Blindness and the Interactional Emergence of Disability

By analysing interactions between blind and sighted individuals, this article shows that, under certain circumstances, the process of social interaction can identify blind participants not only as impaired, but also as disabled. When blind individuals interact with sighted strangers, they often need to ensure that the sighted people know about their condition. However, the very act of ‘‘communicating the impairment’’ may portray them as limited, either functionally (e.g. being unable to find one’s way), cognitively (e.g. being unable to understand a technical object) or socially (e.g. embarrassing others or oneself). Despite its disabling character, such a portrayal constitutes a rational solution to an interactional problem the problem of ensuring that all participants are mutually aware of the fact that one participant is impaired. The article concludes that blind interactants are routinely confronted with a dilemma, as they find themselves forced to trade maximal social respect and empowerment for improved co-operation and efficiency. An important component of the current agenda of social-scientific disability research is that of explaining the relationship between impairments and disabilities (Michailakis 2003). This agenda is motivated by the observation that impairments medical identifications of permanent defects of body or intellect and disabilities social identifications or treatments of individuals as somehow permanently defective to a certain extent vary independently (Tössebro & Kittelsaa 2004). While some impairments may essentially be inconsequential in most social situations (e.g. mild cardiovascular diseases), some disabilities need not be anchored in actual impairments (e.g. a person who is perceived as someone who ‘‘has a limp’’ may simply be someone who temporarily has his or her leg in plaster). Yet the contingency of the link between the two phenomena is only partial depending, among other factors, on the type of impairment (a comparative point made most forcefully by Goffman (1990)). This fact makes it all the more necessary to ask why certain types of impairments are so likely to be accompanied by certain types of disabilities, while others are not (Low 2001). Focusing on the case of blindness, this article explores one specific factor that increases this likelihood. I argue that, under certain circumstances, the logic of social Correspondence: Wendelin Reich, Associate Professor, Uppsala University, Department of Sociology, PO Box 624, SE-751 26 Uppsala, Sweden. Email: wendelin.reich@soc.uu.se 1501-7419 Print/1745-3011 Online/06/04000247 16 # 2006 Taylor & Francis DOI: 10.1080/15017410600630933 Scandinavian Journal of Disability Research Vol. 8, No. 4, 247 262, 2006 interaction (what Goffman (1983) calls the ‘‘interaction order’’) can contribute to the portrayal of blind individuals not only as impaired, but also as disabled. In applying the distinction impairment/disability, this article takes a constructivist stance. My assumption is that both concepts involve identifying or treating an individual as somehow deviant, abnormal, limited or dysfunctional, but that they describe this condition in different ways (see Michailakis & Reich 2005). The observation of an impairment reduces the individual’s condition to perceived abnormalities or dysfunctions in his or her body, psychology or intellect. The observation of a disability, on the other hand, describes the individual as permanently unable to fulfil certain functions that (statistically speaking) ‘‘normal’’ individuals could be expected to fulfil. It is important to see that both constructs refer to an individual (i.e. it is not the individual’s environment which is observed as impaired or disabled), but that this does not mean that they are ontologically given or created by the individual himor herself. Whatever an individual’s physical condition may be, without society he or she could neither be diagnosed with, say, tinnitus (which requires a medical system able to attribute such an impairment) nor declared, say, ‘‘25% work-incapacitated’’ (which requires a social security system able to attribute such a disability). In the following, I shall use the term ‘‘disablement’’ to denote the informational process of attributing a disability to a person (as opposed to the physical process of ‘‘hampering’’ the person, e.g. by placing boardwalks and stairs in his or her way). The type of disablement I am interested in may be termed (public) interactional disablement insofar as it occurs in, and remains confined to, social interaction between unacquainted impaired and non-impaired individuals. From a social-scientific point of view, blindness is a type of impairment that seems particularly suited for analysing the distinction between impairment and disability. In everyday life, blind individuals tend to be forced to rely to a significant extent on help from sighted individuals. For example, blind individuals report that it is usually impossible for them to orient themselves alone in parts of an urban area that they are not familiar with (Berndtsson 2001:205). Similarly, managing not fully routine everyday tasks, such as buying groceries, often requires some form of interactionally provided and not necessarily reciprocated help from others help that goes beyond the typical forms of co-operation that people (e.g. shop assistants) are expected to afford others in contemporary Western society (e.g. informing the customer about the location of a product). For this help to be possible and effective, it is usually necessary that sighted individuals be aware of the fact that they are dealing with blind individuals. For the sake of interactional trust, it is just as important that blind individuals can be certain about sighted individuals’ awareness of their impairment. If blind individuals do not have this certainty, they are less able to predict the subsequent course of the interaction and may experience distress (Berndtsson 2001). At the same time, this certainty is not a matter of course. For example, some but not all cases of blindness are accompanied by 248 W. Reich

An important component of the current agenda of social-scientific disability research is that of explaining the relationship between impairments and disabilities (Michailakis 2003). This agenda is motivated by the observation that impairments Á medical identifications of permanent defects of body or intellect Á and disabilities Á social identifications or treatments of individuals as somehow permanently defective Á to a certain extent vary independently (Tö ssebro & Kittelsaa 2004). While some impairments may essentially be inconsequential in most social situations (e.g. mild cardiovascular diseases), some disabilities need not be anchored in actual impairments (e.g. a person who is perceived as someone who ''has a limp'' may simply be someone who temporarily has his or her leg in plaster). Yet the contingency of the link between the two phenomena is only partial Á depending, among other factors, on the type of impairment (a comparative point made most forcefully by Goffman (1990)). This fact makes it all the more necessary to ask why certain types of impairments are so likely to be accompanied by certain types of disabilities, while others are not (Low 2001). Focusing on the case of blindness, this article explores one specific factor that increases this likelihood. I argue that, under certain circumstances, the logic of social interaction (what Goffman (1983) calls the ''interaction order'') can contribute to the portrayal of blind individuals not only as impaired, but also as disabled.
In applying the distinction impairment/disability, this article takes a constructivist stance. My assumption is that both concepts involve identifying or treating an individual as somehow deviant, abnormal, limited or dysfunctional, but that they describe this condition in different ways (see . The observation of an impairment reduces the individual's condition to perceived abnormalities or dysfunctions in his or her body, psychology or intellect. The observation of a disability, on the other hand, describes the individual as permanently unable to fulfil certain functions that (statistically speaking) ''normal'' individuals could be expected to fulfil. It is important to see that both constructs refer to an individual (i.e. it is not the individual's environment which is observed as impaired or disabled), but that this does not mean that they are ontologically given or created by the individual him-or herself. Whatever an individual's physical condition may be, without society he or she could neither be diagnosed with, say, tinnitus (which requires a medical system able to attribute such an impairment) nor declared, say, ''25% work-incapacitated'' (which requires a social security system able to attribute such a disability). In the following, I shall use the term ''disablement'' to denote the informational process of attributing a disability to a person (as opposed to the physical process of ''hampering'' the person, e.g. by placing boardwalks and stairs in his or her way). The type of disablement I am interested in may be termed (public) interactional disablement insofar as it occurs in, and remains confined to, social interaction between unacquainted impaired and non-impaired individuals.
From a social-scientific point of view, blindness is a type of impairment that seems particularly suited for analysing the distinction between impairment and disability. In everyday life, blind individuals 1 tend to be forced to rely to a significant extent on help from sighted individuals. 2 For example, blind individuals report that it is usually impossible for them to orient themselves alone in parts of an urban area that they are not familiar with (Berndtsson 2001:205). Similarly, managing not fully routine everyday tasks, such as buying groceries, often requires some form of interactionally provided and not necessarily reciprocated help from others Á help that goes beyond the typical forms of co-operation that people (e.g. shop assistants) are expected to afford others in contemporary Western society (e.g. informing the customer about the location of a product).
For this help to be possible and effective, it is usually necessary that sighted individuals be aware of the fact that they are dealing with blind individuals. 3 For the sake of interactional trust, it is just as important that blind individuals can be certain about sighted individuals' awareness of their impairment. If blind individuals do not have this certainty, they are less able to predict the subsequent course of the interaction and may experience distress (Berndtsson 2001). At the same time, this certainty is not a matter of course. For example, some but not all cases of blindness are accompanied by easily visible physical cues. In addition, even when such cues are present, interaction partners may initially overlook them. This constitutes a practical problem for blind individuals because the very condition of blindness entails that they do not have the possibility of observing other individuals' (lack of) non-verbal responses to their impairment Á e.g. a prolonged gaze or a surprised expression.
It can be concluded that blind individuals are forced to rely largely on verbal cues, provided by interaction partners or by themselves, in order to get or provide a definitive answer to the question of whether others have noticed their condition. In a set of 16 interactions between blind customers and sighted shop assistants that I audio-recorded in two larger Swedish cities in 2004, 10 interactions contain clearly identifiable utterances that mark a transition from uncertainty to certainty regarding this issue. These utterances turn the fact that one participant is blind into knowledge that is (potentially) mutual to all participants, that is, known by all, known to be known by all, known to be known to be known by all, and so on. 4 Remarkably, in all 10 cases, the utterance in question, or its immediate sequential context, describe the blind participant as disabled. Specifically, the blind participant becomes identified as someone who not only is unable to see, but also is unable to fulfil either: . certain practical functions (e.g. to find his or her way around a building), . certain cognitive functions (e.g. to comprehend what others are saying), or . certain social functions (e.g. to avoid embarrassing his or her interaction partners).
This article analyses the interactional emergence of these identifications by way of a broadly conversation-analytical approach (Have 1999). I shall start by giving a brief overview as well as an example application of this method. I will then apply it in order to provide focused analyses of the 10 utterances in question, as well as of their immediate sequential contexts. To avoid expectable objections, I state explicitly that my argument will not be that the portrayal of impaired individuals as disabled is likely to occur under certain circumstances Á this is a question of representativeness that a future study may address. Rather, my argument will be that such a portrayal is, under certain circumstances, reasonable. Put simply, it constitutes a solution to an interactional problem faced by both blind and sighted participants: the problem of having to explain to listeners why the information that one participant is impaired should be considered interactionally relevant. I thus try to show that the very act of turning the fact that one participant is impaired into mutual knowledge can proceed according to a pattern that constructs disability and thereby motivates or justifies potentially disempowering behaviour on the part of the nonimpaired. Because this interactional construction of disability is not the result of participants' intentions, one may say that disability emerges in the interactions considered below.

Data and Method
During a pre-study to a larger project on everyday interaction experiences of blind and severely vision impaired individuals, I audio-recorded interactions (n 0/16) between blind shoppers and sighted salespersons in Sweden. 5 For both cultural and economic reasons, sales encounters in Sweden and many other countries tend to be highly co-operative types of social interaction. In contrast to cultures in which bargaining is the norm, the successful sales encounter in these countries is expected to focus on the wishes of the customer rather than the prices of products and to proceed in a friendly, matter-of-fact manner. Salespersons tend to respond to this expectation, no matter if motivated by themselves or by their employers, by displaying overt politeness toward the customer. Through informal complementary interviews carried out in connection with the recordings, I confirmed: (i) that interactions of this type are a normal and important part of the everyday life of blind Swedish citizens, (ii) that all participants went to shops they would potentially visit in their everyday lives (e.g. some blind individuals report that they do not buy clothes for themselves), and (iii) that none of the participants was previously acquainted with the salespersons with whom they interacted.
The social situation underlying the recordings has an important property. Because one person is blind, and because the interactants are unacquainted, we know that the impairment is initially not mutual knowledge. It is the case that some forms of blindness are accompanied by visible physiological deformations of the eyes or the eye region, and many blind Swedish individuals use a thin white cane. Nevertheless, all such cues can be overlooked or misunderstood, and blind individuals are unable to tell from the mimic or gestural behaviour of others whether their condition has been overlooked or misunderstood. If it takes place at all, the transition from nonmutual to mutual knowledge is therefore almost guaranteed to be observable in the audio-recordings.
In 10 out of the 16 recorded interactions, the transition is clearly identifiable. In the following section, I will analyse relevant portions of them. The audio recordings of these interactions were treated according to the standard procedures of the most well-known and widely used methodological framework for detailed qualitative analysis of communicative interaction, that is, ''conversation analysis'' (Have 1999). Raw transcripts were made in order to be able to identify portions in which the transition from non-mutual to mutual knowledge about the vision impairment occurs. These portions were then transcribed according to conversation-analytical conventions, 6 and they are reproduced below. Conversation analysis does not define (or exclude) specific guiding questions by means of which relevant communicative patterns can be discovered in the material. The guiding question for the analysis reported here was: ''What disability-related assumptions about the blind participant are conveyed when his or her impairment is turned into mutual knowledge?'' An early finding of the analysis was that this knowledge can be arrived at by both explicit and implicit means. While statements such as, ''I am blind'' or, ''I understand that you are blind'' obviously create mutual knowledge about the vision impairment, any statement that expectably allows all listeners to infer this information fulfils the same function (Grice 1989b, Reich 2003. It is always possible to imagine additional, unobservable factors that prevented the emergence of mutual knowledge in the actual situation (e.g. the listener was inattentive or misunderstood the speaker). However, usage of such factors in the interpretations provided below is prevented by Occam's razor: given the information available, I shall argue that it is most reasonable to assume that by way of certain utterances, mutual knowledge of the impairment emerged.
A first example of an implicitly created transition from non-mutual to mutual knowledge may serve to illustrate both the conversation analytical approach and the type of problem that can ensue if a sighted interactant fails to take account of the fact that he or she is dealing with a blind person. The following transcript reproduces the first 25 seconds of an encounter between a shop assistant (S) and a blind customer (C). Three things are worth noting. Firstly, in giving directions (lines 5'/7), S fails to display orientation to the fact that the customer is blind. As it is predictable that a blind person is unable to resolve references such as ''left to the information desk'' (line 7), both C and we, the observers, can conclude that S is not aware of the fact that he is interacting with a blind person. C responds hesitantly by first uttering a token of shy confusion (line 8) and then Á after a long, noticeable pause Á asking abashedly for guidance (line 11). The Blindness and the Interactional Emergence of Disability 251 psychological seriousness of embarrassing situations such as these should not be underestimated. In addition, they are difficult to anticipate. In this concrete case, C's personal assistant had left C standing about 2 metres away from S, and a small bookstand happened to be between them, which could explain why S did not seem to notice that he was dealing with a blind individual.
Secondly, line 11 contains not only a request, but also an explanation (Antaki 1994). C justifies his request by explaining that he has ''a little trouble understanding''. By way of referring to his inability to ''understand'', C effectively implies that he has some form of unusual condition that renders him unable to make use of directions. In addition, he reveals that he assumes that S is not yet aware of this condition. Thus, line 11 suggests that C himself attributes S' failure to orient to his vision impairment to S' lack of knowledge of this condition (to C, this lack of knowledge is obviously unexpected). For this reason, the transcript illustrates a situation where a blind individual clearly orients to the fundamental problem that I discussed in the introduction, that is, ensuring that co-interactants are aware of his impairment.
Finally, C justifies his request by referring to a self-attributed inability. A wealth of other justifications would be conceivable Á for instance, C could claim that the architecture of the shop (or S' description of it) is too complex. Nonetheless, C chooses to describe himself as someone who is unable to comprehend simple spatial descriptions, and he does so although he initiated the encounter by demonstrating his general knowledgeability. 7 In practice, C justifies his request by implying that he is not just physically impaired, but also disabled relative to the complex physical environment in which he tries to orient himself.

Patterns of Interactional Disablement
In order to turn an impairment (whether one's own or that of a coparticipant) into mutual knowledge, a speaker needs to establish some kind of communicative reference to it. As said above, this reference can be both explicit and implicit, where ''implicit'' is defined as relying on an inference that all normal, adult, interaction-competent listeners can be expected to carry out. Two types of implicit references can be distinguished: hints, where the speaker must be assumed to intend to convey the information in question, and slips, where the speaker appears to reveal this information unintentionally. It is worth mentioning that the distinction between hints and slips is not a psychological one, as participants' actual intentions are never observable for anyone but themselves. The distinction thus refers exclusively to communicative artefacts (Luhmann 1995:ch. 4). Information counts communicatively as un-/intentionally conveyed if the listener can be expected to infer that it was conveyed un-/intentionally, irrespective of the speaker's real intentions. In the following, I will treat the three resulting types of references (explicit, implicit and intentional, implicit and unintentional) separately.

Mutual Knowledge Through Explicit References
In three of the interactions of the corpus, the transition to mutual knowledge about one participant's blindness is produced by means of explicit references to this fact. In all cases, it is the blind participant who initiates the transition. Consider transcripts 2 to 4.   The three focal lines are marked with arrows ( 0/) and share several interesting properties. Firstly, they occur very early in each interaction (counted from the initial greeting, all take place within the first 10 seconds). Secondly, they are not elicited by S, the salesperson (e.g. by means of a question), but occur seemingly unmotivated within the interactional sequence. Thirdly, they are dignified by an accepting reply on the part of S, which confirms to C, the customer, that the vision impairment has been turned into mutual knowledge, and which implies that S will take it into account. Fourthly, they involve a brief and simple description of the condition (''vision impaired'', ''do not see'', ''see poorly'', respectively). Finally, they are followed (rather than preceded) by an explanation, and the explanation comes immediately after the focal statement (rather than later on) Á in transcripts 2 and 4, statement and explanation are even tied together syntactically by way of a conditional.
Explanations that interlocutors provide for their behaviour justify this behaviour not only socially, but also sequentially Á in Grice's terms, they specify the (interactional) relevance of the explained statement (Grice 1989a, see also Antaki 1994, Heritage 1988. As listeners generally expect any interactional contribution to be somehow relevant to the interaction (Sperber & Wilson 1995), explanations thus rationalize the occurrence of otherwise seemingly surprising or even irrational statements. This means that we need to consider why an unsolicited statement such as, ''I am vision impaired'' would be odd if it stood by itself. The answer lies in the sequential context. In transcripts 2Á4, explanations would hardly have been necessary if the focal statements had been elicited by the salespersons. Not having been asked by the salespersons about their condition, the blind customers are forced to provide reasons why the fact that they are blind matters to the interaction. Of course, these reasons contain information (e.g. the customer's wish to purchase a stereo that is easy to use) that may have been the (unobservable) motive of mentioning the vision impairment in the first place. Yet, their sequential positioning after the statements that mention the impairment marks them interactionally as justifications rather than mere motivations.
Given that different types of explanation would have fulfilled the purpose of making the focal statements relevant, it becomes meaningful to ask why all three customers choose justifications that point out functional restrictions on the articles they are interested in: the customer in transcript 2 needs a steam iron that ''works for me''; the one in transcript 3 wants to have a music player with ''more buttons''; the one in transcript 4 needs a stereo that is ''easy to use''. These functional restrictions mark the customers as individuals who require them and who are therefore, compared with individuals who do not have need of them, disabled. My interpretation of this choice is that it would be difficult to frame the interactional relevance of statements similar to, ''I am vision impaired'' in the given contexts as anything else than a disability. The only actual choice the blind customers of transcripts 2Á4 seem to have is to mention other types of disabilities Á for instance, cognitive disabilities (e.g. by saying ''easy to understand'' instead of ''easy to use''). This option does not appear to be very attractive Á being a person who wants his or her music player to have many buttons is usually preferable to being a person who, say, does not understand their functions in the first place. Furthermore, reference to a functional disability keeps the interaction focused on the socially standardized purpose of sales encounters; that is, inquiry about, and purchase of, an article. In sum, although the explanations in transcripts 2Á4 portray their speakers as disabled individuals, they also constitute rational solutions to an interactional problem.

Intentional Forms
Mutual knowledge may be created reliably by implicit means. In principle, there is always an unlimited number of ways of alluding to a fact without mentioning it. Analytically, they can be distinguished into forms that are observably intentional and forms that are not. From the interactions of the corpus, three can be assigned to the former category (discussed in this subsection) and four to the latter category (discussed in the next subsection).
A first example of an implicit yet observably intentional reference to a speaker's vision impairment could already be seen in transcript 1, line 11: ''you wouldn't be able to show me i (.) actually have a little trouble understanding''. This utterance repeats the statement-explanation format discussed earlier, although the statement does not mention the vision impairment, but contains a direct request. As before, the justification of the statement mentions a lack of ability Á here, the speaker's inability to comprehend what the salesperson was saying. Interpreted literally, the utterance does not imply a vision impairment, but a cognitive failure; after all, inability to comprehend what another person is saying is primarily a cognitive, not a visual problem. However, as the solution which the customer suggests himself (''show me'') is not the kind of solution a purely cognitive failure would require (e.g. ''please speak slower/more simply''), the salesperson can be expected to infer that the customer has a physical impairment. (The question of what kind of impairment this may be, if it comes up at all, is answered as soon as the salesperson is asked to lead the customer slowly to the paper section.) A second, observably intentional reference to the speaker's vision impairment is reproduced in the following transcript. The format is unusual because the vision impairment is communicated only toward the end of the 10-minute encounter, where communicating it hardly fulfils the aforementioned function of reassuring the blind customer that his condition will be taken into account in the emergent interaction. However, it is worth noting that the statement's failure to fulfil this function is observable in two ways. Firstly, C seems to be using the implied reference to his vision impairment for a different purpose, namely, for moving the topic of the conversation to ''digital radio''. Secondly, S' ''no no'' (line 3) marks C's statement as expected, hence redundant, thereby implying both that he already knows of C's vision impairment and that he assumes that C already knows that he knows of it. Yet again, the vision impairment is referred to by means of communicating a lack of ability, specifically, C's inability to make use of ''home cinema'' equipment.
The final example from the data which involves an observably intentional reference to the customer's vision impairment is also the only case in which the salesperson is responsible for creating this reference and, thereby, for turning the impairment into mutual knowledge.
These two interactions involve the same customer, but different shops and salespersons. At first sight, the interaction problem exhibited in them appears to be the direct opposite of the one shown in transcripts 7 and 8 Á rather than ensuring that she is being addressed, C tries to ensure that she is addressing someone. Yet on the level of implicitly conveyed information, the two situations are virtually identical. Even in transcripts 9 and 10, C forces S to rationalize C's behaviour by assuming that C is vision impaired. A noteworthy but subtle difference is that in the latter two interactions, C's ''slip'' occurs in the form of a self-interruption, rather than in direct response to a preceding turn by S. This could lead S to assume that C planned her ''slip'', in other words, that she intends to hint at the fact that she is vision impaired. This, however, would remain a private conclusion drawn by S, and not mutual knowledge.
In sum, the sequences discussed here distinguish themselves from those discussed in the previous subsection by turning the customers' vision impairments into mutual knowledge without implying that this result was intended. It seems tempting to assume that this subtle distinction is irrelevant in the context of a co-operative, guileless sales encounter. However, by revealing their vision impairment through interactional confusions, errors and slips instead of direct or intentionally indirect statements, the customers describe themselves (probably unintentionally) as the kind of persons who create confusing interactions and who are, therefore, a small but noticeable burden on other interactants. In addition, the customers identify themselves as persons who are not attentive enough to anticipate the interaction problems they create. Having to inform others about one's vision impairment by way of social slips and accidents is unlikely to count interactionally as a particularly dignified (if perhaps entertaining) form of behaviour, yet transcripts 7 and 8 show clearly that it can sometimes be unavoidable.

Conclusion
Social interaction between individuals with and without physical disabilities is frequently asymmetrical (Hydén, Nilholm & Karlsson 2003, Michailakis 2004. Whereas the disabled participants become defined as needy and inferior, the non-disabled participants are assigned the role of helpers and supporters (Goffman 1990, Loseke 2003. The debate on ''empowerment'' has focused largely on structural and cultural reasons behind this phenomenon Á for instance, lack of knowledge, negative attitudes, oppressive political-economic conditions and inaccessibility of material infrastructure (Charlton 1998, Ramcharan 1997; see also the contributions to the edited volume by Kovarsky, Duchan & Maxwell 1999). Taking blindness as an example of general relevance, I tried to show that asymmetry, disablement and disempowerment can sometimes be emergent products of the sequential order of social interaction.
As blindness is a physical condition that tends to have a strong impact on the relationship between the individual and his or her surroundings, a normal component of social interaction between unacquainted blind and sighted individuals is the event of turning the impairment into mutual knowledge. If sighted individuals are aware of the fact that they are dealing with a blind individual and if the blind individual is aware of their awareness, interaction in many social situations can be expected to proceed more smoothly and more securely. However, the very procedure of turning the impairment into mutual knowledge can portray the blind individual as disabled. Within the limitations of the corpus, several interesting patterns could be discerned. When the impairment was communicated explicitly, it turned out to be interactionally necessary to justify why this explicit mentioning should be relevant, and participants solved this problem by way of presenting the impaired participant as functionally limited. When the impairment was communicated implicitly and intentionally, presenting the impaired participant as someone who is functionally or cognitively disabled appeared to be a very Á technically speaking Á ''natural'' and efficient way of implying the impairment. When the impairment was communicated implicitly and unintentionally, this occurred due to an interactional failure for which the blind participant seemed responsible.
In principle, it is never possible to establish with certainty the psychological and social reasons that underlie any concrete instance of behaviour (for a general sociological discussion of this problem, see Luhmann 1995:ch. 3). Nonetheless, I have tried to show that all instances of interactional disablement considered in this article can plausibly be understood as interactional constructions that are fully reasonable given the social situation in which they took place. In no case was it necessary or even plausible to make use of non-generalizable, idiosyncratic causes, such as ''perceived inferiority'' or ''hostile attitudes'' . In each interaction, the disablement of the blind participant can be observed as motivated by good reasons (in the intentional cases) or at least as resulting from reasonable responses to surprising circumstances (in the unintentional cases). In the intentional cases, disablement is a corollary of the co-operative, helpful spirit in which sales encounters are to take place. Here, disablement turns out to be the price blind individuals pay for eliciting or receiving interactional support, understanding and security. In the unintentional cases, disablement is caused by interactants' attempts at resolving interactional confusions and problems.
Because disability results from these processes of interactional disablement and not from properties inherent to the individual, the ''existence'' of this result is a rather lofty and purely attributional one. However, disability is ''real in its consequences'' (Thomas & Thomas 1928), and the impaired participant is thus left with the added burden of carrying a socially inferior identity. Admittedly, the patterns of interactional disablement that I have analysed in this article are subtle, but since the groundbreaking studies of Goffman, we know that even subtle forms of disablement and disempowerment, experienced routinely, can spoil a life.
In sum, the interactions analysed in this article reveal that there can be an intrinsic tension between co-operative, problem-free interaction and empowering interaction. In everyday life, selectively impaired but otherwise fully competent interactants are sometimes forced to appear as not only impaired, but also disabled individuals. They are confronted with a dilemma, as they find themselves forced to trade maximal social respect and empowerment for improved co-operation and efficiency. From a theoretical point of view, contemporary social science has hardly an alternative to understanding disability as a social construction. However, at least in the everyday lives of blind people, disability may often be likely to be perceived socially as a direct causal consequence of the impairment.