Representations of disability from the perspective of people with disabilities and their families: a critical discourse analysis of disability groups on Facebook

Contrary to the growing body of research into views about and attitudes towards disability which has focused on the perspectives of non-disabled people, recent disability studies have been directed to the voices of people with disabilities and/or their families, by acknowledging an ‘epistemic privilege’ to them. In the present study, we use the posts of sensory and physical disability groups on Facebook as a lens to uncover the voices and experiences of people with disabilities and/or their families. The critical discourse analysis of the posts of disability groups on Facebook suggests that people with disabilities and/or their families do not constitute a homogeneous group of people being connected because of their common (bodily) condition. On the other hand, different representations of disability also suggest a different use of Facebook.


Introduction
The growing body of research into views about and attitudes towards disability has tended to focus on the perspectives of non-disabled people, such as students (e.g. Favazza and Odom 1997;Nikolaraizi et al. 2005) and teachers (e.g. Buell et al. 1999;Sharma, Forlin, and Loreman 2008), by mostly adopting a quantitative orientation with the use of structured questionnaires containing various attitude scales (Antonak and Livneh 2000). However, more recently, disability studies have been also directed to the voices of people with disabilities (e.g. Cardona 2013) and/or their families (e.g. Bosteels, Van Hove, and Vandenbroeck 2012), with respect to how they conceptualize disability, by acknowledging an 'epistemic privilege' to them, namely by considering them 'experts of their own experiences' (Welsby and Horsfall 2011, 796). In contrast to the studies on nondisabled people's views, this research has mainly employed a qualitative methodology, such as semi-structured/unstructured interviews and focus groups, while participants' views have been mostly approached through some kind of thematic analysis (e.g. Bosteels et al. 2012;Low 1996;Welsby and Horsfall 2011).
Quite expectedly, the experiences of people with disabilities and/or their familiescontrary to non-disabled peopletend to be less stereotypical, by conceptualizing different types of disability (e.g. physical, sensory, intellectual) as primarily an issue of social exclusion (e.g. Cardona 2013; Soulis and Andreou 2007;Ytterhus, Wendelborg, and Lundeby 2008). However, some studies have reported a more complex picture. Specifically, they have noted the opposing forces that are at play in the shaping of the voices of people with disabilities and/or their families. In fact, they often need to oscillate between a social and a medical account of disability, by negotiating insights gained from their own life experiences with mainstream cultural messages about disability (e.g. media portrayals, professional accounts and diagnostic practices). For example, Bosteels et al. (2012) have reported how parents of deaf children talk about their children's disability by drawing upon both a discourse of deafness as a medical category and a discourse of deafness as a responsibility of society. Similarly, the views of people with disabilities and/ or their families may both include compliance as well as opposition to medical and normalizing discourses. For instance, Low (1996) has shown that university students with mobility problems tend to reject the deviant (medical) identities placed on them by others, but that they sometimes need to take on the same deviant identities in order to cope with university life.
Taking into account the central role of the Internet in the life of many people with disabilities and/or their families (Van Kraayenoord 2010), in the present study, we use the posts of sensory and physical disability groups on Facebook as a lens to uncover the voices and experiences of people with disabilities and/or their families. Drawing upon the theoretical resource of critical discourse analysis (e.g. Fairclough 2003), we see text and talk as forms of social practice, which sustain as much as they transform relations of power and ideological dominance. In particular, critical discourse analysis views language as being intrinsically linked to society. Therefore, it focuses on the linguistic features of texts, as well as on the social structures underpinning those texts. The concept which captures the language-society interface is that of 'discourse'. Consequently, in critical discourse analysis, discourse is not only treated as a representational resource, but it is rather viewed as having both linguistic and social aspects. In this respect, discourse is a representation that is built through language, as well as a representation that offers images of social reality (e.g. disability). This conceptualization of discourse has important consequences for the way it is approached: both a detailed linguistic analytical framework is adopted (e.g., systemic-functional linguistics) and a thorough social theory on discourse is drawn upon (e.g. Western Marxism; see also Section 3 below).
In this way, our study distances itself from relevant research on the views about disability, which have only focused on the content of the discourses of disability, and thus they have not also accounted for their structural description (e.g. Corker and French 1999). On the other hand, by choosing Facebook posts as data of the present study, we propose a complementary empirical source to interviewees' talk, in order to consider the shaping and perpetuation of hegemonic representations of disability in people's everyday (digital) practices.
Critical discourse analysis has gained currency in the last years as a research tool, as a result of the 'linguistic turn' of the social sciences, namely due to the acknowledgement of the central role of language and other semiotic systems in the shaping of social life. Critical discourse analytical research has explored the discursive construction of various social issues, ranging from racism (e.g. Van Dijk 1991), to neo-liberalism (e.g. Fairclough 2000) and environmentalism (e.g. ). Yet, there are a limited number of studies on the representation of disability adopting a critical discourse analytical perspective (Grue 2011). On the other hand, taking into account that critical discourse analysis has focused on institutionalized forms of talk and text, since these are the sites where issues of social dominance and inequality tend to occur (Luke 1997), it comes without surprise that most of the relevant research on disability has also concentrated on official and bureaucratic texts, such as policy (Grue 2009), legislative (Liasidou 2008) and referral documents (Nukoosing and Haydon-Laurelut 2010), or special education meetings (Rogers 2002). In contrast, much fewer studies have analyzed the representations of disability coming from less 'elite' social actors, such as teachers (Ferri et al. 2005) or student teachers (Stamou and Padeliadu 2009). To our knowledge, this is the first critical discourse analytical study to explore the representations of disability through the eyes of less institutionalized, ordinary voices, such as disability groups on Facebook.

Conceptual models of disability
The discursive construction of disability embodies a well-known contrast between a biological and a social understanding of this issue. Three conceptual models of disability could be mainly distinguished: an 'individual/medical', a 'minority/social' and an 'affirmative' one. Specifically, the individual/medical conceptualization of disability stresses bodily 'impairment' and limitations, which are considered either as a medical problem or as a personal experience/tragedy (e.g. Scior 2003). Consequently, in this model, medical and psychological accounts of disability prevail, and thus the social exclusion of people with disabilities is thought to be due to bodily dysfunctions, namely to the condition of disability itself. Under this view, society should invest in health services in order for people with disabilities to live a more 'normal' life.
From the 1970s onwards, however, the medical construction of disability has been heavily criticized, as a product of the rise of the disability rights movement at social level and the appearance of social constructionism at academic level. The criticism of the medicalization of disability has taken the form of alternative accounts of disability. Actually, it changed in a variety of ways: from a personal tragedy theory into a social oppression theory; from care to rights, and from individual adaptation to social change (Alevriadou and Lang 2011). Specifically, in the USA, a minority model has developed, according to which, people with disabilities are seen to have analogies with other oppressed social groups (Hahn 1996). Hence, in the minority model, disability is put sideby-side to other sources of discrimination, such as ethnicity, class and gender, and it is mainly understood as a form of minority culture. This conceptualization of disability seems to better fit the case of deafness (Grue 2009). On the other hand, in Britain, a social account of disability has been shaped, which focuses on the disabling social conditions (e.g. poverty, unemployment) which lead to the isolation of people with disabilities (e.g. Barnes, Mercer, and Shakespeare 1999;Oliver 1990Oliver , 1996. The common thread behind both approaches is that disability is not seen as a medical category (i.e. construed as 'impairment') but it is rather considered a phenomenon with social, political and cultural implications, which acts oppressively upon the life of people with disabilities. Consequently, under the minority/social view, people become disabled not because of their bodily dysfunctions but because of social barriers. This changing point of view was and remains very liberating for individuals with disabilities (Oliver 1996).
However, the minority/social model has been also attacked for still considering disability negatively, as a problem which is not located in the body this time but which resides in society. Moreover, it has been criticized for focusing on the social conditions of people with disabilities but neglecting people with disabilities themselves. Therefore, an affirmative view of disability has emerged, which puts a stress on the experience of people with disabilities and considers them in terms of their abilities rather their limitations (Alevriadou and Lang 2011;Swain and French 2000). Hence, contrary to the medical/individual model, which makes people with disabilities desire to become nondisabled, the affirmative view attempts to shape a positive identity of being disabled, focusing on the beneficial aspects of such a life. In this way, the dominant value of being 'normal' is challenged, by showing a positive attitude towards disability. Consequently, the affirmative model mostly relates to the common view of deafness as a different culture (e.g. Reagan 1985).
Despite the conceptual shift from a medical to a social understanding of disability among activists and scholars, the medical model remains a hegemonic discourse in society. Specifically, this is the dominant account of disability one still finds in mainstream cultural (e.g. Gold and Auslander 1999;Valentine 2001) and special education practices (Ferri et al. 2005;Vehkakoski 2004). This means that the medical model remains the prevailing cultural script which people with disabilities and/or their families need to challenge and resist. Therefore, although most research has shown that people with disabilities and/or their families tend to focus on a social explanation of disability (e.g. Cardona 2013), there are some studies which have underscored how the two models are at play in the formation of these people's representations, since they often need to negotiate between a view of disability as emerging from their own experiences with the one they receive from mainstream cultural messages (e.g. Bosteels et al. 2012). In other words, the two models of disability may not be as contradictory and mutually exclusive as they are often described in the literature, but rather to be at interplay (Grue 2009).

Critical discourse analysis: the sociocultural approach
Critical discourse analysis is not a monophonic and coherent framework but is a theoretical nexus consisting of heterogeneous strands (e.g. Fairclough's sociocultural, Van Dijk's sociocognitive and Wodak's discourse-historical approaches, among others: for an overview, see Wodak and Meyer 2001). Despite this diversity, all approaches of critical discourse analysis put a stress on the ideological role of language in the way the media and other institutional texts construct particular representations of the world. From the variety of critical discourse analytical approaches, the framework adopted here draws in particular upon the sociocultural strand of Fairclough (e.g. 1992Fairclough (e.g. , 2003). Fairclough's approach involves a three-dimensional framework for the analysis of discourse, as an attempt to link the micro-level of text and talk to the macro-level of underlying power relations in society, through the discursive practices upon which language use is drawn (meso-level).
First, discourse is seen as 'text', including the linguistic aspects of discourse. Hallidayan linguistics is a major analytical resource employed for textual analysis (Halliday 1994). It is particularly concerned with providing a social explanation of lexicogrammatical phenomena, and therefore it constitutes a valuable resource for critical discourse analysis. In particular, language is seen as a network of options among which language users make selections (e.g. lexical, syntactic) which are always meaningful. Thus, what a speaker chooses to say or not, and how to say it, has important consequences for the meanings conveyed. Moreover, language is seen as being multifunctional, fulfilling three functions simultaneously. Specifically, when we use language, we produce a particular version of the world ('ideational function'), we express our attitudes and build particular social relationships with others ('interpersonal function'), and we organize the ideational and interpersonal information of the text in a coherent whole ('textual function').
Second, discourse is viewed as 'discursive practice', which involves the processes of text production and consumption, and in particular, the norms and conventions employed to make and interpret a text. The most important set of rules that have an influence on what people think and speak/write about a particular topic (informed especially by the thought of Foucault) are 'discourses', that is, representations of social reality from a certain perspective. Hence, a consideration of discursive practice mainly concerns an 'interdiscursive' analysis, namely an exploration of what discourses are drawn upon in texts. Interdiscursive analysis allows the inclusion of elements of context into the analysis of texts, and therefore showing the relationship between occasional linguistic features of texts with more durable social structures.
Third, discourse is considered a 'social practice', which alludes to the broader social context that bears upon the text and the discursive practice, including the ideological effects and power relations with which texts are implicated. For the analysis of social practice, critical discourse analysis mainly draws on Western Marxist thinkers, such as Althusser and Gramsci. Althusser has stressed the role of discourse as the material expression of ideology in the reproduction of capitalist social relations, while Gramsci has developed the notion of 'hegemony' in order to define the consensual exertion of capitalist power, achieved through ideology.
The combination of the micro-linguistic with the macro-social level that critical discourse analysis offers via the meso-level of interdiscursivity, as defined by Fairclough, allows seeing the posts of disability groups on Facebook as textual material which inscribes, through specific lexico-grammatical choices, particular representations (or discourses) about disability (e.g. the medical, the social and the affirmative model), and eventually contributes to the construction, legitimization and perpetuation of dominant ideological meanings about disability issues.

Analytical resources for the micro-level textual analysis
In the micro-level of text, we focused on the analysis of ideational and interpersonal meanings of posts, drawing upon Hallidayan linguistics. Specifically, the ideational aspects of the posts were considered by means of the tools of 'lexicalization' and 'transitivity'. Vocabulary is the level of language that constructs par excellence the world. There are always multiple or alternative lexicalizations (Fairclough 1992) of experience, which shape different versions of social reality. Lexis offers various and contrasting depictions of reality; it may also indicate the amount of knowledge one possesses of a specific field (in the distinction between technical and colloquial vocabulary ;Fowler 1996). Moreover, lack (underlexicalization) or abundance (overlexicalization) of words for a specific area may mark the speaker's/writer's indifference or preoccupation with it, respectively (Fowler 1996).
On the other hand, through transitivity, language users give causal meaning to the world, namely by determining what groups of people do. On an analytical level, transitivity involves the interpretation of experience in terms of 'processes', that is, actions in a broad sense (typically realized through the grammatical category of verbs), and 'participants', namely animate and inanimate entities linked to processes (typically realized through the grammatical category of nouns). In other words, transitivity analyzes the world in terms of agency. An important use of transitivity is to determine the type of processes used for representing the world. Halliday (1994) has distinguished among 'material' (i.e. processes of doing/acting), 'mental' (i.e. processes of feeling/seeing/ thinking), 'relational' (i.e. processes of having an attribute/identity), 'verbal' (i.e. processes of saying), 'behavioural' (i.e. processes of behaving) and 'existential' (i.e. processes of existing) processes, based on the way these processes are worded. Another crucial element of transitivity is to determine which participant has an active role in the process, namely, initiates a process (i.e. the 'agent') and which participant has a passive role in the process, namely receives the impact of the process (i.e. the 'medium').
The interpersonal analysis of the posts included an account of 'personal deixis', 'speech acts' and 'evaluation'. Specifically, personal deixis involves any references made about the writer/speaker and/or reader/hearer and the social relationship established between them. These references are made through the selection of the person in personal and possessive pronouns, through which a personal (e.g. selection of first singular personal) or impersonal style (e.g. selection of third singular or plural person) can be built.
On the other hand, each post was also analyzed in terms of the 'speech acts' it performed (Austin 1962;Searle 1969). According to the theory of speech acts, when using language, people do not merely refer to the world but they also express their intentions and perform certain acts: they may express their belief about ('assertive act': e.g. statement, notification, information) or attitude towards something ('expressive act': e.g. congratulation, complaint, protest), they may ask the addressee to do something ('directive act': e.g. request, urge, recommendation), commit themselves to do something in the future ('commissive act': e.g. promise, threat, offer), or make a change in the world ('declarative act': e.g. nomination, blessing, renunciation).
Finally, 'evaluation' involves the construction of the speaker's social role through the expression of his/her feelings and subjective judgments in the text. Hence, 'evaluative statements are statements about desirability and undesirability' (Fairclough 2003, 172). The most obvious linguistic mechanism through which evaluation is manifested is the use of evaluative adjectives/adverbs and of emotional lexis.

Data collection
Digital technologies in general and the Internet in particular have been both seen as opportunities (e.g. improvement of the self-image and self-esteem) and obstacles (e.g. lack of assistive technologies, lack of learning support) for people with disabilities (for a discussion, see e.g. Bowker and Tuffin 2003;Van Kraayenoord 2010). In any case, they have changed the way both the non-disabled and people with disabilities communicate. In particular, the social media, such as Facebook, have given the opportunity to freely interact with and discuss several issues, including disability.
Specifically, in the present study, we analyzed the posts of sensory and physical disability groups on Facebook. Along with making 'friends', the participation as a member of several groups is one of the major activities for social networking. The 'Wall' is the core area of a group, where the recent activities of the group are posted and where discussions among members are made. 1 Since Facebook is illocal, it does not make sense to try to categorize groups in terms of nationality. Therefore, it was decided not only to focus on groups using Greek language, but also to include groups using English and French, namely to languages which we speak.
Data collection was made in thirteen disability groups, which were randomly selected. Specifically, focusing on sensory (blindness, deafness) and physical disability, groups were determined through a search with relevant keywords (e.g. 'disability', 'handicap', 'blindness', 'deafness', 'mobility problems'). Joining the groups found through this procedure as members, we were able to have access to the posts of each group. The disability groups which were finally selected to be included in the present study were the most active ones in terms of the comments exchanged and posted (e.g. members exchanged posts regularly). In contrast, groups with low activity (e.g. one post in every ten days) as well as groups posting comments which were irrelevant to the purposes of the study (e.g. socialization and arrangement of dates among members) were excluded. From the thirteen groups selected, four talked about issues of blindness, two groups discussed deafness, and seven groups referred to mobility problems (see Table 1). Data collection took place from June to December 2010. Considering that critical discourse analysis focuses on language form, and thus it cannot be easily applied to a large amount of textual data, 235 posts from a total of 1175 comments posted by the groups in this period of time (i.e. amounting for 20% of the data) were randomly selected to be included in the present study.

Overview
Depending on the general topics and the approach adopted to talk about disability, posts were categorized under a 'medical' (e.g. excerpt 1), a 'social' (e.g. excerpt 2) and an 'affirmative' (e.g. excerpt 3) conceptualization. Although most of the posts were unproblematically classified, there were few posts for which there was some doubt. In those limited cases, we consulted two colleagues of special education expertise, and posts were finally classified into the category for which there was unanimous consensus. After the classification of each post into the relevant conceptualization of disability through thematic analysis, linguistic analysis followed for each post classified into the relevant model (see Section 5.2. and 5.3. below).
As it can be seen in Table 2, most posts were framed according to a social understanding of disability. However, it seems that the type of disability plays a crucial role in the way people with disabilities and/or their families represent disability. Specifically, blindness was equally represented through a medical model. On the other hand, deafness was equally constructed by means of an affirmative model, while physical disability was massively conceptualized according to a social account of disability. This indicates that people with disabilities and/or their families in general tend to talk about disability by underlining the social aspects of being disabled. However, they also seem to be influenced by the way particular types of disability are dominantly framed in society. Blindness is often approached as a personal experience and a medical condition, while deafness has a strong tradition in focusing on the positive aspects of deafness (e.g. the view of deafness as a different minority culture). Finally, physical disability has been stereotypically linked to a social understanding of disability (e.g. by often highlighting how people with mobility problems lack access to pavements etc.). These results show the ambivalent way in which issues of disability are discussed: (1) Hi, my cousin went blind about 2 years ago. He got into an accident and he cannot see at all. The doctors said that can't fix his both eyes. However, he is still having a hope that he gets better. And he is very upset that he cannot see, he cries and pray. His life is very hard, because he went from the seeing world to total darkness. Some people go crazy when this happens to them, but his hope is keeping him up. He is home nowadays and waits to go to a blind school in September. His hobbies are using computer, chatting and making new friends. It would be great if you add him and share your stories (Blind People Unite) 2 (2) My sister is totally blind but this does not stop her studying for a law degree and being the mother of 3 kids. Unfortunately, some people do treat her as though she is brain dead or are reluctant to use words like 'see' when talking to her. She often gets asked how she has managed to have children. She usually replies 'I had sex'. She uses lots of different aids including a mad guide dog called Usher (Blindness, Perceptions and Prejudice) (3) I myself am not deaf, but my sister is. She is a wonder to be around! My closest friends are deaf as well. They can flow through life just as well as those of us without impairments, it's quite amazing (I'm just Deaf, not an Alien) More specifically, most of the posts categorized under the social model focused on issues with social discrimination people of disabilities receive (e.g. excerpt 4) as well as on the lack of proper services and education (e.g. excerpt 5). In the case of physical disability, many posts also put a stress on the environmental obstacles caused by the way cities are structured, which exclude people with mobility problems (e.g. excerpt 6). On the other hand, the posts framed according to a medical understanding of disability underlined how disability is an obstacle for the development of the person and his/her family (e.g. excerpt 7) as well as on the medical treatment of disability (e.g. excerpt 8). Finally, the posts representing disability following the affirmative model highlighted the benefits of being disabled. However, in most of the cases, the construction of disability as ability seemed to be more like an attempt to subvert the negative construction of disability as deficit, namely more like a strategy on the part of people with disabilities and/or their families to manage a negative identity (e.g. excerpt 9). Furthermore, some members of the groups analyzed considered disability as a blessing and divine gift given to some selected people with special abilities (e.g. excerpts 10, 11): (4) The dog doesn't quite do it for me. Sometimes I feel that I can't have any respect for a creature that I can't see. People typically beep at me when I cross crosswalks because my dog is slow, so it takes a while to get to the other side. Sometimes it's even worse in other ways because my dog pissed on a pedestrian and I didn't even know it and the pedestrian got mad at me because they were wearing nice clothes. So now I'm getting sued and the only way they'll drop the suit is if I buy them a new jaguar fur coat of equal or higher value (Blind People Unite) (5) Her learning at school is yet to begin, without the right guidance, care and specialized teachings. What will the future hold for little Mary? Are they going to teach Braille at state school now?!!!! (Save the Education of Blind People and Services to the Blind) (6) It is sad. I have a brother with spastic quadriplegia and the State doesn't give him the opportunity to go out for a ride in the city (Respect People with Disabilities) (7) All she hears when you speak is a whisper. She will never hear a bird sing. She'll never hear the wind blowing the leaves. Every time you whispered a lullaby she didn't hear a thing. She may never learn to read or write. She may never learn to speak. These things I hear every day, that 'she won't'. Make my insides weep each week (I'm just Deaf, not an Alien) (8) Good morning to everyone! My daughter makes me very happy. Life is not always easy but she is healthy and has a progress at her own pace. We haven't found a medicine yet for comforting her crises. These last days, her crises were very painful, but I tried to be strong. We have to see her doctor again to change the treatment. I didn't really want to put her on a treatment, but we have to help her become calmer because she suffers a lot. Well, I wish patience to all of you!! (A Handicap is first of all a Human Being]) (9) It still is a bit fun sometimes. You can't hear annoying sounds from outside and at night to keep you awake, during 4th of July you don't get a headache, you can always turn off your hearing aids so you don't have to listen to things you don't want to hear, and a bunch of other stuff. And with certain aids you can switch between hearing or not. So it's kind of fun a bit to be deaf and I 'm proud of who I am (I'm just Deaf, not an Alien) (10) God gives them intelligence and abilities that are beyond our imagination. So, we have to accept them and try to help them, this is the most important thing, isn't it? (Blind People in Tunisia) (11) When my daughter was born and the doctor said that she was hearing, I cried, not of joy but of sadness (Be Proud of Being Deaf)

Ideational aspects of the posts
Since naming is a key process in the segregation of people, what words are and should be used for referring to people with disabilities is an ongoing debate in both disability studies and activist groups (e.g. Devlieger 1999). Regarding the lexical selections made for defining themselves and the others, it is observed that despite some commonalities, these were to some extent also affected by the type of disability and the conceptual model according to which the posts were framed (see Tables 3-5). Specifically, in all three models of disability, the neutral words 'blind people', 'deaf people' and 'disabled people/people with physical disability' were used. However, posts drawing upon the medical model also tended to adopt labels which stressed the view of disability as deficit (e.g. 'visually impaired people', 'people with hearing loss'), or as tragedy (e.g. 'people in total darkness'). On the other hand, posts on deafness which were framed according to the social and affirmative models focused as well on the representation of deafness as a distinct culture (e.g. 'deaf community', 'the world of deaf people'). In contrast, in the case of physical disability, posts understanding disability in social terms also tended to overlexicalize people with disabilities based on the social barriers they face (e.g. 'people with mobility problems', 'inferior citizens'), while posts framed according to the affirmative model also defined them in a very positive light, by underlining their capacities (e.g. 'exceptional people/people with unique abilities'). With respect to non-disabled people, much fewer lexis were employed (underlexicalization), and to the most part, affirmative (e.g. 'the seeing world', 'hearing people') rather than negative labels (e.g. 'people without impairments') were adopted. It is noteworthy, though, that posts from groups of physical disability used an abundance (overlexicalization) of extremely negative (e.g. 'pigs', 'donkeys') or political terms (e.g. 'the State') to define non-disabled people. This indicates the highly pejorative character of the majority of the posts on physical disability that were against the disabling social barriers.
Regarding the roles of people with disabilities constructed in the posts analyzed, the transitivity analysis showed that in the case of blindness and physical disability they were mostly represented in 'medium' role, namely as recipients of processes initiated by nondisabled people (37.7% and 36.4%, respectively). In contrast, in posts coming from groups on deafness, in the majority of the processes (58.3%), people with disabilities were put in agentive roles. This could be explained by the different conceptual models which prevail in the representation of each type of disability. The medical and social models dominating in the posts of blindness and physical disability underscore the agency of non-disabled people, namely doctors (e.g. 'doctors messed up my dad's eyes') and society (e.g. 'people typically beep at me when I cross crosswalks'), respectively. Conversely, the affirmative model upon which many posts of deafness drew focuses on the agency of people with disabilities themselves (e.g. 'you can always turn off your hearing aids so you don't have to listen to things you don't want to hear').
The transitivity analysis of the processes used indicated that those posts approaching disability in medical terms mostly employed mental and material processes (Table 6). In this way, disability was constructed with a focus on what people with disabilities are not able to perceive (e.g. 'she will never hear a bird sing') as well as on what actions they receive from doctors (e.g. 'the doctors said that can't fix his both eyes'). In the social model, mental and material processes also prevailed, through which a reference was made to what society does to people with disabilities (e.g. 'my big sister, who is deaf, was rejected from a school she had applied to') and to how they feel because of such a behaviour (e.g. 'I hate human stupidity, those people looking at people with disabilities as if they were mysterious beasts'). Finally, in the affirmative model, disability was mainly represented through relational and mental processes, namely a stress was put on how people with disabilities describe and define themselves (e.g. 'they are people like us, and in fact with more abilities'), and on what they are able to perceive (e.g. 'they point to things that people have forgotten').

Interpersonal aspects of the posts
The analysis of the use of personal deixis in the posts showed that people with disabilities and/or their families resorted to different selections in personal deixis, depending on the perspective from which they approached disability (Table 7). Specifically, in the posts which medicalized disability, the third singular and plural person prevailed (e.g. 'he got into an accident and he cannot see at all'). In this way, an impersonal style was adopted, through which text producers could make their ideas seem less subjective and more emotionally neutral and generally acceptable. In contrast, in the posts of the social model, the first singular and plural and the second singular person were mostly selected. 3 Thus, a highly personal style was adopted, which was sometimes confessing (e.g. 'The music fills you without even having to work at listening to it', 'Oh! how I wish I had learned Braille as a child. My story is not unique or an exception. Hundreds of blind adults feel they missed out on a proper education'), and in other times signaled collectivity, universality and political action (e.g. 'We live in a country that offers nothing for these people'). Finally, in the affirmative model, both a third and a first singular person were mainly selected. The impersonal character of the former generalizes and naturalizes the abilities of people with disabilities (e.g. 'they can flow through life just as well as those of us without impairments'), while the latter gives a highly emotional and personalized overtone in the account of disability (e.g. 'I am deaf and I love it').
Regarding the speech acts performed in the posts, assertive ones dominated (Table 8). This is quite expected, given the fact that the posts mainly concerned statements and descriptions about disability. Nevertheless, in the case of the posts of the medical model, assertive acts were much more numerous, because their aim was simply to describe how difficult the situation is when you are a person with disability (e.g. 'his life is very hard'). Conversely, posts conceptualizing disability in social terms also contained several directive acts, through which text producers directly addressed readers and members of the group by asking them for action against the harsh social conditions people with disabilities face (e.g. 'we should protect people belonging to vulnerable groups of our society'). Finally, in the affirmative model, except for assertive acts, there were many expressive ones, through which a desirable account of disability was provided (e.g. 'it gives me pleasure to see them make these delicate signs'). In the posts analyzed, different means of evaluation were extensively used, through which people with disabilities and/or their families showed their emotional engagement in the issue. Evaluation was primarily expressed through punctuation marks, such as ellipsis and exclamation marks. Ellipsis was employed to express emotional movement (e.g. 'I'm proud of being deaf…') as well as some insinuation ('it's harder when I am the only deaf person in my school…'). Exclamation marks were used to express more intense feelings, such as enthusiasm (e.g. 'it's wonderful to be deaf!'), exhortation (e.g. 'we must move forward, and education is the key!') or outrage (e.g. 'municipal infrastructures are insufficient!').
Another central means for the expression of evaluation in the posts were several adjectives, adverbs and emotional lexis. In the case of posts framing disability from a medical perspective, evaluation was employed in order to underline the personal tragedy people with disabilities experience (e.g. 'the life of a blind person is very hard and some people go crazy when this happens to them'), in order to cope with their 'sickness', as they often characterized it. This sometimes leads to a fatalist attitude (e.g. 'sometimes it's very difficult, but what can you do?'). On the other hand, in the posts which represented disability in social terms, evaluative words referred to the misconceptions held about disability (e.g. 'they think I'm crazy') and the discrimination people with disabilities experience from society (e.g. 'people must stop looking at deaf people indiscreetly'). Moreover, as it was mentioned in the section of lexical choices above (4.2.), in the posts from groups of physical disability, extremely negative labels were employed for alluding to non-disabled people, who physically restrict people with disabilities (e.g. parking on the pavements). Finally, in the case of affirmative model, there were many evaluative words in order to underline the abilities and benefits of being disabled (e.g. 'angel', 'amazing and hard work of God'). These words were usually accompanied by intensifiers (e.g. 'very smart and intelligent people', 'it's quite amazing', 'deafness is cool').

Conclusions
The present study on the posts of disability groups on Facebook suggests that people with disabilities and/or their families do not constitute a homogeneous group of people being connected because of their common (bodily) condition. Instead, they tend to hold diverse views and ideological standpoints from which they interpret disability and attempt to find solutions to their problems. This means that 'disability' should not be considered as an all-inclusive label but rather as an umbrella term encompassing different (and often contrasting) representations of what disability is and of what it means to be a disabled person.
Specifically, although the social model of disability seems to have currency among people with disabilities and/or their families, the type of disability also plays a central role in the way disability is conceptualized across the different disability groups on Facebook. Thus, the analysis of relevant posts indicates that blindness is equally framed on medical terms, while deafness is also represented through an affirmative model of disability. Physical disability groups, on the other hand, seem to be more coherent from an ideological perspective, by approaching disability mostly as a sociopolitical issue. These findings confirm other studies, which have revealed how the different models of disability are not necessarily mutually exclusive, but are often at interplay (e.g. Bosteels et al. 2012;Grue 2009;Low 1996).
As emerging from our analysis, the contribution of the present study to this discussion is that the type of disability plays a significant role in this interplay among the different models of disability. Hence, blind people and/or their families need probably more to challenge the medical model of disability, as a prevailing cultural script, not only of disability in general, but also of blindness in particular. On the other hand, deaf people, under the influence of the affirmative model, tend to focus on a positive but largely individualistic conceptualization of disability. In this way, it would be probably more difficult for them to be involved in collective actions against the social exclusion of people with disabilities, contrary to people with physical disability, who seem to frame disability as a mostly social problem.
On the other hand, the critical discourse analysis of the posts unveiled that each model of disability does not only mean the discussion of different topics but also the selection of distinct lexicogrammatical features (e.g. transitivity, personal deixis, speech acts) on the part of post producers. Thus, the medical model involves linguistic features (e.g. third person, assertive acts) which provide a commonsensical account of disability, as is the case with well-established and dominant discourses (of disability). Conversely, the social model concerns linguistic selections (e.g. first and second person, directive acts) which give a politicalized and activist account of disability, whereas the affirmative model constructs an idealized and personalized conceptualization, through the selection of first person and expressive acts. Interestingly, each model of disability also suggests a different use of Facebook. For post producers medicalizing disability and representing it as a personal tragedy, Facebook is a sort of postmodern diary, in which they can confess their problems stemming from their bodily limitations. For those approaching disability on social grounds, Facebook is mainly a tool for promoting their political positions, while for those defining disability in an affirmative manner, Facebook is an outlet for making known their distinct culture and lifestyle.
Research on Facebook has several advantages, such as easy access to the data and collection of spontaneous and unplanned (at least for the researcher) talk. However, it has also some limitations. First and foremost, since Facebook is a particularly fluid and elusive form of communication, the researcher has the feeling of taking an instant photo of the phenomenon he/she investigates. On the other hand, it does not easily allow the consideration of the social profile (e.g. gender, age, education, social class) of the participants of the study. Hence, in our case, we were not able to explore to what extent the socioeconomic background of people with disabilities and/or their families probably affects the conceptualizations they hold about disability. In particular, departing from our observation that more female than male relatives of people with disabilities participated with posts in the disability groups, a future study could perhaps investigate whether women are more willing to talk about the disability of some member of their family, and to what degree this probably relates to the gendered aspect of the care of people with disabilities.
Finally, as a first attempt to investigate posts of disability groups on Facebook, the present study left some issues unexplored, which could be addressed by future research. On the one hand, we did not consider any possible differences between representations from people with disabilities as a group compared with their families as a group. On the other hand, although we acknowledge the fact that identities and positioning are coconstructed in interaction, posts were analyzed in isolation rather than as an interactional exchange, for which other analytical resources should be combined with critical discourse analysis (e.g. the model of 'identities in interaction' by Bucholtz and Hall 2005).