Language direction has long been a subject of debate among spoken language conference interpreting educators. Seleskovitch (1978: 100), the pioneer of the Paris school of thought, claims that “simultaneous interpretation can only be done properly into one’s native language.” This view is primarily because native language (L1) production is spontaneous, idiomatic, clear, and smooth, which therefore ensures that the users easily understand the interpreter’s interpretation. By contrast, Denissenko (1989) of the former Soviet Union contends that simultaneous interpretation from L1 into L2 is more favourable, on the grounds that native language (L1) comprehension would ensure accuracy in interpretation. Due to a lack of empirical evidence, the two opposing views are largely based on personal experience, ideology, and tradition (Gile 2005). Both views appear to be oversimplified, because they focus on language competence only and they do not consider non-linguistic parameters that may have an impact on interpreting performance.

Some researchers now look at the directionality issue from a broader perspective. Gile (2005) points out that interpreting performance depends on numerous factors such as the interpreter’s language proficiency, familiarity with subject matter, working memory capacity, state of health, motivation, and professionalism. Gile therefore maintains that directionality effects on interpreting performance might be small or even negligible. In addition, Kalina (2005) argues that the question of which language direction results in a higher quality of interpreting performance should be addressed in specific terms such as language pair, type of conference, nationalities and cultural backgrounds of the speakers and the listeners, and language distribution. Seel (2005) asserts that research into directionality effects on interpreting performance should take into account parameters such as non-verbal discourse patterns of the source culture, working conditions, specific contextual and extralinguistic factors, the speakers and the listeners, subject matter, and cognitive demands on the interpreter. Similarly, Pavlović (2007) endorses that variables such as the directionality norm (that spoken language conference interpreters should interpret into their L1), interpreter education, and the interpreter’s previous work experience may confound directionality effects on interpreting performance. All these variables may jointly influence simultaneous interpreting performance in each language direction, thus confounding or even offsetting directionality effects on simultaneous interpreting performance. Neither language direction is intrinsically easier than, or superior to, the other language direction (Kalina 2005; Martin 2005; Padilla 2005; Seel 2005).

Extensive survey research has concentrated on interpreters’ feelings about directionality. Professional spoken language interpreters typically prefer, and feel more proficient in, L2-to-L1 interpreting and perceive L1-to-L2 interpreting to be more stressful and tiring (Bartłomiejczyk 2004; Choi 2008; Donovan 2004; 2005; Lim 2005; Martin 2005; Nicodemus and Emmorey 2013). Nonetheless, Al-Salman and Al-Khanji (2002) found that professional Arabic/English interpreters felt more comfortable working from Arabic (L1) into English (L2) than the other way around, which may be due to their insufficient mastery of Arabic (Gile 2005). It is worth noting that interpreters’ directionality preferences are subjective and may be related to their personal and professional profile, age of L2 acquisition, self-perceived bilingual proficiency, language combinations, and interpreting experience (Bartłomiejczyk 2004).

Despite a large number of survey studies on interpreters’ directionality preferences, few empirical studies have examined directionality effects on actual interpreting performance. Tommola and Laakso (1997) and Tommola and Helevä (1998) found that student interpreters’ propositional accuracy in simultaneous interpretation was significantly affected by the delivery speed and linguistic complexity of the source text rather than by directionality. These findings support the aforementioned view of Gile (2005) that other factors rather than directionality may have a greater impact on simultaneous interpreting performance.

Nevertheless, results from student interpreters may not apply to professional interpreters. Daro, Lambert et al. (1996) found that 16 professional French/English interpreters[3] made significantly more omissions when simultaneously interpreting difficult texts from L1 into L2 than vice versa. Chang and Schallert (2007) found that six professional Mandarin/English interpreters not only achieved significantly higher propositional accuracy but also made significantly fewer target-language-use errors (grammatical and lexical errors) when simultaneously interpreting from L2 (English) into L1 (Mandarin) than vice versa. By contrast, Chang and Schallert did not observe significant directionality effects on the propositional accuracy or the target-language-use errors for another three professional Mandarin/English interpreters who had reported being dominant in English or being equally competent in both languages. These results indicate that interpreters’ relatively balanced bilingual competence may mitigate directionality effects on their simultaneous interpreting performance. Further empirical research is needed to verify these findings.

Although many studies have investigated directionality in spoken language interpreting, very few studies have explored the same topic in signed language interpreting. To date, three survey studies have examined the directionality preferences of signed language interpreters. Napier, Rohan et al. (2005) found that certified Auslan/English interpreters preferred working from English (L1) into Auslan (L2), with less experienced interpreters expressing a stronger preference for this language direction. Similarly, Xiao and Yu (2009) found that signed language interpreters in China predominantly considered it more difficult to interpret into Chinese (L1) than into Chinese Sign Language (L2). More recently, Nicodemus and Emmorey (2013) found that ASL/English interpreters not only preferred but also felt more proficient working from their L1 (English) into L2 (ASL). As in Napier, Rohan et al.’s (2005) study, novices reported a stronger preference for interpreting from L1 into L2 (English-to-ASL) than experts did, suggesting that extensive interpreting experience may result in equal preference for L2-to-L1 and L1-to-L2 interpreting. Furthermore, Nicodemus and Emmorey found that ASL/English interpreters who were native signers largely reported no directionality preferences, indicating that balanced bilingual proficiency may mitigate directionality preferences.

Interpreters’ directionality preferences may not match their actual interpreting competence. Nicodemus (2011), in the study mentioned earlier, investigated directionality effects on the simultaneous interpreting performance of 32 ASL/English interpreters, all of whom had English as L1 and learned ASL (L2) in adulthood. Her participants comprised two groups: 16 experts who were certified interpreters with over 10 years of interpreting experience, and 16 novices who were pre-certified interpreters with less than five years of interpreting experience. Regarding the experts, Nicodemus found no significant directionality effects on the accuracy, fluency, speed, or prosody of their simultaneous interpreting performance. The novices scored significantly higher in terms of accuracy, speed, and prosody when simultaneously interpreting from ASL into English (L2-to-L1) than vice versa; however, they scored significantly lower in terms of fluency when simultaneously interpreting from ASL into English (L2-to-L1) than the other way around. It is important to note that although novices typically preferred to interpret from L1 (English) into L2 (ASL) (Nicodemus and Emmorey 2013), they actually performed significantly better when interpreting from L2 (ASL) into L1 (English) (Nicodemus 2011).

Moreover, van Dijk, Boers et al. (2011) explored the effects of directionality and age of signed language acquisition on the simultaneous interpreting performance of 25 experienced Sign Language of the Netherlands (SLN)/Dutch interpreters. The participants consisted of 10 native signers and 15 non-native signers. Van Dijk, Boers et al. found that the native signers’ Dutch-to-SLN simultaneous interpreting performance was significantly better than their SLN-to-Dutch simultaneous interpreting performance. By the same token, the non-native signers’ Dutch-to-SLN (L1-to-L2) simultaneous interpreting performance was significantly superior to their SLN-to-Dutch (L2-to-L1) simultaneous interpreting performance. Participants’ inferior performance on SLN-to-Dutch simultaneous interpreting was attributed to the following two reasons:

• In real life, they had less experience interpreting from SLN into Dutch than vice versa;

• They may have encountered SLN comprehension problems due to extensive linguistic variation in SLN.

Furthermore, van Dijk, Boers et al. found that the native signers were similar to the non-native signers in their overall simultaneous interpreting performance (combining Dutch-to-SLN and SLN-to-Dutch simultaneous interpreting performances), in part due to the small sample size of the study. Given that the native signers were compared with the non-native signers in terms of their overall simultaneous interpreting performance, it remained unclear whether the two groups differed in their performance in each language direction, which merits further research.

The experts (non-native signers) in Nicodemus’s (2011) study had a similar amount of interpreting experience to the non-native signers in van Dijk, Boers et al.’s (2011) study; however, these two groups exhibited different directionality effects on their simultaneous interpreting performance. The different results between the SLN/Dutch interpreters and the ASL/English interpreters might reflect greater lexical variation in the SLN data compared with the ASL data. Future research is needed to identify the reasons for such divergent findings. In terms of evaluating accuracy in interpretation, one concern is that both studies conducted the evaluation using selected sentences from the source text instead of using the entire source text. Even though the selected sentences were accurately transferred, the entire source text may still be inadequately interpreted. Further empirical research therefore is needed to shed more light on directionality in signed language interpreting.

Given that signed language interpreters consist of native signers and non-native signers, one may wonder whether the two groups differ in their interpreting performance. Apart from van Dijk, Boers et al.’s (2011) study, only the following two studies have explored the similarities and differences between native signers and non-native signers in actual interpreting performance.

Napier (2006) required two professional Auslan/English interpreters (a native signer and a non-native signer) to interpret university lectures from English into Auslan (a different lecture for each interpreter but on a similar topic). She noted that the native signer used mouthing and fingerspelling[4] appropriately and effectively in the English-to-Auslan interpretation, while the non-native signer exhibited more English interference and used more marked mouthing and fingerspelling in the English-to-Auslan interpretation (for example, in terms of inclusion of function words). Napier suggested that the non-native signer might still be thinking in English rather than in Auslan during the English-to-Auslan interpreting, which may have affected their Auslan output. This speculation on the part of Napier is consistent with Nicodemus and Emmorey’s (2013) view that novice interpreters may easily default to fingerspelling and transcoding (literal word-for-sign translation) when interpreting from a spoken language into a signed language.

Moreover, Goswell (2011) investigated the use of role shift[5] in English-to-Auslan interpretations produced by four accredited Auslan/English interpreters (two native signers and two non-native signers). She found that the interpreters were highly variable in the amount of role shift produced, with no clear advantage on the part of the native signers. Thus, more research is needed to provide further insight into how important the nativeness effect may be in relation to accuracy in interpretation.

In sum, much evidence shows that spoken language interpreters often prefer to interpret from their non-native language (L2) into their native language (L1); in stark contrast, signed language interpreters typically prefer to interpret from their L1 (a spoken language) into their L2 (a signed language). It is important to note that interpreters’ directionality preference may not align with their actual interpreting competence. To date, very few empirical studies have explored directionality effects on the actual interpreting performance of either spoken or signed language interpreters. Hence, the current mixed methods study makes a contribution to the body of literature reviewed thus far, as it aims to investigate the effects of directionality and age of signed language acquisition on the simultaneous interpreting performance of professional Auslan/English interpreters.

Due to its focus on conference level signed language interpreting in Australia, this study required that participants be professional level Auslan/English interpreters qualified by the National Accreditation Authority for Translators and Interpreters (NAATI).[6] A total of 31 NAATI-accredited professional level Auslan/English interpreters self-selected to participate in the study, consisting of 14 native signers and 17 non-native signers. The native signers acquired both Auslan and English from birth – the first from their signing deaf parents, the second through interaction with the surrounding hearing population.[7] The native signers were thus considered to be native users of both English and Auslan (balanced bilinguals). According to participant self-reports, the non-native signers acquired English from birth from their hearing parents; they often had no family connections to the Australian Deaf Community, but started to learn Auslan at or after age 10 by receiving formal education in Auslan and/or associating with deaf signers through work or social networks. The non-native signers were thus considered to be unbalanced bilinguals with English as their native language (L1) and Auslan as their non-native language (L2). Data were collected in four major cities across Australia: Sydney, Melbourne, Brisbane, and Perth.

Participants’ demographic and professional data are summarized in Table 1. The majority of participants were female (87%), which is representative of the working population of Auslan/English interpreters (see Bontempo and Napier 2007; Napier and Barker 2003). The native signers were similar to the non-native signers in age and interpreting experience. Compared with the native signers, the non-native signers received more general education and interpreter education. Although participants generally expressed no directionality preference, they reported practising English-to-Auslan interpreting more often than Auslan-to-English interpreting.

Note. Interpreting experience means the years of working as a paid interpreter, irrespective of NAATI accreditation.

-> See the list of tables

The main goals in creating the English and Auslan source texts were achieving comparability (of context, topic, presentation rate, and length) and authenticity. Both texts were created for a notional national conference themed “Interpreting and Human Rights”[8]. To make the texts as realistic as possible, each presenter was briefed in terms of the conference theme and the intended audience of their presentation. The audience for both presentations was described as including English-speaking attendees who did not understand Auslan, and deaf professionals who were bilingual in both Auslan and written English. It was assumed that the audience was familiar with human rights issues. Each presenter created their own presentation based on their professional expertise, rather than actors being given a speech to read out, thus increasing authenticity and content validity.

For the English source text, a solicitor (a native Australian English speaker based in Sydney) was filmed giving a presentation entitled “Interpreting and Human Rights.” He was filmed giving his presentation, using printed PowerPoint slides as a prompt. The manager of a Deaf Community service organization (a deaf native Auslan signer based in Sydney) provided the Auslan source text, giving a signed presentation. The topic was “Deaf People and Human Rights,” reporting findings from an international research project. The deaf presenter was also filmed giving his presentation based on PowerPoint slides.

Each presenter started with a brief introduction (about three minutes), paused for a few seconds, and then gave their formal presentation (approximately 17 minutes). The researcher served as each presenter’s audience and filmed each presentation as an mp4 video, using an Ultra HD Flip camera. A native Auslan signer was also present as an audience member for the filming of the Auslan presentation. Due to environmental constraints, the camera captured the presenters but did not capture their PowerPoint slides.

In order to replicate a real conference interpreting experience and encourage pre-task preparation, the following materials were sent to participants two weeks prior to data collection: task instructions (for example, conference theme, description of the intended audience, each presenter’s name and background, presentation title), glossary (legal terms in the English source text, fingerspelled words in the Auslan source text), and copies of PowerPoint slides for both presentations.

Participants were tested individually, predominantly in Deaf Community service organizations[9] in their respective cities between March and July 2011. After completing a consent form and a demographic questionnaire, each participant was filmed simultaneously interpreting the source text videos shown on a laptop computer; first, the English source text, then the Auslan source text. Each participant watched the first three minutes of each video to warm up and then simultaneously interpreted the remaining 17 minutes of formal presentation. During each task, the researcher sat next to the camera as the participant’s audience, coordinating the filming. After each interpretation, the researcher engaged the participant in a short semi-structured interview. Participation took approximately one and a half hours in total.

Two external raters were selected to assess participants’ simultaneous interpreting performances using pre-designed rubrics. For the purpose of this paper, they have been named as Morgan and Jamie. Each rater met the following criteria: native English speaker, hearing native signer, and NAATI-accredited professional level Auslan/English interpreter with considerable interpreting experience. In terms of assessment experience, both Morgan and Jamie were full-time signed language interpreter educators in an academic interpreter education program. They were also experienced assessors for national NAATI accreditation exams.

Two rubrics were developed for assessing participants’ simultaneous interpreting performances, one rubric for one language direction. Each rubric comprised the following four macro-level assessment criteria:

• accuracy (a maximum score of 50), consisting of the following five sub-criteria: minimal unjustifiable additions, minimal unjustifiable omissions, minimal unjustifiable substitutions, minimal unjustifiable intrusions, and cohesion and coherence;

• target text features (a maximum score of 20), including the following three sub-criteria: grammaticality, appropriate vocabulary, and conveying speaker/signer register and affect;

• delivery features (a maximum score of 15), consisting of the following three sub-criteria: clarity of articulation, fluency, and natural prosody;

• processing skills (a maximum score of 15), comprising the following five sub-criteria: use of time lag, a free-literal interpretation continuum with appropriate shifts in between, strategic additions, strategic omissions, and successful anticipations.

A sub-score was assigned to each macro-level assessment criterion according to the following five levels:

• poor (less than 50%);

• average (50%-64%);

• good (65%-74%);

• very good (75%-84%);

• excellent (85%-100%).

Four sub-scores were then summed up to form the total score for a participant, with the maximum total score being 100. For example, in relation to English-to-Auslan simultaneous interpreting performance, if a participant received a score of 43 for accuracy (86%, in the excellent range), 15 for target text features (75%, in the very good range), 11.25 for delivery features (75%, in the very good range), and 11.25 for processing skills (75%, in the very good range), the participant’s total score for this language direction was 80.50.

Morgan assessed all 31 participants’ simultaneous interpreting performances in both language directions. For inter-rater reliability purposes, Jamie evaluated 16 randomly chosen participants’ simultaneous interpreting performances in both language directions. Pearson correlation coefficients were used as inter-rater reliability estimates. Regarding English-to-Auslan simultaneous interpreting performance, inter-rater reliability was r = .87, N = 16, p < .001. With respect to Auslan-to-English simultaneous interpreting performance, inter-rater reliability was r = .92, N = 16, p < .001. Given that the absolute minimum of acceptable inter-rater reliability is .7 (McNamara 2000), both inter-rater reliability estimates of this study were considered satisfactory. Quantitative results based on Morgan’s ratings are reported below.

Apart from the above main analysis, paired-samples t-tests were used to compare each group’s English-to-Auslan simultaneous interpreting performance with their Auslan-to-English simultaneous interpreting performance. As shown in Table 2, there was no significant difference between the native signers’ English-to-Auslan simultaneous interpreting performance and their Auslan-to-English simultaneous interpreting performance in terms of the total score, t(13) = .60, p = .56. Regarding sub-scores, there were no significant differences between the native signers’ English-to-Auslan simultaneous interpreting performance and their Auslan-to-English simultaneous interpreting performance in terms of accuracy, target text features, delivery features, or processing skills, with all p values larger than .05.

As shown in Table 3 below, in terms of the total score, there was no significant difference between the non-native signers’ English-to-Auslan simultaneous interpreting performance and their Auslan-to-English simultaneous interpreting performance, t(16) = -1.74, p = .10. However, there was a substantial directionality effect on target text features: the non-native signers’ English-to-Auslan (L1-to-L2) simultaneous interpreting performance was significantly inferior to their Auslan-to-English (L2-to-L1) simultaneous interpreting performance in terms of target text features, t(16) = -3.16, p = .006, η² = .38. There were no significant differences between the non-native signers’ English-to-Auslan simultaneous interpreting performance and their Auslan-to-English simultaneous interpreting performance in terms of accuracy, delivery features, or processing skills, with p values above .05.

Independent-samples t-tests were conducted to compare the native signers with the non-native signers in terms of simultaneous interpreting performance in each language direction. As can be seen in Table 4 below, the native signers were similar to the non-native signers in the total score of English-to-Auslan simultaneous interpreting performance, t(29) = 1.76, p = .09. Nonetheless, the two groups differed substantially in two sub-scores: the native signers were significantly superior to the non-native signers in terms of both the target text features and the delivery features of English-to-Auslan simultaneous interpreting performance, p = .039 and p = .03, respectively. However, the native signers were similar to the non-native signers in terms of the accuracy and processing skills of English-to-Auslan simultaneous interpreting performance, with both p values above .05.

As shown in Table 5, the native signers were comparable to the non-native signers in the total score of Auslan-to-English simultaneous interpreting performance, t(29) = .13, p = .90. Additionally, the two groups were similar in terms of the accuracy, target text features, delivery features, and processing skills of Auslan-to-English simultaneous interpreting performance, with all p values larger than .05.

A significant, moderate, positive correlation was identified between all 31 professional interpreters’ English-to-Auslan simultaneous interpreting performance (the total score) and their Auslan-to-English simultaneous interpreting performance (the total score), r = .62, N = 31, p < .001, r² = .38. There was also a significant, moderate, positive correlation between the native signers’ English-to-Auslan simultaneous interpreting performance and their Auslan-to-English simultaneous interpreting performance, r = .75, N = 14, p = .002, r² = .56. Finally, there was a significant, moderate, positive correlation between the non-native signers’ English-to-Auslan simultaneous interpreting performance and their Auslan-to-English simultaneous interpreting performance, r = .53, N = 17, p = .03, r² = .28. These findings are consistent with van Dijk, Boers et al.’s (2011) results that there was a significant, moderate, positive correlation between experienced SLN/Dutch interpreters’ Dutch-to-SLN simultaneous interpreting performance and their SLN-to-Dutch simultaneous interpreting performance. The findings suggest that in the present study both simultaneous interpreting tasks tapped into similar cognitive resources and required similar interpreting skills.

In order to replicate Nicodemus’s (2011) empirical study noted earlier, we repeated all the above statistical analyses on participants who had over 10 years of interpreting experience. These participants comprised 11 native signers (9 female and 2 male; mean age 44, SD = 12; with an average of 20 years of interpreting experience, SD = 5) and 11 non-native signers (10 female and 1 male; mean age 45, SD = 9; with an average of 18 years of interpreting experience, SD = 6). The result pattern remained the same, apart from the change that there were no significant differences between the native signers and the non-native signers in the target text features or delivery features of English-to-Auslan simultaneous interpreting performance. This finding suggests that native signers are comparable to non-native signers in simultaneous interpreting performance, as long as both groups are highly experienced interpreters. The quantitative results will be discussed in a later section.

Results reveal that all 31 professional interpreters’ English-to-Auslan simultaneous interpreting performance was similar to their Auslan-to-English simultaneous interpreting performance in terms of the total score. This finding suggests that there was no significant directionality effect on the professional interpreters’ simultaneous interpreting performance. The result is encouraging for signed language interpreting students and practitioners, because anecdotally many of them are not confident about their abilities to interpret from a signed language into a spoken language. This finding may be attributable to an offset between ineffective self-monitoring of Auslan output in English-to-Auslan simultaneous interpreting and difficulties in comprehending Auslan input in Auslan-to-English simultaneous interpreting, discussed as follows.

When working from a spoken language into a signed language, signed language interpreters cannot see their own facial expression or mouthing; they view their hands in the lower visual field (or their hands may also fall out of their eyesight); and they do not look directly at their hands while signing. This means that signed language interpreters cannot actively monitor their signed output for comprehensibility or grammaticality. Visual feedback may be less effective than auditory feedback in online error detection (Nicodemus and Emmorey 2013). Hence, signed language interpreters often rely on deaf clients’ feedback (in the form of the deaf clients’ facial expression, head nodding, etc.) to gauge whether their interpretations are comprehensible and whether they need to make adjustments. Given that there was no real deaf audience in the English-to-Auslan simultaneous interpreting task in the current study, participants could not have their Auslan output monitored effectively. When working from Auslan into English, on the other hand, participants could immediately monitor their English lexical choices, grammar, coherence, register, pace, and prosody, as they could immediately hear their own interpretations. However, the majority of participants encountered Auslan comprehension problems largely due to unfamiliarity with the deaf presenter and the constraints of the research setting (see the qualitative results in Table 6). Hence, a lack of self-monitoring of the Auslan output and difficulty in comprehending the Auslan input may have resulted in no significant directionality effect on all 31 participants’ simultaneous interpreting performance.

As expected, results reveal that the native signers’ English-to-Auslan simultaneous interpreting performance was as good as their Auslan-to-English simultaneous interpreting performance in terms of the total score and sub-scores. This finding indicates that there was no significant directionality effect on the native signers’ simultaneous interpreting performance. This outcome conflicts with van Dijk, Boers et al.’s (2011) results that native signers’ Dutch-to-SLN simultaneous interpreting performance was significantly better than their SLN-to-Dutch simultaneous interpreting performance. The discrepant findings may be due to methodological differences between the two studies (for example, type of source text, method for assessing simultaneous interpreting performance). Interestingly, the finding in the present study echoes the prior results obtained by Chang and Schallert (2007), who found no significant directionality effect on the simultaneous interpreting performance of professional interpreters with equal proficiency in English and Mandarin. Our finding also lends some credence to the claim that balanced bilingual interpreters are free from the rule of directionality because they work between two native languages (see Denissenko 1989; Seleskovitch 1978).

In addition, results show that the non-native signers’ English-to-Auslan (L1-to-L2) simultaneous interpreting performance was similar to their Auslan-to-English (L2-to-L1) simultaneous interpreting performance in terms of the total score. This finding indicates that there was no significant directionality effect on the non-native signers’ simultaneous interpreting performance. Again, this contradicts the results obtained by van Dijk, Boers et al. (2011), who found that non-native signers’ Dutch-to-SLN (L1-to-L2) simultaneous interpreting performance was significantly superior to their SLN-to-Dutch (L2-to-L1) simultaneous interpreting performance. Nevertheless, our finding is in accordance with Nicodemus’s (2011) results that there was no significant directionality effect on the simultaneous interpreting performance of experienced ASL/English interpreters (non-native signers). Our finding is also in line with the results of Tommola and Helevä (1998) and Tommola and Laakso (1997), who found no significant directionality effect on the simultaneous interpreting performance of student interpreters. A number of factors may explain our finding. To begin with, given that the majority of the non-native signers in the present study had over 10 years of interpreting experience, they may have developed relatively equal proficiency in English and Auslan as well as relatively balanced competence of interpreting in each language direction. Alternatively, compared with other variables (for example, interpreters’ interpreting skills and interpreting experience), directionality may have less impact on simultaneous interpreting performance (see Gile 2005; Kalina 2005; Pavlović 2007; Seel 2005).

Interestingly, the non-native signers’ English-to-Auslan (L1-to-L2) simultaneous interpreting performance was significantly inferior to their Aulsan-to-English (L2-to-L1) simultaneous interpreting performance in terms of target text features. This corroborates the results of Chang and Schallert (2007), who found that professional Mandarin/English interpreters made significantly more target-language-use errors when simultaneously interpreting from L1 (Mandarin) into L2 (English) than vice versa. The finding highlights the importance for non-native signer interpreters to continue to improve their signed language (L2) proficiency (see Taylor 1993; 2002 for a similar claim). This result supports Donovan’s (2005) view that both student interpreters and professional interpreters should make constant efforts to refine their non-native language skills.

Further, as in van Dijk, Boers et al.’s (2011) study, the present study found that the native signers were similar to the non-native signers in overall simultaneous interpreting performance (collapsing English-to-Auslan and Auslan-to-English simultaneous interpreting performances). Another revealing finding is that the native signers were similar to the non-native signers in the total score of English-to-Auslan simultaneous interpreting performance. This new finding suggests that there was no significant age of signed language acquisition effect on the professional interpreters’ English-to-Auslan simultaneous interpreting performance. The finding indicates that signed language nativeness may not be crucial for determining the simultaneous interpreting performance of professional signed language interpreters.

As expected, the native signers were significantly superior to the non-native signers in terms of the target text features and delivery features of English-to-Auslan simultaneous interpreting performance. These findings are consistent with Napier’s (2006) results that a native signer interpreter’s Auslan target text was more idiomatic and natural than that of a non-native signer interpreter. The findings also support Nicodemus and Emmorey’s (2013) results that native signer interpreters’ self-perceived proficiency in a signed language (ASL) was substantially higher than that of non-native signer interpreters. Again, these findings highlight the need for non-native signer interpreters to further develop their signed language (L2) skills. Regarding differences between native signers and non-native signers, native signer Rachael commented:

I believe that there is a big difference between native signers and non-native signers when they interpret from English into Auslan. I truly believe that. I am often told that I use a lot of space and many non-native signers do not exploit space as much (space is used for linguistic purposes in signed languages). I also use a lot of facial expression and sometimes that is not so obvious in non-native signers (facial expression has important functions in signed languages, such as to mark questions and to show emotional state).

Similarly, native signer Lauren commented:

I think it is just more natural for native signers to use more space in English-to-Auslan interpreting. Spreading it out gives you a little bit more flexibility. … Once I have put the information into Auslan in my head, it just comes out on my hands naturally. So, while that is happening, I can still be listening. I don’t necessarily have to concentrate on making sure my hands are in the right areas. That sort of just happens.

Participants in this study exhibited marked variability in Auslan proficiency. In the English-to-Auslan simultaneous interpretation data, several non-native signer interpreters showed inappropriate signing style and use of Auslan (for example, exaggerated facial expression, unclear production of signs, inappropriate prosody). By contrast, many other non-native signer interpreters demonstrated near-native proficiency in Auslan, in that their Auslan output was fluent, smooth, and clear.

Moreover, this study found that the native signers were similar to the non-native signers in terms of the total score and sub-scores of Auslan-to-English simultaneous interpreting performance. These findings suggest that there was no significant age of signed language acquisition effect on the professional interpreters’ Auslan-to-English simultaneous interpreting performance. There are some possible explanations for the results. First, both the native signers and the non-native signers encountered some Auslan comprehension problems when interpreting the Auslan source text into English, partly due to unfamiliarity with the deaf presenter, no opportunity to meet him beforehand, and no PowerPoint slides on show. Second, Auslan-to-English interpreting performance may be more related to an interpreter’s educational background, English vocabulary bank, and interpreting experience than to Auslan nativeness. These assumptions merit further research.

There is a considerable degree of signed language variation in the Deaf Community of a country (Napier, McKee et al. 2006/2010; Nicodemus and Emmorey 2013; Schembri, Johnston et al. 2006; Schembri, McKee et al. 2009). In other words, deaf Auslan signers often differ from one another in terms of signing style and lexical choice. An analysis of the qualitative interview data reveals that many interpreters in the present study encountered Auslan comprehension problems due to unfamiliarity with the deaf presenter. Several interpreters who were familiar with the deaf presenter in real life produced highly accurate Auslan-to-English interpretations in this study. These findings support the paradigm that deaf professionals should work with designated interpreters (see Hauser, Finch et al. 2008). Even highly skilled interpreters find it difficult to interpret for deaf professionals with whom they have never worked before (Hurwitz 2008). Familiarity, regular contact, mutual trust, rapport, and teamwork between a deaf professional and a designated interpreter, contribute to the interpreter’s optimal interpreting performance (Hauser and Hauser 2008). Napier, Carmichael et al. (2008) found that in a monologic seminar presentation a deaf presenter and two designated interpreters strategically used pauses, nods, and eye contact as contextualization cues in order to signal comprehension, mark episodes, and control the pace of the presentation. Therefore, if the professional interpreters in the current study had been in the room with the deaf presenter and had met him beforehand, they would have been able to negotiate, agree on, and implement similar strategies. Similarly, Gile (1995/2009) maintains that interpreters’ good knowledge of the speaker and subject matter often enables them to anticipate information in the source text.

This study has some limitations that need to be acknowledged. To begin with, given the relatively small sample size and considerable variability in participants’ demographic characteristics, findings of the study may not be generalized to the broader community of signed language interpreters. Further research should include larger samples of professional signed language interpreters who are relatively homogenous in demographic data (for example, educational background, level of interpreter education, interpreting experience). Secondly, the researcher’s presence as participants’ audience in both interpreting tasks, the researcher’s linguistic background (a native speaker of Mandarin with near-native proficiency in English and intermediate proficiency in Auslan), and the lack of real conference audience may have affected the participants’ simultaneous interpreting performances in this study. Thirdly, the research setting in this study prevented participants from using the following coping strategies that are available in real-life conference interpreting assignments:

• meeting the hearing and deaf presenters beforehand to clarify issues and to familiarize with their pace and style;

• receiving feedback from the audience;

• referring to visual aids (e.g., PowerPoint slides);

• working with a team interpreter.

Therefore, participants’ simultaneous interpreting performances in this artificial testing situation may not represent their optimal interpreting competence in real-time assignments. Fourthly, as signed language interpreters in Australia do not work full-time as conference interpreters due to a lack of job availability, the majority of participants work primarily in community settings in real life. This means that some participants’ interpreting performance in the present study may not represent their optimal interpreting performance in community settings.