The Interpreter’s Paradox
INTERPRETERS, like other participants, bring their own frames and schema to interpreted encounters. Moreover, analysis of interpreter-generated utterances that are not renditions of other-party discourse indicates that interpreters’ contributions can be categorized into two types: relayings and interactional management. Regardless of which type of footing an interpreter employs, each utterance requires that a choice be made about the code in which the utterance is conveyed. The fact that three-quarters of the interpreters’ nonrenditions are accessible only to the Deaf participants raises questions about interpreter partiality. Examination of interpreters’ footing types reveals that interpreter neutrality is a complex notion.
What is interpreter neutrality? If interpreters add no interpretergenerated contributions to the interaction, certain information that is normally accessible in interactive discourse would be missing. If equivalency is supposed to be a marker of neutrality, certainly the omission of this type of information would detract from neutrality. Thus, an interpreter’s offering “extra” information or utterances in order to provide equal information actually minimizes the interpreter’s influence on the interaction.
Despite suggestions that an interpreter’s presence is best left unnoticed (Fink 1982), introducing and identifying participants’ roles within interaction is an accepted practice. Providing no introduction of an interpreter can leave participants feeling confused about who to talk to and how the interaction should proceed. Introducing an interpreter provides the opportunity not only to clarify the interpreted encounter frame, but to try to align participants’ schema regarding that frame prior to the development of interactional problems. Who should be responsible for such introductions and how they might best be carried out is an area for further investigation.
Interpreters responding to questions or interference have the opportunity to provide no response, minimal conventional responses, or longer explanations. It appears that nonresponses represent marked behavior and actually cause more interactional problems than do responses. Minimal responses appear to fulfill conventional requirements with the least amount of interference to either the interaction itself or the interpreter’s ability to provide renditions of it. Further investigation is needed not only with regard to types of responses in various situated encounters, but also regarding which participants receive explanations. If interpreters provide more explanations to Deaf participants than to hearing participants, there is a question as to the impact of this apparent partiality on interpreted interactions.
Interpreters’ nonrenditions cannot be assumed a priori to be extraneous information. Some aspects of the discourse are conveyable only through interpreter-generated nonrenditions. Moreover, some interpreter-generated nonrenditions could provide procedural information in an attempt to avoid interactional problems evolving from the interpreter’s presence later in the interaction. Thus, certain footing types can be seen as avenues for minimizing the interpreters’ influence on the discourse. Moreover, interpreters influence the interactive discourse to different degrees depending, in part, on their marked or unmarked responses. Although the interpreters appeared to attempt to limit their influence through nonresponses (at least to the hearing participants), findings suggest that minimal responses are actually less marked, and therefore less disruptive, than either nonresponses or explanations. Apparently, the concept of “neutral” interpreter behavior is better defined by situated interactional norms and expectations than by predefined constraints resulting in marked interactional behavior.
In order to determine how the interpreters impacted the encounters in which they interpreted, it is possible to examine the ways in which various types of footing within the interpreters’ utterances influenced the interaction. Some of the interpreters’ footing types created an opportunity for the discourse to proceed in a manner not dissimilar from monolingual interactive discourse. Conversely, other interpreter contributions to the discourse actually misrepresented the footing within the interaction.
Similarities
In monolingual interactive discourse, the fact that participants share both a language and the mode in which it is transmitted means that participants generally have access to not only the content of utterances, but also the fact that an utterance is occurring and from where it originates. Moreover, participants in monolingual interaction most likely either know one another, or have the means to discover the identity and relevance of another participant’s presence. That is, people generally introduce themselves (or one another) and include pertinent information regarding relationships to one another and, hence, to the interactive event (e.g., “This is my sister, Wendy”). Thus, introductions of interpreters in interactive events can be seen as similar to introductions of any unknown participant. Once an event has gotten under way, conventions exist for addressing unplanned interruptions within interactions. Interpreters following these conventions exhibit unmarked behavior similar to that encountered in noninterpreted interaction. All of these areas that occur within monolingual interaction also can be found in the interpreted encounters examined here. Discussion of these issues can help to elucidate some of the complexities associated with the notion of interpreter neutrality.
Turn-Initiation and Voice Recognition
In monolingual interactive discourse, when an interlocutor begins an utterance, addressees are generally able to determine that a turn has been initiated and who is the source of that turn, in addition to receiving access to the content of the utterance. How this is accomplished in ASL and English discourse is somewhat different, however. When interpreting between two distinct modalities, information about the occurrence and source of an original utterance might not be accessible to participants without an interpreter-generated contribution. Therefore, for ASL-English interpreters, the rendering of all three parts of an utterance is an important consideration. Because this discourse-relevant information is not directly available to participants who are native to languages conveyed in two different modes, there is the potential for participants to experience confusion regarding who is the original source of a given utterance, or even when another participant begins a turn.
In English monolingual interaction the interlocutors are generally able to hear when someone begins a turn at talk. On the basis of prior exposure to the speaker’s voice, the addressee or overhearer can generally identify the speaker (as a familiar person or someone unknown). Similarly, in ASL turns are initiated in part on the basis of eye gaze. When an addressee is not gazing in the direction of an interlocutor who is initiating a turn, the potential addressee is summoned until eye contact is made, allowing the turn to begin.
The interpreters convey this information by generating utterances. These utterances (e.g., a summons or source attribution) provide the information normally accessible in monolingual interaction. Thus, the interpreter-generated contributions can be seen to function in a manner that makes the interpreted interaction similar to monolingual interactive discourse. That is, each participant receives information similar to the information that would be accessible in a monolingual interview. It is important to reiterate that the interpreters did not always provide this information consistently. The issue here is that such contributions seem to represent a similarity to monolingual interaction. Clearly, a difference is that the interpreter has the power to omit information that is, by necessity, always present in monolingual discourse.
Identifying Participant Roles
The only time one of the interpreters provides an identification of her role in the interaction is in the mock encounter. The student interpreter introduces herself as the “sign language interpreter.” Although noninterpreted encounters do not have interpreter introductions, they often include introductions of unknown participants. Such introductions often include reference to the participant’s role within the encounter (as the doctor or nurse, husband or wife, etc.). Thus, introduction of an interpreter can be seen as similar to noninterpreted encounters, where participant identification is not uncommon. Nevertheless, the way in which an interpreter’s role is identified can make a difference in the progression of the interaction.
One issue with regard to identification of the interpreter’s role is who controls the introduction. In the mock medical interview, the interpreter introduction is handled by the interpreter herself. In the actual medical interview there is no formal introduction, but when the interpreter’s participant status is questioned by the doctor, the Deaf participant explains the interpreter’s role. It is conceivable that any of the participants could introduce the interpreter, and each participant introduction would convey a certain participant alignment within the interaction from that point on. For example, when the Deaf mother explains the interpreter’s function, the “patient” and the interpreter appear to be aligned and the doctor excluded; he is the one who does not understand. This Deaf participant-interpreter alignment is created despite the fact that the doctor and Deaf participant may have met before, whereas the Deaf participant and the interpreter have not. It is conceivable that the sharing of a common language (ASL) creates a unique bond. It is also likely that where the doctor has never worked with an interpreter before, the Deaf participant has experienced interpreted encounters for many years. Thus, both the Deaf participant and the interpreter share a familiarity with the social roles, if not with one another as individuals. One of the potential problems with an introduction generated by the Deaf participant as a “patient” (or parent of patient) is that it could threaten the doctor’s position as the expert and higher-status interlocutor by placing him or her in the less experienced, naive, or student role. Although such a challenge might not be undesirable, it could influence the quality of care received by a Deaf patient.
It is also conceivable that the doctor could introduce the interpreter or explain her function. Frequently, it is the institution that is responsible for hiring the services of an interpreter, and as professional service providers the doctor and interpreter could share a bond (in addition to that of a shared language, in this case, English). One of the potential problems in this case is that the Deaf person is the minority group member. Since hearing people represent an oppressor group (and most interpreters are hearing), seeing an alignment between the doctor and the interpreter could be uncomfortable for the Deaf participant.
A third possibility, and the one seen in the mock interview, is that the interpreter introduces herself. The issue inherent in this option is that the interpreter must make code choices. When the student interpreter attempted to introduce herself while both speaking and signing at the same time, her utterances were awkward and inaccurately produced (at least the signed portion). On the other hand, if an interpreter chooses to introduce herself in one language at a time there is still a question of who receives the introduction first. Once again, the interpreter’s choice could reflect some partiality with regard to one or another participant. Treating either participant in a partial manner can influence the subsequent interaction and the quality of medical care. The doctor might feel threatened, and the patient might feel unwilling to share details of personal history. Clearly, further investigation regarding the ways in which interpreter introductions are handled and their impact on the medical interview is necessary.
A second issue with regard to the identification of the interpreter’s role is related to the interaction of frames and schema. The student interpreter introduces herself as the sign language interpreter. However, she does not elaborate on the function of an interpreter within an interaction. Thus, while all the participants share the interpreted encounter frame, there is a mismatch in their schema for that frame. Because there is a mismatch in the schema related to the interpreted encounter frame in both cases, it is possible that all the participants in interpreted encounters could benefit from briefly addressing the interpreter’s function. As Mclntire and Sanderson (1995) point out, interpreters do not always function in a consistent manner. An interpreter might function differently depending on the situation and the participants. Thus, it is even more critical that all participants share a similar schema for the interpreted encounter frame, as it is situated within a particular interpreted encounter.
Because of the social roles adopted by the Deaf and hearing participants in the two cases under examination here (doctor as hearing, and patient, or parent of patient, as Deaf), this discussion has focused on the implications of introductions based on hearing doctor and Deaf patient roles. In future research it would be interesting to examine this issue in medical interviews in which the doctor is Deaf and the patients are either Deaf or hearing. Nevertheless, for the present discussion it is clear that the inclusion of interpreter introductions in interpreted interaction is similar to the introductions that are frequently a part of interactive, multiparty discourse. Who should be responsible for the introductions, and what information should be included are areas that warrant further research. The results of such research could provide a basis for both professional guidelines and for interpreter education programs.
Following Discourse Conventions
Interpreters, as participants within an interaction, are subject to interference from the physical environment, as are all the participants. In addition, other participants are capable of addressing the interpreter at any time during the interaction. When faced with such circumstances, the interpreters have the option of not responding, responding minimally, or providing lengthy responses and explanation to one or more of the other participants. There are often certain conventions for participant responses. For example, a request for information conventionally receives some sort of response, even if it is an indication that the information is unknown or unavailable. Whether or not interpreters follow these conventions can determine the ways in which the interpreter influences the interactive discourse.
It has become clear that both interpreters avoid responding when addressed by the hearing participants. In both cases, the nonresponses lead to awkward or problematic moments in the discourse. For example, the student interpreter avoids answering the hearing participant’s question about how to sign “ulcer.” The result is not only confusion, but some of the discourse is not interpreted into ASL (while the interpreter elicits a response from the Deaf participant). In this example, the hearing participant is left with an unfilled slot in a question-answer adjacency pair. Conversely, when the student interpreter provides minimal, noncomplying responses to the Deaf participant’s requests, there is no unfilled slot and the discourse proceeds. The interpreter’s use of a minimal response is a way of fulfilling the need for a second part in the adjacency pair.
The use of adjacency pairs can also occur as a result of interference. Like Schegloff’s (1972) discussion of telephone conversations in which the ring of the phone is the first part of an adjacency pair (the summons), interference in the environment can create first parts that conventionally require a second part. At one point in the mock medical encounter, the interpreter accidentally touches the hearing participant. The interpreter, perhaps recognizing the potential interpretation of the touch as a summons, says, “‘Scuse me.” It appears that the touch is interpreted as a summons by the hearing participant. She stops speaking, gazes at the interpreter, and asks, “Is there a problem?” These two utterances overlap with one another. The interpreter’s apology is brief and indicates that the touch is not a summons, but an accident. The interpreter’s minimal response allows the interaction to resume.
It is interesting to note that while the interpreter’s response to the hearing participant is minimal in the preceding example, her response to the situation with regard to the Deaf participant is not. She provides an explanation of the brief interaction, during which time she is not interpreting the doctor’s continuing discourse. The use of explanations, as in this case, sometimes causes interruptions within the interaction. However, other explanations, such as explanations about the nurse coming to the door as a way of introducing the upcoming speaker, do not seem to interfere with the discourse. In fact, in the latter case, the explanation seems to function as a summons and a source attribution; without them, some of the interactional information normally a part of monolingual interactive discourse would be missing.
The use of various types of responses to interlocutors, or to interference, seems to impact differently on the ways in which the interactive discourse itself is influenced. In the two cases under examination here, the use of minimal responses (as opposed to no response or lengthy explanations) allows the interaction to resume with the least amount of influence from the interpreter. It appears that nonresponses, which do not follow conventions of interactive discourse, cause more interruptions than responses. The interpreters’ use of minimal responses creates a situation most similar to noninterpreted interactive discourse.
Differences
Interpreter footings, including source attribution, summonses, introductions, interference, and explanations could be perceived as contributing to the interaction in such a way that the interpreted discourse is similar to noninterpreted discourse. Nevertheless, these footings are not always consistently maintained by the interpreters. For example, source attribution is not included in every interpreter utterance, despite the fact that each utterance has the potential to represent utterances by at least two participants (the interpreter or the “other” language user). Further, while some footing types clearly have the potential to contribute to discourse so that it is similar to noninterpreted interaction, these same footings can distinguish the interaction as an interpreted one. For instance, depending on who controls an introduction and what information it contains, the presence of an interpreter introduction can create interactional alignments unique to interpreted discourse. In addition to these potential differences, the two cases under examination exhibited a profound difference between the interpreted interviews and noninterpreted interviews in the representation of footing within renditions.
Misrepresentation of Footing
Neither the student interpreter nor the professional interpreter responds directly to the hearing participants. In the actual medical interview, one of the strategies used to avoid responding to questions directed toward the interpreter is to relay the question. In this case, the Deaf participant responds to the question and the interaction proceeds. However, upon closer examination of the interpreter’s rendition, it becomes clear that she alters the footing from the original utterance. As will be seen shortly, this has an important impact on the interpreted interactive discourse.
In the following example (seen earlier in example 4.24), the doctor directs a question to the interpreter. He is requesting clarification regarding the interpreter’s rendition of the mother’s utterance that indicates she tried to give her son something after he vomited:
Example 5.1
55
D: You (points to interpreter) tried or she (points to mother) tried?
In the doctor’s utterance it is clear that he is addressing the interpreter because of his point to her while he uses the second-person pronoun. Despite the fact that the doctor uses the second-person pronoun to refer to the interpreter, and refers to the mother in the third person, the interpreter’s rendition of this utterance does not:
Example 5.2
56
I: (taps mother’s knee) PRO.1 TRY OR PRO.2 TRY?
I tried or you tried?
In this example, the interpreter has altered the footing of the utterance from that in the doctor’s original utterance. She might have a reason for doing so. For example, it might be a clearer way of rendering a complicated moment in the discourse. Regardless of the rationale, however, the rendition does contain a misrepresented footing; that is, in this particular utterance, it is not clear to the Deaf participant that the doctor is speaking about her rather than to her. In this instance, the Deaf participant begins an explanation of the interpreter’s function. This example represents a point of interactional confusion in the discourse.
The fact that the interpreter relays an utterance with a change in footing at this point in the interaction is not surprising. Confusion about the process of interpreted discourse and the function of the interpreter is evident in this encounter (the doctor is not certain whether the interpreter is also a caretaker of the child). However, if the footing in interpreted discourse is misrepresented in other contexts as well, then the interpreter can be seen to impose a tremendous influence on the interaction. That is, if the Deaf participant does not know that she is being talked about, she has no power to request that she be spoken to, or to otherwise address the situation. Schiffrin (1993) discusses various possible reasons for speaking about another participant. For example, one might speak about another as a way of showing support for them. Conversely, assumptions of communicative incompetence can accompany such a footing. Given these very different alignments, it is significant that a participant might not have access to the fact that he or she is being talked about. Because this participant-participant footing is misrepresented, it becomes clear that it is worth examining the interpreters’ relaying of footing within renditions.
Interpreter Pronoun Use within Renditions
Pronominal reference differs among languages. Therefore, it is important to recognize that the use of a pronoun in an original utterance in one language does not always require the use of a pronoun within the rendition in the other language. For this reason, only the interpreters’ explicit use of pronouns is examined. Each pronoun that occurs within a rendition is categorized as either consistent with the original, inconsistent with the original, or both. Consistent pronouns are those that match person1 (i.e., when a first-person rendition is used to relay a first-person pronoun in the original). Inconsistent pronouns refer to interpreter’s pronouns that do not match the person of the original utterance (as in example 5.2). When pronouns are categorized as both, some vacillation within the rendition, and more than one pronoun occurs. These categories are examined for renditions in ASL and English produced by the student interpreter and by the professional interpreter.
Table 5.1 Occurrences of Pronominal Reference in Student Interpreter’s ASL Renditions in Mock Medical Interview
| N | Consistent (%) | Inconsistent (%) | Both (%) | |
| Number of Occurrences | 13 | 6 (46.1%) | 5 (38.5%) | 2 (15.4%) |
An examination of the student interpreter’s use of pronouns within ASL renditions indicates that the rendered footings are not always consistent with the original utterances they are intended to relay (see table 5.1). Within the ASL renditions, the student interpreter produces a total of thirteen pronouns. Of these thirteen, only half are consistent with the original utterance. In example 5.3 below, the student interpreter can be seen to use the first-person singular pronoun in ASL to render the doctor’s use of first person in the original utterance:
Example 5.3
27
P:
D: Oh, I had an emergency this morning, and I need to excuse myself, just for a second. Okay?
I:
I: EMERGENCY MORNING PRO.1 EXCUSE
It is important to remember that the frequency of pronominal occurrence is not at issue here. Rather, the relevant question is whether or not the interpreter uses a pronoun consistent in person with the pronoun used in the original. In example 5.3 the interpreter clearly does so. Nevertheless, more than half of the ASL pronouns are either inconsistent or contain both consistencies and inconsistencies in the multiple pronoun use. An example of the former can be seen in the following:
Example 5.4
54
P: = AWKWARD PRO.1
D: Well- y’know it just depends on whether you really =
I: = If I don’t have it, I’m a wreck.
I:
UH =
55
P:
D: = wanna heal or not, uh, r- as I said before, right now this isn’t a big problem =
I:
I: = TRUE DEPEND PRO.1 WANT HEALTH
Example 5.4 begins with the end of the patient’s response to the doctor’s suggestion that he not drink a lot of coffee. The patient is explaining that he really depends on coffee, or he is a “wreck.” The doctor then indicates that whether or not the patient decides to continue to drink coffee is really a matter of his interest in healing. In line 55, the student interpreter can be seen to render the doctor’s second-person pronoun (seen in line 54) from the original English utterance as a first-person pronoun in the ASL rendition. This is an example of an inconsistent rendering.
The student interpreter also produces a mixture of both consistent and inconsistent pronouns within her renditions. This can be seen in example 5.5. This example takes place in the beginning of the mock interview, just after the interpreter has introduced herself:
Example 5.5
2
P:
D: Oh, you’re the interpreter for today.
I: … and I’m gonna be the sign langu-language interpreter for today.
I: … POSS.1 SIGN LANGUAGE INTERPRETER LANGUAGE NOW. PRO.1 PRO.2 PRO.1 OH =
3
P: (nods)
D: It’s nice to meet you.
I:
I: = INTERPRETER NOW PRO.1 #OH #OK. PRO.1 - NICE MEET PRO.1
In this example, the interpreter is rendering the doctor’s response to her self-introduction. The doctor’s use of the second-person pronoun results in a vacillating pronoun use on the part of the interpreter. This vacillation could indicate indecision on the part of the student interpreter as to whether or not she should render the doctor’s utterance with the use of constructed dialogue. Indecision is likely the result of limited interpreting experience on the part of the student interpreter. Thus, one possible explanation for the high occurrence of inconsistency and misrepresentation of footings is that the interpreter is an interpreting student. This being a potential cause, it is even more striking to note the contrast in English pronoun use within English renditions (see table 5.2).
The student interpreter produces a total of sixteen English pronouns. All of the pronouns within the English renditions are consistent. An example of this can be seen in 5.6 (p. 171), in which the patient is commenting on his feelings about getting the results of a medical test that he supposedly underwent a week prior to this encounter:
Table 5.2 Occurrences of Pronominal Reference in Student Interpreter’s English Renditions in Mock Medical Interview
| N | Consistent (%) | Inconsistent (%) | Both (%) | |
| Number of Occurrences | 16 | 16 (100.0%) | 0 (0.0%) | 0 (0.0%) |
Example 5.6
8
Patient: PRO.1 NERVOUS PRO.1
Doctor:
Interpreter: Yeah, I’m nervous.
Interpreter:
The interpreter clearly uses a first-person-singular pronoun in her rendition, just as the patient did in the original utterance. Every instance of pronoun use within the interpreter’s English renditions is consistent with the pronoun in the original ASL utterance.
There are several possible explanations for the significant difference between the student interpreter’s pronoun use within the ASL and English renditions. One explanation is that the interpreter is better able to provide consistent renditions into her first language. This would support the contention that “simultaneous interpretation can only be done properly into one’s native language” (Seleskovitch 1978, 100). A second possibility is that the footing alterations are due to translation issues unique to English-to-ASL renditions. For example, where an English source utterance refers to someone in second or third person, an ASL translation might incorporate constructed dialogue in which the interpreter would take on the “role” of the other and refer to him or her by referring to herself, using a first-person pronoun. A third possibility is that the difference reflects the interpreter’s schema regarding interpreted encounters. The student interpreter interacts in different ways with the hearing and Deaf participants. For example, she answers the Deaf participant’s questions while refraining from answering questions posed by the hearing participant. It is possible that the interpreter’s schema allows her to communicate directly with the Deaf participant but not with the hearing participant. Thus, some of the discourse is relayed as one person telling another what the third person said. Similarly, if the interpreter feels more partial toward the Deaf participant, it is possible that she is more comfortable representing his words as her own, whereas the doctor’s words are more comfortably represented as the words of another (e.g., “she said”). Finally, it is possible that the student interpreter is simply demonstrating her current status as an interpreter in training. In order to assess the latter possibility, it is useful to examine the professional interpreter’s use of pronouns in renditions.
Examination of the professional interpreter’s pronoun use within the ASL renditions reveals that, as in the case of the student interpreter, the rendered footings are not always consistent with the original utterances they represent (see table 5.3). The professional interpreter produces a total of eighty-one ASL pronouns.2 Unlike the student interpreter, only about one-fourth of the pronouns are inconsistent with the original utterance. An example of an inconsistency was seen earlier in examples 5.1 and 5.2, in which the doctor clearly addresses the interpreter with a second-person pronoun (pointing to her simultaneously), which she then renders with a first-person pronoun in ASL. The professional interpreter clearly produces fewer inconsistencies within her ASL renditions. Nevertheless, despite the smaller proportion of inconsistency, the fact that one-quarter of the pronouns are inconsistent is significant, especially considering that the professional interpreter grew up with ASL as a native language. Once again, the variation in consistency between the ASL and English renditions is striking.
Table 5.3 Occurrences of Pronominal Reference in Professional Interpreter’s ASL Renditions in Actual Medical Interview
| N | Consistent (%) | Inconsistent (%) | Both (%) | |
| Number of Occurrences | 81 | 61 (75.3%) | 18 (22.2%) | 2 (2.5%) |
The professional interpreter produces a total of 115 English pronouns within her renditions. As in the case of the student interpreter, these pronouns are consistent in every case with the original utterances they are intended to relay (see table 5.4). The professional interpreter is a native bilingual, while the student interpreter has learned ASL as a second language. In addition, although the student interpreter is still an interpreter “in-training,” the professional interpreter is certified and has years of experience. Thus, on the basis of the two cases under examination here, it appears that the different representations of footings within the ASL and English renditions are not simply due to language background or years of experience. Rather, the difference could be the result of some aspect of the process of rendering messages from English to ASL or could reflect the interpreters’ schema in which Deaf participants and hearing participants are treated differently.
Table 5.4 Occurrences of Pronominal Reference in Professional Interpreter’s English Renditions in Actual Medical Interview
| N | Consistent (%) | Inconsistent (%) | Both (%) | |
| Number of Occurrences | 115 | 115 (100.0%) | 0 (0.0%) | 0 (0.0%) |
An additional possibility is that the interpreters’ pronoun use reflects consistencies and inconsistencies in their input. The Deaf mother in the actual medical interview refers to the hearing doctor and nurse in the second person consistently throughout the interview. This could account for the fact that the professional interpreter consistently refers to the hearing participants in the second person. However, the hearing participants are not consistent in their reference to the Deaf mother. The doctor refers to the mother using second person in less than half of his references to her, using third-person pronominal reference just over half of the time. Similarly, three-fourths of the nurse’s pronominal references to the mother are third-person pronouns. Thus, the professional interpreter is faced with recurring shifts in footing in the original English utterances. These shifts could create “leaky” frames for the interpreter. Tannen and Wallat (1987, 1993) describe “leaky” frames in a pediatrician’s discourse, when she shifts register among consulting with medical students, with the patient’s mother, and with the child patient undergoing examination. While managing these different frames, the doctor occasionally uses a register from her talk with one of the participants in her talk with another, as in using more technical terms while engaged in the “motherese” talk she reserves for the child patient. For the interpreter, “leaky” frames could be the cause of the inconsistencies in her ASL pronoun use.
Unlike the actual medical interview, the Deaf and hearing participants in the mock interview both consistently refer to their interlocutor in second person. Despite this consistency, the student interpreter exhibits inconsistencies in her pronoun use only within the ASL renditions. The fact that the input differs in each case while the output is so similar raises an interesting question for future research: Is this difference the result of language differences or differences in learning on the part of the interpreters. To examine this issue effectively, a larger body of data is necessary.
It has become clear that the interpreters inconsistently relay footings in ASL renditions from a quarter to over half of the time. Whatever the reason that interpreters might alter a footing from the original, the interpreter is (consciously or not) exercising his or her power to withhold information from participants. The misrepresentation of footings within the interactive discourse is inextricably related to the issue of empowerment. For example, if the Deaf participant in the actual medical interview knew she was being talked about, she might say something to change that. Without access to the information, she does not have the power to influence the ways in which others align themselves to her through their talk. Whether or not the alteration of footing is inherent in the process of interpreting, it represents a distinct difference between interpreted discourse and noninterpreted discourse.
Implications of Similarities and Differences
Interpreter footings can influence interactive discourse in one of two ways. Some footing types consist of interpreter-generated nonrenditions that actually contribute to the monolingual-interactional quality of the interpreted interactive discourse. Examination of pronoun use in interpreter-generated renditions reveals that the footing within participants’ utterances is not always consistently rendered. Discussion of the notion of interpreter neutrality has been based, in part, on the notion of equivalency. As a result, several issues regarding neutrality have been addressed, including strategies for minimizing the differences in interpreted discourse (e.g., by providing information that is normally a part of monolingual discourse), providing minimal responses, and following discourse conventions (e.g., filling second-part slots in adjacency pairs). Clearly, the notion of neutrality in interactive discourse is complex. The primary focus here has been on the interpreters’ neutrality with regard to their own utterances. The examination of misrepresented footings indicates that even interpreter renditions are generated by interpreters, though they function to retell utterances originally generated by others. Thus, the interpreters are both participants in the interaction and conveyors of the discourse. How does this paradoxical role manifest itself in terms of the structure of the interactions?
Unique Triadic Structure of the Interpreted Encounters
Interpreters not only generate their own contributions to the discourse, but also generate their renditions of utterances originated by others. Examination of pronoun use in interpreter renditions reveals that interpreted utterances do not necessarily reflect the original author’s footings. This supports the findings of other sociolinguistic research on interpreters, which indicates that interpreters function as participants within the interaction rather than as conduits between other interlocutors. Yet, equally clear is the fact that the interpreters are far more constrained in their participation than are other participants. That is to say, a large proportion of interpreter utterances are retellings or reports of what other participants have said. This unique position within the interaction has raised questions about the nature of three-party interpreted interaction. Interpreted encounters can be seen as dyadic communication between the Deaf and hearing participant, or as a triadic encounter that also includes the interpreter as a participant. In order to better distinguish between these two, it is necessary to examine the issue of interpreters as reporters of other participants’ discourse.
Constructed Dialogue
Examination of the interpreters’ pronoun use within renditions indicates there is not always consistency in the manner in which renditions are authored by the interpreters. Some of the renditions use pronouns in the target language that match (in person) those in the source utterance, while others do not. Interpreters, as reporters of others’ utterances, clearly do not always report exactly what was said. That reported “speech” is not always an exact duplicate of the original utterance is not a new concept to discourse analysts.
Tannen (1989) points out that reported speech is frequently not representative of a “report” at all. Based, in part, on Voloshinov’s ([1929] 1986) and Bakhtin’s ([1975] 1981) discussions of reported speech as inextricably situated within the context in which the reporting occurs, Tannen suggests that “reported speech” represents creatively constructed discourse of the reporter rather than an untainted, neutral report of another’s utterance. For this reason, Tannen proposes the term constructed dialogue, as opposed to reported speech. The notion of constructed dialogue fits well with the findings here regarding interpreters’ renditions. The interpreters’ ”reports” of what others have said frequently represent the interpreter’s footings, rather than consistently representing the footings of the original utterances.
When interpreters “report” what others have said, they are not reproducing exact duplicates any more than the speakers in the noninterpreted discourse (discussed in Tannen 1989). If the interpreters did attempt to produce exact duplicates, the addressees would not understand them, for they would be repeating a duplicate in the same language as the original. The fact that interpreters are not only “reporting” what another has said, but also doing so in another language, indicates that interpreter renditions are clearly the “words” of the interpreter. This is parallel to the construction of dialogue in noninterpreted discourse, except that for the interpreter the reported context and the reporting context occur simultaneously. As Hamilton points out (personal communication), the words of others can be reported either directly or indirectly, serving different functions within the discourse. Evidence from the current data suggest that interpreters, like people telling narratives, make such choices for what could be different purposes within the interaction (example 5.5 seems to represent just such a choice on the part of the interpreter). This would be an interesting area for future research.
All of the interpreters’ English renditions use pronouns that are consistent in person with the original utterance. When the ASL includes a first-person pronoun, if the English rendition includes a pronoun then it is also first-person. This can be seen in the following example (seen earlier as example 4.17), in which the mother is responding to the doctor’s request for her son’s shot records:
Example 5.7
138
M: YES (+VOICE) PRO.1 FUTURE GET #IT =
D:
N:
C:
I:
139
M: = FROM INDEX (right)+ POSS.1 DOCTOR PRO.1 GET CL:G FROM #GT UNIVERSITY =
D:
N:
C:
I: = I will get it- I’ll get from my other doctor. I have an appointment - =
I:
140
M: = P-E-D-lto interpreter) #GT UNIVERSITY P-E-D-I-A-T-R-I-C CENTER INDEX (right) =
D: All right
N:
C:
I: = (head tilt left) (head nods) from Georgetown, uh, the pediatric center there?
I:
In this example, the use of first-person pronouns in the original ASL utterances (identifiable in the transcript by the use of PRO.1) are rendered in the first-person by the interpreter in her English constructions of the mother’s discourse. It also interesting to note that, in line 139, the interpreter’s rendition includes a false start: “I will get it—I’ll get it from my other doctor.” This false start does not represent a false start in the original utterance. Rather, it reflects the interpreter’s false start. Clearly, the interpreter is author of the utterance in the sense that she is originating the words she is uttering in a language and in a manner in which they have never been uttered before. Nevertheless, the content of her utterance is clearly motivated by another’s utterance. That is, it is not the interpreter who will be contacting her previous doctor. Thus, while the interpreter is authoring, not merely animating, her renditions, this is still qualitatively different from the authorship of the original utterer. One way to account for this difference would be to examine the conditions for authorship, possibly identifying a continuum of authorship based on a range of features such as ownership of content and form of an utterance. A second possible explanation is that every participant within an encounter is influenced by a range of factors, including their social role and responsibilities and their personal interests in outcomes. Given the latter, the current data indicate that interpreters are not external conveyors of these social and personal aspects of other interactants. Rather, they are themselves participants who bring their own personal interests and social responsibilities to the interaction. For interpreters, attempting to repeat what others utter is a part of their social role.
A second point worth noting is that the interpreter halts her construction of the mother’s utterance at the end of line 139. The interpreter is newer to the area than either the mother or the doctor. If the mother and the doctor were communicating in a Deaf participant-hearing participant dyad, they each would have the relevant background knowledge regarding the referent (the pediatric center). However, the interpreter does not share this background knowledge, and she cannot render what she does not understand. This also serves as evidence that the interpreter’s renditions are her own.
Tannen’s (1989) notion of constructed dialogue has also been applied to ASL discourse (Roy 1989b; Winston 1991, 1992, 1993; Metzger 1995). Yet, even earlier researchers analyzing “reported speech” in ASL describe the use of body shifts, head shifts, eye gaze, and so forth as associated with the representation of another person’s utterances in ASL. In this study, attribution of source related to the ASL renditions consists of both body shifts and pointing. These strategies on the part of the interpreter are similar to the strategies used by ASL signers when constructing dialogue in ASL discourse (this has also been referred by numerous terms in the literature, including role shifting, role playing, and identity shifting). In the following excerpt, the interpreter is first constructing the nurse’s utterance, and then shifts her body as she constructs the doctor’s utterance:
95
I: #OH D-A-R-N #rr Ishifts to side) #HE HAVE FOUR EYE TEETH C-U-T
The interpreter shifts her body position to indicate that she is relaying the words of a different person. This is a common strategy found in ASL constructions of dialogue (Winston 1991, 1992, 1993; Metzger 1995; Liddell and Metzger 1995). While constructed dialogue is often used to represent the real or imagined words of another from a distinct time and place, the interpreter appears to be constructing the dialogue of participants in real time.
Once the interpreters are seen to be constructing dialogue, their role within the interaction becomes somewhat clearer. The interpreter is having a direct conversation with each of the participants, who are unable to have a direct conversation between each other. Thus, three-party interpreted interaction is not a dyadic conversation between the Deaf and hearing participants. Nor is it a triadic conversation between the Deaf participant, the hearing participant, and the interpreter. Instead, the three-party interpreter-interactive discourse is comprised of two overlapping dyads.
Overlapping Dyads
While spoken-spoken or signed-signed language interpreters deal with participants who are presumably not fluent in one another’s languages, interpreters who work in signed-spoken language settings encounter a unique phenomenon. That is, the interlocutors not only cannot understand the other language, but even prosodic information, or the fact than an utterance has occurred at all, might be totally unknown to a participant without the interpreter’s contributions. Perhaps for this reason, while interpreter-Deaf participant dyads exist and interpreter-hearing participant dyads exist, little or no Deaf participant-hearing participant dyads occur in the data. This contrasts with the assumption that many people have about the interpreted encounter frame, which relegates the interpreter to a passive, conduit role. Here, the interpreter can be seen to be the pivotal player, the one consistent participant providing the needed overlap between the two separate dyads (see figure 5.1). Contrary to the dyadic and triadic views of interpreting addressed in chapter 1, figure 5.1 represents the overlapping dyadic structure of the two cases of interpreted encounters examined here. Each dyad represents an interaction between two people in one language (interpreter-Deaf participant in ASL; interpreter-hearing participant in English). While it is true that Deaf and hearing participants have opportunities to communicate directly through gestures, paper and pencil, and possibly speechreading, there was little direct communication in this study. On those rare occasions where direct communication was attempted, when the doctor tried to gesture or the mother tapped the doctor’s arm, communication was not particularly successful. No significant linguistic dyadic interaction occurred between the hearing and Deaf participants.
Figure 5.1. Overlapping-dyad view of interpreting.
As the pivotal point between the two dyads, the interpreter is in a unique position. The interpreter’s role within each dyad is essentially to understand what the interlocutor says (which is an interactive task) and to construct information taken from the other dyadic interaction.
The existence of the two overlapping dyads can be seen in the structure of greeting exchanges. In the actual medical encounter, the doctor initiates a greeting when he first enters the examination room:
40
Mother: HELLO
Doctor: (enters) Hello.
Nurse:
Child:
Interpreter: Hi.
Interpreter: HELLO
This example clearly has four parts. When analyzed in the structure of adjacency pairs, a unique form of embedding can be seen (G refers to Greeting, thus G1 is the first part of the adjacency pair, and G2 is the second part). As can be seen in figure 5.2, the exchange of greetings reflects the two dyads. When the doctor initiates a greeting exchange in the English dyad (G1), the interpreter immediately initiates a greeting in the ASL dyad (G1a). When the Deaf participant provides a second part pair (G2a), the interpreter fills the slot within the English dyad (G2). According to Schiffrin (1987), “An initial greeting constrains the next available interactional slot.” Yet, though the initial greeting is followed by a greeting generated by the interpreter, the interpreter’s utterance is not seen as a second greeting. Instead, the interpreter’s greeting is treated as an initial greeting itself. Thus, in this example, an adjacency pair is initiated by the hearing participant in English, and the ASL adjacency pair, initiated by the interpreter, is embedded within the English exchange. The existence of the embedded exchange structure serves as evidence of the overlapping dyadic structure of the interpreted encounters.
Figure 5.2. Interpreted greeting exchange.
The existence of overlapping dyads can also be seen in the following example, in which the doctor attempts to clarify the referent of a pronoun found in the English discourse. This example was seen earlier, and is repeated below for convenience:
Example 5.10
54
M: VOMIT EVERY TIME PRO.1 TRY GIVE PRO.3 (neutral) SOMETHING VOMIT … #SO =
D: And the fever’s -
C:
I: And he threw up, um, and every time he threw up I tried to give him something, so- =
I:
55
M: = PRO.1 GIVE-UP … E-V-E-N L-I-Q-U-I-D-S PRO.1
D: You (points to interpreter) tried or she (points to mother) tried?
N:
C:
I: = I don’t know: I- (looks at doctor) I tried =
I: (waves)
56
M: = PRO.1 ++ (glances to interpreter) NO (taps doctor’s arm) PRO.3 INTERPRETER-
D:
N:
C:
I: = to give him liquids too.
I: (taps mother’s knee) PRO.1 TRY OR PRO.2 TRY?
M: = LET PRO.1 EXPLAIN, WHEN INTERPRETER TRUE INTERPRET, PRO.3 WILL =
D:
N:
C:
I: = Now let me explain. When the interpreter is interpreting, she will be speaking =
I:
58
M: = BE TALK A-S I-F PRO.3 NOT I-N ROOM PRO.3. TALK FOR PRO.1, #SO (?off screen) =
D: (chuckles) Okay, gotcha.
N:
C:
I: = as if, she’s not in the room. She’s speaking for me. So, if it =
I:
59
M: = CONFUSE TRUE TALK PRO. 1, A-S I-F PRO.3 PRO. 1
D: I’m just interested in the pronoun, that’s all.
N:
C:
I: = confuses you, the interpreter’s really speaking as if
I: PRO.1 (REALLY) INTEREST =
60
M: #OK PRO.1
D: You’re not taking care =
N:
C: (cries)
I:
I: = WORD “I” (i on chest) PRO.3 (left hand) I (i on chest)
M: NO PRO.1 RECENT MEET- (interpreter) #HER FIRST TIME =
D: = of him, you’re just interpreting.
N:
C: (crying continues)
I: I just met =
I: NOT PRO.1, PRO.1 NOT TAKE-CARE -
62
M: = NOW.
D: okay I’ve been in practice thirty-five years, I’ve =
N:
C: (screams then crying stops)
I: = her for the first time today, so
I: #OK PRO. 1 … PRO.1 WORK++ =
This example has provided evidence of mismatches in the participants’ schemes regarding interpreted encounters. Evidence of the overlapping dyads can best be seen here by examining the English and ASL portions separately. The ASL portion of the interaction can be seen in example 5.11 below:
Example 5.11
54
M: VOMIT EVERY TIME PRO.1 TRY GIVE PRO.3 (neutral) SOMETHING VOMIT #SO =
and every time I gave him anything he just threw it up, so =
I:
55
M: = PRO.1 GIVE-UP … E-V-E-N L-I-Q-U-I-D-S PRO.1
I stopped giving him anything, even liquids Me…
I: (waves)
excuse me …
M: = PRO.1+ + (glances to interpreter) NO (taps doctor’s arm) PRO.3 interpreter -
It’s me, I did … No, excuse me, she’s the interpreter
I: (taps mother’s knee) PRO.1 TRY OR PRO.2 TRY?
Excuse me, I tried (giving him food) or you tried?
57
M: = LET PRO.1 EXPLAIN, WHEN INTERPRETER TRUE INTERPRET, PRO.3 WILL =
Let me explain. When the interpreter is interpreting, she will =
I:
58
M: = BE TALK A-S I-F PRO.3 NOT I-N ROOM PRO.3. TALK FOR PRO.1, #SO (?off screen) =
= be talking as if she is not in the room. She’s talking for me. So, (!) =
I:
59
M: = CONFUSE TRUE TALK PRO.1, A-S I-F PRO.3 PRO.1
= confused, she’s really talking for me, as if she were me.
I: PRO.1 (REALLY) INTEREST =
I’m really interested =
60
M: #OK PRO. 1
okay, I
I: = WORD “I” (i on chest) PRO.3 (left hand) i (i on chest)
in the word, “I.” He said, “I.”
61
M: NO PRO.1 RECENT MEET-(interpreter) #HER FIRST TIME =
No, I just met her for the first time =
I: NOT PRO.1, PRO.1 NOT TAKE-CARE -
It’s not me, I’m not taking care of-
M: =NOW
today
I: #OK
okay.
In the discourse, the first evidence of an interactional problem can be seen in lines 55–56, when the mother produces utterances that do not make sense within the ASL portion of the interaction (translated as follows): “Me … it’s me, I did … No, excuse me, she’s the interpreter. …” This utterance can be understood by referring back to example 5.6. The doctor has just asked the question designed to clarify who is the caretaker of the child: “You (point to interpreter) tried or she (point to mother) tried?” Apparently, the mother has understood the doctor’s confusion here, either as a result of the doctor’s visible gestures (pointing), or by speechreading, or both. The mother has understood the doctor despite the fact that the interpreter has not yet rendered it in ASL and is still rendering the mother’s prior utterance into English. Though it is rare to see such an occurrence within the data, the fact that the Deaf participant has understood the hearing participant without a rendition seems to argue for the notion of a triadic structure. However, examination of responses to the mother’s attempts to respond to the doctor indicates that the link between the mother and the doctor is short-lived.
The interpreter, who is juggling the different frames, does not respond in any way to the mother’s utterances in lines 55–56. That is, the interpreter does not respond to the mother, nor does she render the utterances into English for the doctor to hear. Instead, the interpreter completes her renditions of the mother’s prior discourse, and attempts to get the mother’s attention (by waving at her and tapping her knee in lines 55–56) in order to render the doctor’s request for clarification. Once the interpreter successfully elicits the mother’s attention and renders the doctor’s question, the mother begins to explain the role of the interpreter, and the interpreter renders this into English. Because the mother does not successfully communicate directly with the doctor, even after tapping him on the arm (in line 56), no stable connection between the mother and the doctor can be found within the discourse. The interaction as a whole proceeds smoothly only after the mother and the interpreter resolve their interactional problems.
Further evidence of interactional problems can be seen in the interpreter’s pronoun use as shown in lines 59–61. The interpreter, who is rendering into ASL utterances initiated by the doctor, uses the first-person pronoun found in line 59 to refer to the doctor (translated as follows): “I’m really interested in the word ’I.’” In line 60, however, the interpreter shifts footing, and uses a third-person pronoun to refer to the doctor, “He said ’I.’” This utterance is generated by the interpreter, and looks like an attempt to clarify the confusion about who is authoring first-person pronouns. Finally, in line 61, the interpreter renders the doctor’s question about whether she (the interpreter) is caretaker of the child, or simply an interpreter. The interpreter uses a first-person pronoun, in this case, to refer to herself, “It’s not me, I’m not taking care of—.” Whether the mother recognizes the shift as part of a rendition, or perceives the utterances as additional clarification from the interpreter, she clearly understands the point, as is evidenced by her response in line 61. The mother indicates that she has first met the interpreter that very day, and refers to the interpreter with an emphasized English pronoun, finger-spelling H-E-R to assist in alleviating the confusion over pronominal reference.
Part of what characterizes the ASL portion of the interaction is a shifting of footing on the part of the interpreter, sometimes using first person to refer to the doctor and other times to refer to herself. Despite the fact that this shifting could be confusing, the mother exhibits no difficulty in responding appropriately to the questions that are raised. That is, the mother and the interpreter have successfully interacted.
Having examined the ASL portion of the interaction, it is useful to examine the English portion as well. The English interaction, as accessible to the monolingual doctor, is provided as example 5.12.
54
D: And the fever’s -
I: And he threw up, um, and every time he threw up I tried to give him something, so … =
55
D: You (points to interpreter) tried or she (points to mother) tried?
I: = I don’t know: I- (looks at doctor) I tried =
56
D:
I: = to give him liquids too.
57
D:
I: = Now let me explain. When the interpreter is interpreting, she will be speaking =
58
D: (chuckles) Okay, gotcha.
I: = as if, she’s not in the room. She’s speaking for me. So, if it =
59
D: I’m just interested in the pronoun, that’s all.
I: = confuses you, the interpreter’s really speaking as if-
60
D: You’re not taking care =
I:
61
D: = of him, you’re just interpreting.
I: I just met =
62
D: okay
I: = her for the first time today, so
In this portion of the interaction, the doctor hears the interpreter say, “I tried to give him something” (line 54). This sounds as if it is the interpreter who is caring for the child, and the doctor asks for clarification, as seen in line 55. Within the English portion of the interaction, a brief pause occurs after his request, at which point the interpreter says, “Now let me explain. When the interpreter is interpreting, she will be speaking as if she’s not in the room.” The fact that the interpreter uses third person to refer to herself is very odd, and the doctor chuckles at this point in line 58. When the interpreter utters, “She’s speaking for me,” there is again the opportunity for confusion, since the interpreter is referring to herself in the third person, and to the mother in the first person. Though the doctor indicates that he understands in line 58, “Okay, gotcha,” he requests further clarification in lines 60–61 when he asserts, “You’re not taking care of him, you’re just interpreting.” This utterance is clearly directed to the interpreter. From the perspective of the doctor, the interpreter appears to finally respond directly to his request in lines 61–62, when she says, “I just met her for the first time today, so.” However, access to the complete interaction (example 5.10) makes clear what the doctor cannot know from the English portion of the interaction alone. The interpreter did not initiate the utterance in lines 61–62. This is merely another rendition of what the mother has originated. The fact is that the interpreter never responds directly to the doctor’s request. This information is irrelevant to the progression of the English dyadic encounter.
Through examination of the interaction with regard to the two separate languages, it is clear that a dyadic interaction occurs within each. Yet, these two dyadic interactions are not separate from one another. Evidence of interactional problems exists in both dyadic segments. The interpreter is the link between the two dyads, either in the conveying of information (e.g., to the doctor, that mother and interpreter had first met that very day), or in the relaying of information (e.g., when the mother’s explanation is not relayed). Clearly, these examples demonstrate that two separate dyads, one in each language, are connected via the interpreter. The interpreter functions as the pivotal point between the two, either by providing access to the content of one dyad to the addressee in the other, or through failure to do so (as in the mother’s unrendered utterances). In fact, confusion over the referents of first-person pronouns is in itself evidence that the interpreter serves as the pivotal link. The doctor recognizes a question about to whom the first-person-singular pronoun refers, and addresses his dyadic partner (the interpreter) in an attempt to discover the answer.
In order to test the existence of overlapping dyads it is also useful to examine a segment of the interaction in which no interactional problems are apparent. The following three examples occur just before the physical examination of the child begins. The doctor and the mother are discussing the nature of the child’s visit to the same doctor’s office the prior week. Example 5.13 represents the ASL portion of the interaction:
Example 5.13
78
M:
I: PRO.3 (baby) HERE BEFORE PRO.3
He was here before.
79
M: #WAS HERE LAST-WEEK THURSDAY- FRIDAY, PRO. 1 FORGET PRO.3 NOT =
He was here last Thursday or Friday, I can’t remember which day. He doesn’t =
I:
80
M: = SEEM IMPROVE BETTER HAVE F-E-V-E-R (palm up)
= seem to have gotten better, he still has a fever…
I: #WELL TRUE =
Well, that =
M:
I: = DIFFERENT THING BEFORE. PRO.3 HAVE FAT-NOSE NOSE RIGHT? =
was for a different thing. He had a stuffed nose, right!
82
M: NO NOT- NOW- NO BEFORE LAST-WEEK HAVE F-E-V-E-R PRO.1 =
No not- today- No, last week he had a fever =
I: = S-T-U-F-F-Y NOSE, NOW TALK ABOUT F-E-V-E-R =
= a stuffy nose. Now you’re talking about a fever.
83
M: = BRING I-N BETTER F-E-V-E-R STILL OBSERVE SAY THAT PRO.3 SEEM FINE =
= I brought him in and his fever had improved but they examined him and said he was fine =
I:
84
M: = NONE SEEM WRONG AND THAT TRUE SEEM (blocked by nurse) BECAUSE =
= that nothing seemed to wrong, and that he really seemed (!) because =
I:
85
M: = TEETH (palm up)
= of his teething, so …
I: #OK
okay
Within the ASL discourse in the preceding example, the interaction proceeds smoothly. No questions regarding referents occur in the discourse. Though there is disagreement, responses are appropriate to the questions that precede them. Thus, this segment of the interaction serves as evidence of the existence of the ASL dyad. Evidence of the existence of the English dyad can be seen in example 5.14. This example represents the English portion of the same segment of the interview:
78
D: He’s been here before by the way, uh Doctor um, =
I:
79
D: = (?) saw him, uh
I: Yeah, he was here last week on Thursday, er- it was Thursday or Friday, =
80
D: Well, it was a total =
I: = I’m not sure but, it doesn’t seem like he’s gotten any better.
81
M:
D: = different kind of a thing the last time. He had a stuffy nose? he had, y’know, he had-now =
I:
82
D: = we’re talkin’ about fever and vomiting. Last time we were talkin’ about cold symptoms.
I: No, last week, he had fever too, =
83
D: Okay, I see
I: = I brought him in and his fever had gotten better but then … they- and the doctor said =
84
D:
I: = that it seemed like he was fine, and it seemed like he was unhappy =
85
D: Okay.
I: = because he was teething, (shrugs to doctor)
As in example 5.13, the English discourse in example 5.14 proceeds smoothly and without apparent interactional problems. This segment of the English discourse functions in a manner similar to dyadic, noninterpreted interaction. In the previous two examples, both the ASL and the English dyadic interactions function clearly as if they were whole and complete. In example 5.15 below, both portions will be shown as they occur in the data:
Example 5.15
78
M:
D: He’s been here before by the way, uh, Doctor um, =
N: = let me just get one out of two.
C:
I:
I: PRO.3 (baby) HERE BEFORE PRO.3
79
M: #WAS HERE LAST-WEEK THURSDAY- FRIDAY, PRO. 1 FORGET PRO. 3 NOT =
D: = (?) saw him, uh
N:
C:
I: Yeah, he was here last week on Thursday, er- it was Thursday or Friday, =
I:
80
M: = SEEM IMPROVE BETTER HAVE F-E-V-E-R (palm up)
D: Well, it was a total =
N:
C:
I: = I’m not sure but, it doesn’t seem like he’s gotten any better.
I: #WELL TRUE=
M:
D: = different kind of a thing the last time. He had a stuffy nose? he had, y’know, he had-now =
N:
C:
I:
I: = DIFFERENT THING BEFORE. PRO.3 HAVE FAT-NOSE NOSE RIGHT? =
82
M: NO NOT- NOW- NO BEFORE LAST-WEEK HAVE F-E-V-E-R PRO.1 =
D: = we’re talkin’ about fever and vomiting. Last time we were talkin’ about cold symptoms.
N:
C:
I: No, last week, he had fever too, =
I: = S-T-U-F-F-Y NOSE, NOW TALK ABOUT F-E-V-E-R
83
M: = BRING I-N BETTER F-E-V-E-R STILL OBSERVE SAY THAT PRO.3 SEEM FINE =
D: Okay, I see
N:
C: (crying)
I: = I brought him in and his fever had gotten better but then … they- and the doctor said =
I:
84
M: = NONE SEEM WRONG AND THAT TRUE SEEM (blocked by nurse) BECAUSE =
D:
N: (enters)
C: (wails)
I: = that it seemed like he was fine, and it seemed like he was unhappy =
I:
M: = TEETH (palm up)
D: Okay. Ask her to just =
N:
C:
I: = because he was teething (shrugs to doctor)
I: #OK
The two dyads overlap within the same stretch of discourse, pivoting around the single participant who is engaged in both dyads: the interpreter. The overlap is best described in terms of the exchange structure (see figure 5.3).
In line 78, the doctor initiates an exchange with his assertion about a prior visit (A1). The interpreter then initiates an embedded assertion (A1a), which receives a confirmation from the mother as response (R2a). The response within the original exchange is then rendered by the interpreter (R2). As can be seen in example 5.11, the exchanges overlap in time with one another. That is, A1a begins before Ai is completed. Similarly, R2 begins before R2a is completed. The interpreter, as the only participant whose utterances appear in both the original and the embedded exchanges, is the link between the two overlapping dyads within the interpreted discourse.
Despite views of interpreted interaction in which the interpreter is seen as a passive conduit or bridge conveying messages between a Deaf-hearing dyad, or in which the interpreter is seen as a participant in a three-party interaction, the findings here indicate that both the interpreted mock and actual medical interviews actually consist of two overlapping dyads. The use of constructed dialogue by the interpreters and the structure of the greeting exchange in the actual medical interview provide linguistic evidence in support of the overlapping-dyad view of interpreting. This analysis indicates that the situated utterances by participants within the interaction and the sequential structure of the interaction as a whole is not the same for the interpreted interviews as it would be in noninterpreted discourse. While an interpreter can pursue strategies to minimize the intrusiveness of this difference (e.g., through filling second-pair slots with minimal responses), the fact remains that interpreters do influence interactive discourse. Thus, as Baker-Shenk (1991) suggests, interpreters ought to be making intentional choices about those influences.
Figure 5.3. Interpreted Exchange
Implications
It has been seen here that both the student and professional interpreters do influence the medical interview discourse. Further investigation with a larger body of data is warranted in order to seek broad patterns within interpreted medical interviews, other genres of discourse, and across language types (that is, interpreted encounters other than ASL-English). Nevertheless, the analysis here has implications regarding interpreter practice, interpreter education, and, of course, linguistics.
ASL-English Interpreters in Medical Settings
In practice, ASL-English interpreters in medical settings should be alert to the fact that each participant brings a unique schema to the interpreted encounter. As a result, medical interpreters should be prepared to follow two important steps to assist in reducing the impact of mismatches in schema. First, it would be useful for interpreters to monitor research regarding the most effective strategies for achieving the goals of interpreted encounters. As Mclntire and Sanderson (1995) indicate, ASL-English interpreters currently apply different models of interpreting at different times. Moreover, Wadensjö (1992) suggests that interpreters often must improvise strategies for coping with the unpredictable circumstances that arise in actual interpreted encounters. However, systematic research regarding the nature of interpreters’ strategies and the impact of various strategies in a variety of settings could assist in identifying whether or not certain models of interpreting are best applied or avoided in a given setting, and what types of interpreter strategies are likely to yield desired outcomes.
Medical interpreters who participate in or keep abreast of such research will be best prepared to do the job and to perform the second step, which involves interpreters’ providing information to Deaf and hearing consumers about interpreted interviews so that all participants can begin to share similar schema regarding the interpreted encounter frame. Because of the difficulties inherent in providing such information at the time of an interview, it would be useful to develop more global strategies for informing Deaf and hearing consumers.3 This information sharing should be available not only to medical practitioners and patients, but also to the relatives of patients. Most directly related to the findings here, interpreters should be aware of the potential impact of the presence or absence of interpreter introductions as well as of who provides such introductions and what kind of information is included.
In addition to these two steps, interpreters should be aware of the potential impact of their choices with regard to footing types. Interpreters in medical settings should be aware that code choice can create an appearance of partiality. Patients who feel excluded might not provide as much information as needed to elicit appropriate medical care. Similarly, the services provided by doctors who feel excluded might be influenced. Further research will assist interpreters in this area. Interpreters should also consider the potential impact of added footing types, the omission of discourse-relevant information, and the impact of following certain discourse conventions. For example, it could be useful for interpreters to be aware of the potential impact of nonresponses, minimal responses, or lengthy responses. In this study, some response types create awkward, somewhat tangential exchanges. Given findings regarding doctor-patient communication problems resulting, in part, from doctors’ hectic schedules, it seems especially important that interpreters find response-strategies that do not create a need for time-consuming repairs.
Finally, recognition of interpreted encounters as overlapping dyads can help to clarify the two footing types addressed in this study: relayings and interactional management. For example, recognition of the multiple tasks falling on the interpreter as the pivotal point between the two dyads indicates potential benefits for consecutive, rather than simultaneous, interpretation. Consecutive interpretation would allow the interpreter to focus on one dyad at a time, and thus reduce the occurrence of interactional problems. Although it might intuitively seem that there is more time involved in the process of consecutive interpreting, the potential reduction in interactional confusion or misrepresented footings could make consecutive interpretation more time-efficient than simultaneously interpreted interactive discourse. Further investigation of the differences between the two could assist in identifying whether or not certain settings significantly benefit from one or the other.
This study indicates that interpreters influence interactive discourse. The reality of the interpreters’ influences is at odds with professionally defined goals (e.g., in professional codes). Interpreters working in medical settings should be aware of this discrepancy. While this study indicates that interpreters’ goals are at odds with the reality of interactive discourse, it is beyond the scope of this study to identify a tangible, prescribed solution to this dilemma. What can be seen from this study is that some interpreter-generated contributions are an essential part of the interpretation of interactional equivalence. For example, summoning a participant so they know they are being addressed, and attributing the source of an utterance (information that two monolinguals would know), are essential components of the interaction. A patient might respond differently to medical advice depending on whether it comes from a doctor or from a nurse. These contributions from an interpreter should be a part of the interpretation process, perhaps even more than that seen in the cases examined here. Clearly, it is critical to continue empirically based analyses of naturally occurring interpreted encounters in order to investigate the impacts of this divergence between the ideal and reality. Research should also be gathered to identify the various schema that interpreters and consumers bring to interpreted interaction. If professional interpreters, consumers of their services, and researchers work together, they can identify realistic goals for interpreted interaction and determine which interpreter strategies support those goals.
ASL-English Interpreter Education
Investigation of the frames and schema within interpreted medical interviews requires a larger body of data. Such research should include an examination of many student interpreters from a variety of interpreter education programs. Nevertheless, this comparison of the student interpreter case with the case of the professional interpreter does have some implications with regard to interpreter education. These implications relate to how student interpreters are trained to make choices associated with code and footing, as well as to the use of role plays as a training tool.
On the basis of code choices made by the two interpreters, code choice is an issue that should be addressed in interpreter education. Despite the fact that some have recommended simultaneous code production as a way for interpreters to communicate with Deaf and hearing participants simultaneously to avoid leaving anyone out (Earwood 1983), in the findings here the simultaneously produced utterances are faulty when produced by the student interpreter and nonexistent on the part of the native bilingual professional interpreter. More research with a larger body of data could address this issue. Nevertheless, the findings here support recent research that indicates that simultaneously produced utterances are problematic. Further investigation is needed to determine whether a high proportion of single code utterances (in this case, signed utterances), creates a feeling of exclusion among unratified participants.
Interpreter education programs could also discuss the implications of certain footing types. Introductions are an area that can be problematic and that require further investigation, so this is one area worth addressing in interpreter education. In addition, the student interpreter in the mock interview case occasionally did not follow certain discourse conventions in one or the other of the dyads. For example, she left an unfilled slot in question-answer pairs and provided lengthy nonresponses (e.g., explanations), which created interactional problems. According to this study, the use of minimal responses seemed to be the least problematic in the flow of the discourse. This is an area worth addressing with interpreting students as well.
An additional area found to be lacking in the student interpretation in this study is the summoning of participants. This could be the result of the artificial nature of the role play. Nevertheless, students of ASL-English interpretation should have the opportunity to practice this particular type of footing. One way of gaining such practice and reducing the artificial nature of medical interview role plays is to work with medical students. Although the mock medical encounter has much in common with the actual medical encounter in terms of the interaction of frames and schema and the two categories of interpreter-generated footings, there are also some differences that result, in part, from the fact that a technical redoing is not the same as an actual encounter. For this reason, and in an effort to prepare medical practitioners to work in interpreted encounters, it might be useful for interpreter education programs to team up with medical education programs. The students in both programs could assist one another in learning about the realities of interpreted medical interviews.
The examination of both a student interpreter and a professional interpreter has provided some information relevant to educating interpreters to work in medical settings. Nevertheless, it should not be overlooked that this is a case study. More research is needed to provide detailed information for interpreter educators. Both cross-sectional and longitudinal studies of interpreting students are needed. Such studies should include these interpreters working in classrooms, role plays, and any other settings in which they work. A systematic investigation of student interpreters in such settings can help to determine the effectiveness of such training strategies as role plays. In addition, such research could help to determine the relevance of temporal, sociolinguistic and other factors that impact on interpreter education.
As an applied sociolinguistic examination of interpreted interaction, this study ultimately contributes to the broader field of linguistics. The examination of interactive discourse in any setting provides useful information to those interested in understanding language. For example, while non-data-based or empirically based experimental studies of interpreting comprise the bulk of research on interpreted discourse, the information obtained from this study regarding the interaction of frames and schema, the types and functions of footing produced in interpreter-generated utterances, and the overlapping dyadic structure of the discourse have become clear as a result of the interactional approach and the use of naturally occurring data. The application of sociolinguistics to studies in other areas, such as the study of language changes resulting from such disorders as Alzheimer’s or aphasia (Hamilton 1994; Goodwin 1995), provides useful insights regarding language unavailable from noninteractional studies. Similarly, this study provides some information about language not available from noninteractional studies of interpreting. This information includes nonlinguists’ assumptions about language as well as information regarding frame theory and potential sequential structures of discourse that could apply in noninterpreted discourse.
According to Shuy (1995), “Tacit theories of language use could not be discovered outside of an interactive context.” The examination of interactive discourse can offer insights into such tacit theories held by participants within the interaction. The fact that interpreters generate nonrenditions in order to provide information regarding the fact that an utterance has been initiated and who is the source of the utterance reflects the interpreters’ tacit understanding that an utterance contains three parts: its existence, its source, and its content. The fact that this study focuses on interpreted discourse, and that the languages involved occur in two distinct modes (one visual, one acoustic), make the discovery of this tacit assumption about language possible.
This study also proposes a definition of the often conflicting terms applied within frame theory. The definition proposed here is based on the constructs that recur within the literature on frames, schema, and scripts. Building on the linguistic applications of frame theory within single language discourse such as English and ASL, this study expands the application to not only a multicultural context (Watanabe 1993), but to a multilingual one as well.
Examination of the footing functions and types within the interpreted encounters provides support for footing discovered in other settings, and for new types that might yet be found in different contexts. For example, the interpreters’ use of constructed dialogue provides a rare opportunity for future investigation of this phenomenon. As Voloshinov ([1929] 1986) indicates, in order to truly understand constructed dialogue it is necessary to examine the relationship between the constructed dialogue and the original utterances upon which it is based. Since interpreters in this study have been found to construct the dialogue of participants in real time, both parts are available for examination. This is an area of linguistic research worth further examination, in both simultaneous and consecutive interpreting contexts.
Finally, another implication for linguistics relates to the sequential structure of the discourse itself. The interpreted encounters can be seen to consist of two overlapping dyads, on the basis of the embedded adjacency pairs. While insertion sequences within adjacency pairs have been addressed in noninterpreted interaction (Schegloff 1972; Merritt 1976), these embeddings are generally accessible to and between the interlocutors. However, the embedded greetings discussed in this study indicate two separate interactions that overlap. This is a unique structure that influences the sequence of the discourse in both the ASL and English dyadic discourse. The implication here is that some types of noninterpreted interaction might include similar embeddings. Schiffrin (1993) and Aronsson (1991) both address the issue of speaking for another (e.g., speaking for a friend or a child). In addition, Rosenfeld (1996) finds an overlapping dyadic structure in her examination of therapeutic discourse. Future research could help to determine whether similar embedded structures occur in other types of discourse.
Examination of the questions raised in chapter 1 of this study has demonstrated that interpreters do influence interactive discourse. The ways in which interpreters and other participants frame the interpreted encounters were presented in chapter 3. The types and functions of interpreter contributions to the discourse were addressed in chapter 4. Thus, the initial question regarding the linguistic differences between a monolingual face-to-face conversation and an interpreted conversation has become clear. In interpreted discourse, the interpreter has the power to influence the interaction not only through interpreter-generated utterances that are not renditions or constructions of others’ discourse, but also through a misrepresentation of the source message footings within renditions.
While this study has revealed that interpreters have the power to influence discourse, it has only begun to examine the interpreter’s ability to not influence interactive discourse. That is, the findings here indicate that some interpreter strategies result in less marked influences within the interaction. Given the professional goal of not influencing discourse, more research regarding an interpreter’s ability to limit or constrain their influences in interpreted encounters is needed. In the meantime, this study supports previous sociolinguistic research that shows that interpreters are participants within interactive discourse and not merely conduits to it. Thus, the question for the field of interpreting becomes clear: should interpreters pursue full participation rights within interpreted encounters? Or should interpreters attempt to minimize, where possible, their influence within interpreted interactions? Herein lies the paradox of neutrality.