2

Analyzing Interpreted
Medical Interviews

IN ORDER to examine ways in which ASL-English interpreters influence interactive discourse, it is important to recognize that not all interactive discourse is identical. Interactive discourse, as it is used here, refers to dialogic or multiparty interactions. Certain aspects of interaction are applicable to most encounters, while others are situationally determined and as varied as the background and goals of the participants. For instance, conversation analysts, in the ethnomethodological tradition of Garfmkel (1967, 1974), have identified stable, structural features of interactive discourse, such as adjacency pairs (Schegloff and Sacks 1973) and the organization of turn-taking (Sacks, Schegloff, and Jefferson 1974). An adjacency pair is a feature of conversation that has two parts, such as a greeting followed by a greeting or a question followed by an answer. These stable, structural features can be found in one form or another in almost all interactional discourse.

Hymes (1972) has identified consistent features of communicative events, indicating that setting, participants, goals, and discourse genre are some of the important components of interactive discourse. Taking an anthropological perspective, Hymes’s work addresses the need to understand the larger context in which interaction occurs, as well as features that vary from one interaction to another. For example, Hymes identifies a taxonomy regarding interactive discourse, including the speech situation, the speech event, and the speech act. Speech act refers to the function of a particular utterance, such as a joke. The speech act takes place within a particular speech event, such as a conversation. This event takes place within a larger speech situation, such as a party (Hymes 1972, 56). This is one way in which Hymes is able to examine both the larger context and the local context within an interaction.

Interactional sociolinguists have studied how people understand (or misunderstand) one another in specific interactions. Evidence regarding the situated nature of interactions and the potentially significant influence of such contextual clues as code choice (e.g., choice of language or dialect) can be found in Gumperz (1982). In one example, Gumperz demonstrates that when a speaker chooses to switch from one dialect to another, that code switch can be interpreted by different people to mean different things:

Following an informal graduate seminar at a major university, a black student approached the instructor, who was about to leave the room accompanied by several other black and white students, and said:

a.Could I talk to you for a minute? I’m gonna apply for a fellowship and I was wondering if I could get a recommendation?

The instructor replied:

b.O.K. Come along to the office and tell me what you want to do. As the instructor and the rest of the group left the room, the black student said, turning his head ever so slightly to the other students:

c.Ahma git me a gig!

(Rough gloss: “I’m going to get myself some support.”)

(Gumperz 1982, 30)

Gumperz indicates that the code switch from Standard English to Black dialect was interpreted by some to indicate a rejection of both the white instructor and the institution. Others, however, interpreted the change in code choice to suggest that the student was addressing other students in the group, demonstrating that he was merely going along with the system, as many do in a white-dominated environment. A variety of social factors influence the interpretation of this interaction, including the time, people, and location in which it occurred.

Given that interactive discourse is highly influenced by the situation in which it occurs, it is important to narrow studies of interpretation to specific genres of discourse. This allows specific encounters to be analyzed with regard to the same macro- and micro-analytical issues that have been applied to analyses of noninterpreted discourse. As a result of the serious nature of the outcomes associated with the genre, and despite methodological challenges in collecting the necessary data, the focus here will be on interpreted medical interviews.

Interviews As a Discourse Genre

Among the various types of interactive discourse, interviews constitute a particular genre of discourse. Interviews serve a variety of functions, including elicitation of information for research purposes or for news reports. In general, interviews involve two or more people. The interviewer generally has more control over the structure of the interaction than the interviewee (for example, control over turn-taking). On the other hand, the interviewee has knowledge or information that is sought by the interviewer.

With regard to institutional interviews, Labov (1984) discusses the existence of asymmetry between interviewer and interviewee. Moreover, he indicates that the structure of such interviews is relatively fixed, as compared with other types of interaction, both in terms of topic selection and question-response format. Hohenberg (1983), McDowell (1986), and Mischler (1986) discuss media interviews, information interviews, and research interviews, respectively, indicating that in all three cases the interviewer maintains some measure of control over the structure of the interaction. Nevertheless, it is worth noting that the ways in which interviewers manifest their control might vary. While media interviewers are generally encouraged to thoroughly research their topic in advance and prepare a well-organized sequence of questions, a technical writer interviewing for information is likely to follow a more spontaneous question format based on interviewee responses rather than adhering to a preset sequence of inquiry. Despite differences in approach, the relatively fixed and formal structure of interviews as described by Labov (1984) remains consistent.

The formal nature of interviews, and their effectiveness as a means of eliciting information, has been of interest to sociolinguists since the first sociolinguistic interviews were conducted almost thirty years ago. Shuy, Wolfram, and Riley (1968), Labov (1972), Kibrik (1977), and Briggs (1986) were among the first to address such issues. In a more recent discussion of the sociolinguistic interview, Schiffrin (1993) applies discourse analysis techniques in her analysis of a sociolinguistic interview. She concludes that just as sociolinguists recognize variations in speaker style, sociolinguistic studies can “incorporate the idea that identity is dynamic and is mutually constitutive with the organization of talk” (259). Schiffrin points out that people quite possibly demonstrate a similar dynamism with regard to their identity and participant structures within other types of interviews, not just in sociolinguistic interviews. The issue of how participants respond to interview discourse has been critical in analyses of medical interviews.

Analyses of Medical Interview Discourse

In general, a medical interview can be described as a professional interview, similar to those conducted by attorneys and accountants. Donaghy describes medical interviews as a type of diagnostic interview, “an interview between an interviewer who possesses special knowledge and skills and a respondent who provides information so that the interviewer can analyze a specific situation or problem” (1984, 301). The purpose of most medical interviews is for the interviewee, or patient, to consult an expert in order to relieve some physical ailment. As a result, the focus of such interviews is on eliciting a medical history by collecting information about the current illness and symptoms as well as the past medical history (MacKinnon and Michels 1971). The elicitation of this information is critical to the effective diagnosis of a patient’s condition. However, as Shuy (1972, 1976, 1983) and others since have discovered through applied sociolinguistic analyses, the elicitation of such information in medical interviews is often problematic.

Problems associated with doctor-patient communication have been attributed to many factors, including differences in language, culture, background knowledge, and goals. These differences have led to further problems, including ineffective medical policies and unnecessary operations. Perhaps because of the serious implications resulting from problems in doctor-patient communication, research has focused on identifying the causes for the problems in this particular setting.

Many researchers have studied the effects of doctors’ use of medical jargon on a patient’s ability to comprehend medical discourse. For example, Shuy (1972, 1976, 1979, and 1983), Ford (1976), and Fisher (1983) address problems found in the use of specialized vocabulary. Words relating to anatomy and illness are not the only kind of terminology that creates problems in understanding, however. The application of everyday terms to specialized purposes is also cited as causing confusion for some patients. For example, the term “infection” has a meaning in everyday language that might not extend to the description of such ailments as pneumonia or blood poisoning. Yet, according to Shuy (1972), medical specialists use this everyday term in reference to these ailments.

Research also indicates that problems in communication are not limited to patient confusion. A doctor’s inability to understand patients’ ways of speaking can cause communication problems in the medical interview as well. In recent work, Bonanno (1995) analyzes the unbalanced use of approximators by doctors and patients. Bonanno’s study of approximators is based on a study of the typology of hedges, which are words or phrases that tend to weaken or qualify a statement ("It was sort of ugly”) or a speaker’s relationship to it (“I think she is coming”) (Prince et al. 1982). Approximators are a subcategory of hedges that “create fuzziness within the prepositional content” (Bonanno 1995, 46), for example, describing a symptom as “sort of a spinning sensation” (132). Bonanno finds that patients use approximators more than twice as often as doctors. Bonanno suggests that since patients’ use of approximators can leave the doctor confused, patients who answer questions more directly, even by indicating when solicited information is unknown, will likely experience more successful medical interviews and diagnoses.

Communication breakdowns are not limited to vocabulary choices. Shuy (1972, 1976, 1983) discusses ways in which cultural differences create barriers in doctor-patient understanding. For example, Shuy (1972) discusses the use of questions that reflect the doctors’ middle-class lifestyles, rather than the reality of most patients’ lives. A question about exercise specifically asked patients about how much time they spend exercising (i.e., jogging or playing tennis), when most of these patients spent over twenty minutes a day walking and were not concerned about such issues. Similarly, Mishler (1984) discusses how the distinct voices of the medical world and the “lifeworld” interact within doctor-patient communication. His analysis, like those of Tannen and Wallat and Cicourel, indicates that the patient and the doctor bring different experiences and backgrounds to the medical interview.

Cicourel and Tannen and Wallat discuss differences in register use between doctors and patients. For example, the presence of the mother in a pediatric interview causes the doctor to shift between registers and interrupts the flow of the medical examination (Tannen and Wallat 1982, 1983, 1987, 1993). Cicourel compares a doctor’s factual notes from a medical interview with the imprecise and emotional language actually provided by the patient. In this study the shift in registers was shown to result in factual errors in documentation (1983).

Bonanno (1995) and Fisher (1983) both discuss the fact that doctors and patients are attending to different tasks. For example, while doctors conduct interviews in a familiar environment, using familiar language, and following their own busy schedules, patients must face unfamiliar surroundings and unknown jargon, and are generally in the weakened position of being ill at the time of the interview. Moreover, while a doctor is seeing one of many people with a given condition, patients might be facing a decision that could impact their daily routine or even their lives (Fisher 1983, 153). The different tasks faced by doctors and patients can influence who controls the interview (Bonanno 1995).

Asymmetry within the Medical Interview

The medical interview is often an asymmetrical interchange. Most research indicates that doctors control the interaction, as evidenced by a variety of linguistic features, including topic initiation and regulation of turn-taking.

Shuy (1979) suggests that the situation of doctors’ controlling medical interviews is analogous to teachers’ managing classroom discourse or police investigators’ running interrogations. Shuy and others have found that professionals’ control of an interaction can negatively impact the effectiveness of communication. Skopek (1975) found that the structural features controlled by doctors include openings, closings, turn-taking, and topic initiation. In an examination of topic initiation by doctors, Shuy (1983) indicates that of three physicians who control the selection of topic, the one who uses a more casual conversational style elicits information from his patient more successfully. Cicourel (1983) supports the contention that misunderstandings between doctors and patients can result from a patient’s weakened ability to communicate.

Despite the fact that doctors have been found to have the more powerful role in medical interviews, patients influence medical discourse as well. For instance, work described earlier by Tannen and Wallat (1982, 1983, 1987, 1993) indicates that the presence of the parent in a pediatric examination has an impact on the doctor’s discourse. Aronsson (1991) reports similar findings, indicating that a mother’s use of pronouns to refer to her child affects the involvement of the child, casting the patient as a side-participant or a nonparticipant in the interaction (Aronsson 1991, 71).

West (1983) and Frankel (1984) examine the question-answer format within medical interviews and suggest that the asymmetrical role is not simply assumed by the doctor. West suggests that patients who stammer while asking questions demonstrate that both patients and doctors contribute to the asymmetry of medical interview discourse, since stammering might represent a less confident or powerful way of speaking. Still, Shuy (1983) finds that while some patients attempt to interrupt doctors, more often than not the attempts fail to elicit a turn. As Shuy points out, since patients are the ones with the information critical to the diagnosis and treatment of medical problems, doctors would better elicit such information by not following the traditional structure of medical interviews. Instead, Shuy suggests that doctors use a more conversational type of discourse, encouraging patients to share the information so critical to the doctors’ task.

Interpreters in Medical Settings

A special package designed for medical practitioners who might be working with Deaf patients was developed in the late seventies. In this package, DiPietro (1979) offers suggestions for working with sign language interpreters. Medical practitioners are encouraged to communicate directly with deaf patients, in terms of both eye gaze and pronominal reference. Interestingly, only two of thirty-three pages are devoted to working with interpreters. The majority of the information presented by DiPietro assumes that no interpreter will be present and that medical practitioners are primarily hearing and have little experience with either the American Deaf community or ASL. Perhaps researchers share this assumption, for very little research has been conducted in the area of interpreting in medical settings.

Cicourel (1981) and Cokely (1982) address the issue of problematic doctor-patient communication as it extends to interpreted medical interviews. For example, Cicourel examines a doctor-patient interview in which a third person is acting as interpreter. As in his work on uninterpreted medical interviews, Cicourel finds that as a result of control over the interactive exchange, the doctor might not be effective in eliciting patient information.

Cokely (1982) conducted an experimental study in which a hearing nurse interviewed a Deaf patient on two separate occasions. In each case the interview was interpreted by certified, professional ASL-English interpreters. Cokely found that four factors interfered with communication, beyond the normal communication problems addressed in literature on doctor-patient communication: perception errors, memory errors, semantic errors, and performance errors. Perception errors occur when the interpreter believes he/she understood an original utterance (such as a proper name), but, in fact, did not understand it correctly. Memory errors are identifiable when small portions of the original utterance unintentionally fail to appear in the interpreter’s rendition. When an interpreter incorrectly uses certain lexical items or syntactic structures within a target language rendition, Cokely terms this a semantic error. Finally, performance errors include extraneous behaviors or errors in the production of the utterance. For example, a false start with no repair can appear to the receiver as if it were a word different from the one intended. In Cokely’s example, the interpreter fingerspells the name of a medication, interrupts that to spell the abbreviation of the term, then attempts to spell the name of the medication a second time: “E-D-E-T-H-O-S-D-E-S-D-E-T-H-O-S-T-A-B-E-T-H-A-L.” Cokely points out that without any pauses or repairs, it looks as though the name of the medication is “edethosdesdethostabethal,” when in fact the medication being referred to is “diethylstilbestrol” (9). Cokely concludes that potential communication problems specific to interpreting increase the likelihood of miscommunication in medical interviews.

In a study of spoken language interpreted medical interviews, Wadensjö (1992) examined twenty Swedish-Russian interpreted encounters, thirteen of which occurred in medical settings. Wadensjö refers to Swedish studies of interpreted medical interviews that indicate that interpreted conversations are unnatural dyadic communication (Kulick 1982), based, in part, on findings of unusual backchanneling in the interpreted interviews (Englund Dimitrova 1991). Wadensjö herself finds that interpreters function not only as translators, but as negotiators or coordinators of the interactive discourse.

Interpreted medical settings are influenced not only by linguistic issues. Prince (1986) analyzed medical interviews between English-speaking doctors and Spanish-speaking patients. In addition to those interviews that were conducted in Spanish, Prince discusses the interviews for which a clinic employee or patient’s friend or relative served as an interpreter. She concludes that the doctors lacked both linguistic and cultural awareness of these patients and that this negatively influenced their ability to provide effective medical care.

Interactive ASL-English Interpreting in Medical Interviews

Collecting linguistic data from medical interviews is never a simple matter, and the need for the presence of video cameras to capture the signed portion of the interpreted encounter makes data collection even more challenging. These challenges range from human issues, such as informant privacy, to technological constraints such as how many cameras are needed to capture the signed utterances of both the Deaf participant and the interpreter. These issues influence the choice of data to be analyzed, as outlined below.

The Data

In initial data gathering, more than eight hours of interpreted interactions were collected. This data consists of two types. One type of data was collected through live videotaping of the interactive encounters, with the researcher present as a technician to run the camera, check the lighting and sound, and so forth. The second type of data was elicited by contacting interpreter education programs. These programs were asked to suggest or submit videotapes of interpreted encounters. Both types of data have advantages and disadvantages.

Videotaped Data

Videotaping interpreted encounters involves problems of both a human and a technological nature. These issues include the intrusiveness of the recording equipment, the limitations of the recording equipment, and the added risk to informants’ confidentiality as a result of the type of recording being made.

Technologically, it is necessary to set up video equipment that will record both the visual and acoustic linguistic information. Thus, lighting, sound, and physical space might need to be manipulated in an effort to record “naturally occurring” data. Clearly, the presence of a researcher is enough to alter the natural flow of an interaction, what Labov (1972) terms the Observer’s Paradox. Add to that the presence of, and space occupied by, the video equipment and the logistical manipulations of participants and equipment, and the result is somewhat less than ideal. Wadensjö, in her study of spoken language interpreted encounters, describes this as problematic and avoids using video recorders for this very reason (1992, 58).

Aside from the intrusive nature of video technology, the problem of recording the necessary data also exists. While an audio recorder is capable of capturing all of the sounds in a given environment, albeit in an overlapping way, a video camera is necessarily limited to a restricted scope of “vision.” That is, the only way to record all interactants is to be far from the center of events, and to risk capturing some participants from behind. Neither of these conditions is conducive to the work of the signed language linguist, who requires access to detailed manual, body, and facial articulations in order to accurately transcribe the linguistic data. One solution to this problem is to use additional video cameras. However, this adds to the level of intrusiveness, and is difficult to coordinate for later viewing and transcription. A more realistic solution is to adapt the single camera to the extent possible, sometimes risking the loss of valuable pieces of information as participants move away from one another and force the technician to select which parts of the interaction to record.

An additional problem inherent in the collection of visual data is the issue of informant confidentiality. Although voice recognition is always a possibility with audio recorded data, video recordings leave no question as to the identity of participants involved in such research (Winston and Ball 1994). When collecting data in confidential settings, such as medical interviews, many potential informants are uncomfortable with the prospect of videotaping these encounters. Especially problematic is the need to maintain privacy while attempting to publish transcriptions of the ASL data. Although conventions have been established for transcribing ASL through the use of English glosses and other mechanisms, often still photographs are used to most accurately represent the linguistic forms under discussion (for examples, see Locker McKee 1992; Winston 1993).

Finally, an unavoidable issue with regard to the collection of live videotaped ASL data is the influence of the technician and setting on the participants’ language choice. Lucas (1994) and Lucas and Valli (1989, 1991, 1992) discuss sociolinguistic factors that influence language use among native ASL signers. They argue that certain situational factors, such as formality of the situation or the linguistic background of interlocutors, can influence whether or not Deaf native ASL signers actually produce ASL or a type of contact signing.

In their examination of language contact situations, Lucas and Valli found that what had previously been labeled Pidgin Sign English was not, in fact, a pidgin, but rather, a form of contact signing. Moreover, they found that hearing or deaf status is not the influencing factor regarding the output of contact signing. Rather, sociolinguistic factors, such as language background, had a greater influence on language choice. For example, Deaf informants signed ASL with hearing native ASL users in some settings and used contact signing with other native Deaf signers in other situations.

The contact variety of signing described in Lucas and Valli (1992) includes unique lexical forms as well as morphological and syntactic structures. Lexical signs are used in contact signing to represent English prepositions, conjunctions, and so forth. In addition, mouthing, whispering, or even audible voicing of English words occurs sometimes with and sometimes without accompanying signs. Fingerspelled signs used to represent English inflectional and derivational morphemes such as #ING and #MENT are one type of morphological structure found in contact signing (84). Other morphological structures include the use of some ASL verb inflections and classifier predicates. Interestingly, Lucas and Valli also found some ASL indicating verbs are produced without subject and/or object marking in the contact variety. Lucas and Valli also discuss a variety of syntactic structures found within the contact variety of signing in their data. Some of these structures include the use of prepositions, “that” constructions, and sentence constructions that follow English word order. Some of the syntactic structures found in the contact signing followed neither English nor ASL word order. Clearly, contact signing differs linguistically from standard ASL.

Lucas and Valli’s findings have implications regarding the collection of data in medical settings. Whatever influences the medical settings and interpreter’s background might have on the Deaf participant’s language choice (as well as the influence that the setting and participants all have on each other’s language choices), the presence of the researcher and video camera clearly lessen the likelihood of collecting ASL, rather than contact signing, data from the Deaf participants. The extent to which code switching within the ASL discourse might differ without the presence of researcher and camera cannot be determined. Nevertheless, this is an especially relevant issue since ASL-English interpreters are expected to use language “most readily understood by the personjs) whom they serve” (RID Code of Ethics). Thus, it would be useful to have the opportunity to analyze data that was clearly not influenced by a researcher and the presence of the intrusive video camera.

Prerecorded Data

The second type of data collected for this study came from interpreter education programs. The data consist primarily of interpreting students engaged in interpreting role plays simulating a variety of real-life encounters. Most of the participants are Deaf and hearing interlocutors who have been invited to the interpreting program to interact in a real or imagined area of expertise. Because the resulting scenarios are mock interviews, the interpreting students have the opportunity to interpret these interactions with less concern about questions and errors that arise as part of the learning process.

The benefits regarding this prerecorded data are twofold. First, the student interpreters and the deaf and hearing participants have all given permission after the fact for the videotapes to be used for research purposes. That is, during the encounters, no researcher was present and participants had no knowledge or expectation that the videotape would be used for the purpose of research. Second, since videotaping these interactions was a part of the classroom routine, the video camera was not perceived as an outside intrusion. Moreover, because the original purpose of videotaping was for students to have the opportunity to observe and critique their interpreting performance, the videotapes were designed to include all of the participants in the recording for its duration. In some cases, this was accomplished through careful arrangement of participant seating. In other cases, a two-camera simultaneous recording was made using a device that allows participants to face one another while appearing on the recording via a split screen. Because the prerecorded videotapes were produced in a unique setting for a unique purpose, they reduce the impact of the Observer’s Paradox on the participants.

The prerecorded data also offer a unique benefit resulting from the analysis of student interpreters. Since student interpreters are presumably not fully proficient in the task of interactive interpreting, it is precisely the areas where students have difficulty that could indicate relevant issues in the analysis of real-life interpreted interaction. That is, the analysis of student interpreters can benefit not only interpreting pedagogy, but also can provide useful insights regarding aspects of the task so often overlooked or taken for granted by skilled professionals.

The student interpreters in the prerecorded data were engaged in the real task of interpreting. They were providing access to an interaction between two interlocutors who had been invited to attend a day of class in order to interact with someone not using their native language. Because the invited participants do not know both languages fluently (or at least do not have access to them), the student interpreters were actually engaged in the task of interpreting the interaction between the two languages. Nevertheless, it cannot be denied that classroom role play differs from real-life interaction. In the case of a medical interview, the stakes of participants who are playing the roles of doctor and patient do not necessarily reflect the concerns of patients and doctors in real-life medical encounters.

Data Selection

In an effort to capitalize on the benefits and to reduce the limitations of each of the types of data collected for this study, two cases were selected for analysis and comparison. One case consists of a prerecorded role play of a medical interview as interpreted by an interpreting student. The second case consists of a professional interpreter interpreting a real-life medical interview in a pediatric setting. The examination and comparison of these two cases provides useful information regarding the influences interpreters have on interpreted interactions. In addition, examination of these two cases provides useful information regarding the application of role plays in interpreter education.

Case 1: Mock Medical Interview

The mock medical interview took place in a fairly large classroom where interpreting is taught. Video cameras and related equipment are available and used in this room on a daily basis. The mock medical interview lasted for 7.16 minutes. The interview included three participants who were essentially facing one another, although the interpreter was seated more to one side of the “doctor,” who was across from the Deaf “patient” (see figure 2.1). The participants remained seated throughout the encounter. The instructor and other students were present and observing the interaction as a part of the class but were not visible on the videotape and did not interact with the visible participants.

All three participants were white and middle-class; two were hearing women, and one was a Deaf man. One of the hearing women was an interpreting student in the class. She was in her early to midtwenties and was the designated interpreter for the mock encounter. The other hearing woman was also interested in interpreting, and although she was not a student in the class, she was invited to participate in the role play of a medical encounter. She was in her early fifties and played the part of the doctor. English was the first language of both hearing women, and both were studying ASL as a second language. The Deaf man, who was in his late thirties, was a native signer who attended residential schools and who often taught ASL as a second language.1 He primarily signed ASL throughout the encounter.

Figure 2.1. Mock Interview: Logistics

Case 2: Actual Medical Interview

The actual medical interview took place in a pediatrician’s office in a suburban, middle-class neighborhood. The encounter lasted for 26.31 minutes. A total of six people participated in this encounter: the doctor, the nurse, the interpreter, the mother, the child, and the researcher. The examination room was quite small and, as a result, no more than four participants were visible at a time on the recording. The researcher and camera were situated just inside the door (almost behind the open door, in fact) and very close to the wall. As the researcher faced the room, a weight scale was located to her right against the same wall. This wall intersected with the back wall, which had windows leading to the outside of the building. This back wall was directly across from the wall housing the door to the examining room. An examining table was located along the wall across from the camera. Against the front wall, between the end of the examining table and the door, was a counter with papers on it, used by the doctor and nurse to make notes. Although the participants frequently moved around within the room, for much of the interview the mother faced the camera, with the interpreter located to her left (see figure 2.2). The interpreter was facing the doctor (who was to the mother’s right). The child was either being examined or in his mother’s arms.

Figure 2.2. Actual Interview: Logistics

All six participants were white and middle class. The doctor was a hearing male in his mid- to upper fifties. The nurse was a hearing female in her late thirties to early forties. Both medical practitioners were native speakers of English and did not know ASL. Although the doctor indicated that he has cared for other deaf patients, he had not worked with an interpreter before, and indicated prior to this interview that he normally communicated with deaf patients by writing notes back and forth. The researcher was a hearing female in her early thirties who was a native speaker of English and who learned ASL as a second language. She was also a professional, certified ASL-English interpreter. The interpreter in the actual medical interview was a hearing female in her mid-thirties. She was bilingual in both ASL and English and had Deaf parents who were fluent ASL signers. She also has had extensive experience as an interpreter educator. The mother, a Deaf native signer of ASL, was almost thirty. She attended residential schools and taught ASL as an adult. Her son was hearing and was about eighteen months old. The mother was approximately five months pregnant with her second child.

Two Cases Compared

Two aspects of the cases under examination are worth noting here. First, based on the aspects of contact signing identified by Lucas and Valli (1992), the mother in the actual medical interview primarily used a contact variety of signing in her discourse. This could be due to a number of factors, many of which have already been discussed. An additional reason for her signing choice could be that she is holding her son for much of the interview. Thus, one hand is not available for communication, and so she uses a lot of fingerspelling and other one-handed signs. Since language choices are often unconscious, it is impossible to know for certain why she chose a contact variety of signing. Nevertheless, her language choice influenced the signing of the interpreter, who incorporated contact signing into many of her renditions in the examples that follow. Unlike the Deaf participant in the actual medical interview, the Deaf participant in the mock interview primarily signed ASL. He had no reason to sign one-handed (and he does not do so), and he was not in an actual medical office or faced with the presence of a researcher. Whatever the reasons for their language choices, it is worth noting that there was a difference in language use by participants in the two encounters.

A second issue worth noting has to do with the length of the encounters. As stated earlier, the actual medical interview lasted for 26.31 minutes. It is interesting to note that the actual medical interview, in which the medical practitioners have the time constraints often cited in literature regarding problems in medical discourse, is almost four times longer than the mock interview occurring in a classroom. One might expect that a mock interview would be the encounter that is not constrained by time. One possible reason for the time differential is that the participants in the mock interview simply ran out of fabricated material. A second possibility is that the teacher ended the interview for pedagogical purposes. It is also possible that the doctor in the actual medical interview extended the encounter as a result of the presence of the researcher and camera. Similarly, the doctor might have extended the interview because of the presence of the interpreter. The mother mentioned after the interview concluded that it had lasted longer than normal. If it was the presence of the interpreter that caused this change in behavior, it is worth investigating whether extended length of time is a feature of interpreted medical interviews. However, if it was the presence of the researcher that caused the time extension, it is worth pursuing other avenues for collecting more realistic examples of how doctors interact with deaf patients via interpreters under normal time constraints.

Analysis of Two Interpreted Interviews

One of the difficulties in collecting data from real-life interpreted medical encounters involves technological challenges. For example, in order to interpret between ASL and English, interpreters must be positioned so that they can both see, and be seen by, any Deaf consumers. Yet, analysis requires that both the Deaf participants’ and the interpreter’s signing be accessible to the camera view. Bringing additional cameras into an interview setting is not only increasingly intrusive, but challenging to coordinate for later viewing. Thus, one advantage of the prerecorded mock medical interview data used here is the relatively unobtrusive presence of multiple cameras in the classroom environment. Another is the availability of advanced technology to make all parties visible on the screen. Because intrusiveness is significantly less of an issue in the mock medical encounter, the advantages of advanced technology can be used to obtain data that offers access and insights potentially unavailable in the actual interpreted medical encounter. Nevertheless, it cannot be denied that while simulating a medical interview, the mock interview will clearly differ from medical encounters that occur in the real-world context of hospitals and clinics.

As anticipated, analysis of the actual medical interview did exhibit evidence of the influence of camera and researcher presence. For example, at times during the pediatric interview the mother and child can be seen waving to the camera. In fact, the mother refers to the camera and researcher in an apparent attempt to distract the child from crying. In addition, the doctor, who essentially ignores the camera and researcher throughout the interview, makes a comment as he leaves the office indicating his awareness of their presence. Finally, the interpreter, who occasionally gazes in the direction of the camera (for example, in response to the mother and child waving), commented after the interview session that she had been tempted once to ask the researcher (also a certified professional interpreter) for assistance with the interpreting task. Thus, all informants were affected by the presence of camera and researcher/interpreter, unlike the role play, which was videotaped as a routine part of class and prior to any consideration of submission for research purposes. Nevertheless, the actual interview offers authenticity in the medical setting that the role play does not. Hopefully, through a comparison of these two encounters, both authenticity and naturalness will be represented in the data, and the impact of the Observer’s Paradox, to some extent, will be minimized.

Transcription Issues

Transcription of signed language data is never easy. Part of the difficulty arises from a lack of standardization. Winston and Ball indicate that the need to include visual and spatial information, and the need for consistency in the use of transcription symbols, are among the concerns to be addressed in the process of standardization (Winston and Ball, 1994). Because ASL is not a written language and, perhaps even more important, because of the existence of multiple articulators (including fingers, hands, arms, shoulders, neck, head, mouth, cheeks, eyes, and eyebrows), written transcription can lose more of the original than it captures.

Mischler (1991) compares transcription to photography, indicating that “what we assume to have been ’really’ there, and how the photographer selected and framed the event, and how the photograph is presented and located within the flow of other information … all influence our understanding.” A transcription reflects and emphasizes what the transcriber thinks is relevant within the data. Perhaps for this reason, Ochs (1979) suggests that transcripts should be based on theoretical goals. Because the focus here is to analyze the interpreters’ utterances as they interact with the discourse as a whole, a musical-score format of transcription is used.

The musical-score format of transcription is one way of representing the simultaneous and overlapping nature of interactive discourse. As described by Ehlich (1993), the musical-score format allows the sequence of events to unfold from left to right on a horizontal line, while the list of participants occurring from top to bottom allows each person’s utterances to be captured within a single moment of overlap. For example, in figure 2.3, an excerpt of the transcript from the actual medical interview illustrates the musical-score format. In this example, the first event in the time line (from left to right) is that of the doctor entering the examining room. Upon his entry he says, “Hello.” Overlapping with the latter part of this brief utterance, the interpreter initiates her ASL rendition, indicated with the English gloss HELLO (see Transcription Conventions in Appendix 2 for information regarding the use of glosses and other transcription features). Just after the interpreter signs HELLO, the mother signs, HELLO. Once again, during the latter part of her utterance, the interpreter begins her English rendition of this when she says, “Hi.” Throughout this exchange, the child is silent. This example illustrates that the musical-score format not only indicates the occurrence of these four utterances, but also, at least roughly, their relationship to one another in time.

Figure 2.3. Sample of musical-score transcription.

The musical-score format is particularly useful for an interpreted encounter. Since most of what is uttered by noninterpreter participants (i.e., the doctor, nurse, and mother) is reuttered by the interpreter, there is a tremendous amount of overlap throughout the data. Moreover, because the interpreter not only uses two languages but two modalities as well, it is possible that she will produce utterances in more than one modality at a time (both signing and speaking). In order to accurately and unambiguously represent the interpreter’s utterances, two lines of transcription are ascribed to the interpreter throughout the data analyzed: one for English utterances and the other for ASL utterances (as marked in figure 2.3).

Two additional points are relevant with regard to the transcription format. First, because this is a preliminary study of the influence of one participant in deference to her social role (interpreter), the transcript identifies all of the participants on the basis of their social roles within the medical interview context (e.g., doctor, mother, interpreter, and so forth). This practice is not intended to dehumanize the participants, nor is it intended to suggest that a single social role can capture the identity of any given individual. In fact, as Schiffrin (1993) points out, social identities are situated within activities and are dynamic, not static, in nature.

Second, although there is no researcher present during the mock interview, the researcher is present throughout the actual medical interview. The main task of the researcher is to operate the video camera, and thus, she is not visible on the screen. Because all of her utterances are in ASL, they are not accessible for transcription purposes. Moreover, the participants rarely address or respond to the researcher during the course of the medical interview. Therefore, no line of the transcript is provided for the researcher as one of the potential interlocutors. This is not intended to hide her presence, however, and all of her audible contributions (such as laughter) are indicated within the appropriate line.

Although the musical-score format allows for the indication of discourse phenomenon such as pausing (by leaving appropriate portions of the line blank), certain transcription conventions related to both spoken and signed language transcription are adhered to within the transcript. (For details regarding the transcription of both the ASL and English discourse, see Transcription Conventions in Appendix 2.) However, as Locker McKee (1992) points out regarding indication of some aspects of ASL, transcription can provide an etic or an emic description of the signs. She describes etic transcriptions as those that focus on the form of the sign, and emic transcriptions as those that emphasize the meaning of a given sign. The transcription for this study uses some etic features (such as INDEX [rt] for a point to the right) and some emic (PRO.3 [baby]) to indicate the meaning of a pronoun rather than simply the form, depending on the issue being discussed. In addition, nonmanual and contextual information is included where necessary to enhance readability of the transcript.

Translations of the ASL are provided where necessary to make the transcript more accessible. In addition, when specific parts of an example are particularly relevant, the relevant portion of the transcript is either provided in isolation or highlighted in bold within the transcribed example. The transcript was originally produced by the researcher, with assistance from both native signers and qualified interpreters (both hearing and Deaf) to promote accuracy.

The transcription of discourse in any language involves consideration of a variety of theoretical issues. However, the problems facing signed language transcription are even greater than those faced by spoken language linguists, as a result of both the unwritten status of signed languages and their articulatory complexities. Nevertheless, an attempt has been made to address these theoretical issues, especially as they interact within multiparty, bilingual, interpreted discourse.

Examining Interpreted Medical Interviews

The Interpreter’s Paradox exists because interpreters are faced with the goal of providing access to interaction of which they are not a part, while they are, in fact, physically and interactionally present. In order to examine the interpreter’s involvement in interpreted encounters, it is first necessary to understand the nature of such encounters as they occur without the presence of interpreters. Thus, this chapter has provided a discussion of medical interview discourse as well as a discussion of findings regarding the influence of interpreters in such settings. Because the collection of ASL-English interpreted medical interviews is challenging for both humanistic and technological reasons, the focus here will be on the analysis and comparison of two cases of interpreted encounters: a mock medical interview interpreted by a student and an actual medical interview interpreted by a native bilingual, certified professional. Theoretical issues regarding the transcription of both spoken and signed languages, with an emphasis on the issue of interpreted encounters, provides the rationale for the use of a musical-score format of transcription.

Analysis of the data is based on consideration of a variety of factors including identification of the originator of each utterance and whether or not that utterance was accessible to other participants, identification of the addressee on the basis of linguistic or paralinguistic cues, and documentation of turn-sequences, false starts, repairs, occurrences of constructed dialogue, use of question forms, reference, and so forth. Needless to say, the data reflect the complexity of interpreted discourse. As a data-driven study, the relevant features of the discourse have become clear through examination of the problem at hand: what is the interpreter’s influence on the interactive discourse?