use of language where there is a specific mismatch between the literal meaning of the individual words of the phrase and the expected interpretation (e.g., “he was such a turtle” means that “he was slow” and not “he was a reptile with a shell”). Types of nonliteral language include figurative language (e.g., metaphors, such as “the girl was a butterfly”), sarcasm (e.g., saying “good job” when someone does poorly), and indirect requests (e.g., “I want all eyes on the board”). Importantly, various aspects of nonliteral language may vary in how close the intended meaning of the phrase is to the individual words.

The majority of research on both pragmatic language and nonliteral language abilities in children with ASD has used group-matching designs (often averaging across large age ranges). However, the overall heterogeneity in language abilities in children with ASD can be problematic for group matching designs (Tager-Flusberg, 2004). In particular, little is known about the development with age of these abilities in children with ASD. In addition, little is known about how vocabulary and/or syntax development relates to either pragmatic language or nonliteral language development for children with ASD, and whether or not these developmental relations differ between children with ASD and TD. Research has recently begun to use more developmentally sensitive approaches to examining language abilities for children with autism (e.g., Rundblad & Annaz, 2010b), but there are still remaining questions about the relative progress in development of pragmatic language and figurative language. The current study uses a cross-sectional developmental trajectory analysis (Thomas et al., 2009) to examine how pragmatic language and nonliteral language abilities develop in children with ASD and TD with regard to their age, syntax, and vocabulary abilities.

Pragmatic language abilities develop across childhood and into adolescence for TD children (Nippold, 2000, Ryder and Leionen, 2014). For example, Ryder and Leinonen (2003) suggest that 5-year-old children are able to understand and answer more complex questions than 3- or 4-year-old children based on their increasingly more complex ability to use contextual information. Lokusa, Leinonen, & Ryder (2007) examined pragmatic language abilities in children, ages 3–9, and found that while three-year-old children were able to use some context for resolving potential ambiguity when answering questions, the use of context increased substantially among the older children. Research suggests that children with ASD have difficulties in their pragmatic language development, including difficulties with the use language in the context of social conversations (Loukusa and Moilanen, 2009, Reichow et al., 2008, Volden et al., 2009, Young et al., 2005). Lam and Yeung (2012) found that children and adolescents with ASD, ages 8–15 years, performed worse than individuals with TD when using an observational scale to measure the pragmatic language abilities, such as having more instances of “out-of-synchrony communicative behavior” (Lam & Yeung, 2012). Children with ASD do show some developmental progress in pragmatic abilities. Specifically, Loukusa, Leinonen, Kuusikko, et al. (2007) found that children with ASD who were 10–12 years old performed better at pragmatic language measures than children with ASD who were younger (7–9 years), suggesting that pragmatic language abilities and the use of context increases with age and experience for individuals with ASD.

Nonliteral language comprehension also shows a protracted developmental period across childhood in typical development. Some TD children, ages 4–6 years, are able to demonstrate an emerging knowledge of metaphors, with older children and adolescents showing greater levels of metaphor comprehension (Le Sourn-Bissaoui et al., 2012, Rundblad and Annaz, 2010a, Vosinadou and Ortony, 1983, Winner et al., 1976). Additionally, comprehension of indirect requests (Bernicot et al., 2007, Ledbetter and Dent, 1988) and sarcasm (Glenwright & Pexman, 2010) also increase with age for TD children. Children with ASD may also have difficulties with understanding various aspects of nonliteral language compared to their same-age typically developing peers (Dennis et al., 2001, MacKay and Shaw, 2004, Martin and McDonald, 2004, Rundblad and Annaz, 2010b). Research has found that children with ASD (diagnosed with Asgerger's syndrome or high functioning autism) perform worse than TD children on measures of metaphor comprehension (Dennis et al., 2001, Nikolaenko, 2004). In addition, MacKay and Shaw (2004) found that children with Asgerger's syndrome showed difficulties on a variety of nonliteral language measures (including irony, metonymy, and indirect requests) compared to TD children. While many studies find differences, at the group-level, in pragmatic and nonliteral language abilities between TD children and children with ASD, important questions still remain about the development of these language abilities both in individuals with TD and ASD.

Factors supporting understanding of pragmatic language and nonliteral language for children have been examined in the literature, including the potential contributions of basic language abilities and theory of mind (TOM) for both language domains, though these relationships have been more commonly studied with regard to nonliteral language. For TD children, there is some evidence that language abilities develop together in parallel, with basic language abilities (such as vocabulary) correlating strongly with nonliteral language abilities (Johnson, 1991, Rundblad and Annaz, 2010a). Research suggests that understanding of nonliteral language is supported by the ability to use the linguistic and social context to abstract meaning, and that vocabulary and syntax abilities are likely to support this process (Gernsbacher and Pripas-Kapit, 2012, Rundblad and Annaz, 2010a, Vosniadou, 1987). With regard to pragmatic language, Volden et al. (2009) found that a composite of expressive and receptive basic language abilities predicted a large portion of the variance in pragmatic language ability for children with ASD. Eisenmajer and Prior (1991) found that pragmatic language abilities correlated with several basic language measures, supporting the relationship between basic language abilities and pragmatic language. There is, however, some controversy in the literature as to whether or not nonliteral language abilities are disproportionately impacted for children with ASD, beyond their potential differences in basic language abilities. For example, Rundblad and Annaz (2010b) found that receptive vocabulary abilities only predicted performance on metonyms, but not metaphors, in their sample of 10 children with ASD.

The relevance theory has driven research focused on the potential role of theory of mind for contributing to pragmatic language (Le Sourn-Bissaoui et al., 2011, Loukusa and Moilanen, 2009) difficulties for children with ASD. Loukusa and Moilanen (2009) suggest that receptive vocabulary skills alone may be insufficient for understanding various aspects of pragmatic language for individuals with ASD, due to their likely difficulties with regard to the use of the social and linguistic context for understanding the meaning of these utterances. This line of research largely suggests that individuals with ASD struggle with these more advanced language abilities, even when they have good basic language abilities (e.g., receptive vocabulary), potentially due to weaknesses in utilizing the social and linguistic context for understanding the meaning of these utterances (Loukusa and Moilanen, 2009, Noens and van Berckelaer-Onnes, 2005).

Thus, limitations with theory of mind abilities have been implicated in difficulties with understanding nonliteral language, especially with regard to understanding metaphors (Happé, 1993). While primarily examined with regard to performance on first- or second-order false belief measures (e.g., Happé, 1993), there have overall been mixed results with using false belief measures with older children, with some studies failing to find a unique contribution of false belief performance on nonliteral language abilities (e.g., Norbury, 2005, Rundblad and Annaz, 2010b). Other advanced measures of theory of mind, that are more developmentally sensitive across childhood, may instead better elucidate these relations. For example, performance on the children's “reading the mind in the eyes” task (RMTE; Baron-Cohen, Wheelwright, Spong, Scahill, & Lawson, 2001) is correlated with idiom comprehension abilities (e.g., understanding phrases such as “raining cats and dogs”) in children with autism (Whyte, Nelson, & Scherf, 2014), but it is unknown how this measure relates to measures of pragmatic language or nonliteral language abilities.

Other research suggests that receptive vocabulary may be an area of relative strength for individuals with ASD when compared to their syntax abilities (Eigsti, Bennetto, & Dadlani, 2007). It is possible that syntax abilities, specifically, are predictive of their relative pragmatic (Volden et al., 2009) and nonliteral language abilities (Gernsbacher and Pripas-Kapit, 2012, Norbury, 2004, Norbury, 2005, Whyte et al., 2013) for individuals with ASD, rather than a diagnosis of ASD in particular. For nonliteral language, Norbury (2005) found that syntax comprehension is a better predictor of metaphor comprehension than receptive vocabulary in particular, for children with language impairments, regardless of whether or not they exhibited social impairments associated with ASD. Additionally, children with ASD (especially those with good syntax abilities) were able to benefit from the use of context in understanding unfamiliar idiomatic phrases (Norbury, 2004). This line of research overall suggests that some children with ASD may not show impairments in pragmatic language or nonliteral language if they show significant progress in the development of their basic language abilities (Gernsbacher and Pripas-Kapit, 2012, Norbury, 2004, Norbury, 2005, Whyte et al., 2014). Thus, understanding the role of expressive vocabulary and syntax abilities in supporting the development of pragmatic language and nonliteral language may be important for choosing appropriate matching variables to prevent biased conclusions about the nature of language abilities for individuals with ASD. Further, any theoretical account of language development in children with TD as well as those with ASD needs to take into account how the varied language domains may develop together across childhood, and how this also may interact with other domains, such as theory of mind.

One approach to examining potential relations between language domains across development, arising from the neuroconstructivist theory, is the cross-sectional developmental trajectory analysis (Thomas et al., 2009). The developmental trajectory analysis described by Thomas et al. (2009) allows for examining differences in the rate of growth of language abilities across development, even for cross-sectional samples, using ANCOVA models to directly compare the potential differences in the intercepts and slopes of the trajectories. It is possible to look at age-related trajectories for a group of children with ASD compared to TD children to see if the children with ASD have a different rate of development with regard to chronological age. It is also possible to examine the rate of development in a behavior as a function of various possible predictors, such as vocabulary or syntax (Annaz et al., 2008, Thomas et al., 2009). This analysis approach can detect group differences in the intercept of the trajectory (defined as differences at the youngest point of overlap in scores on the predictor variable between the groups), slowed rate of development (defined as differences in the slopes of the trend lines between groups), or both (Annaz et al., 2008, Thomas et al., 2009) for each potential predictor variable. This stands in contrast to traditional group matching approaches that are only sensitive to overall group differences.

For group matching approaches, the choice of matching variables needs to be theory-driven, as improperly matched groups runs a risk of leading to biased conclusions about the nature of symptoms in autism (Gernsbacher and Pripas-Kapit, 2012, Jarrold and Brock, 2004, Thomas et al., 2009). However, the trajectory approach does not require this same a priori group matching. Instead of controlling for differences in vocabulary and other potential matching variables between groups, this approach can tell us important information about how these variables may actually provide important contributions to the rate of development of pragmatic and nonliteral language abilities. While this approach does not reduce the need for longitudinal research, it is more developmentally sensitive than group matching designs, and can potentially inform future longitudinal research (Annaz et al., 2008, Thomas et al., 2009).

This cross-sectional developmental trajectory approach has been previously used in several studies to examine the development of vocabulary, nonliteral language, and face processing abilities of children with ASD (Annaz et al., 2009, Kover et al., 2013, Rundblad and Annaz, 2010b). For example, Rundblad and Annaz (2010b) used a developmental trajectory analysis to examine two different aspects of nonliteral language (metaphors and metonyms) in 11 children with ASD, ages 5–11 years. For the children with ASD, receptive vocabulary predicted performance on metonyms, but not metaphors (Rundblad & Annaz, 2010b). However, additional measures of language abilities (including syntax abilities and expressive vocabulary abilities) and larger sample sizes are needed to further investigate development of nonliteral language for children with ASD.

The current study uses the trajectory analysis methods of Thomas et al. (2009) to examine how age, vocabulary, syntax, and theory of mind predict performance on the pragmatic language and nonliteral language subtests of the Comprehensive Assessment of Spoken Language (CASL; Carrow-Woolfolk, 1999) for children with ASD and TD. The two subtests from the CASL (pragmatic language and nonliteral language) were chosen due to their potential use in measuring nonliteral and pragmatic language in clinical settings (e.g., Reichow et al., 2008). Additionally, these outcome measures are sensitive across a wide age range of development and are unlikely to show ceiling or floor effects for either individuals with TD or ASD.

For the current study, the age range of 5–12 years was chosen such that it would be likely that TD children aged 5–6 may perform poorly on these measures, with increasing mastery of pragmatic and nonliteral language with increasing age for the TD children. Thus, it is hypothesized that there may not be a difference in the youngest age of overlap between groups for the current sample (representing the intercepts of the trajectories). The primary outcome of interest, instead, is the comparison of the rate of development between groups for each trajectory. The current study hypothesizes that children with ASD will have slower rates of development of nonliteral and pragmatic language trajectories from the TD group when examining chronological age, but the two groups will not have significantly different nonliteral or pragmatic language trajectories when examining relations with syntax age-equivalence scores, vocabulary age-equivalence scores, or TOM scores. Additionally, it is predicted that TOM abilities will significantly relate to individual differences in pragmatic and nonliteral language abilities, above and beyond the contributions of basic language abilities.

Section snippets

Children with ASD

A total of 26 children (21 m, 5 f), ages 5–12 (M = 9.07, SD = 1.87) years, who were previously diagnosed with an autism spectrum disorder (ASD), participated in the study (see Table 1). Highly verbal children were included if they had a previous diagnosis by a clinician under the DSM-IV criteria (American Psychiatric Association, 1994) of autism or high-functioning autism (n = 8), Asperger's syndrome (n = 13), or pervasive developmental disorder (PDD-NOS; n = 5), based on parent report. The current study

Pragmatic language trajectories

Pragmatic language scores from the CASL increased with chronological age, both for TD children, R2 = .66, F(1,66) = 126.20, p < .001, and children with ASD, R2 = .32, F(1,23) = 10.72, p < .01. Comparing the groups, the TD children and children with ASD did not significantly differ at the youngest age of overlap between groups (representing no differences in the intercept of the trajectories), p > .05. Children with ASD exhibited a slower rate of development for pragmatic language with chronological age

Discussion

The current study used a cross-sectional trajectory approach, following the methods of Thomas et al. (2009), to examine the development of pragmatic language and nonliteral language (metaphor, sarcasm, indirect requests) for children with ASD and TD children. For both TD children and children with ASD, ages 5–12 years, chronological age, basic language abilities (syntax and vocabulary), and TOM abilities, were significantly predictive of raw scores on both the pragmatic language and nonliteral

Funding

A liberal arts dissertation improvement award from Penn State provided funding for this research.

Conflict of interest

None declared.

References (54)

  • S. Baron-Cohen et al.

    Are intuitive physics and intuitive psychology independent? A test with children with Asperger syndrome

    Journal of Developmental and Learning Disorders

    (2001)
  • E. Carrow-Woolfolk

    Comprehensive assessment of spoken language

    (1999)
  • J.N. Constantino

    Social Responsiveness Scale

    (2002)
  • M. Dennis et al.

    Inferential language in high-functioning children with autism

    Journal of Autism and Developmental Disorders

    (2001)
  • I. Eigsti et al.

    Beyond pragmatics: Morphosyntactic development in autism

    Journal of Autism and Developmental Disorders

    (2007)
  • R. Eisenmajer et al.

    Cognitive linguistic correlates of ‘theory of mind’ ability in autistic children

    British Journal of Developmental Psychology

    (1991)
  • M.A. Gernsbacher et al.

    Who's missing the point? A commentary on claims that autistic persons have a specific deficit in figurative language comprehension

    Metaphor and Symbol

    (2012)
  • M. Glenwright et al.

    Development of children's ability to distinguish sarcasm and verbal irony

    Journal of Child Language

    (2010)
  • C. Jarrold et al.

    To match or not to match? Methodological issues in autism-related research

    Journal of Autism and Developmental Disorders

    (2004)
  • J. Johnson

    Developmental versus language-based factors in metaphor interpretation

    Journal of Educational Psychology

    (1991)
  • A.S. Kaufman et al.

    Kaufman Brief Intelligence Test

    (2004)
  • A. Klin et al.

    Social and communication abilities and disabilities in higher functioning individuals with autism spectrum disorders: The Vineland and the ADOS

    Journal of Autism and Developmental Disorders

    (2007)
  • S.T. Kover et al.

    Receptive vocabulary in boys with autism spectrum disorder: Cross-sectional developmental trajectories

    Journal of Autism and Developmental Disorders

    (2013)
  • P.J. Ledbetter et al.

    Young children's sensitivity to direct and indirect request structure

    First Language

    (1988)
  • C. Lord et al.

    A multisite study of the clinical diagnosis of different autism spectrum disorders

    Archives of General Psychiatry

    (2012)
  • C. Lord et al.

    Autism Diagnostic Observation Schedule – WPS (ADOS-WPS)

    (1999)
  • S. Loukusa et al.

    Use of context in pragmatic language comprehension by children with Asperger syndrome or high-functioning autism

    Journal of Autism and Developmental Disorders

    (2007)
  • Cited by (0)

    View full text