Discourse analysis as a tool for uncovering strengths in communicative practices of autistic individuals

Abstract

This article aims to show how discourse analysis can help identify and reinterpret the communicative practices of individuals with autism spectrum disorder, presenting them as co-constructed by the neurotypical interlocutor. The data described in the article come from three interviews with autistic adolescents. The participants completed two tasks: picture description and narrative production. The interviews were further analysed with the use of discourse analysis. The study demonstrates how the participants oriented to the interviewer’s utterances and what communicative strategies they used throughout the interview. Discourse analysis is presented as an approach to the study of autistic communication, which can substantially contribute to the current state of knowledge about autism spectrum disorder, and be an invaluable help for practitioners.

Keywords

Adolescents autism spectrum disorder case study communicative practices discourse analysis verbal communication

Introduction

An interaction with an autistic individual tends to constitute a challenge from the perspective of a neurotypical interlocutor. Numerous psychological and linguistic sources described non-verbal and verbal deficits of people with autism spectrum disorder (ASD), such as lack of eye contact (Ames and Jarrold, 2007; Pisula, 2010; Wiklund, 2016) and echolalia (Sterponi and De Kirby, 2016; Tager-Flusberg, 1999), which may affect the perception of a person with this condition by non-autistic people, and impede communication between these two groups. Few of the papers on autistic communication took into consideration the interactional contribution of the other, usually neurotypical, participant in the conversation. Since the act of communication requires the cooperation of at least two parties, the input of both interlocutors needs to be considered in order to gain a holistic view of their conversation and see how participants orient to each other’s utterances and co-construct their interaction (Dobbinson et al., 1998; Stiegler, 2007).

A possible way of illuminating this collaboration is the use of discourse analysis (DA) in research. DA is an approach to the study of language whose goal is to recognise and describe repeated patterns in discourse (Herring, 2004). This method allows one to observe regularities in a communicative exchange and apply this knowledge to enhance the process of communication (Stiegler, 2007).

This article aims to identify, with the use of DA, the communicative practices of people with ASD, showing them as co-constructed by the typically developing interlocutor. The change of perspective enables the reinterpretation of selected phenomena observed in autistic communication, revealing their possible function and meaning, which are alternative to what has been described in the psychological literature.

The analysed data come from three audio-recorded interviews with autistic adolescents. The tasks completed by participants included picture description and narrative production. DA was applied to analyse the recordings. The study identified the participants’ communicative strategies and the ways they oriented to interviewers’ contributions throughout the interview.

The following article presents DA as a valuable tool in the study of autistic communication and shows how it can contribute to the current state of knowledge about ASD and its treatment.

Theoretical background

Approaches to language and communication in autism research

Atypical language and communication are among the key elements that constitute the autistic dyad, which is the basis for diagnosing ASD (American Psychiatric Association, 2013). The term is used to describe the areas of life that are affected by this condition. Apart from social communication impairments, the dyad includes repetitive and restricted patterns of behaviour (American Psychiatric Association, 2013). Given the important role that communication and language play in establishing the diagnosis of ASD, it comes as no surprise that many studies have been devoted to exploring this area of the autistic spectrum. Although their subjects of interest are varied, there are some features that most of these studies have in common, namely focus on an individual, experimental paradigm and quantitative analysis.

Individual-focused work concentrates on language and communication of individuals diagnosed with ASD, without discussing the interactional input of the other interlocutor in a conversation. The language of autistic participants is perceived solely as a reflection of their communicative and psychological competencies (O’Reilly et al., 2016). As a result of this decontextualised approach, individual-focused studies pay attention to deficits connected with ASD (Stiegler, 2007). Alternatively, language can be seen as an interactional accomplishment (Sidnell, 2012; Sterponi and De Kirby, 2016), the product of interaction between a speaker and a hearer. This perspective makes both interlocutors responsible for the conversation (Dobbinson et al., 1998). For instance, possible interruptions in the flow of a dialogue are no longer considered a sign of incompetence of one or both participants, but rather seen in a broader context of the interaction itself (Antaki and Wilkinson, 2013). This is how communication is viewed in discourse analytic approaches.

Numerous studies use an experimental paradigm to uncover some unrecognised aspects of the autistic spectrum or verify new or existing hypotheses (Colle et al., 2008; Vanmarcke et al., 2016). These are usually comparative studies, which aim to identify differences in people on the autism spectrum and their neurotypical controls. Experiments concern various domains: genetics, cognition or neurology. Possible discrepancies are further examined in the context of their potential influence on language performance of individuals with ASD (O’Reilly et al., 2016).

Currently, the quantitative approach dominates in the studies on autistic language and communication. This kind of research involves many participants who are tested against hypotheses that have been created in advance. Quantitative studies use tests or specific procedures that are supposed to ensure an objective assessment of individuals’ abilities, in order to compare their results with the rest of the population. The final outcome of these studies is a statistical report with correlations, which shows relations between different elements that have been analysed.

The contribution of quantitative research to the study of ASD cannot be neglected. Many characteristic features of this condition have been identified and described with the use of quantitative methods, which can provide useful information regarding a given population (for instance, the frequency of some phenomena in the group under scrutiny). Nevertheless, since the core impairments in ASD concern social communication, it seems that language-oriented qualitative methods, which enable examining the actual conversational practices, will be more relevant in this case (O’Reilly et al., 2016).

In contrast to the quantitative approach, which focuses on whole populations, qualitative research concentrates on an individual. It aims to understand and interpret the phenomenon under study. The number of participants in this case is very limited, and these are often case studies. Instead of a statistical report, one gets a narrative report as a final product of a qualitative study.

Despite the fact that qualitative methods still seem to be the secondary choice among researchers, there is a growing number of studies that adopt them. This refers especially to the areas where the access to participants is restricted, for example people with rare conditions, and it would be impossible to get a sufficient amount of data for a quantitative analysis or where standard testing procedures do not apply, for example due to some limitations in the group under study (Tager-Flusberg, 1999). ASD is no exception here.

Qualitative methods in autism research

As the population of people with ASD is extremely diversified, the analysis of single cases seems to be the only way to understand this condition (Markiewicz, 2004). Qualitative methods enable an in-depth analysis of the subject and thus go beyond the statistics and show how something works or can be achieved (O’Reilly et al., 2016).

What is more, the qualitative approach may be a ‘way of giving a “voice” to participants’ (O’Reilly et al., 2016), resulting in both empowering them and changing the perspective in research. For instance, when analysing the natural speech samples of a particular population, for example people with ASD, one can see a given phenomenon from the point of view of people whom it concerns, which may lead to conclusions different from those of an outsider. It is also worth mentioning that due to their focus on naturally occurring data, the qualitative methods are considered to have a high internal validity and provide an accurate depiction of the scrutinised issue (Pope et al., 2002).

Among language-based qualitative approaches in autism research, one can distinguish DA and conversation analysis (CA). They are both data-driven methods of analysis (O’Reilly et al., 2016), which means that the data generate new hypotheses or theories, not the other way round.

DA

DA is a language-focused analytical method whose main aim is to identify and describe regular patterns in the use of language (Herring, 2004). It is an umbrella term that encompasses different approaches, for example DA proper, interactional linguistics or CA (Roberts and Sarangi, 2005). DA assumes that interlocutors co-construct any communicative event (Schiffrin et al., 2001), so it studies the utterances of both participants. The focus of DA is on language use in context, thus it examines mainly naturally occurring interactions; however, it is also employed in interviews and text analysis. Because the patterns observed in language use are produced both consciously and unconsciously (Goffman, 1959), DA can be useful for practitioners, as it can uncover the practices of which they were unaware before and thus help them use these practices deliberately (Antaki and Wilkinson, 2013; Roberts and Sarangi, 2005).

Contrary to the psychological approach, where language is viewed as a representation of one’s thoughts and cognitive abilities (Sterponi and De Kirby, 2016), which means it is context independent and individual, DA considers language to be an interactional accomplishment that takes place in a given context. This contextual embeddedness of utterances is very important, since it may influence their meaning (Sterponi and De Kirby, 2016).

CA

CA is a subgroup of DA (Roberts and Sarangi, 2005) that concentrates on the study of talk in interaction (O’Reilly et al., 2016). Every act of communication follows a certain set of rules; for instance, each turn in a conversation is usually preceded by a previous action and followed by another relevant action (Sidnell, 2012; Solomon, 2004; Sterponi and De Kirby, 2016). CA focuses on such regular patterns and illuminates the organisation and purpose of a given interaction. This method of analysis is more restrictive than DA in terms of data type. Based on the view of social interaction as a joint contribution of co-participants, which is situated in a local context, it ‘insists on the study of naturally occurring activities as they ordinarily unfold in social settings’ (Mondada, 2012). CA, being a data-driven approach, allows an unbiased analysis of a communicative interaction and can help identify phenomena (O’Reilly et al., 2016) that tend to be overlooked in more traditional (theory-driven) approaches to language research.

When comparing both methods, one can observe that DA uses a wider analytic frame than CA (Solomon et al., 2016). The latter focuses on a more technical, utterance-by-utterance analysis of a given speech sample. According to Sidnell (2012), CA, as a method, is restricted to ‘the talk and other conduct in interaction’, meaning that the analysis involves the context of the interaction itself. The former, DA, goes beyond the talk-in-interaction and takes into consideration also the external context of a conversation.

Methodology

The methodological approach in this article employs the view of language as a contextualised, interactional accomplishment (Sterponi and De Kirby, 2016), where language is situated in an interaction co-constructed by both interlocutors who orient to each other’s utterances. The data were collected in the form of audio-recorded interviews and thick notes that include descriptions of participants’ non-verbal behaviour and their comments after the interview. The data obtained in this way were further analysed with the use of DA, where the researcher identifies ‘recurring features and structures in a corpus of data that point to the relationship between the structural organisation of language and its functional interpretation in context’ (Pawelczyk, 2011).

Aims

The goal of this study is to show how DA can help uncover the communicative practices of individuals on the autism spectrum, which tend to go unnoticed by the dominant approaches discussed above. While analysing the data, the author reveals phenomena observed in autistic communication and provides their function-focused interpretation that is alternative to the common, deficit-oriented approach. Moreover, taking into consideration the turns and utterances of both parties, the author depicts how the above-mentioned practices are co-constructed by the neurotypical interlocutor.

Ethics

The project was approved by the appropriate authorities to ensure it met the standards of the Human Research Ethics Committee at Adam Mickiewicz University in Poznań (Poland). The parents of all participants received a detailed description of the project, including the planned procedures. They were also informed (on behalf of their children) about their right to decline to participate or withdraw from the study at a later stage. Similarly, the participants were informed about the study and their rights as participants; however, the information was adjusted to their comprehension abilities. Written parental consent was obtained. All data were anonymised to ensure confidentiality and protect the privacy of participants and their families.

Participants and data

The study described in this article is a case study of three autistic adolescents (2 girls, 1 boy) aged 17–18 years old. All participants were monolingual, native Polish speakers. The study qualitatively analyses, with the use of DA, audio-recordings of two tasks, namely picture description and narrative production, which were the elements of an individual interview with adolescents on the autism spectrum. The interaction of the researcher with each participant lasted for about an hour. Both tasks were conducted in a home setting. The recordings were transcribed according to transcription conventions adapted from Jefferson (2004, see Appendix 1).

Picture description

The participants were asked to describe a picture from the book ‘Moje ciało’ (‘My body’) by Agnès Vandewiele (2005). In the picture, a group of people are on a beach and they are occupied with different activities. The open-ended question from the researcher was ‘What is happening in the picture?’. The same question was repeated with every participant. There was no time limit to accomplish this task and no further instructions were provided. It was left to the participants what they wanted to focus on in their descriptions. The researcher asked additional questions if a given participant was struggling to give an answer.

Narrative production

The second task in the study was narrative production, which was intertwined with a semi-structured, dyadic interview. At the beginning, all participants were asked the same question – ‘What did you do at school today?’. What followed depended on the participants’ answers, therefore each interview covered different topics. Again, no time limit was set to complete the assignment.

The examples discussed in the article were chosen because of the common features of autistic language that they represent (Tager-Flusberg, 2004). In order to provide alternative interpretations of the actions observed in the analysed communicative interactions, the aspects of autistic language that are widely referred to in the literature, for instance listing and repetitions, were shown in the context of a conversation.

Findings

This section is organised according to prototypical features of autistic language (Tager-Flusberg, 2004), which are illustrated in the examples from the data. At the beginning of each subsection a brief introduction is provided, showing the perception of the selected aspects of autistic communication, as presented in the literature. This perspective is further juxtaposed with the discourse analytic view.

The analysis of the data illuminated a number of phenomena that could have been observed in communicative interactions between the person diagnosed with ASD and the interviewer. Taking the interactional contributions of both parties into consideration and situating them in the conversational context allowed the provision of an alternative interpretation of what happened in the analysed stretch of talk.

Listing

Individuals on the autism spectrum are said to have problems with producing narrative descriptions (Tager-Flusberg, 1999). Instead of joining different elements into a coherent unity, they have a tendency to name objects without embedding them in any context. A possible explanation for this phenomenon would be weak central coherence (WCC) (Frith, 2008), which refers to autistic preference for local over global information. People with ASD tend to focus on details and do not pay attention to context (either conversational or general). They seem to lack the holistic view. This feature is troublesome in context-dependent situations, such as having a conversation, but can be valuable in tasks that require observation and attention to detail (Dawson et al., 2007; Solomon, 2004), for instance doing jigsaws. The inclination towards naming instead of producing a coherent narrative is referred to as listing in this article.

Example 1 comes from the picture description task, where the participant was presented with a picture and asked to describe it. The person in the first example was an 18-year-old girl diagnosed with ASD. Throughout the interview she demonstrated extensive vocabulary, and the utterances she produced were well organised and to the point, though rather short. This tendency for precision, understood here as restricting the answers to the exact requirements of the assignment, is also visible in the picture description task. The excerpt below illustrates the literalness (Sterponi and De Kirby, 2016) of the participant’s answers, resulting from the way she oriented to the interviewer’s questions.

Example 1: Listing - verbs

1 I: co się dzieje na tym obrazku?

what is happening in the picture?

2 P1: (na tym) ludzie są na plaży i (3.0) robią babki z piasku (1.0) rozmawiają (2.0) siedzą (5.0) kopią, jeden trzyma koło ratunkowe

(in the) people are on the beach and (3.0) are making sand cakes (1.0) talking (2.0) sitting (5.0) digging , one is holding a life ring

To begin with, the wording of the question asked by the researcher seems to play a major role in this task. She asks (line 1) ‘what is happening in the picture?’, thus giving her autistic interlocutor a sign that she is interested in the actions that are visible in the picture. That is exactly how this question is oriented to by the individual with ASD. One can see that instead of describing the whole picture (which was in fact what the researcher expected), the participant focuses on the actions, saying, ‘(…) people are on the beach and (3.0) are making sand cakes (1.0) talking (2.0) sitting (5.0) digging’ (line 2). The result is a list of activities rather than a picture description.

One could claim that the preference for local information combined with the participant’s precision led to listing the actions in the picture instead of describing it, thus seeing this interaction from the deficit perspective. Although these factors cannot be neglected, it is nevertheless important to highlight the interviewer’s contribution to the task. Clearly, the way she asked the question determined the participant’s orientation to it, and consequently constrained her answer. Taking this into consideration allows one to see the participant’s response as an accommodation of the researcher’s requirement (naming actions in the picture) rather than a demonstration of her deficits.

Interestingly, the listing was task-specific, as during a more spontaneous interaction (interview part) the participant was capable of taking part in a conversation, using complete sentences, taking her turns and providing both detailed and general information on various topics. This, again, draws attention to the phrasing of the question asked by the researcher and the extent to which it shaped the response of the participant.

Schema

Among the characteristic features of ASD, one can find insistence on sameness (Seligman et al., 2003). People on the autism spectrum are afraid of unexpected changes, and therefore they try to perform different activities always in the same way (Szatmari, 2007). This helps them lower their level of anxiety. This strategy can also be observed in autistic communication, where individuals tend to start a conversation or a narrative description in the same way, as if they were using some schema or templates.

Another participant was a 17-year-old girl with ASD. She also had rich and sometimes sophisticated vocabulary and was much more detailed in her answers than her predecessor. However, this could be seen only during the interview, as in the picture description task her replies were very short and precise. Her performance on this task is analysed in Example 2.

Example 2: Using templates

1 I: co się dzieje na tym obrazku? (…)

what is happening in the picture?(…)

2 P2: ((odchrząknięcie)) tutaj (.) dzieci (.) y tutaj dzieci i dorośli bawią się na plaży. (…)

((clearing her throat)) here (.) children (.) erm here children and adults are playing on the beach. (…)

3 I: mhm↑

mhm↑

4 (7.0)

5 I: coś jeszcze mi możesz opowiedzieć na tym obrazku? (.) co się dzieje?

is there anything else you can tell me about this picture? (.) what is happening?

6 P2: tutaj dzieci: robią zamki z piasku.

here childre:n are making sand castles.

7 I: °zamki z piasku°. okej, co dalej?

°sand castles°. OK, what else?

8 P2: tutaj dorośli się opalają i odpoczywają.

here adults are sunbathing and resting .

The participant starts all her answers with the same word ‘here’ (lines 2, 6 and 8), which is followed by selected actions visible in the picture. Interestingly, the word is not accompanied by any gesture, for instance pointing to the elements being described at the moment. ‘Here’ appears to be a template in her narrative production. The discursive strategy of repetition of the item ‘here’ that begins her utterances seems to help the girl structure and organise her answers. This allows one to see it as a resource more than a deficit. By relying on her strategy, the participant manages to build complete sentences instead of a list of verbs, as in the previous example.

Again, one can observe that the wording of the interviewer’s question results in focus on actions presented in the picture. All utterances are structured in the same way and refer to what people are doing, for instance ‘here childre:n are making sand castles’ (line 6).

Taking into consideration the researcher’s contribution, there is a marker of active listening, namely ‘mhm↑’ with rising intonation (line 3). However, it is not oriented to by the participant as an invitation to continue her description, which is indicated by a 7-second gap in line 4. It seems that the autistic individual considers her answer to be complete and therefore is not willing to proceed. However, when the interviewer asks additional questions (lines 5 and 7), the participant replies without any pauses or interruptions, knowing that she is expected to produce an answer. Moreover, every time she points to new elements in the picture, which suggests that she does not have any problems with recognising what is presented there, and could possibly provide a much more detailed description if she knew that this was her task. What is also worth noticing is the way the typically developing interlocutor formulates her questions, in order to maintain the conversation. She keeps asking exactly the same question, thus getting similar answers from the participant, focusing on actions. Perhaps if the researcher had tried to rephrase her question, the participant would have described different elements in the picture.

This observation is important in terms of the effectiveness of communicative cues used in an interaction with an autistic person. Sometimes, the utterances of a neurotypical interlocutor fail to convey information successfully (e.g. what the aim of a given task is), resulting in an underestimation of the abilities of an individual on the autism spectrum.

Repetitions

Another prototypical feature of ASD is echolalia, which involves repeating the words and utterances of others (Eigsti et al., 2007). Interestingly, it refers to echoing not only the language of another person but also his or her intonation (Local and Wootton, 1995; Tager-Flusberg, 1999). Echolalia can be associated with repetitive behaviour, which is one of the core symptoms of the autistic spectrum. Although limiting on the surface, as it puts the words of others into autistic individuals’ mouths, echolalia can be valuable for people with ASD, since it enables them to take part in a conversation even if they do not fully understand it (Tager-Flusberg, 1999). What is more, as has already been acknowledged by Kanner (see Sterponi and de Kirby, 2016), echoes can be functional and work as affirmative responses, which can be observed in Example 3.

Example 3: Repetitions

1 I: jak myślisz, dobrze się tam bawią?

what do you think, are they having fun there ?

2 P2: dobrze się bawią.

they are having fun.

3 (11.0)

4 I: i co, coś jeszcze byś chciała dodać czy to już koniec?

so, would you like to add anything or is that all?

5 P2: >nie, niekoniecznie. nic nie chcę dodawać.<

>no, not necessarily. I don’t want to add anything .<

The third example presents another excerpt from the picture description task of participant number 2. Trying to obtain a more detailed description of the picture from the participant, the researcher asks an additional question (line 1) – ‘what do you think, are they having fun there?’ – to which the participant replies by partially echoing the words of her interlocutor ‘they are having fun’ (line 2). Although repetitive, the answer fits perfectly in context, becoming a relevant reply to the interviewer’s question. A long, 11-second gap follows, which indicates that the autistic participant perceives her answer as complete (line 3). The neurotypical interlocutor makes another attempt to get a more detailed reply from the participant, asking ‘so, would you like to add anything or is that all?’ (line 4). The participant answers with a mitigated echo (Local and Wootton, 1995), that is, by repeating the interlocutor’s words and adding her own (line 5): ‘no, not necessarily. I don’t want to add anything’. Her reply, again, seems to be an appropriately fitted conversational move.

From the discourse analytic perspective, repetitions can be viewed as a discourse strategy (Local and Wootton, 1995), where a person with ASD uses the words of the other party to construct his or her answers. It is also a sign that the autistic individual is monitoring the conversation and orienting to the prior turn, since echoes tend to occur in the sequential position where the interlocutor with ASD is required to take a turn (Local and Wootton, 1995).

It is worth noticing that in the above example, the researcher uses closed-ended questions. These are questions that can be answered with some specific information or with a simple yes/no. As illustrated by Example 3, closed-ended questions encourage repetitions, and usually do not help maintain a conversation. What is more, this type of question can cause a given utterance to directly determine the interlocutor’s reply, as the respondent can echo the wording introduced by the interviewer (Local and Wootton, 1995). Therefore, in order to explore the linguistic abilities of people with ASD and allow them to use their communicative competence fully, it seems salient to refrain from closed-ended questions and leave space for autistic individuals to speak with their own words. Still, the usefulness of echoing cannot be denied because, as presented in Example 3, it can be an effective strategy that helps people on the autism spectrum to actively participate in a conversation and provide conversation-relevant input to it.

Question–answer exchange

A conversation with an autistic person has been identified to resemble a question–answer exchange rather than a dialogue (Seligman et al., 2003). Individuals with ASD are usually very literal in their replies and provide little input to an interaction. They do not pro-actively co-construct the dialogue, mostly restricting their contributions to instances when they are directly asked for one, thus gaining the status of rather passive interlocutors (Local and Wootton, 1995). This can be connected to the turn-organisation of a conversation, which tends to be challenging for an autistic person (Eigsti et al., 2007; Hale and Tager-Flusberg, 2005). They find it difficult to recognise when the interlocutor’s turn ends and they are expected to contribute to the interaction. Nevertheless, they rarely have problems with direct questions (e.g. wh- questions, where it is clear that the speaker awaits an answer), as can be seen in Example 4. This kind of utterance usually allows one to have an uninterrupted and relatively fluent conversation with an individual on the autism spectrum. However, the type of question matters here. As has been described in the previous section, open-ended questions tend to be more effective in maintaining an interaction.

Example 4 comes from the narrative production task. This time the participant is the same girl as in the first example. After completing the picture description task, participants were asked to tell the researcher about their day at school. As illustrated below, although the girl does not refuse to take part in the dialogue, she keeps her answers to a minimum, giving the impression that she is unwilling to talk. The result is a question–answer exchange.

Example 4: Question–answer exchange

1 I: (…) jak dzisiaj było w szkole?

(…) how was it at school today?

2 P1: dobrze.

fine.

3 I: a coś więcej mi powiesz?

will you tell me something more?

4 P1: nie.

no.

5 I: a słyszałam że: (3.0) że chodzisz na kółko (.) filmowe, tak? (…)

and I’ve heard tha:t (3.0) you are attending a film (.) club, right? (…)

6 P1: tak.

yes.

7 I: i jak tam jest?

and what is it like?

8 P1: fajnie.

nice.

The researcher starts with a closed-ended question in line 1: ‘how was it at school today?’. The participant gives a single-word reply: ‘fine’ (line 2). The interviewer attempts to continue the conversation, which is visible in line 3: ‘will you tell me something more?’; however, she does not succeed because the question is again closed-ended and does not require an elaborate answer. The participant says ‘no’ (line 4), refusing to continue the conversation. In this case, the neurotypical interlocutor tries to change the topic (line 5), finishing her utterance with a question tag (‘right?’). The result is an affirmative, one-word answer: ‘yes’ (line 6). Another question follows (line 7), which again leads to a single-word reply (line 8). All the questions in this exchange were closed-ended. This allowed the participant to keep her answers short and made it impossible for the typically developing interlocutor to maintain the conversation.

Interestingly, this reluctance of the participant to speak with the researcher was limited to school topics only. After switching the subject of the conversation to baking (which was one of the girl’s interests), she was willing to talk and gave quite detailed descriptions of what she could bake.

This observation points to another possible strategy used by autistic individuals. It seems that withdrawing from a conversation may not be connected to lack of communicative competence, but can be a sign of disinterest or unwillingness of an autistic individual to talk about a particular subject.

To sum up, the above example shows that non-fluent and laboured conversations may not be the result of autistic deficits, but of the type of the interlocutor’s questions or the topic of a conversation.

Topic management – shifting

Digressions are the last aspect of autistic communication to be discussed in this article. People with ASD have a tendency to include off-topic comments in a conversation (Eigsti et al., 2007; Hale and Tager-Flusberg, 2005) or start new subjects without preparing their interlocutors for the change (Dobbinson et al., 1998), which is referred to as shifting in this article. Although going back to a previous topic or introducing a new one are common elements of every conversation, doing so without using connective markers, such as ‘well’ (Ochs et al., 2004), may be confusing for co-participants.

The extract below illustrates an example of shifting. It comes from an interview with the third participant, a 17-year-old boy with an ASD diagnosis. Overall, his speech was slow, and the answers he provided were usually short and exact. Nevertheless, he presented a wide vocabulary and was engaged in the conversation, which was visible in his rather detailed answers.

Example 5: Topic management (shifting)

1 I: mhm. (.) ale tych prób dużo pewnie musi być, [żeby tak grać?=

mhm. (.) well there has to be a lot of these rehearsals [to play like this?=

2 P3: [ta:k =tak.

[ye:s =yes.

3 (2.0)

4 I: i- i jak wygląda taka próba? możesz mi opowiedzieć o niej?

an- and what does such a rehearsal look like ? can you tell me about it?

5 P3: no: jest- (4.0) próba, że spotyka się kilka osób i potem (.) nieraz (.) w zeszłym roku też byliśmy na targach poznańskich

well, there is- (4.0) a rehearsal, so a couple of people meet and then (.) sometimes (.) last year we were also at the trade fair in Poznan

6 I: mhm

mhm

7 (3.0)

8 P3: edukacyjnych znaczy się, przepraszam y:: które były- na któ:- gdzie też graliśmy jakieś przedstawienie.

education ((fair)) I mean, I’m sorry erm:: which was- at whi:- where we also played ((sic!)) some theatre performance.

This final example addresses the phenomenon of topic management. In line 1, the researcher introduces the topic of theatre rehearsals, suggesting that many rehearsals are required in order to perform well. The participant interrupts her stretch of talk twice (line 2). First, an overlap can be observed, which is a situation where both interlocutors speak simultaneously. The participant confirms with a short ‘ye:s’ that indeed many rehearsals are needed. Then, the interviewer specifies her thought, saying ‘to play like this’, which is followed without any break (latched) by the autistic participant’s ‘yes’, showing his agreement with the interviewer, but no further elaboration is provided. This part of the exchange illustrates the aforementioned difficulties with turn-taking among individuals with ASD; however, it also depicts the participant as an active listener who, with his short confirming interruptions, assures the interviewer that he is following the conversation. In line 3 there is a short pause, after which the neurotypical interlocutor tries to continue the conversation, asking the participant to describe a rehearsal (line 4). In response to this request, the autistic individual starts his description with a discourse marker ‘well’ (line 5), signalling an intention to build a coherent answer: ‘well, there is- (4.0) a rehearsal, so a couple of people meet and then (.) sometimes (…)’. There are numerous pauses in his reply, implying that he is struggling with the construction of the answer. Suddenly, the participant shifts the topic (line 5) and starts talking about a trade fair. The interviewer does not clarify this change in any way, encouraging the participant to proceed with a continuer ‘mhm’ (line 6). Another gap follows (line 7), and the participant continues his response, describing a performance at an education fair (line 8). Two cut-offs can be observed in his turn, where the participant attempts to self-repair his utterances, changing the structure of the sentence: ‘(…) which was- at whi:- where’. This presents him as a self-conscious speaker who tries to control his stretch of talk. In terms of content, his answer seems to be irrelevant on the surface; however, when analysed in the local interactional context, it becomes meaningful. The participant refers back to the word ‘play’, used by the researcher in line 1. Therefore, his utterances can be seen as proximally relevant (Solomon, 2004), that is, not explicitly connected with the previous talk. The concept of proximal relevance is based on the assumption that relevance is not absolute but relative.

It is worth noticing that this time the researcher used an open-ended question (line 4), which, in contrast to the closed-ended one (line 1), resulted in a detailed and quite long answer. This confirms the importance of question types in an interaction.

As with the previous examples, the seemingly unrelated utterances turned out to be purposeful and meaningful when analysed in the context of the conversation.

Discussion

The study revealed a number of communicative practices used by individuals with ASD, such as listing, repetitions and topic shifting. Although these practices could have been interpreted in terms of communicative deficits of autistic participants, seeing the utterances in the conversational context allowed the identification of their function and relevance. The alternative interpretations of the participants’ answers, presented in this article, included an accommodation to task requirements (listing), a discursive strategy (repetitions) and the proximal relevance (shifting). It could be observed that in many cases the answers were determined by the preceding utterances of a neurotypical interlocutor. What is more, participants’ performance differed, depending on the task. Some of them dealt better with the more structured one (picture description), and others preferred to engage in an interview – their narratives were more natural and spontaneous. This confirms the importance of the variety of tasks when assessing the communicative abilities of an individual on the autism spectrum (Tager-Flusberg, 1999).

Comparing DA to experimental paradigm, where prototypical characteristics of autistic language are treated as a manifestation of a neurological or cognitive disorder, one can observe that the DA framework allows researchers to see these features as responses situated in the local, interactional context and strategies that help individuals to overcome the challenges of being involved in an interaction (Damico and Nelson, 2005; Sterponi and De Kirby, 2016; Stiegler, 2007). By focusing on both interlocutors in a conversation, rather than solely on the autistic participant, DA broadens the interpretation and reveals the alternative meaning that can only be discovered when the whole dialogue is seen in context. This change of perspective enables going beyond deficit interpretations and points to the abilities and resources of people with ASD, which are neglected in traditional, individual-focused approaches to autism research.

DA can complete the existing knowledge of ASD, shedding new light on certain aspects of autistic language and showing their possible, alternative functions. The aim of this method is not to deny the deficits observed in autistic communication, but to shift the focus from deficits to abilities. Seeing prototypical features of autistic language as coping strategies and a demonstration of competencies may lead to novel approaches in professional treatment of ASD, where these features would not be rejected but treated as important aspects around which interventions can be built (Sterponi and De Kirby, 2016; Stiegler, 2007). Therefore, discourse analytic findings can become the basis for re-assessment of current therapies of ASD, raising awareness among professionals and making them use various communicative practices more consciously (Bottema-Beutel, 2017; Dobbinson et al., 1998; O’Reilly et al., 2016; Sarangi, 2013). By monitoring, with the use of DA, their own strategies for maintaining a therapeutic interaction, therapists can assess the effectiveness of particular communicative cues and modify the unsuccessful ones. Moreover, shifting from the deficits approach, which is encouraged by DA, makes therapists see their clients as competent interlocutors and treat the unique features of their language as potential resources that can facilitate the therapeutic process.

This study involves a number of limitations that have to be mentioned. First, being a case study, the project engaged a small number of participants, which means that the findings may not refer to the whole autistic population. The observed tendencies require further investigation in order to conclude whether the described features are typical for individuals with ASD. Other limitations are connected to the qualitative approach, which has been used in the current project, for instance the risk of subjectivity of the author in her analysis and interpretation. Finally, the individual interviews described here were conducted at participants’ homes, therefore each person responded in different conditions, which could have influenced their responses. All the above-mentioned factors point to the possible drawbacks of this study and encourage more research in this field. Future studies could analyse conversational practices of people with ASD in various environments and verify to what extent these practices differ depending on the context or interlocutor.

The number of discourse analytic studies of autistic communication is still limited; therefore, further exploration of the topic is required. Due to the increase in the number of ASD diagnoses (O’Reilly et al., 2017), it seems to be especially important to shift from the ‘nomenclatures of deficits and deviance’ (Solomon, 2004) and focus on resources and strategies that autistic people use to facilitate their communicative interactions. Following Happé (1999), success is more interesting than failure.

Footnotes

Appendix 1 Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

Author biography

Eliza Maciejewska is a doctoral student at the Faculty of English, Adam Mickiewicz University in Poznań, Poland. She has been awarded MA degrees in both Psychology and English. During her Psychology studies, she worked with children with autism spectrum disorder (ASD) in various settings. Her research interests include ASD, communication, autism therapies, discourse analysis and conversation analysis. In her current research project, she is planning to investigate one of the relationship-based therapies of ASD, using discourse analysis.

References

American Psychiatric Association (2013) Diagnostic and Statistical Manual of Mental Disorders, 5th edn (DSM-5). Washington, DC: American Psychiatric Association.

Ames

Jarrold

(2007) The problem with using eye-gaze to infer desire: A deficit of cue inference in children with autism spectrum disorder? Journal of Autism and Developmental Disorders 37: 1761–1775.

Antaki

Wilkinson

(2013) Conversation analysis and the study of atypical populations. In: Sidnell

Stivers

(eds) The Handbook of Conversation Analysis. Oxford: Blackwell, pp. 533–550.

Bottema-Beutel

(2017) Glimpses into the blind spot: Social interaction and autism. Journal of Communication Disorders 68: 24–34.

Colle

Baron-Cohen

Wheelwright

et al . (2008) Narrative discourse in adults with high-functioning autism or Asperger syndrome. Journal of Autism and Developmental Disorders 38: 28–40.

Damico

Nelson

(2005) Interpreting problematic behavior: Systematic compensatory adaptations as emergent phenomena in autism. Clinical Linguistics and Phonetics 19(5): 405–417.

Dawson

Soulières

Gernsbacher

et al . (2007) The level and nature of autistic intelligence. Psychological Science 18(8): 657–662.

Dobbinson

Perkins

Boucher

(1998) Structural patterns in conversations with a woman who has autism. Journal of Communicative Disorders 31: 113–134.

Eigsti

I-M

Bennetto

Dadlani

(2007) Beyond pragmatics: Morphosyntactic development in autism. Journal of Autism and Developmental Disorders 37: 1007–1023.

10.

Frith

(2008) Autyzm. Wyjaśnienie Tajemnicy [Autism: Explaining the Enigma]. Gdańsk: Gdańskie Wydawnictwo Pedagogiczne.

11.

Goffman

(1959) The Presentation of Self in Everyday Life. Garden City, NY: Doubleday.

12.

Hale

Tager-Flusberg

(2005) Brief report: The relationship between discourse deficits and autism symptomatology. Journal of Autism and Developmental Disorders 35: 519–524.

13.

Happé

(1999) Understanding assets and deficits in autism. Why success is more interesting than failure. The Psychologist 12: 540–547.

14.

Herring

(2004) Computer-mediated discourse analysis: An approach to researching online behavior. In: Barab

Kling

Gray

(eds) Designing for Virtual Communities in the Service of Learning. New York: Cambridge University Press, pp. 338–376.

15.

Jefferson

(2004) Glossary of transcript symbols with an introduction. In: Lerner

(ed.) Conversation Analysis: Studies from the First Generation. Philadelphia, PA: John Benjamins, pp. 13–23.

16.

Local

Wootton

(1995) Interactional and phonetic aspects of immediate echolalia in autism: A case study. Clinical Linguistics and Phonetics 9: 155–184.

17.

Markiewicz

(2004) Możliwości komunikacyjne dzieci autystycznych [Communicative Abilities of Autistic Children]. Lublin: Wydawnictwo Uniwersytetu Marii Curie- Skłodowskiej.

18.

Mondada

(2012) Analytic approach to data collection. In: Sidnell

Stivers

(eds) The Handbook of Conversation Analysis. Chichester: John Wiley & Sons, Ltd, pp. 32–56.

19.

O’Reilly

Lester

Muskett

(2016) Discourse/conversation analysis and autism spectrum disorder. Journal of Autism and Developmental Disorders 46: 355–359.

20.

O’Reilly

Lester

Muskett

et al . (2017) How parents build a case for autism spectrum disorder during initial assessments: ‘We’re fighting a losing battle’. Discourse Studies 19(1): 69–83.

21.

Ochs

Kremer-Sadlik

Sirota

et al . (2004) Autism and the social world: An anthropological perspective. Discourse Studies 6(2): 147–183.

22.

Pawelczyk

(2011) Talk as Therapy: Psychotherapy in a Linguistic Perspective. Berlin; Boston, MA: Mouton de Gruyter.

23.

Pisula

(2010) Autyzm. Przyczyny. Symptomy. Terapia [Autism: Causes. Symptoms. Therapy.]. Gdańsk: Wydawnictwo Harmonia.

24.

Pope

van Royen

Baker

(2002) Qualitative methods in research on healthcare quality. British Medical Journal Quality & Safety 11: 148–152.

25.

Roberts

Sarangi

(2005) Theme-oriented discourse analysis of medical encounters. Medical Education 39: 632–640.

26.

Sarangi

(2013) Practising discourse analysis in healthcare settings. In: Bourgeault

Dingwall

de Vries

(eds) The SAGE Handbook of Qualitative Methods in Health Research. London: SAGE, pp. 397–416.

27.

Schiffrin

Tannen

Hamilton

(eds) (2001) The Handbook of Discourse Analysis. Malden, MA: Blackwell.

28.

Seligman

MEP

Rosenham

Walker

(2003) Psychopatologia [Abnormal Psychology]. Poznań: Wydawnictwo Zysk i S-ka.

29.

Sidnell

(2012) Basic conversation analytic methods. In: Sidnell

Stivers

(eds) The Handbook of Conversation Analysis. Chichester: John Wiley & Sons, Ltd, pp. 77–99.

30.

Solomon

(2004) Narrative introductions: Discourse competence of children with autistic spectrum disorders. Discourse Studies 6(2): 253–276.

31.

Solomon

Heritage

Yin

et al . (2016) ‘What brings him here today?’: Medical problem presentation involving children with autism spectrum disorders and typically developing children. Journal of Autism and Developmental Disorders 46: 378–393.

32.

Sterponi

de Kirby

(2016) A multidimensional reappraisal of language in autism: Insights from a discourse analytic study. Journal of Autism and Developmental Disorders 46: 394–405.

33.

Stiegler

(2007) Discovering communicative competencies in a non-speaking child with autism. Language, Speech, and Hearing Services in Schools 38(4): 400–413.

34.

Szatmari

(2007) Uwięziony Umysł. Opowieści o ludziach z autyzmem [A Mind Apart: Understanding Children with Autism and Asperger Syndrome]. Kraków: Wydawnictwo Znak.

35.

Tager-Flusberg

(1999) The challenge of studying language development in children with autism. Available at: http://www.abiliko.co.il/Media/Uploads/challenge_lang_dev_autism.pdf (accessed 25 March 2018).

36.

Tager-Flusberg

(2004) Strategies for conducting research on language in autism. Journal of Autism and Developmental Disorders 34(1): 75–80.

37.

Vandewiele

(2005) Moje ciało [My body]. Warszawa: Wydawnictwo REA.

38.

Vanmarcke

Van Der Hallen

Evers

et al . (2016) Ultra-rapid categorization of meaningful real-life scenes in adults with and without ASD. Journal of Autism and Developmental Disorders 46: 450–466.

39.

Wiklund

(2016) Interactional challenges in conversations with autistic preadolescents: The role of prosody and non-verbal communication in other-initiated repairs. Journal of Pragmatics 94: 76–97.