This article reviews recent ethnographic research on how children with autism spectrum disorders (ASD) use language in their everyday lives. Conducted at the Ethnography of Autism laboratory at the University of California, Los Angeles (UCLA), the research studies reviewed in this article were based on participant observation and extensive video and audio recording of everyday social interaction. These studies differ from most investigations that constitute autism research because they frame autism more ethno-methodologically than clinically: more as an experience and a way of being in a social world and less as a disorder in need of an intervention. These studies examined language as it is used in situ by children with ASD rather than seeing their language as a disembodied cognitive process awaiting remediation. Moreover, from the ethnographic perspective, children with ASD are seen as actively participating in the co-construction of their life worlds through communication with others, rather than being acted upon by these others to be managed or treated. Last, the ethnographic perspective affords a privileged view of how the children and their family members engage with each other in their everyday lives, how they go about their daily activities together, and what kinds of personal experiences they share with one another. This ethnographic perspective makes visible aspects of language use and everyday experiences of children with ASD and their families that are usually obscured in other theoretical approaches to autism.

Autism is a lifelong neurodevelopmental condition that disrupts sociocommunicative development. In the United States, over 1.5 million people currently live with some form of autism, and the prevalence of autistic disorder in children is estimated at 6 to 7 per 1,000 (Centers for Disease Control and Prevention, 2007; Fombonne, Reference Fombonne2003). Laboratory studies of cognitive, sociocommunicative, and linguistic functioning across the autism spectrum have resulted in a taxonomy of impairments that cluster to form an “autistic phenotype,” such as language delay, pronoun reversal, atypical use of gesture and eye gaze, echoing of own or others' utterances and reciting stretches of talk from videos and films, lack of imaginative play, lack of reciprocity and shared enjoyment of objects or events, and inability to form age-appropriate friendships (Lord & Spence, Reference Lord, Spence, Moldin and Rubenstein2006; Micali, Chakrabarti, & Fombonne, Reference Micali, Chakrabarti and Fombonne2004).

A theory of mind (ToM) account of autism was proposed to integrate impairments in pragmatics of language use and in nonverbal social behavior (Walenski, Tager-Flusberg, & Ullman, Reference Walenski, Tager-Flusberg, Ullman, Moldin and Rubenstein2006). According to this account, ToM impairments prevent persons with ASD from being able to perceive a cause-and-effect relationship between mental states and actions, whether their own or other people's, thus resulting in “abnormalities in understanding other minds” (Baron-Cohen, Reference Baron-Cohen, Baron-Cohen, Tager-Flusberg and Cohen2000, p. 3). Cognitive psychology sees ToM as “one of the quintessential abilities that make us human” (Baron-Cohen, p. 3), positioning individuals diagnosed with autism on the margins of human experience. Unlike cognitive psychology, developmental psychopathology does not ask the question of uniqueness (i.e., “How are people with autism different from other people?”) or specificity (i.e., “What is the specific cause of their difference?”), but rather it pursues an understanding of “the behavior of persons with autism and their relationships with others,” of “parent-child relationships and . . . of the child within the context of various aspects of the community” (Burack, Charman, Yirmiya, & Zelazo, Reference Burack, Charman, Yirmiya, Zelazo, Burack, Charman, Yirmiya and Zelazo2001, p. 5). In this theoretical world, such questions are addressed through direct tests narrowly targeting specific areas or evaluating a whole domain of functioning. In both cognitive psychology and developmental psychopathology, the issue is whether a behavioral deficit or strength is the source of autistic symptomatology, and to what extent its association indicates primacy of a particular disfunction in autism (Burack et al., Reference Burack, Charman, Yirmiya, Zelazo, Burack, Charman, Yirmiya and Zelazo2001).

“Autism as a subject touches on the deepest questions of ontology, for it involves a radical deviation in the development of brain and mind,” wrote neurologist Oliver Sacks (Reference Sacks1995, p. 246). “Our insight is advancing, but tantalizingly slowly. The ultimate understanding of autism may demand both technical advances and conceptual ones beyond anything we can now dream of” (p. 249). Thus, the biomedical definition of autism casts it as an ontologically complex disorder that forces affected persons to live in “some unimaginable, alien world” (p. 249). To understand the alien world of autism, we need to develop new, previously inconceivable technologies and conceptual frameworks, Sacks suggested. This position presents a serious epistemological dilemma laden with institutionally ratified assumptions and ideologies permeating basic autism research. The enigma view of autism is sustained by its ambiguous clinical status: incurable but not life-threatening, autism constitutes a particular, liminal kind of clinical condition, a syndrome but not a disease (e.g., Muhle, Trencoste, & Rapin, Reference Muhle, Trencoste and Rapin2004).

Language use of those affected by autism touches on the most fundamental concerns of applied linguistics and linguistic anthropology: communicative competence, conversational relevance, intentionality, and intersubjectivity. What individuals with autism do with their language is often unconventional and sometimes difficult to interpret, falling in the range from the slightly off, “proximally relevant” quality (Ochs & Solomon, Reference Solomon2004), to entirely cryptic utterances impossible to perceive as contingent with any object, linguistic or physical, in the interactional context. The meaning of the words and the illocutionary force of the speech acts (Austin, Reference Austin1975) of those affected with autism are at times uninterpretable. The study of language and autism also inevitably foregrounds the notions of intentionality and language use. For example, a child who routinely says “Can I call you right back, sweetie?” when she does not wish to answer a question (Lord & Spence, Reference Lord, Spence, Moldin and Rubenstein2006) succeeds in deflecting that question. It is difficult to determine, however, whether persons affected by autism, especially those affected more severely, intend to use language in this way and whether they understand what they themselves are attempting to communicate to others. An alternative explanation, although not the one held by this author, is that such utterances are produced not by a desire to convey an intentional, albeit obscure meaning, but are the voice of neurological impairments, as if “broken mirrors” (Ramachandran & Oberman, Reference Ramachandran and Oberman2006) are haphazardly reflecting the musings of a mind gone astray.

Through participant observation and extensive naturalistic data collection involving video and audio recording of everyday interaction, ethnographically informed studies reviewed in this article shed light on these questions of intentionality and language use, and provide an important contribution to the understanding of language use in autism more broadly. They do so by examining language use in situ, identifying patterns of the use of linguistic forms and linking these patterns to particular cultural practices, making language use of children with autism more intelligible and interpretable.

Ethnography of Autism Project: Study of High-Functioning Children

Several studies discussed in this article were conducted as part of UCLA Ethnography of Autism Project, an integrated ethnographic and clinical research project on everyday interactions of children with ASD (autistic disorder and Asperger's disorder; see American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders, 4th ed., 1994). Sixteen families with 8- to 12-year-old high-functioning children with ASD were recruited to participate in the project through clinicians' referrals and presentations at parent group meetings. All 16 families' socioeconomic status was middle class, and all but one lived in two-parent households. The recruitment criteria included, besides a previous diagnosis of autistic disorder or Asperger's disorder, a full scale IQ above 70 and full inclusion in regular classroom. The ethnographic data collection involved documenting the children's conversational interactions with family members at home and with peers and teachers at school. To ensure ethnographically informed data collection, the children were observed on numerous days at school and video-recorded for up to 1 month while interacting with their peers and teachers in class during instruction and on the playground during recess and lunch break. The parents were asked to audio-record their children's interactions during breakfast before school, and in the car on the way to and from school for 5 days. Additionally, each child was video-recorded before, during, and after family dinnertime. A total corpus of 320 hours of video and 60 hours of audio data was collected and transferred to digital format. Analytically relevant stretches of talk were transcribed using conversation analytic conventions (Atkinson & Heritage, Reference Atkinson and Heritage1984; see Appendix).

The clinical component of the project was carried out by a research team directed by clinical psychologist Lisa Capps at the University of California, Berkeley (UCB). The clinical data corpus consisted of approximately 50 hours of video- and audio-recorded psychological evaluations of the participating children and their parents. The clinical psychology team confirmed each child's diagnosis through the Autism Diagnostic Interview-Revised (ADI-R; Le Couteur et al., Reference Le Couteur, Rutter, Lord, Rios, Robertson and Holdgrafer1989) and the Autism Behavior Checklist (ABC; Krug, Arick, & Almond, Reference Krug, Arick and Almond1978). The children's evaluations included Wechsler Intelligence Scales (WISC-III; Wechsler, Reference Wechsler1992); ToM tests (Baron-Cohen, Reference Baron-Cohen1989; Baron-Cohen, Leslie, & Frith, Reference Baron-Cohen, Leslie and Frith1985; Happé, Reference Happé1994); and tasks measuring empathy, emotion expression, and emotion recognition. Results of these clinical measures were used as complementary information in video and audio data analysis.

Study of Children with Severe Autism

Several studies discussed in this article have been conducted as part of a larger ethnographic language socialization project on communication of children with severe autism. A data corpus of approximately 200 hours of video-recorded interactions of 16 children and teens with severe autism ages 3 to 18 was compiled. The video data documented these children and teens' interactions with therapists, family members, teachers, and peers in everyday activities. Additional video data consisted of ethnographic, person-centered interviews with family members, teachers, and therapists. The data corpus consists of (1) approximately 100 hours of video recordings that were selected by the author from a larger video archive collected by the Cure Autism Now Foundation and (2) approximately 100 hours of video recordings collected by the author for a research project that followed families participating in the Cure Autism Now foundation's project, as well as families subsequently recruited in Los Angeles and Chicago.

Overview of Research

The overview will first discuss four larger-scale theoretical investigations on (1) how children with autism are socialized into sociocultural competence to become “speakers of culture” (Ochs, Reference Ochs and Kramsch2002); (2) how children with autism participate in the social world as members of families, classrooms, and communities (Ochs, Kremer-Sadlik, Sirota, & Solomon, Reference Solomon2004); (3) how children with autism draw upon structural properties of social interaction to participate in everyday talk (Ochs & Solomon, Reference Solomon2004); and (4) how child-directed communication (CDC) may not have a solely nurturing influence on children's language development, but rather may impede such development for children affected with neurological conditions such as severe autism (Ochs, Solomon, & Sterponi, Reference Ochs, Solomon and Sterponi2005).

Following discussion of these four studies, the article will then review five more narrowly defined studies that examined specific areas of language use and social behavior in autism: (1) ToM and question-answer sequences (Kremer-Sadlik, Reference Kremer-Sadlik2004); (2) bilingualism (Kremer-Sadlik, Reference Kremer-Sadlik, Cohen, McAlister, Rolstad and MacSwan2005); (3) accountability and morality (Sterponi, Reference Sterponi2004); (4) politeness (Sirota, Reference Sirota2004); and (5) narrative (Solomon, Reference Solomon2004). Together, these nine studies create a collective picture of language use, sociality, and communicative abilities of children with ASD.

Autism and Language Socialization: Becoming a Speaker of Culture

A central argument of language socialization is that human beings develop the ability to speak a language as a way of becoming competent members of societies, and that the development of cultural competence is accomplished through language (Ochs, Reference Ochs1982, Reference Ochs1988; Ochs & Schieffelin, Reference Ochs, Schieffelin, Shweder and LeVine1984; Schieffelin, Reference Schieffelin1990; Schieffelin & Ochs, Reference Schieffelin and Ochs1986).

Guided by enduring interest in the cultural organization of children's social and communicative practices, Ochs (Reference Ochs and Kramsch2002) applied language socialization theory to the study of the knowledge and skills required of high functioning children with ASD to become competent members of society and culture.

Through close examination of the children's engagement in ordinary activities such as a game of softball or a math test, Ochs (Reference Ochs and Kramsch2002) mapped the playing field of cultural knowledge. It consists of the categories and rules of activities; situational expectations and strategies for acting and positioning self in the course of these activities; goal-directed acts and participants' psychological stances towards these acts; links of acts and stances to expectations of particular participants' actions; and the contextualization of the actions, stances, and participants in temporal enfolding of what just happened and what is about to happen next. Social participation of children with autism is complicated by the multiplicity of activities that may be co-occurring. Like neurologically unaffected members, children with autism need to monitor turn by turn the emergent, contingent interactional construction of social realities. They need to learn that they have to act within established conventional parameters to carry out their activities and to realize their social identities. The space within these parameters implies a certain degree of spontaneity and improvisation, something that children with autism often find challenging.

The challenges faced by children with autism are similar to the challenges encountered by second language learners, and thus the theoretical framework described in this study has important implications for second-language acquisition and socialization. It is useful for understanding cross-cultural similarities and differences in how language relates to action, stance, activity, and social identity. Although there is considerable overlap in how speakers across speech communities signal actions and psychological stances, there are considerable differences in how they use actions and stances to realize particular activities and identities. Although these commonalities may help second language learners, the cross-cultural differences often thwart their efforts to become competent members of their second cultures. Ethnographic studies of language and autism, Ochs (Reference Ochs and Kramsch2002) wrote, illuminate how

These insights are useful for a more general understanding of the processes underlying first and second language acquisition and socialization.

Autism and the Social World

Ochs et al. (Reference Ochs, Kremer-Sadlik, Sirota and Solomon2004) extended the scope of inquiry on language, social interaction, and autism from the interpersonal domain in which autism has been defined as a disorder of theory of mind (Baron-Cohen, Reference Baron-Cohen1995; Baron-Cohen et al., Reference Baron-Cohen, Leslie and Frith1985) to the sociocultural one. Arguing for a distinction between social as interpersonal and social as sociocultural, Ochs et al. proposed that persons with autism need to be viewed as participating members of social groups and communities, and not only as individuals in relation to other individuals as has been the case in psychological research. To this end, this ethnographically informed study considered persons with autism as sociocultural beings who act displaying both social abilities and difficulties that are relative to socioculturally organized expectations of competence and conduct. This reframing of autism in anthropological, rather than psychological terms afforded a new sociocultural approach to understanding challenges in perspective-taking faced by individuals with autism in social interaction.

The sociocultural perspective-taking entails not only knowledge and skills necessary for interpersonal engagement, but also sociocultural knowledge and skills that make it possible for members of social groups to perform and interpret conventional social behaviors, roles, activities, and norms organizing a range of institutions, from families and schools to workplace and religious communities. It involves the production, interpretation, and prediction of human social behavior beyond one-on-one social interaction. While interpersonal perspective-taking has been defined as the understanding of another person's, or even one's own, intentions, beliefs, knowledge, or feelings (e.g., Happé, Reference Happé2003), sociocultural perspective-taking requires the understanding of other members' expected intentions, beliefs, knowledge, or feelings that are conventionally linked to socioculturally organized practices, roles, institution, and membership in a social group.

Having outlined the boundaries of the sociocultural perspective-taking, Ochs et al. (Reference Ochs, Kremer-Sadlik, Sirota and Solomon2004) examined how children with autism display it across three interactional domains: conversational turn-taking and sequences; formulating situational scenarios involving other persons, and interpreting the sociocultural meanings of indexical forms and behaviors. Analysis of video-recorded spontaneous social interactions involving children with autism spectrum disorders evinced a cline of competence, ranging from most success in anticipating conversational moves and participating in conversational turn-taking to least success in inferring indexical meanings of forms and behaviors.

For example, the 8- to 12-year-old high-functioning children with ASD participating in the Ethnography of Autism Project had few difficulties taking their turns at talk at expected transition-relevant places. It is important to remind the reader that all children in the study had full IQ scores in the “nonretarded” range, which spanned from 73 to 139. Moreover, the majority of the children (15 out of 16) passed the first-order ToM task, and exactly half of the children (8 out of 16) were able to pass the second-order ToM test. In relation to the competence in conversational turn-taking, similar findings were reported by Frith, Happé, and Siddons (Reference Frith, Happé and Siddons1994). Their caregiver-report study was based on Vineland Adaptive Behavior Scales to examine real-life adaptive skills of children with autism who pass standard ToM tests. Frith and colleagues reported that the “passers” of false belief ToM tasks engaged in social behavior that was characterized by better than expected skills in language use and demonstrated their understanding of others' mental states.

Examples of this kind abound in the ethnography of autism data corpus. In the following interaction, Connor, an 8-year-old boy with Asperger's syndrome, disagrees with his father about the choice of musical instruments that he plans to play in a music band at his school (see Appendix for transcription conventions):

Besides evincing Connor's understanding of other minds who “don't really care which instrument” he chooses to play, the seamlessness of this interaction suggests that conversational turn-taking is well mastered by high-functioning children with ASD. Not all domains of social interaction, however, presented so few challenges: A less accessible area was formulating situational scenarios (e.g., after a party, not getting into a friend's car if the driver is drunk), which rests on awareness of conventional arrangements of situational roles, actions, and dispositions to infer participants' knowledge, intentions, and feelings.

The most challenging domain of sociocultural perspective-taking for children with ASD, Ochs et al. (Reference Ochs, Kremer-Sadlik, Sirota and Solomon2004) argued, was indexicality. The ability to infer indexical meanings rests on representing and recognizing social situations, and indexical sense-making relies on the members' knowledge of conventional associations between entities in social contexts. Membership in a community requires understanding the ways in which particular forms of behavior, appearance, artifacts, and physical environment index the relevant practices, identities, dispositions, and institutions. The children in the study had least success in inferring practices, dispositions, identities, and other sociocultural meanings from conventional indexical forms, although these subtle misreadings were usually ratified by their interlocutors. Inferring indexical meanings may have been especially challenging because their potential is wide-ranging and requires awareness of co-occurring salient properties of circumstances at hand. Ochs et al. (Reference Ochs, Kremer-Sadlik, Sirota and Solomon2004) suggested that the high level of competence in conversational turn-taking was due to the local orderliness of conversational sequences and turn-taking that was more accessible to a cognitive and information-processing style of those with autism who thrive on orderliness and tend to focus on details.

Language, Practical Logic, and Autism

Intrigued by the often seamless participation of high-functioning children with autism in certain everyday social interaction, Ochs & Solomon (Reference Solomon2004) considered Garfinkel's and Bourdieu's perspectives on practice (e.g., Bourdieu, Reference Bourdieu1990a, Reference Bourdieu1990b; Garfinkel, Reference Garfinkel1967) to examine children's engagement in social encounters that require fluid, contingent, and practical strategies and behavior. Their study explored the relation between structure and agency, and between disposition and practice, showing how autism refracts these relations in theoretically important ways. Ochs & Solomon discussed evidence from ethnographic observations and audio and video recordings showing that high-functioning children with ASD have a heightened awareness of how the mastery of certain social practices is critical for being perceived as competent at “being a kid.” In the next example, 9-year-old Karl summons his classroom aide to help him during a math test because he feels that “he does not know anything.” As the aide approaches, Karl arches his neck to glimpse an answer from the multiplication table on the wall behind her. She reprimands him:

Similar awareness of practical competence was displayed by other children in the study, especially when they were made accountable by their peers for the hand flapping and rocking behaviors symptomatic of autism. In these instances, the children with autism attempted to reframe these behaviors as intentional and humorous, a kind of a clown act, a disposition that typically developing peers often did not share. Thus, the children with ASD were able to reflect upon practices as linked to identities, and attempted to strategically accommodate to normative expectations, or voiced their frustrations at what they considered their shortcomings.

Because theoretical accounts of autism (i.e., weak central coherence, impaired executive function, and anomalous ToM) aim to articulate cognitive processes by which a qualitatively established pattern of impairments is justified, Ochs & Solomon (Reference Solomon2004) considered these theoretical accounts to examine how autistic impairments may organize children's social practices when they engage in everyday interaction. For example, the weakness of central coherence hypothesis is a theoretical account of autism that attributes its characteristic cognitive and sociocommunicative profile to an imbalance in information processing that inhibits integration of information into a coherent hierarchical organization (Frith, Reference Frith1989; Happé, Reference Happé1996; Plaisted, Reference Plaisted, Baron-Cohen, Tager-Flusberg and Cohen2000; Shah & Frith, Reference Shah and Frith1993). An ability to perceive and construct central coherence is critical to understanding cultural systems of classification and social order, and to participating in everyday conversational discourse. Such a limitation in central coherence was identified in the discourse of a number of high-functioning children in the study who engaged in constructing conceptual paradigms and listing and contrasting members of sets, as illustrated in the following example. Mary, a 9-year-old girl diagnosed with autistic disorder, is having a conversation with her mother and father at dinner:

Excited by her mother's explanation that a “nickname” is “like a little fun name,” Mary proposes “Golden State” as a “nickname state,” a candidate member of the “nickname” category. Having been asked by Mother a fill-in-the-blank question with a rising intonation (“The Golden State is a nickname for?”), Mary enthusiastically replies “State! CaliFORnia!.” This just co-constructed paradigmatic set (the Golden State is a nickname for California) triggers Mary's mentioning of another, similar paradigmatic set that she articulates independently: “the golden poppy is the state flower for CaliFORnia.” In this dinner interaction, the parents turn Mary's interest in categories and conceptual paradigms into a language game (Wittgenstein, Reference Wittgenstein1958), supporting Mary's interactional success.

Asking how most of the children in the study seemed able to navigate the flow of social exchanges, Ochs & Solomon (Reference Solomon2004) offered two interrelated explanations: actor-based and practice-based. An actor-based explanation concerns properties of the practical actors in the social practices observed, such as the tendency of family members and other participants to be “generous interlocutors”: to design their talk and conduct to be comprehensible and interesting to children with ASD, and to richly interpret the talk and conduct of the children. Such participants usually make certain that they secure the child's attention, clarify possible misunderstandings, fill in missing information, and otherwise promote the child's social involvement.

A practice-based explanation considered certain fundamental properties of the social practices underway. The social practices that exhibited certain properties appeared well within the grasp of high-functioning children with ASD, while other social practices were more challenging. The children's social fluency resided primarily in their ability to act relevantly and generatively in response to locally prior and upcoming actions. Linking their predications to the propositional content of locally prior and anticipated utterances was somewhat more challenging. Linking actions to their own and others' actions over a more extensive span of social interaction was significantly more difficult, while the greatest difficulty lay in grasping more global themes constructed across an extended series of utterances.

Radical incoherence was rarely observed in the spontaneous interactions involving high-functioning children with ASD, Ochs & Solomon (Reference Solomon2004) wrote, explaining that the children's ability to maintain some semblance of relevance was aided by the relative rather than absolute quality of the concept. Profiting from the fuzziness of situational relevance, the children in the study routinely maintained social practices by expressing ideas that were proximally relevant, not quite irrelevant but also not quite in synch with the focal concern of conversation. Their two prevailing strategies for making their utterances proximally relevant were to (1) make the interactional contribution locally relevant to what was just said, but not to the more extensive concern or enterprise under consideration; and (2) to shift the focus away from personal states and situations to topically relevant impersonal, objective cultural knowledge.

Some children mixed the two strategies, proximally relating objective knowledge to a locally prior move. The examination of social practice through the prism of autism, Ochs & Solomon (Reference Solomon2004) suggested, allows us to see how practical logic applies to the flow of local and extended actions and propositions which may present different degrees of accessibility for speakers with and without neurodevelopmental disorders such as autism.

Language, Autism, and the Limits of Habitus

Ochs et al. (Reference Ochs, Solomon and Sterponi2005) continued to examine language and autism, drawing on the theoretical frameworks of practice theory and language socialization to analyze the impact of habitus (Bourdieu Reference Bourdieu1977, Reference Bourdieu1990a, Reference Bourdieu1990b) on the communicative development of children, and especially severely impacted children with ASD. In the practice theory framework, habitus, a system of socially organized dispositions that enables people to interpret and creatively engage in the flow of social practices, affords both regularity and improvisation in social life (Bourdieu, Reference Bourdieu1990b). In their discussion of child-directed communication (CDC), Ochs et al. (Reference Ochs, Kremer-Sadlik, Sirota and Solomon2005) brought together the language socialization premise that society and culture organize communication with children and the practice theory argument that habitus organizes how members perceive, appreciate, and act in relation to specific situations.

Ochs et al. (Reference Ochs, Kremer-Sadlik, Sirota and Solomon2005) proposed a model of CDC that consisted of analytic dimensions relevant to the dispositions that constitute members' habitus in particular sociocultural contexts. These dimensions, variably realized across situations and communities, include CDC ideologies, habitats, participation frameworks, activities, and semiotic repertoires. The authors used the model of CDC to analyze how certain sociocultural configurations of CDC may limit or enable children's communicative potential. The analysis illustrated how certain dispositions of European American CDC habitus may compound, rather than minimize, the communicative difficulties associated with severely autistic children's impairments: face-to-face body orientation, speech as the primary semiotic medium for the child, and caregivers' slowed speech tempo and profuse praise. Face-to-face orientation, for example, is important for emotion recognition in typical development, but is an area of challenge in autism (e.g., Baron-Cohen, Campbell, Karmiloff-Smith, Grant, & Walker, Reference Baron-Cohen, Campbell, Karmiloff-Smith, Grant and Walker1995). Slowing down tempo of speech may also be problematic because of the difficulties with extended courses of actions and with higher-order informational structures. Ochs et al. (Reference Ochs, Kremer-Sadlik, Sirota and Solomon2005) argued that when a child manifests neurodevelopmental impairment such as autism, the habitus of a speech community may poorly serve the communicative development of the child, yet mature speakers may find themselves at a loss to improvise alternative strategies and persist with their default CDC practices.

Alternatively, Ochs et al. (Reference Ochs, Kremer-Sadlik, Sirota and Solomon2005) wrote, because habitus is open-ended, it may be open to potentially radical transformation in crisis situations. When members feel helpless to cope with situational exigencies presented by symptoms of severe autism, the members' sense of order may be undermined. Amid this experiential crisis that renders all existing CDC practices inefficient, a disciplined, empirical orientation toward the circumstances sometimes emerges and brings about a shift in habitus. On a small but significant scale, such a transformation took place first in India and then in the United States when Soma Mukhopadhyay, a mother of a severely impacted boy with autism, revised commonly held assumptions about this disorder and initiated an alternative set of CDC practices attuned to severe autism. These practices included side-by-side body orientation, pointing to symbols as the primary semiotic medium for the child, and caregivers' rapid prompts and restrained praise.

Although not a magic cure that eradicates all symptoms of autism, Soma Mukhopadhyay's method positively transformed the life worlds of children and families. Some of the children with whom Mrs. Mukhopadhyay worked, although still severely autistic, were able to communicate independently and engage in studies that were appropriate to their grade level. It was difficult, however, for families to adopt this culturally different set of CDC characteristics, and some families were reverting to their original CDC practices. Examining how family members, teachers, therapists, and others communicate with children with severe autism offered insight into the limitations of habitus and the capacity of members to transform and to restructure their CDC dispositions.

Language and Autism: Specific Areas of Language Use

The studies reviewed in this section had distinct analytic foci that examined language and autism from a range of vintage points. All these studies, however, were motivated by a specific assumption about language use of children with ASD.

Kremer-Sadlik (Reference Kremer-Sadlik2004) pursued contextualization of the theory of mind account of autism by examining children's social-cognitive understanding of others' minds. Her analysis focused on the children's linguistic performance when answering family members' questions during dinnertime. Drawing on spontaneous video-recorded data, this article analyzed the children's abilities in reading speakers' communicative intentions, knowledge, and beliefs, which are embedded in questions, and in adjusting their responses accordingly. Kremer-Sadlik found that, contrary to findings in cognitive psychological research, the majority of the time the children were able to detect their interlocutors' communicative intentions and produce relevant responses that were marked by these interlocutors as acceptable. Furthermore, this research suggested that parents play an important role in facilitating these children's access to interlocutors' intentions, and it examined the different strategies that parents employ to improve their children's communicative skills.

In another study, Kremer-Sadlik (Reference Kremer-Sadlik, Cohen, McAlister, Rolstad and MacSwan2005) critically examined a belief often held by clinicians that children with autism should not be exposed to multilingual linguistic environment. Kremer-Sadlik considered the common albeit empirically unsupported clinical practice of advising multilingual families to speak English only to their children with ASD regardless of the parents' English proficiency. Conducting a series of ethnographic interviews about children's developmental history and family language use, Kremer-Sadlik established that the rationale for such clinical recommendations, as it was understood by the parents, was to ensure their child's exposure to the same language inside and outside of the home. This English-only strategy appeared to be recommended by clinicians to promote a homogeneous and simplified linguistic environment that in theory was to facilitate language development in a child with ASD. These recommendations, however, had a detrimental effect on the children's opportunities for language development. Ethnographic observations, video recordings, and interviews with the parents documented that most families ceased to speak their native languages only when addressing their child with ASD, but not when speaking among other family members. Video recordings of family dinnertime unequivocally captured the children with ASD in multilingual families that followed the English-only recommendation being linguistically excluded from family conversation. Family members routinely used their native language with one another while addressing their child with ASD in heavily accented and grammatically flawed English. Kremer-Sadlik concluded that such “simplified register” strategy is neither realistic nor productive in that it places children with ASD at an additional disadvantage by excluding them from social interactions and denying them the opportunity to actively engage in activities with others or overhear others' conversation.

Focusing on children's engagement in the social lives of their families through examination of positive politeness, Sirota (Reference Sirota2004) expanded current understandings of the sociocommunicative capabilities and challenges of children with ASD. Through analysis of spontaneous interactions in family and community settings, Sirota examined a range of discursive resources used by the children to accomplish reciprocal positive politeness practices in coordination with others. Sirota's practice-based analysis described positive politeness of children with ASD as a discursive process encompassing both sociocultural and interpersonal knowledge and skills.

Sterponi (Reference Sterponi2004) examined how children with ASD oriented to norms and expectations of social conduct. Focusing on rule violation episodes in everyday life, she showed how high-functioning children with ASD were actively engaged in a Wittgensteinian language game of accountability, a discursive practice composed of a set of interrelated moves (Wittgenstein, Reference Wittgenstein1958). Analysis of social rule violations and transgressions illuminated the children's mastery in the deployment of social rules as guides for appropriate conduct. Moreover, Sterponi convincingly showed that high-functioning children with ASD were sensitive to the moral dimensions of interpersonal conduct and competently carried out actions evincing sequentially based understanding of other people's beliefs and emotions. Sterponi concluded that participation in discursive social rule violation episodes provides a unique opportunity for children with ASD to be socialized into socially acceptable conduct and to learn to think about their own and others' social behavior.

Solomon (Reference Solomon2004) critically reexamined the assumption that children with ASD are not interested or able to competently participate in narrative discourse. She analyzed conversational narrative interactions involving high-functioning children with ASD and their family members to consider the children's ability to attend to and make use of contextual information when introducing narratives into ongoing conversation. The article examined narrative introduction practices to illuminate what the children oriented to in the flow of conversation, how they interpret it, and what specific discursive resources they employed to bring about narrative co-telling. The analysis focused on the children's management of thematic continuity or discontinuity with prior talk in narrative introductions relating to personal and fictional (originating in video games, films, books, etc.) experience. Both the children's own contributions in bringing about narrative interaction and their ability to build upon contributions of others were considered. The article delineated conversational practices used by the children to introduce narrative interactions into ongoing talk. The main finding of the study was that high-functioning children with ASD oriented to thematic continuity as to a locally driven phenomenon and had preference for narrative introductions that were linked to immediately prior discourse.

Solomon (Reference Solomon2004) illustrated that children with ASD were able to proactively engage in narrative activity with family members, establish themselves as focal co-participants, and effectively shape their participation over the course of narrative introduction. Some of the children with ASD adopted highly conventionalized introduction formats, especially when introducing fictional narratives derived from books, television programs, feature films, and computer games. Fictional narrative introductions appeared to be well within reach of children with lower verbal ability who competently and successfully used the procedurally stable formats afforded by the global preorganization of these narratives by their modalities of expression (video recording, printed text, etc.). Narrative co-telling over the extended course of propositions, however, was more challenging, lending a degree of support to the theory of weak central coherence. It may be suggested that even when a narrative introduction as a hierarchically implicative action was achieved successfully, its global function may not be successfully maintained over the projected propositional flow. In addition to considering narrative introduction practices, Solomon identified adaptive conversational practices used by the parents to counterbalance the limitations of autism that organizes their children's participation in everyday narrative discourse.