For bilingual families of children diagnosed with autism spectrum disorder (ASD), it can be difficult to know what language or languages they should use and encourage their children to use (Hambly and Fombonne 2012; Ohashi et al. 2012; Yu 2013). As a result, families are faced with many challenging questions, including the following: Do families concentrate on one language and which one do they choose? Or, do they provide access to two or even more languages? Is it better to focus on a home language or the language that will be used by the majority of people in their child’s world, including in education? What part should practitioners play in contributing to this decision? These questions are pertinent for bilingual families of children with ASD because social communication impairment has been identified as one of its core features (American Psychiatric Association 2013). Some children with ASD will not develop any spoken language, while some will have a restricted range of verbal communication skills, and yet others may experience subtle difficulties in the use of spoken language (Osterling et al. 2001; Prizant and Wetherby 2005).

Speaking two or more languages is the norm for nearly two thirds of the world’s population (Crystal 2006). In Canada, for example, both English and French are national languages and there has been a growth in its migrant population. As a result, nearly 6.6 million people speak a language other than English or French at home (Statistics Canada 2011). Similar trends have been identified in other Western countries. For example, estimates from the 2008–2012 American Community Survey (United States Census Bureau 2012) indicated 20.5 % of people used a language other than English at home, the Australian Bureau of Statistics (2012) reported that 20 % of second-generation migrants speak a language other than English at home, and the 2013 New Zealand census reported Hindi as the fourth most spoken language, which has tripled since 2001 (Statistics New Zealand 2013). However, data on ethnicity and ASD have at times been conflicting, with some literature suggesting that there is no difference in prevalence of ASD across ethnicities, while other studies suggest an increased prevalence of ASD in immigrant families compared to nonimmigrant families (Dyches et al. 2004). Dyches et al. caution that there has been a lack of rigor in the differentiation of ethnicity, immigrant, and minority status. Furthermore, little research has analyzed ethnicity in the ASD population. Using data from the Minnesota Ministry of Education, Estrem and Zhang (2010) found an increase in the prevalence of students with ASD across all language groups with the highest increase found in Somali students compared with English, Spanish, Mhmong, and other language speakers. Together, these statistics suggest an increase in the number of children and families from culturally and linguistically diverse backgrounds in Western countries. Given this, it is perhaps not surprising that families and professionals endeavor to make the right decision regarding language use for children with ASD. There is a need for high-quality educational and clinical service provision for this growing population of clients (Soto and Yu 2014).

Bilingualism has been defined as the use of two or more languages (American Speech Language Hearing Association [ASHA] 2014; Baker 2011; Bhatia and Ritchie 2006). Most definitions also discuss the complexity of language use including language contexts and proficiency (Baker 2011; Bhatia and Ritchie 2006). Bilingualism is further defined by onset of exposure to languages with simultaneous (SIM) bilingualism occurring from birth and sequential (SEQ) exposure occurring after the first three years (ASHA 2014).

In the past, there has been considerable debate surrounding the impact of bilingualism on general educational outcomes for children with typical development (Greene 1997; Rossell and Baker 1996). While details of the debate are far beyond the scope of this paper, recent research has suggested that supporting both (or multiple) languages in the preschool years is related to cognitive, academic, and social advantages (Kohnert 2010; Soto and Yu 2014). Bilingual children score as well as and in some cases better than monolingual children across domains, including cognitive and reasoning tasks (Barac and Bialystok 2012; Kapa and Columbo 2013; Kovaks 2009), mathematical tasks (Han 2012; Leikin 2013), reading outcomes (Davison et al. 2011), and language skills (David and Wei 2008; MacLeod et al. 2011; Nicholls et al. 2011; Unsworth 2013).

Although there is considerable evidence to support bilingualism in children with typical development, parents and professionals are still fearful of speaking more than one language with children who have specific language impairments and other developmental disabilities (Yu 2013). Professionals working in the area of communication and language disorders have not always felt competent with how to work with bilingual families (D’Souza et al. 2012; Moore and Perez-Mendez 2006). D’Souza and colleagues found that while many speech-language pathologists in Canada work with families from different linguistic backgrounds, more than half of them only work with the families in the pathologist’s own language. ASHA (2012) reported that less than 5 % of speech-language pathologists certified by their organization speak a second language in addition to English. Practitioners are often advising parents to concentrate on one language in order to support a child’s language development (Moore and Perez-Mendez 2006).

Contrary to this, there is a growing body of research suggesting a bilingual approach to working with children with communication impairments (Barrera 1993; Gutierrez-Clellen et al. 2008; Hand 2011; Kay-Raining Bird et al. 2005; Kay-Raining Bird and Trudeau 2009; Paradis 2007; Paradis et al. 2003). For example, Kay-Raining Bird et al. (2005) compared four groups: (a) children with Down syndrome being raised in a bilingual environment, (b) children with Down syndrome being raised in a monolingual environment, (c) children with typical development being raised in a bilingual environment, and (d) children with typical development being raised in a monolingual environment. They did not find any evidence of a negative effect for children with Down syndrome being raised in a bilingual environment. With results like this, Soto and Yu (2014) suggest there is no evidence to support the idea that bilingualism is confusing or interferes with language-learning abilities of children with communication impairment; learning the home language does not impair or significantly delay learning of a second language.

However, it cannot be assumed that research on bilingualism in children with communication impairments will translate to children with ASD (Dyches et al. 2004). Impairments in social interaction skills and the presence of restricted, repetitive, and stereotyped behaviors in children with ASD impact and potentially complicate their communication and language development (Wilkinson 1998). With little research specific to bilingualism in children with ASD, professionals struggle to develop informed language recommendations for these children. As a result, families often do not know where to seek information on bilingualism for their child with ASD. National ASD guidelines and organizations in New Zealand (e.g., http://www.autismnz.org.nz ; Ministries of Health and Education 2008), Canada (http://www.autismsocietycanada.ca), the UK (http://www.autism.org.uk), and Australia (http://www.autismspectrum.org.au) do not provide information or support for bilingual families of children with ASD. In the USA, Autism Speaks (http://www.autismspeaks.org) does not provide specific advice on bilingualism but does list bilingual providers, resources, and some research.

With this lack of knowledge and understanding, the objective of the present systematic review was to provide an up to date assessment of the extant literature on bilingualism in children with ASD in order to facilitate and inform evidence-based practice. Descriptions of the studies are provided with particular attention given to the evidence supporting bilingual language development and advice given by practitioners to families about whether to adopt the use of one or more languages. The review also endeavors to identify gaps in the existing literature, which might guide future research investigating bilingual language development in children with ASD.

Methods

A systematic search was conducted to identify articles that provided a language assessment, intervention, and/or parental perceptions concerning second/additional language learning in children with ASD. Studies that met the inclusion criteria were summarized in terms of participants, languages and communication level, assessment/intervention, instruments, findings, and evidence of bilingual language development.

Search Procedures

An initial search of three electronic databases (ERIC, ProQuest, and PsycINFO) was undertaken in August 2012. These databases were searched using a combination of the following free-text terms (with truncation and Boolean operators): bilingual, multilingual, multiple language, autism spectrum disorders (ASD), autism, and Asperger. The search was limited to English-language, peer-reviewed journal articles. Therefore, theses or dissertations were not included in the review. No other restrictions (e.g., publication date) were utilized. To increase the likelihood that all relevant studies were identified, the following two additional search strategies were applied: (a) reviewing the reference lists of articles identified in the database search and (b) searching the databases for additional studies by authors of studies identified in the initial search. The search was updated in July 2013 and February 2014.

Screening and Inclusion Criteria

To be included in the review, the study had to provide a language assessment and/or intervention for children diagnosed with ASD growing up in a bilingual family. Studies that evaluated parental perceptions of raising their child with ASD in a bilingual environment were also included. Bilingual was defined as the use of two or more languages spoken in the home environment. The definition of assessment included the use of either standardized or nonstandardized measures of language ability, parent reports of language ability assessed through questionnaire and/or interview, and observation by the researchers of language ability. Intervention was defined as the teaching of two or more languages spoken in the home to an individual with ASD. Parental perceptions had to be evaluated through interviews and/or questionnaires. The language assessment and/or intervention had to involve spoken language. Children with ASD identified as communicating with augmentative and alternative communication (AAC) systems (e.g., manual signing) as an additional/second language were excluded from the review.

Twenty-one studies were identified through the search strategy. The abstracts for these articles were screened against the inclusion criteria by two independent reviewers (second and third authors), resulting in seven studies retained for review. Agreement as to whether or not the studies met the inclusion criteria was 100 %. The second author independently replicated this search and screening process in July 2013, which resulted in the identification and subsequent inclusion of one additional study. Another update of the search procedures in February 2014 did not identify any further studies. This process resulted in a total of eight studies included for review.

Data Extraction and Coding

The eight studies that met the inclusion criteria were summarized in terms of the following: (a) participants (family background, number, age in years, diagnosis), (b) languages spoken at home and communication level, (c) assessment/intervention procedures, (d) instruments used, (e) main findings, and (f) evidence of bilingual language development.

Data was extracted from the included studies by the second author and checked by an independent rater (first author) for accuracy. To check the accuracy of extraction and the resulting summary for each study, the first author followed a checklist that included six questions:

  1. 1.

    Is this an accurate description of the participants?

  2. 2.

    Is this an accurate description of the languages and communication level?

  3. 3.

    Is this an accurate description of the assessment/intervention?

  4. 4.

    Is this an accurate description of the instruments?

  5. 5.

    Is this an accurate description of the main findings?

  6. 6.

    Is this an accurate description of evidence of bilingual language development?

This approach has been used been used elsewhere (e.g., Sigafoos et al. 2009) as a measure of inter-rater agreement when extracting data. In terms of inter-rater agreement, there were 48 items on which there could be agreement or disagreement (8 studies by 6 questions per study). Agreement was obtained on 45 items (94 %). In the three cases of disagreement on extracted data, the articles were re-examined until consensus (100 % agreement) between the first and second author regarding accuracy was reached. Changes were then made as a result of the consensus reached. Specifically, description of the assessment/intervention in the Hambly and Fombonne (2012) study was clarified; the main findings in the Petersen et al. (2012) study were clarified, and description of a Likert scale was added to the instruments section in the Kay-Raining Bird et al. (2012) study.

Results

Table 1 provides a summary of the eight studies that involved language assessment/intervention of children with ASD from bilingual families and/or parental perceptions surrounding bilingualism. In Table 1, the studies are categorized by focus of the study, listed by the authors, and summarized in terms of (a) participants, (b) languages spoken and communication level, (c) assessment/intervention procedures, (d) instruments used, (e) main findings, and (f) evidence of bilingual language development. A narrative summary of findings and recommendations regarding the adoption of bilingualism from each study, including an evaluation of the quality of the research methodologies utilized, is provided below to supplement the summaries provided in Table 1.

Table 1 Summary of the key features for each of the eight studies
Full size table

Studies Focused on Comparing Monolingual and Bilingual Language Development

Four articles compared the language skills of monolingual and bilingual children with ASD. Hambly and Fombonne (2012) found significantly different amounts of bilingual exposure across all groups (p ≤ 0.001) and significantly stronger social interaction scores in the SIM group compared to SEQ on the Vineland-II (Sparrow et al. 2005) Interpersonal subdomain (p = 0.025), but no significant group differences in language level. However, SIM and MON children consistently scored higher than SEQ children. The first language scores were better than the second language scores. Based on the results, the authors concluded that children with ASD exposed to more than one language do not experience additional language delays when compared to monolingual children. Hambly and Fambonne suggested that, because there is little research in the field of bilingual language development for children with ASD, parents should base their decision on whether to provide a bilingual environment on their child’s language experience and history.

Ohashi et al. (2012) found no statistically significant differences between monolingual and bilingual groups in terms of ASD-related communication impairment, age of first words, age of first phrases, receptive language scores, expressive language scores, and functional communication scores. The authors concluded that a bilingual environment does not have an adverse impact on the early language development of children with ASD. Ohashi et al. provided support for bilingual exposure, but recommended that individual child characteristics should be taken into account.

Results of Petersen et al. (2012) indicated bilingual and monolingual groups had equivalent scores in all but one language measure of language and vocabulary, including English production vocabulary, conceptual production vocabulary, and vocabulary comprehension. When the two languages of the bilingual group were compared, there were no significant differences in production vocabulary size or vocabulary comprehension scores. English scores tended to be higher than Chinese scores. When nonverbal IQ (NVIQ) was controlled for, the bilingual group had larger total production vocabularies than the monolingual group. Conceptual vocabulary and English vocabulary sizes were equivalent to those of the monolingual children. They suggested bilingual English-Chinese preschool children with ASD have the capacity to function successfully as bilinguals. The authors concluded that bilingualism does not have a negative impact on the language development of young children with ASD. Peterson et al. suggested that when a child is diagnosed with ASD, bilingual exposure should not be limited and that it is appropriate to use the native language at home and the other language elsewhere.

Valicenti-McDermott et al. (2013) reviewed the database of a university-affiliated developmental center for children under 3 years diagnosed with ASD from 2003 to 2010. The records of bilingual English-Spanish and monolingual English children were analyzed for data on qualitative and quantitative language skills. They found no significant difference between groups in expressive language (number of words, presence of word combinations, babbling, or vocalization). Bilingual children cooed more frequently than monolingual children. For communicative means, bilingual children used gestures including pointing, leading to desired objects, and pretend play more than monolingual children. There were no significant differences for receptive language. The authors concluded bilingual children are not more prone to additional language impairments than monolingual children. Valicenti-McDermott et al. did not specify any recommendations regarding whether parents should provide a bilingual environment, but did suggest that bilingualism does not seem to confer vulnerability.

Although these studies used sound methodologies, including rigorous selection criteria for participants and attempts to ensure equivalency between groups (e.g., matched for age), use of systematic tools for assessing language, and appropriate statistical analyses to obtain group comparisons, findings should be interpreted as suggestive rather than definitive. The studies were primarily exploratory in nature and various potential limitations of the studies might have confounded results. Small sample sizes were acknowledged (Ohashi et al. 2012; Peterson et al. 2012) as well as the use of a retrospective methodology (Valicenti-McDermott et al. 2013), and caution with interpreting any causal nature of the results was highlighted (Hambly and Fombonne 2012; Ohashi et al. 2012). Each study only provided a snapshot of differences in language skills between monolingual and bilingual children with ASD at one point in time.

All of the studies matched groups in terms of age of the participants, and various other methods of equivalency were utilized, including gender (Hambly and Fombonne 2012; Valicenti-McDermott et al. 2013) and nonverbal IQ (Ohashi et al. 2012). However, none of the studies matched participants in terms of ethnicity and language/s of exposure. Each study used a different definition to distinguish SIM and SEQ bilingual exposure, including SIM being defined as occurring before 1 year (Hambly and Fombonne 2012), before 2 years (Ohashi et al. 2012), and before 3 years (Petersen et al. 2012). Valicenti-McDermott et al. did not make a distinction between SIM and SEQ bilingualism. Ohashi et al. only implemented measures in one language, meaning they might have underestimated the language ability of bilingual children. They acknowledged that when all languages are taken into account, bilingual children have been found to have equal or larger vocabularies than their monolingual peers (Pearson et al. 1993). Hambly and Fombonne and Peterson et al. controlled for this by assessing in each language and computing total conceptual vocabulary (TCV), whereby concepts named in both languages were mapped together so that translation equivalent word forms were only counted once (even if they were used in two languages). It was inferred that Valicenti-McDermott et al. assessed in the dominant language only because they stated that a speech-language therapist established each child’s dominant language.

The studies also used a range of criteria for ASD diagnosis, including (a) diagnostic reports (Hambly and Fombonne 2012), (b) the ADOS [Lord et al. 2003] and ADI-R [Lord et al. 1994] (Hambly and Fombonne 2012; Ohashi et al. 2012; Petersen et al. 2012; Valicenti-McDermott et al. 2013), (c) CARS [Schopler et al. 1988] (Petersen et al. 2012; Valicenti-McDermott et al. 2013), and (d) DSM-IV [American Psychiatric Association 2000] (Petersen et al. 2012; Valicenti-McDermott et al. 2013). Finally, a total of 16 different measures were used to assess language skills in children across these four studies (see Table 1).

Studies Focused on Bilingual Language Intervention

One case study was aimed at teaching a second language to a child with ASD (Seung et al. 2006). The speech-language therapist was Korean-English bilingual and conducted a bilingual Korean-English speech-language intervention. Over the intervention period, the child made gains in receptive and expressive language in both languages and was able to complete the final assessment (after 24 months) in English. He followed simple verbal instructions, made two- to three-word combinations to request activities, greeted clinic staff, asked WH questions frequently, and increased in nonverbal communication. There was also a reduction of problem behavior and increased eye contact and social affect. The study demonstrated how an intervention conducted in two languages can develop the bilingual language skills of a child with ASD. Seung et al. recommend that bilingual language development is appropriate for children with ASD and that practitioners should work in culturally appropriate ways with bilingual clients.

However, results of this case study should be interpreted with caution because it did not employ an experimental design or any control which limits the ability to attribute participant gains to the intervention. Furthermore, the intervention was implemented with one participant only preventing generality of the findings.

Studies Focused on Issues Surrounding Bilingualism

Three studies investigated issues surrounding bilingualism and ASD. Jegatheesan (2011) investigated the perceptions of three Muslim families regarding raising a child with ASD in multiple languages and the challenges involved. The parents stated each language used was important and they served specific cultural and religious purposes. English was viewed as a passport to life in mainstream society and parents felt it was important for a better quality of life. It enabled children to survive school and community life. The native languages were also viewed as important because they were vital for communication with family and elders who may not speak English and for family social life. Arabic, in particular, had a significant importance as parents expected their children to be able to take part in religious ceremonies undertaken in Arabic only. Jegathessan described a parent-professional collusion about multiple language use at home, in which all parents reported receiving professional advice against using multiple languages at home. They were advised to only use one language and that was English. Professionals thought children would become confused with multiple language use. The author reported, however, that the children made notable improvements in language acquisition and communication skills in more than one language. The study varied in length for each family ranging from 1.4 to 1.9 years. All children were encouraged to speak English in mainstream society, and one participant (Raqib) was explicitly identified as receiving speech therapy in English. Native language use was encouraged at home. Jalil was nonverbal prior to the study. At end of the study, he had full sentences in Bangla, English, and Arabic. Raqib had 12–15 words in Urdu and English prior to the study. At the end of the study, he had full sentences in English and Arabic, and two-word sentences in Urdu. Aziz was nonverbal prior to the study. At the end of the study, he had expressive vocabulary of 15–20 words in English, Urdu, and Hindi, and could imitate simple words.

Kay-Raining Bird et al. (2012) surveyed parents or guardians of children with ASD in a bilingual family. Children raised as monolingual versus bilingual had equivalent levels of ASD severity and did not differ on age or any language or literacy rating of the first language including language comprehension, production, reading, or writing abilities. The choice to raise the child with more than one language seemed to be influenced by how important the carers felt bilingualism was and what professional advice they were given. Carers reported concerns with regard to access to services or professional help and the child’s ability to learn two languages; they were afraid children would be confused. The majority of carers were advised to restrict input to a single language for children with ASD (43 % = do not use two languages; 16 % = use two languages). Some carers using multiple languages were advised to use different languages in different contexts (people, places, and times). Others reported receiving conflicting advice from practitioners. On the other hand, carers who felt bilingualism was important reported living in a bilingual city/country, the ability to communicate with family/school/neighbors, and the provision of life opportunities as reasons to teach two languages to their child. The majority of carers felt they had some success in raising their child with ASD bilingually.

Yu (2013) investigated issues regarding language practices of Chinese-English bilingual mothers and their children with ASD. Mothers reported that whether their children learned Chinese or English did not matter as long as the language(s) led to success in school and in life. For 9/10 mothers, English was seen as the most important language for their children to learn in order to meet societal demands and access special education services that are usually provided in English only. All mothers had reservations about bilingualism (that it might confuse or further delay their child) and believed a monolingual environment would be better for their child. Despite proficiency in English language, parents’ use of English was more constrained than their use of Chinese, which impacted the use of English with their child. While 20 % of mothers reported that practitioners did support bilingualism, the majority of advice was given against providing a bilingual environment.

Studies evaluating issues surrounding bilingualism and ASD primarily used qualitative methodologies, including ethnography (Jegatheesan 2011) and phenomenology (Yu 2013). These studies employed various strategies to increase the validity of the findings, including the use of triangulation. However, Yu acknowledged that the principal use of interviews meant data captured only the information participants chose to represent at any given time. The same could be assumed for the survey design employed by Kay-Raining Bird et al. (2012). However, quantitative as well as qualitative data were obtained. All studies acknowledged that small (Kay-Raining Bird et al. 2012; Jegatheesan 2011) and nonrepresentative (e.g., highly educated and affluent families; Yu 2013) participant samples potentially reduced the generalizability of the findings. Kay-Raining Bird et al. recognized that the use of an English-only survey may have limited their sample size. The definition of bilingual language exposure across these three studies was vague and variable. Jegatheesan did not make a distinction between SIM and SEQ bilingualism, Kay-Raining Bird et al. included participants from a range of bilingual situations, and the definition of bilingualism in Yu might also be questioned because participants immigrated to the USA after the children were preschool-aged and it was not stated if they had been exposed to English before this time.

Discussion

The present systematic review identified eight studies addressing bilingualism in a total of 182 children (aged 20 months to 22 years, mean = 4.8 years) with ASD and 62 parents of children with ASD. The preceding summary and analysis of these studies suggest two main themes of the existing literature that warrant further discussion. Namely, a discrepancy between the ability of children with ASD to learn two or more languages and advice given by practitioners surrounding the adoption of a bilingual environment. While findings from all of the studies suggest bilingualism does not have a negative impact on language development, advice regarding adoption of bilingualism is somewhat conflicting. Most studies concur that individual characteristics and family history should be taken into account and that a culturally sensitive approach should be employed. However, the three studies that investigated issues surrounding the adoption of bilingualism (Kay-Raining Bird et al. 2012; Jegatheesan 2011; Yu 2013) found that the majority of parents reported that practitioners had advised them not to provide a bilingual environment.

Main Findings and Implications for Practice

With respect to the evidence for bilingual language development, five studies (Hambly and Fombonne 2012; Kay-Raining Bird et al. 2012; Ohashi et al. 2012; Petersen et al. 2012; Valicenti-McDermott et al. 2013) reported that children with ASD from bilingual families did not exhibit additional language delays (i.e., results were neutral). However, these studies did not track language development; they only assessed differences in current language skills between bilingual and monolingual groups at one point in time. Two studies (Jegatheesan 2011; Seung et al. 2006) suggested children were making progress in more than one language and supported the adoption of bilingualism. Seung and colleagues proposed children can benefit from intervention that acknowledges the home language and fosters the second language, with outcomes of their study indicating gains were made in both languages. None of the studies reviewed indicated negative outcomes in terms of language development for bilingually raised children with ASD.

Relating to issues surrounding the adoption of a bilingual environment for children with ASD, perceptions of parents were at times contrasting. Parents in the Jegatheesan (2011) study believed it was important for their children to grow up in a bilingual language environment, whereby English would allow for participation in mainstream life and the home language would allow for participation in cultural and religious ceremonies. However, parents in the Yu (2013) study explicitly stated that it was more important for their child to learn English than the home language because English was seen to lead to success in school and in life. These findings reflect how cultural values impact decisions regarding language use. Parents in the study by Yu also acknowledged that it was often awkward, unnatural, and difficult for them to speak English, which in turn influenced their ability to communicate with their child. It is not just the child’s potential language ability, but also the parent’s ability that might impact decisions regarding language use. Yu suggested that discouraging families from adopting both languages can significantly impact interaction dynamics. In fact, rather than causing confusion, the use of more than one language can open up opportunities for communication and thereby shared meaning (Soto and Yu 2014). These examples illustrate the multidimensional and contextualized nature of bilingualism. As a result, families need support and advice from practitioners, yet based on the literature reviewed and other sources (e.g., Baker 2013; D’Souza et al. 2012; Wilder et al. 2004), practitioners are often ill prepared to provide this guidance, resulting in a recommendation to reduce to using one language.

While these recommendations might be well intended, it is hoped that findings from this review of the extant literature on bilingualism in ASD will lead to the adoption of more evidence-based practice. It might seem counterintuitive to encourage bilingualism if children are struggling to learn one language, but the research reviewed consistently indicates that they can in fact learn two languages. Bilingual learners with ASD are not at a greater disadvantage for language learning than monolingual children. Building on this literature, practitioners might advise families that bilingualism does not impede language development, but that individual, family, and cultural circumstances must be taken into consideration when making this decision. Baker (2013) suggests language practices that work well for one family may not be suited for another. As a result, bilingual service provision for children with ASD needs to be individualized. A bilingual approach should start with an accurate assessment of the child’s communicative ability, followed by collaborative goal setting, and intervention considerations can include coaching parents, family members, teachers, and peers to implement language facilitation strategies (Baker 2013; Pickl 2011; Seung et al. 2006; Soto and Yu 2014; Wilder et al. 2004).

Limitations and Future Research

Caution is warranted when interpreting the findings of this literature because the review was limited to English-language journal articles only, potentially restricting a truly international perspective of bilingualism in children with ASD. Half of the studies were conducted in Canada thereby limiting the generalizability of findings to other countries and cultural/linguistic contexts. Language development of children with ASD from bilingual families and perspectives of bilingualism for these children from European countries, for example, have not been captured. Further, in the studies that compared language abilities of bilingual and monolingual children, the groups were not ethnically or linguistically matched. For example, Valicenti-McDermott et al. (2013) compared bilingual English-Spanish to monolingual English, but excluded monolingual Spanish because of the small sample size. Specific language pairings may result in different language outcomes because some have similar semantic and syntactic structures, while others do not (Werker and Byers-Heinlein 2008).

Methodological heterogeneity of the reviewed literature might have impacted results. For example, a uniform definition of SIM and SEQ bilingual exposure was not evident. Hambly and Fombonne (2012) found that that the SIM (exposure before 1 year of age) group consistently scored higher than the SEQ group. However, with variation in definitions across the literature, it would seem difficult to determine the relative importance of SIM versus SEQ exposure and what the optimum age of exposure is for best language outcomes. Further, whether the child is nonverbal, nonverbal but vocal, minimally verbal, or fluent might impact their ability to learn a second language. Although the communication level of children included in the eight studies varied considerably, there were fewer children who were nonverbal, minimally verbal, or used AAC as a predominant means of communication. Ohashi et al. (2012) and Petersen et al. (2012) excluded participants who had fewer than 30 words, for example. Whether primarily nonverbal children have similar outcomes in a bilingual environment therefore remains largely unknown.

A multitude of different measures were used to assess language ability and to diagnose ASD. The vast variation in these criteria of ability no doubt impacts the current state of knowledge on the ability of children with ASD to learn two or more languages. As suggested by Petersen et al. (2012), future research should use the same diagnostic criteria for ASD and comparable measures in all languages. For example, because administration of the ADOS (Lord et al. 2003) is based on expressive language levels, it might be a good assessment to confirm diagnosis of ASD and severity of autism and obtain a general level of language ability. It would also be useful to standardize language criteria, such as competency of expressive and receptive ability in each language. The MacArthur Communicative Development Inventories (CDI; Fenson et al. 1993) was commonly used and might be recommended because it can be used to measure each language. However, this is reliant on parent reports rather than practitioner observation. Valicenti-McDermott et al. (2013) and Hambly and Fombonne (2012) also assessed other communicative (i.e., nonverbal communication) and social (e.g., pretend play) abilities. They found that the bilingual children used more social communicative behaviors than the monolingual children, also shown in previous research with typically developing children (Pika et al. 2006). Consistent measurement of these behaviors is also needed to determine what impact they have on language outcomes.

While the reviewed research is promising, there is still little known about the course and pace of bilingual language development in children with ASD, as acknowledged in much of the literature (e.g., Hambly and Fombonne 2012; Ohashi et al. 2012). Longitudinal studies with larger sample sizes are needed to determine the impact of bilingual exposure on language development over time. Further research is also needed to identify the impact of language therapy/intervention on bilingual language development. Finally, research is needed to identify autism-specific characteristics that might impact outcomes. For example, the lack of flexibility and need for routine, characteristic of children with high functioning autism, might in fact allow for systematic learning of a second language and thereby enhance bilingual ability (Wire 2005). Conversely, Baker (2013) suggested that because it is often thought that theory of mind skills are weak in children with ASD, the perspective-taking requirements of bilingualism might make learning a second language incredibly difficult for children with ASD. Research is needed to assess the interplay of such metalinguistic processes, autism-specific characteristics, and the impact this might have on child-specific variation of bilingual development across the autism spectrum.

In sum, there is no empirical evidence to support the idea that bilingualism can have a detrimental influence on the language development of children with ASD. Seven of the eight studies were published within the last 3 years, suggesting promise of more research in this field. This would enable professionals to make evidence-based child-specific language recommendations for culturally and linguistically diverse families of children with ASD.