Interpreting Between Modes: Navigating Between Signed and Spoken Language

Vicky Crawley[1]
York St John University

Download PDF


This article examines an interpreting challenge faced by interpreters working between spoken and signed languages: the difference in the amount of concreteness (which the author terms “specificity”) between the two languages. This paper outlines the necessity to edit specificity when interpreting from British Sign Language (BSL) to English in order to produce natural-sounding language. Just as important is for specificity to be elaborated upon when interpreting from English to BSL. By examining this challenge, strategies often considered to be “innate” have been extracted from practice. This contribution to theory can then inform interpreter training. The author draws upon their recent research into the phenomenon known as “clarification,” which highlighted “underspecificity” as the most common cause of interpreter participation using clarification.
Keywords: specificity, signed language interpreting, clarification, interpreter participation, Map Task, underspecificity, overspecificity.


Signed language interpreting is a process conducted mostly between languages and between modes. The typical signed language interpreter, working between a visual and an oral language, may be considered by their clients to be only working between modes. The assumption that the interpreter is translating literally is sometimes highlighted when the interpreter uses a type of participation referred to in the literature as “clarification” (Angelilli, 2014, Major, 2014). Clarification involves the interpreter, participating as herself, asking for more information from the source language user in order to interpret accurately into the target language. A recent study by Crawley (2016) analysed the reasons for and the structures of these clarifications using conversation analysis (CA) as a theoretical base. The phenomenon called “clarification” in interpreting is similar to what is referred to in CA as “repair organisation” (Schegloff, Jefferson, & Sacks, 1977). The results of that study showed that the most common reason for the interpreter to participate as herself was a particular aspect of language difference: the amount of concreteness or abstractness internal to both British Sign Language and English. I have called this phenomenon “specificity” (Crawley, 2016). In this article, I show how differences in levels of specificity inhibit the ability of the interpreter to interpret between languages. These differing levels of specificity are a normal part of each source language and cause no problems when conversing monolingually. However, the interpreter is sometimes unable to interpret fully between spoken and signed languages without seeking further information because of differences in specificity. Understanding these differences will inform interpreters and interpreter trainers, and benefit the recipients of signed language interpreting services.

Language Difference Between Modes

The particular demands of working cross-modally have been addressed by several scholars. To begin with, the signed language interpreter must be physically visible in all of her work, including conference interpreting (Brennan & Brown, 1997), as she must be seen by the signed language user. In contrast, the spoken language interpreter is able to interpret over the phone, in a booth (both unseen), as well as visibly present in their community work. Brennan and Brown (1997) described the visual information which is present within the grammar of signed languages but not in spoken languages. Spoken-language interpreters working between English and languages such as Chinese and Hindi experience similar differences between languages; “brother,” for example, will need clarification between “older brother” and “younger brother”. When translating a sentence starting with “I” in English, the Hindi–English interpreter will need to know the gender of the speaker before being able to translate the sentence, due to verb agreement with gender.
I am interested in the linguistic work done by a signed language interpreter in terms of the processes of visual elaboration from spoken to signed language and of the stripping down of visual description when producing English from BSL. Both of these processes need to be performed in order for either target language to be accessible and intelligible. Less has been written about this part of the interpreter’s work (although see Finton & Smith, 2004, for “compression strategies”). Lawrence (1995) has also addressed “expansions” as a way to adequately translate from American Sign Language to English. Janzen and Shaffer (2008), however, believed this method to be a simplification, and they describe how each interpreting challenge should be dealt with individually, according to contextualization, a subject that had previously been written about by Gile (1995).
English relies on the ability to understand language and look at reference material at the same time, thus building up a picture from less-specified speech, and more-specified visual information. Crawley (2016, p. 155) states, “Any number of war films show soldiers referring to maps and giving instructions at the same time. If they were sign language users, they would refer, and then they would give instructions”. BSL relies on the ability to describe detailed visual information within the language. In a BSL conversation, the process of referring to the source material and talking about it would be consecutive. Dively (1998), in her work on repair strategies in ASL–ASL conversations, discussed eye gaze withdrawal by speaker to indicate word searches or recall difficulties.


BSL also differs from English in its use of superordinates. Stone (2010) showed how superordinate terms such as “injection”, “medicine”, and “treatment” are differently organised into superordinate terms in BSL, and therefore “treatment” must be separated into “oral medication – tablet-form”, “oral medication – liquid”, “radiating treatment” “intravenous-”; “intramuscular-”; and “subcutaneous-” injections. Stone also described how a more neutral visual depiction of an injection could be placed on a nonspecific part of the body (the back of a hand) in order for the distinction to be left unspecified in order for accurate representation of the source language.
English when compared with BSL, can be under-specified in that English tends to draw upon superordinates such as “weapons” rather than a specific term: “gun”, “knife”, “club” and so on. To interpret “weapons” into BSL, the interpreter must draw upon her own experience of “weapons”, in order to produce an appropriately visually rich rendition in the target language. She must decide which of all possible weapons she is going to use to depict “weapons”, as well as explain that these are examples and not necessarily the actual “weapons” meant by the speaker. To interpret “weapons” from BSL to English, the interpreter must search for English vocabulary items which evoke the visual imagery created in the source language. The interpreter will need to be careful about what she chooses to depict, and so the need to clarify is understandable. Indeed, Major (2014) found that the primary participants in her healthcare interpreting research reported expecting to have the interpreters clarify. In my own experience of practice, however, most primary participants do not behave in this way and are often surprised when the interpreter speaks as herself. This creates an internal tension within the interpreter between the need to clarify and the need to maintain a good rapport with both clients.
Such language differences are not restricted to cross-modal interpreting; they can also occur between spoken languages and between signed languages. For example, Berk-Seligson (1990) describes how the word camionetacan mean either a large car or a small truck. In order to fully interpret that word into English, an interpreter must consider whether to interpret both meanings or to clarify which meaning was intended.

Meaning Making/Shared Understanding

Unlike in an interpreted conversation, two interlocutors in a monolingual conversation about a camioneta can and often do interact at this level of vagueness. Understanding forms in the mind of the addressee, which may or may not coincide with the meaning expressed by the speaker; only when the addressee’s understanding no longer fits with what is being added might the addressee may ask information gathering/checking questions (Clark, 1996). Even without comment to the addresser, the understanding may simply re-form in the mind of the addressee as she changes her original assumptions. Foppa (1995) concluded that the only way to know that the addressee has understood is by the answers she gives. In a similar vein, Napier, McKee, and Goswell (2010) stated that the recipients of the interpreter’s work can judge its merit only by the output they receive in their language.
When an interpreter reassesses meaning in this way, she displays her understanding in the interpretation. In order to immediately produce meaning in the target language, signed language interpreters must demonstrate understanding earlier in a conversation than monolingual speakers do.  In addition, to ensure her contribution is as informative as is required (Grice’s maxim of quantity; Grice, 1975), the interpreter must use a different strategy for each language, because that which is sufficient in English is lacking in BSL and that which is sufficient in BSL is overly informative in English.


Signed language interpreters are accustomed to producing the sort of visual information which is included in signed languages. In BSL, saying that someone came into a room would include the direction of travel of the person and would show the positioning of the door. The interpreter must allocate positions and directions of travel, and then remember these for additional information as it comes. The BSL user knows that the interpreter’s placement of signs in the sign space does not reflect the real world. Both parties allow for a conversation-specific organisation of the protagonists in that interpretation. It has been widely understood that the interpreter should include what Brennan and Brown (1997) referred to as “visual encoding”, meaning, for example, including a person’s height and level of mobility within the statement “I crossed the road”. It might even include the traffic on the road, and the person’s speed. This type of encoding is done in BSL “as a matter of course” (Brennan & Brown, p. 121).
Clark and Marshall (1981) described two types of mutual knowledge: genericand particular. To interpret a superordinate noun such as “cancer”, an interpreter may designate a conversation-specific but neutral place (either on the body or in the signing space) to denote “cancer” and this place can be referred back to (pointed to) if necessary. By the choice of a neutral placement for “cancer”, the interpreter demonstrates generic knowledge. If the position of the cancer is known, the interpreter will refer to it specifically, demonstrating the interpreter’s particular knowledge. For example, to interpret “liver cancer”, an interpreter may point to a place approximately where the liver is in the body or to a representation of a tumour in that area. However, as the interpreter is not the principal or author (Goffman, 1981) of the English being spoken, she may not know where the cancer is. The speaker may not know either, or he may know but have simply not specified. The interpreter is, however, author of that version of what the English speaker has said, and, as such, she needs to be as specific as she can, if she wishes to be both clear and grammatical. She may need to clarify in this case.
Given the above, it is not surprising to note that visual encoding is often omitted when interpreting into English from BSL. Brennan and Brown (1997) provided the example of producing “window” in BSL, which must be depicted as opening inwards or outwards, left to right, right to left, or sash and so on (Brennan & Brown, 1997). None of these attributes need to be included in the English rendition “window”. Implied within the English word are all the possible types of window, and specification in conversation is rarely needed.
Visual encoding is also necessary in BSL for more abstract ideas as well as concrete items. In BSL, left and right are conventionalised whereby the left of the signer’s signing space is seen as “left,” despite being on the right side of the addressee. Similarly, the right-hand side of the signer signifies east and the left-hand side west. (Languages such as Australian Aboriginal languages use absolute north and south in conversation and would orient to these absolutes when describing the landscape; Seyfeddinipur & Gullberg, 2014).
The signed language interpreter is faced therefore with an imbalance in the specificity required in signed versus spoken modes. Understanding which is usually negotiated over the length of a conversation must be understood and represented with visual detail. Modal differences mean that looking at a map and listening to directions is not optimal in a visual language. Superordinates, plentiful in spoken languages, are differently organised in signed languages. In some instances, English superordinates must be represented in BSL by examples of those items which are the most typical (vehicles – trains, buses, cars). To highlight the discrepancies between the two languages, for which interpreters may have to provide more clarification, I had participants perform the HCRC BSL Map Task. In this task, English- and BSL-using participants had give and receive directions according to a printed map, with signed-language interpretation.

The Study: Methodology

The data for this study were eight BSL–English interpreted interactions captured on video by Gary Quinn of Heriot Watt University for the HCRC BSL Map Task Project (Turner & Merrison, 2016). In the HCRC Map Task (Anderson et al., 1991), an established tool for eliciting natural-language production, each participant is given a map, one with a route drawn on it and the other without. The task of transcribing a route from one destination on a map to another solely through spoken language distracts the participants from their language production and onto the task.
The participants chosen were all female, in order to eliminate potential gender bias. They comprised two deaf women who were BSL users, two hearing women who were non-BSL-using English speakers, and two bilingual BSL–English interpreters. Each of the three members of the interaction (one BSL user, one English speaker and one bilingual interpreter per video) was filmed face on. My study had originally aimed to use a purely conversation-analysis (CA) approach. Coupled with a close analysis of data, CA holds that talk is recipient designed for the immediately relevant parties; as Schegloff (1987) states, utterances are “produced by the parties for one another and were designed, at least in part, by reference to a set of features of the interlocutors, the setting, and so on that are relevant for the participants” (1987:209). Further to the claim that talk is recipient designed, CA describes a “sequential architecture of intersubjectivity” (Heritage 1984:108), that is, each utterance is built upon the utterance which comes before it.  After much consideration, I decided to analyse the videos using a method inspired by CA. (The spoken English part of the videos had been transcribed earlier by another researcher (Jack Wilson)). I had planned to translate the BSL and add that to the English within ELAN, but this proved inappropriate due to the differing ways that the languages are perceived: In spoken languages, an overlap in CA is perceived by all of the participants, but in a BSL–English conversation, only the interpreter knows that there is overlap. I also did not want to convert BSL into English because timings became obsolete, and I was concerned with the interaction between participants. My study aimed to find (a) the most common linguistic environments which caused an interpreter to speak for themselves in order to interpret; (b) how interpreters showed that they were speaking as themselves; and (c) whether interpreters’ clients understood whom the interpreters were speaking as.

Figure 1: HCRC Map Task (Turner & Merrison, 2016).

Map showing route around african savanna
The original Map Task includes discrepancies  among the features and landmarks marked on each of the two maps, to encourage negotiation between the participants. In the BSL Map Task project, the addition of professional BSL–English interpreters meant there were three participants (albeit with differing levels of participatory rights). Each primary participant was talking to someone whom they knew had sight of the map, and not to the interpreter. This is a particularly important consideration when analysing the way the participants framed their talk (in either language).
When transcribing spoken language interpreting using CA, both languages can be written transcribed in their written form, and any further translations can be added as an additional row. Both languages will be present on any audio recording. For signed languages, however, there are two major differences. First, there is no standardized form of writing down signed languages (although there are systems to annotate signed languages). Second, signed language is not heard, so interpreters are routinely faced with having both clients talking at the same time without being aware that the other person is talking. I did not have the option of writing BSL in BSL, and then adding a translation of it below (cf. Wadensjö, 1998), so at first I added a free translation of both the interpreter and the BSL user. As stated above, this did not prove adequate, however, as the timings of the signs used, and those of the translations I had made were incompatible with the usual CA method. I decided to use glosses instead.


CA suggests that the analyst observe the data as like a naturalist identifying new species of animals. In reviewing the videos of the map task, my aim was to find examples of interpreter difficulty, followed by repair sequences. I analysed the data and marked the timings of “interesting” features, including 105 examples of “repair”. Of these—most of which were repair between the participants—77 were repair undertaken by the interpreter, which I then categorised into 11 different general types. From the videos I also produced a method of transcription (described in Fig. 2; cf. Ochs, 1979, “transcription as theory”) that enabled me to observe the behaviour of all three of the participants at the appropriate level of detail as well as allowed for further analysis (an elaborate system of understanding receipts by the interpreter and the BSL user, which is more difficult to perceive when watching three videos at the same time).The figure is separated into three parts, the first for the “hearing follower” (HF), who is describing the route she has followed, to check with the “deaf giver” (DG) that she has gone the right way. The first line is that of her gaze—in a visual language such as BSL, eye gaze denotes attention (Baker, 1977). By noting where the interlocutor is looking, the researcher also notes what is being attended to. The second layer of the transcription musical score is that of the interpreter (I, marked IBSL or IEng depending on language used), and her eye gaze is also marked.


As I have detailed earlier, interpreters elaborate and edit visual information in their everyday work. Interpreting for a task involving a map with features which are drawn and named, the interpreters found themselves in the unusual situation of needing to represent visual information which is generally omitted in spoken English. The decision to not allow the interpreter sight of the map meant that she needed to build up a version of the map with the DG, the BSL-using client. This is routine. However, in this task the presence of the physical map, which was referred to by both primary participants, meant that the interpreter had an external reference. The usual strategy of building a conversation-specific version of what has been said was not available to her.
Visual information from DG, therefore, had to be made more explicit in English than is usual in average interpreting situations. The interpreter had to leave as much visual information in place as possible because she was unable to determine which pieces of information were going to be relevant. When working from English into BSL, the interpreter was not able to create visual information in the way described in Section 2 above, and as a result, the interpreter had to participate as herself—mostly because “leftness” or “rightness” are necessary in descriptions in BSL yet are not usually specified in English.
As mentioned in the methodology, the English half of the BSL Map Task had been transcribed into a CA format by a non-BSL-using researcher. This researcher commented anecdotally that some of the vocabulary used by the interpreter (e.g., the interpretation of the BSL user) sounded slightly unusual, and the researcher wondered if this was an artefact of the interpreting process. I believe that it was, but not due to a failing in the way that it was interpreted. Instead, the interpreter’s creative use of English vocabulary indicated a deliberate strategy to most effectively incorporate the visual information from BSL into English, and this required more poetic language. For example, a route was described as “slightly snaking”, which was a translation of a particularly specific description of the route taken across the page. Another term for a similar description was “meandering,” which was chosen to describe a slow and gentle movement across the paper. It is interesting to note that the English-speaking clients sometimes responded using the same vocabulary been produced by the interpreter, showing understanding that the words held specific meaning.
I have already stated that leftness and rightness were a challenge for the interpreter. In the Map Task, leftness and rightness were often omitted from the English instructions, with the use of “round”, “around the”, “next to” and “under the” being used in combination with sight of the map. Interpreters dealt with this lack of specificity in a number of ways. In the excerpt in Figure 2, the interpreter waits for more English which may give more visual information.

Figure 2: Interpreter waits for more information.


Gaze:                                                                                          +map   

HF:     all thewayroundthehighviewpoint

Gaze:                     +DG                                                              +awayandback x2          


Gaze:                                                                                          +INT     


The language produced by HF is not sufficiently specific for the interpreter to produce BSL. The interpreter uses a conventional strategy of waiting in the hope of gaining more visual information. The interpreter’s eye gaze moves away and back from DG twice, right after “all the way” and “the”. This is significant grammatically, as the points of eye gaze change correlate with two grammatical items which are awkward to interpret: (a) the English word “round”, which in BSL must include direction but does not include direction in the source; and (b) when the English speaker mentions a new landmark on the map. In BSL a place must either be first set up in space and then named, or first given a name that then correlates to an established location in space. Here the interpreter can do neither because the information is incomplete, that is, it is underspecified in the source language. She might try to provide a temporary, conversation-based direction (in the first instance) and a location (in the second instance); but, due to the importance of adhering to a real-life map, neither of these tactics is available. And neither can the interpreter consult the map to check it. Therefore, the interpreter suspends her interpretation, and displays eye gaze withdrawal to show word/sign retrieval (Dively, 1998). She then produces two versions of “going around” a feature (AROUND-LEFT AROUND-RIGHT), meaning that it could be either direction, then reiterates that it could be either left or right (LEFT-RIGHT), and then DG acknowledges (to the interpreter) that it is towards the right. The interpreter does not provide overt clarification, but rather presents the possibility of one of two different directions to the BSL user for her to do with as she wishes.
In the next example, the same interpreter makes sure that the information she does have access to is fully described, and seeks confirmation from DG. She changes her eye gaze at the completion of the English phrase, and starts with a location name, FIELD STATION.
Figure 3: Interpreter locates the landmark in space before moving in relation to it.


Gaze:   +map__________________________________________

HF:       I wen’rou::ndthefieldstation

Gaze:   +HF___________________+DG___________


IBSL:                               FIELDSTATION

Gaze:   +HF_____________________+INT ___________

DG:                                    NOD    NOD

HF uses the sound stretch “rou::nd” to represent an elongated ‘going round’; at the same time that she gestures to represent a significant journey around the landmark. Adverbials of duration, effort expended and direction, as well as trajectories which are nonstatic are all possible in BSL; however, they cannot be interpreted accurately if the interpreter does not know which way, how far, how much energy has been expended or the shape of the trajectory, and all of these are underspecified in HF’s spoken English. INTlooksatHFwhileHF isstating,“Iwentroundthefieldstation”.Atthepointthat INThears“fieldstation”(thelandmarkshewaswaitingfor),she turnstoDGandsigns FIELD,whichreceivesanodfromDG,andSTATION,whichalsoreceivesanod. Shehas thus establishedwithDGthatthereis a landmarkcalled FIELD STATION.
The interpreter handles the direction by again positing to the BSL user that there are two different ways to go around. She uses a feature of BSL grammar: raising or lowering eyebrows to signify a question. To the question is AROUND-LEFT, DG looks immediately down at her map and then back up and nods. To AROUND-RIGHT with questioning eyebrows, DG does not nod. The interpreter then signs AROUND-LEFT more confidently and then nods herself.
The next extract deals with more than one problem resulting from the original utterance.

Figure 4: Dual underspecificity.


Gaze:        +map ________________________________________

HF:            Ithen went     rou::nd             the ropebridge

Gaze:        +HF___________________________________________+up


Gaze:        +INT___________________________________________


The difficulty for the interpreter in this extract is that ‘rope bridge’ is a landmark which had not been mentioned before by either participant. The interpreter has the same challenge as in Figure 3: She needs to know which way to go around the landmark. Up to here, this is compounded by the difficulty of the appearance of a new landmark, specified only by name and not location. The interpreter is faced with these two underspecifications at the same time. For HF, the act of saying the name of the landmark defines where it is in space, because the name is identifiable on the map (which can be accessed at the same time as listening for English speakers). This means that HF has been specific in English but has not been specific enough in terms of enabling the interpreter to produce a fully specific translation in BSL. The complex nature of the interpretation could be the cause of an upward gaze by the interpreter. The upward gaze, or eye gaze withdrawal, is a marker of self-repair initiation (Dively, 1998)—when a signer, in Dively’s work this is an ASL user, believes s/he has made a mistake. For the interpreter here there is no mistake; however, a reformulation needs to be made due to the complexity of the language difference. The interpreter again posits two different directions (left or right) and DF nods after both, With these nods she is agreeing that there is a landmark, and that you go either left or right around it. Her nods do not signify confirmation to the interpreter, only an acknowledgement that it could be either way.
The next extract considers the connotations of the word “edge”.

Figure 5: Underspecificity of ‘edge’.


Gaze:  +map                                                                                                      

HG:      sonowyouareat                   theedge            of thecranebay

Gaze:  +HG                            +DF                           +signspace__+DF

IBSL:                                                                      NOWb-a-yTHEREYOU

Gaze:  +map                                +INT                                                              

DF:                                                                                            NOD

In Figure 5 the interpreter has three underspecified elements to interpret: “So now you are at” (place is underspecified) “the edge” (any landmark will have any number of edges) “of the crane bay” (now the interpreter knows that ‘edge’ must mean border between land and sea, which is even more underspecified than “edge”). The interpreter looks first at HF (signifying to DG that she is “listening”), then she looks back to DG to show that interpretation will be starting soon, then looks into sign space (eye gaze withdrawal and also perhaps to try to place things in the sign space) signing NOW b-a-y THERE YOU. In this way, the interpreter is setting up placement of what is definite, or specified, and elicits a nod from DG. This nod may be to acknowledge understanding up to that point.
Also seen in the data were instances of overspecificity from DG. Interpreters are accustomed to reducing the amount of visual information they produce in English, to create natural-sounding target language. In this task, however, the interpreter is aware that the visual information coming from DG is important not simply for the grammar of BSL, but also for the accomplishment of the task. The way that this has been done in the task is that the vocabulary produced by the interpreter becomes more poetic, presumably in an attempt to capture some of the specificity in BSL. As described above, words such as “meandering”, phrases like “slightly snaking” and “forking back on yourself” are examples of the interpreter’s efforts to reproduce visual information in English.
Conducting and analysing the map task between a BSL user and non-BSL user with interpretation demonstrates that interpreters, when faced with underspecificity, will wait for more information before starting their interpretation; posit two different directions and wait for acknowledgement from the BSL user; posit two different directions without seeking the answer. An elaborate system of nods to confirm understanding is used by both the interpreter and the BSL user to work together to negotiate meaning.


Interpreters routinely navigate the differences between the amount of visual information found between BSL and English. This is done by elaboration of detail from English into BSL and the editing out of separate details, and judicious selecting of vocabulary from BSL into English. Interpreters are often allowed a high degree of latitude when it comes to producing visual information in BSL; clients will know that an “edge” could be on either side, or that “going round” has a direction in BSL, and might not be the direction given by the interpreter.
The map task highlighted the differences in specificity between BSL and English, specifically, differences in the ways BSL users and English speakers refer to maps. For example, English speakers go from place mark to place mark, whereas BSL users describe the route and its direction such that they do not need to rely on the place names as heavily but can show the precise specificities of the route.
Interpreter trainers as well as interpreters in the field may benefit from this analysis of specificity in spoken and signed languages. It may be advantageous to use this or a similar map task at various stages during training. The fact that there is an artefact at the end will help in the process of examination of the difficult parts of the interpreting process. Being able to see the route drawn by the person following the directions will enable students and trainers to identify the exact moments when difficulties arise. Trainers could use the map task to teach particularly difficult concepts, such as found with the meanings of “edge” (“sharp corner”, “straight side”, or “border between water and land”). An interpreter who understands the reasons behind a difficulty in an interpreting assignment may more easily resolve the difficulty in the future, for example when faced with a complicated and underspecified piece of information to interpret. By focussing on the reasons for needing more information, the student can build a repertoire of simple explanations. As noted by Napier et al. (2010, p. 75), “It is ethical to ask for clarification, but it could make the interpreter look incompetent”. Armed with more knowledge about how specificity causes challenges for the interpretation of visual information, the interpreter can confidently explain their difficulties to their clients, and the interpreter trainer can devise teaching methods designed to address this particular difficulty in the ways listed above.
The number of participants in this study was small, and so future research could analyse the strategies used by a larger number of interpreters, in order to expand on these findings. Additional research could also include an action research project, perhaps where interpreters reflect on their practice using the drawn route as a guide. The study revealed “specificity” as the most common reason for interpreters participating as themselves. Interpreter trainers are often faced with answering questions from their students with “Well, it depends …”. Research into the natural strategies interpreters have devised for use in their practice will go a long way to improving the quality of training for students and their eventual clients.


I would firstly like to thank the participants in the study, who very generously gave of their time. I would like to thank Gary Quinn for the video recordings and for organising the data collection; my supervisors Andrew Merrison, Dai O’Brien, Rachel Wicaksono and Graham Turner for their expertise and counsel through the process of my degree; and my examiners Bethan Davies, Lorraine Leeson and Christopher Stone for their guidance on the topic of this article. Lastly, I would like to thank Tim, Felix and Gus for living with me through the process.


Anderson, A. H., Bader, M., Gurman Bard, E., Boyle, E., Doherty, G., Garrod, S., Isard, S., Kowtko, J., McAllister J., Miller, J., Sotillo, C., Thompson, J. S., & Weinert, R. (1991). The HCRC Map Task Corpus. Language and Speech34(4), 351–366.
Angelilli, C. V. (2014). Uh … I am not understanding you at all: Constructing (mis)understanding in provider/patient interpreted medical encounters. In B. Nicodemus & D. Metzger (Eds.), Investigations in healthcare interpreting(pp 1-31). Washington, DC: Gallaudet University Press.
Baker, C. (1977). Regulators and turn-taking in American Sign Language discourse. In L. Friedman (Ed.), On the other hand(pp. 215–236). New York, NY: Academic Press.
Berk-Seligson, S. (1990). The bilingual courtroom: Court interpreters in the judicial process.Chicago, IL: University of Chicago Press.
Brennan, M., & Brown, D. (1997). Equality before the law: Deaf people’s access to justice.Durham, NC: Deaf Studies Research Unit.
Clark, H. H. (1996). Using language.Cambridge, UK: Cambridge University Press.
Clark, H. H., & Marshall, C. R. (1981). Definite reference and mutual knowledge. In A. Joshi, B. Webber, & I. Sag (Eds.), Elements of discourse understanding(pp. 10–63). Cambridge, UK: Cambridge University Press
Crawley, V. L. (2016). Achieving understanding via interpreter participation. Unpublished PhD thesis. University of Leeds, Leeds, England.
Dively, V. (1998). Conversational repairs in ASL. In E. Winston (Ed.), Pinky extension and eye gaze: Language use in Deaf communities(pp. 137–169). Washington DC: Gallaudet University Press.
Finton, L., & Smith, R. (2004).The natives are restless: Using compression strategies to deliver linguistically appropriate ASL to English interpretation. In E. M. Maroney (Ed.), Proceedings of the 15th national convention, Conference of Interpreter Trainers, CIT: Still shining after 25 years(pp. 125–143). Washington, DC: CIT. [AN 1821]
Foppa, K. (1995). On mutual understanding and agreement in dialogues. In L. Markova, C. F. Graumann, & K. Foppa (Eds.), Mutualities in dialogue(pp. 149–175). Cambridge, UK: Cambridge University Press.
Gile, D. (1995). Basic concepts and models for interpreter and translator.Amsterdam, the Netherlands: John Benjamins.
Goffman, E. (1981). Forms of talk. Philadelphia: University of Pennsylvania Press.
Grice, H. P. (1975). Logic and conversation. In Cole, P., Morgan, J. (Eds.), Syntax and semantics 3: Speech acts. pp 41-58. New York: Academic Press.
Heritage, J. (1984). Garfinkel and Ethnomethodology. Cambridge: Polity Press.
Janzen, T., & Shaffer, B. (2008). Intersubjectivity in interpreted interactions. In J. Zlatev, T. P. Racine, C. Sinha, & E. Itkonen (Eds), The shared mind: Perspectives on intersubjectivity.(Converging Evidence in Language and Communication Research 12; pp. 333–355). Amsterdam, the Netherlands: John Benjamins.
Lawrence, S. (1995). Interpreter discourse: English to ASL expansions. In E. A. Winston (Ed.), Mapping our course: A collaborative venture, Proceedings of the tenth national convention, Conference of Interpreter Trainers(pp. 205–214). Conference of Interpreter Trainers.
Major, G. (2014). “Sorry, could you explain that?” Clarification requests in interpreted healthcare interaction. In B. Nicodemus & M. Metzger (Eds.), Investigations in healthcare interpreting(pp. 32–69). ashington: Gallaudet University Press.
Napier, J., McKee, R., & Goswell, D. (2010). Sign language interpreting. Annandale, NSW: Federation Press.
Ochs, E., (1979). Transcription as theory.In E. Ochs & B. B. Scheifflen (Eds.), Developmental pragmatics(pp. 43–72).New York, NY: Academic Press.
Schegloff, E. A. (1987). Some sources of misunderstanding in talk-in-interaction. Linguistics 25(1), pp 201-218.
Schegloff, E. A., Jefferson G., & Sacks, H. (1977). The preference for self-correction in the organization of repair in conversation. Language, 53(2), 361–382.
Seyfeddinipur, M., & Gullberg, M. (2014). From gesture in conversation to visible action as utterance.London, UK: John Benjamins.
Stone, C. (2010). Access all areas—Sign language interpreting: Is it that special? Journal of Specialised Translation, 14,41–54.
Turner, G. H., & Merrison, A. J. (2016). Doing ‘understanding’ in dialogue interpreting: Advancing theory and method. Interpreting: International Journal of Research and Practice in Interpreting, 18(2), 137–171.
Wadensjö, C. (1998). Interpreting as interaction.New York, NY: Longman.